Identification and characterization of the
novel CEACAM-binding adhesin of
Haemophilus influenzae
Dissertation
Zur Erlangung des akademischen Grades
des Doktors des Naturwissenschaften (Dr. rer. nat.)
des Fachbereichs Biologie der Universität Konstanz
vorgelegt von
Arnaud Kengmo Tchoupa
Konstanz 2015
Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-0-310442
Date of the doctoral oral examination: 20.11.2015
Chairperson and oral examiner: Professor Dr. Bürkle
1. Oral examiner: Professor Dr. Hauck
2. Oral examiner: Professor Dr. Reidl, Karl-Franzes-University Graz
Table of Contents
I
Table of Contents
Table of Contents .............................................................................................................................. I
Acknowledgement ........................................................................................................................... III
Summary .......................................................................................................................................... IV
Zusammenfassung .......................................................................................................................... VI
General Introduction ........................................................................................................................ 1
1. Haemophilus influenzae – the bacterium ............................................................................. 1
2. Clinical significance of Haemophilus influenzae ................................................................... 2
3. Pathogenesis of Haemophilus influenzae ............................................................................. 3
4. Virulence factors of Haemophilus influenzae ........................................................................ 5
5. CEACAM-recognition by Haemophilus influenzae .............................................................. 10
6. CEACAMs – a common target for many bacteria ............................................................... 12
Aims of the study ........................................................................................................................... 16
CHAPTER I: SIGNALING BY EPITHELIAL MEMBERS OF THE CEACAM FAMILY – MUCOSAL
DOCKING SITES FOR PATHOGENIC BACTERIA ................................................ 18
I.1 Abstract ............................................................................................................................... 19
I.2 Introduction .......................................................................................................................... 19
I.3 Physiological roles of epithelial CEACAMs ......................................................................... 21
I.4 Role of CEACAM extracellular domains in mediating cis- and trans-oligomerization ......... 23
I.5 Signaling by epithelial CEACAMs ....................................................................................... 24
I.6 CEACAM1 cis-oligomerization sustained by the transmembrane domain ......................... 26
I.7 CEACAM-binding bacteria reveal the lipid raft association of their receptors .................... 27
I.8 CEACAM1 signaling initiated by the IgC2-like extracellular domains ................................... 29
I.9 CEACAM cooperation with integrins and other membrane receptors ................................ 31
I.10 Conclusions ..................................................................................................................... 34
Table of Contents
II
CHAPTER II: OUTER MEMBRANE PROTEIN P1 IS THE CEACAM-BINDING ADHESIN OF
HAEMOPHILUS INFLUENZAE ................................................................................ 35
II.1 Summary ............................................................................................................................. 36
II.2 Introduction .......................................................................................................................... 36
II.3 Results ................................................................................................................................. 38
II.4 Discussion ........................................................................................................................... 55
II.5 Materials and methods ........................................................................................................ 57
II.6 Acknowledgements ............................................................................................................. 60
II.7 Supporting information ........................................................................................................ 61
CHAPTER III: THE OUTER MEMBRANE PROTEIN P1 OF HAEMOPHILUS INFLUENZAE
UTILIZES ITS EXTRACELLULAR LOOPS TO GRASP HUMAN CEACAM ........ 63
III.1 Abstract ........................................................................................................................... 64
III.2 Introduction ..................................................................................................................... 64
III.3 Results ............................................................................................................................ 66
III.4 Discussion ....................................................................................................................... 77
III.5 Materials and methods .................................................................................................... 79
CHAPTER IV: THE HAEMOPHILUS INFLUENZAE ADHESIN OMP P1 IS REGULATED BY
HOST-DERIVED FATTY ACIDS ............................................................................. 83
IV.1 Abstract ........................................................................................................................... 84
IV.2 Introduction ..................................................................................................................... 84
IV.3 Results ............................................................................................................................ 87
IV.4 Discussion ....................................................................................................................... 95
IV.5 Materials and methods .................................................................................................... 98
General Discussion ...................................................................................................................... 101
Declaration of author’s contributions ........................................................................................ 107
List of publications ...................................................................................................................... 108
References .................................................................................................................................... 109
Acknowledgement
III
Acknowledgement
This piece of work would not have been possible without the joint efforts of
numerous people who have gained my deepest gratitude. In the following lines,
not meant to be exhaustive, I want to sincerely thank:
Prof. Dr. Christof R. Hauck, my supervisor. He offered me the unique opportunity
to conduct research in the forefront of Host-Pathogen interaction. I am amazed by
his enthusiasm, fascinated by his patience and grateful for his smart advices.
Prof Dr. Joachim Reidl for accepting to judge this work. Besides, his intensive
support was instrumental for our major findings.
Prof. Dr. Alexander Bürkle for his incessant guidance as member of my thesis
committee and graduate school, and as examiner of my thesis.
Drs. Alex Buntru, Petra Muezner-Voigt, Naja Nyffenegger, Alexandra Roth and
Maike Voges. As experienced (former) members of our lab they have introduced
me to the technics and provided me with numerous tips and tricks.
Julia Delgado-Tascón, Nina Dierdorf, Christof Paone, Alexander Timper and Yong
Shi for being irreplaceable team-players in the lab and beyond. A special mention
to Chris for proofreading my work and to Nina for the abstract in German.
Anne Keller for her guidance and inestimable help with bureaucratic issues.
Susanne Feindler-Boeckh, Claudia Hentschel and Petra Zoll-Kiewitz, for their
expert technical support on a daily basis. Susi, you are simply the best!
The AWOLI football team (Eric, Christian, Patrick, Faustin, Pericles, Daniel,
Nadège…) without which my weekends would have been monotone.
Elke Cubillos, my flat-mate and lessor. She was kind enough to host me during the
last three years of my PhD studies. “Danke für alles Elke”!
My cousin Tchoupa Marlyse for her incessant support.
My parents (Jean-Marie and Madeleine Kengmo), my brother Samsimon and my
sisters (Kévile, Suzy, Clarissa). They have been very supportive and an endless
source of motivation. Merci à vous les K.E.N.G.M.O. L’avenir nous appartient.
Summary
IV
Summary
Haemophilus influenzae (Hinf), a mostly commensal inhabitant of human
respiratory airways, is the causative agent of local infections such as middle ear
infections, sinusitis, bronchitis, conjunctivitis and pneumonia, but can also cause
life-threatening disseminating diseases, including meningitis and septicemia. The
success of Hinf as commensal or opportunistic pathogen relies on the ability of the
bacterium to use a combination of adhesive surface proteins, the so-called
adhesins, which intimately bind to structures on the human mucosal tissue in order
to overcome mechanical clearance and escape immune recognition. In that
respect, Hinf targets carcinoembryonic antigen-related cell adhesion molecules
(CEACAMs), which are glycoproteins of the immunoglobulin superfamily and
which are present on the apical side of nasopharyngeal epithelial cells. The first
chapter of this work presents in detail the numerous advantages for the bacteria
inherent in their interaction with epithelial CEACAMs (e.g.: strong attachment to
the mucosal surface, internalization within epithelial cells and therefore protection
against immune response).
The interaction between Hinf and CEACAMs was postulated to be mediated by the
outer membrane protein (OMP) P5, one of the major OMPs of Hinf. Therefore, we
set out to characterize the molecular requirements of CEACAM-binding by OMP
P5. However, (i) the ability of P5-deficient Hinf to strongly interact with CEACAMs;
(ii) the similarity of P5 sequences from non-CEACAM-binding and CEACAM-
binding Hinf strains, and (iii) the failure of OMP P5 to bind CEACAMs when
heterologously expressed in Escherichia coli questioned the role of OMP P5 as
the Hinf CEACAM-binding adhesin. A screen of a panel of Hinf mutants with
defined deletions in their outer membrane proteins revealed that the depletion of
OMP P1 totally abrogated the interaction of Hinf with CEACAMs. Furthermore,
non-CEACAM-binding E. coli gained CEACAM-binding properties upon ectopic
expression of Hinf OMP P1. Thus, OMP P1 is necessary and sufficient for
CEACAM-targeting.
Following the identification of OMP P1 as the bona fide CEACAM-binding adhesin
of Hinf, we sought to identify the domain(s) of OMP P1 involved in CEACAM-
recognition. Therefore, we took advantage of the high similarity between OMP P1
Summary
V
and its non-CEACAM-binding homolog FadL in E. coli. The solved crystal structure
of FadL helped to model OMP P1, which is predicted to build a 14-stranded β-
barrel. Interestingly, chimeras resulting from the fusion of OMP P1 with FadL
fragments combined with OMP P1 mutants deprived of specific extracellular loops
clearly underlined the involvement of the most prominent surface-exposed OMP
P1 loops (L1, L3, L4 and L7) in CEACAM recognition. The flexibility of these loops
appears important for the proper binding to CEACAM. In contrast to Hinf OMP P1,
P1 homologs of other members of the Pasteurellaceae family were unable to
recognize CEACAMs. Moreover, Hinf OMP P1 bound exclusively to human
CEACAMs, but not to other distantly related mammalian CEACAMs.
FadL, the OMP P1 homolog in E. coli, mediates the transport of long-chain fatty
acids (LCFAs) across the bacterial outer membrane. The high structural homology
shared by FadL and OMP P1 raised the question whether OMP P1 is involved in
fatty acid transport and/or is regulated by fatty acids. Though the addition of fatty
acids to the Hinf medium was able to potently stimulate bacterial growth, the role
of OMP P1 in LCFA transport requires further investigations. However, it was
obvious that LCFAs supplementations upregulated OMP P1 expression in Hinf.
Furthermore, Hinf strains treated with LCFAs were able to better interact with
recombinant CEACAMs and were more invasive in CEACAM-expressing epithelial
cells, compared to untreated bacteria.
In sum, this study not only identifies a novel CEACAM-binding adhesin in a
prominent human pathogen, but also provides insights into a new CEACAM-
binding interface built by 4 flexible loops of a bacterial outer membrane protein.
Zusammenfassung
VI
Zusammenfassung
Haemophilus influenzae (Hinf), ein meist kommensaler Bewohner der
menschlichen Atemwege, ist der Erreger lokaler Infektionen, wie Bronchitis,
Bindehautentzündung und Lungenentzündung. In einigen Fällen löst Hinf
lebensbedrohliche disseminierte Krankheiten wie Meningitis und Septikämie aus.
Hinf ist sehr erfolgreich als Kommensal und Pathogen, da das Bakterium eine
Vielzahl von speziellen oberflächlichen Proteinen besitzt, welche als Adhäsine
bezeichnet werden. Um eine mechanische Eliminierung und Immunerkennung zu
umgehen, binden diese Moleküle eng an Strukturen des menschlichen
Schleimhautgewebes. Hierzu bindet Hinf an CEACAMs (Glykoproteine der
Immunglobulin-Superfamilie, die auf der apikalen Seite nasopharyngealen
Epithelzellen vorhanden sind). Das erste Kapitel der vorliegenden Arbeit wird im
Einzelnen die Vorteile jener Bakterien erfassen, welche mit epithelialen CEACAMs
interagieren. Dazu gehören eine starke Bindung an der Schleimhautoberfläche,
die Internalisierung in Epithelzellen und dadurch Schutz gegen Immunantwort.
Es wurde postuliert, dass die Interaktion zwischen Hinf und CEACAMs vom
äußeren Membranprotein (OMP) P5 – ein der wichtigsten OMPs von Hinf –
vermittelt wird. Deswegen haben wir es uns zum Ziel gesetzt, die molekularen
Voraussetzungen für ein CEACAM – OMP P5 Bindung zu charakterisieren.
Jedoch (i) die Fähigkeit von P5-defizienten Hinf-Mutanten stark mit CEACAMs
zusammen zu interagieren; (ii) die Ähnlichkeit der P5-Sequenzen von nicht-
CEACAM-bindende und CEACAM-bindende Hinf Stämme, und (iii) die Unfähigkeit
der P5-Proteine CEACAMs zu erkennen, wenn sie in E. coli exprimiert wurden,
warfen ernste Zweifel an P5 als Hinf CEACAM-bindende Adhäsin auf. Ein Screen
mit verschiedenen Hinf-Mutanten, welche Deletionen in definierten
Außenmembranproteinen aufwiesen, ergab, dass die Abwesenheit von OMP P1
vollständig die Interaktion von Hinf mit CEACAMs unterdrückte. Des Weiteren
konnte gezeigt werden, dass nicht-CEACAM-bindende E. coli nach ektopischer
Expression von Hinf OMP P1 CEACAM-Bindungseigenschaften erwarben. Somit
ist OMP P1 für eine CEACAM-Bindung notwendig und ausreichend.
Nach der Identifizierung von Hinf OMP P1 als CEACAM-bindendes Adhäsin
versuchten wir, den/die Sequenzabschnitt/e von OMP P1 zu charakterisieren,
Zusammenfassung
VII
welcher/e für die Bindung an CEACAMs verantwortlich ist/sind. Durch Protein
Modellierung, dank dem bereits kristallisierten nicht-CEACAM-bindenden E. coli
FadL (OMP P1 Homolog), konnte eine 14-strängige β-barrel-Struktur für P1 erstellt
werden. Mit Hilfe von Chimären aus OMP P1 und FadL sowie OMP P1-Mutanten,
in denen spezifische extrazelluläre Loops deletiert wurden, konnte gezeigt werden,
dass die größten Loops (L1, L3, L4 und L7) von OMP P1 essentiell für die
CEACAM-Bindung sind. Die Flexibilität dieser Loops ist scheinbar wichtig für die
richtige Bindung an CEACAM. Im Gegensatz zu Hinf OMP P1, P1-Homologe aus
anderen Mitgliedern der Pasteurellaceae Familie konnten nicht an CEACAMs
binden. Außerdem erkennt Hinf P1 ausschließlich menschliche CEACAMs und
nicht solche entfernt verwandter Säuger.
FadL, das OMP P1 Homolog in E. coli, vermittelt den Transport von langkettigen
Fettsäuren (LCFAs) über die äußere Membran. Die hohe strukturelle Homologie
von FadL und OMP P1 warf die Frage auf, ob OMP P1 am Fettsäuretransport
beteiligt ist und / oder von Fettsäuren reguliert wird. Durch die Zugabe von
Fettsäure in das Kulturmedium von Hinf konnte das Wachstum dieser Bakterien
effektiv stimuliert werden. Es sind jedoch weitere Untersuchungen notwendig, um
die Rolle von OMP P1 am LCFA-Transport zu evaluieren. Es war offensichtlich,
dass die LCFA-Zugabe die OMP P1 Expression in Hinf erhöhte. Darüber hinaus
waren Hinf Stämme nach LCFA-Behandlung in der Lage besser mit
rekombinanten CEACAMs zu interagieren und schneller in CEACAM-
exprimierenden Epithelzellen einzudringen.
Zusammengefasst, identifiziert diese Studie nicht nur ein neuartiges CEACAM-
bindendes Adhäsin, sondern liefert auch Einblicke in ein neues CEACAM-
bindendes Interface, bestehend aus 4 flexiblen Loops eines bakteriellen
Außenmembranproteins.
General Introduction
1
General Introduction
1. Haemophilus influenzae – the bacterium
Haemophilus influenzae (Hinf) is a non-motile, pleomorphic, Gram-negative
coccobacillus. Its genus’ name, Haemophilus (from Greek: blood loving),
describes the requirement by this bacterium of blood-derived nutrients, namely
hemin and nicotinamide-adenine dinucleotide (NAD). Haemophilus influenzae,
formerly Pfeiffer’s bacillus, was discovered during an influenza pandemic and was
wrongly thought, during more than 3 decades, to be the causative agent of the flu
(Norskov-Lauritsen, 2014).
Haemophilus influenzae species is phenotypically heterogeneous. For instance,
some strains are shielded by a polysaccharide capsule that is absent in other
strains. The encapsulated Hinf strains (typeable) can be categorized, depending
on the composition of their capsular polysaccharides, into six serotypes,
designated a through f. The unencapsulated strains consist of encapsulated
bacteria that have switched off their capsule expression or bacteria that completely
lack the genetic information for capsule synthesis (non-typeable) (Agrawal and
Murphy, 2011).
At a genomic level, by multilocus sequence analysis, Hinf strains are divided into
phylogenetic groups I and II. The group I represents the core of the Hinf species
and includes all strains of serotypes c and d, most of strains of serotypes a and b,
and the majority of non-typeable isolates. Conversely, the phylogenetic group II is
small and encompasses all strains of serotypes e and f, few strains of a and b, and
some non-typeable isolates. However, there is an increasing number of
unencapsulated isolates that can be assigned to either group I or II (Norskov-
Lauritsen, 2014). DNA-based analyses remain nevertheless the method of choice
when it comes to subtyping Hinf. This is exemplified by the former species
Haemophilus aegyptius that was shown in DNA hybridization studies to belong to
Haemophilus influenzae species and therefore renamed H. influenzae biogroup
aegyptius (Harrison et al., 2008).
General Introduction
2
2. Clinical significance of Haemophilus influenzae
Hinf is a commensal colonizer of the human respiratory tract. The carriage rates
vary with the type of Hinf strains and the age of the healthy carrier. For instance,
the colonization rate is close to zero for Haemophilus influenzae type b (Hib)
where the Hib vaccination is widespread (Agrawal and Murphy, 2011). Since the
Hib vaccine is ineffective against non-typeable Hinf (NTHi) strains, these
organisms can colonize up to 80% of both children and adults (Rao et al., 1999).
By contiguous spread (in the case of NTHi) or invasion of the bloodstream
(encapsulated strains), Hinf reaches privileged anatomical sites where it, under
predisposing conditions (e.g.: age, viral infections, immunodeficiency), can cause
diverse affections. Some examples are listed below:
(i) Meningitis. In the pre-Hib-vaccine era, Hinf meningitis were mostly due to type
b strains. Nowadays (in the post-Hib-vaccine era), non-typeable strains are
responsible for most invasive infections, followed by serotypes f and b (Norskov-
Lauritsen, 2014).
(ii) Otitis media. It is the childhood disease for which medical assistance is mostly
sought. NTHi accounts for 27–37% of acute otitis media and is therefore the
second most common etiological agent of this disease. Moreover, NTHi represents
the most common bacterial cause of chronic otitis media with effusion (Rao et al.,
1999).
(iii) Conjunctivitis. Children are more affected by bacterial conjunctivitis than
adults and NTHi causes 44–68% of cases in children, compared with about 25% in
adults (Van Eldere et al., 2014). Importantly, 20–70% of children with bacterial
conjunctivitis also suffer from acute otitis media. Given the high prevalence of
NTHi in bacterial conjunctivitis, it is likely that NTHi is the causative agent of the
concomitant otitis media.
(iv) Sinusitis. Hinf is accountable for approximately one-third of episodes of acute
and chronic sinusitis. Besides, the usage of pneumococcal conjugate vaccines
seems to have relatively increased the contribution of Hinf as causative agent of
sinusitis (Agrawal and Murphy, 2011).
(v) Exacerbations in adults with COPD (chronic obstructive pulmonary
disease). COPD is a leading cause of death worldwide. NTHi is a major cause of
General Introduction
3
acute exacerbations of COPD by inducing airway inflammation and tissue damage
(Van Eldere et al., 2014).
As numerous studies report the widespread resistance of Hinf to ampicillin, the
main drug of choice in proven Hinf infections, the clinicians still have a lot of
alternative antibiotics. Antibiotic therapy is even not necessary in case of a mild
infection, as the symptomatic therapy is, in that case, enough to accelerate the
healing of the patient. For moderate or severe infections, where the involvement of
Hinf is suspected, empirical therapy (with for example, advance macrolides,
cephalosporin and fluoroquinolone) can be successful. However, bacterial culture
and antimicrobial susceptibility testing enable the clinician to optimize the antibiotic
treatment according to the identified bacterial strain.
3. Pathogenesis of Haemophilus influenzae
As a common commensal inhabitant of the human nasopharyngeal mucosa, Hinf
has to face and withstand host clearance mechanisms, including mucociliary
escalator and local immunity (St. Geme III, 2002). Undeniably, the mucus layer of
the respiratory mucosa efficiently traps most of the incoming bacteria and viruses
(Rao et al., 1999). Although able to bind the mucin, the main glycoprotein of the
mucus (Reddy et al., 1996), Hinf preferentially adheres to non-ciliated cells and to
areas of damaged epithelium via numerous adhesins (St. Geme III, 2002).
Conditions impairing mucociliary clearance or damaging the epithelial barrier
therefore favor Hinf colonization (Fig. 1). However, the bacterium on its own is
able to reduce the ciliary beat frequency, to promote the loss of cilia by epithelial
cells, or to directly damage the mucociliary escalator (Foxwell et al., 1998).
To circumvent host humoral response, Hinf secretes IgA1 proteases which cleave
secretory IgA1, the most abundant mucosal immunoglobulin. A strong support for
extreme importance of IgA1 proteases for Hinf pathogenesis was provided by a
recent study with human volunteers colonized with Hinf. There, the phase variable
igaB gene of IgA1 protease was significantly more switched on after 3–6 days of
human nasopharyngeal colonization, compared to the initial inoculum (Poole et al.,
2013).
General Introduction
4
The strong adherence of Hinf to epithelial cells, following the interaction of Hinf
adhesins with plasma membrane receptors, results in some cases in the invasion
of the human cells by the bacteria (Clementi and Murphy, 2011). Internalized Hinf,
escaping antibiotic therapy and immune response, are thought to constitute a
protective reservoir for persistence and recurrent infections (Clementi and Murphy,
2011). Besides, good proportion of intracellular Hinf is likely released beneath the
epithelium by the so-called transcytosis. Alternatively, Hinf can reach the
basement membrane by paracytosis, which is the passage of bacterium between
cells. It is elusive how Hinf degrades the basement membrane in order to get
access to the capillary enterocytes and thereby invades the bloodstream. Once in
blood, the rigid polysaccharide capsule is an enormous advantage for survival.
This explains why the unencapsulated NTHi isolates cause mainly local infections.
Fig. 1. Pathogenesis of Haemophilus influenzae. (1) Viral infections, exposure to environmentaltoxicants and other factors impairing the mucociliary function increase the likelihood of bacterialcolonization. Bacteria (red ovals), which are not trapped in the mucus and cleared, are able toreduce the ciliary beat frequency or to promote the loss of cilia by epithelial cells. (2) Upon damageof the mucociliary escalator, bacteria can attach to non-ciliated epithelial cells. (3) Adherentbacteria can form aggregates or be internalized. (4) Under certain conditions, bacterialmicrocolonies break up to disperse organisms, spreading within the respiratory tract (modified from(Rao et al., 1999)).
General Introduction
5
The presence of Hinf on the mucosal surface of its host does not remain
undetected. There is a robust production of proinflammatory cytokines in response
to Hinf. Fascinatingly, the resulting influx of neutrophils, monocytes, eosinophils,
and macrophages, accentuates the epithelial cell damage, stimulates mucus
production and therefore fosters the spreading of the bacteria (Rao et al., 1999).
Eventually, the outcome of acquisition of a Hinf strain will depend on both host
fitness and pathogen virulence.
4. Virulence factors of Haemophilus influenzae
In an apparently redundant manner, Hinf utilizes numerous factors to accomplish
the same function, which may be attachment to and invasion into epithelial cells or
immune response evasion. These virulence factors are summarized in the Fig. 2.
Fig. 2. Overview of Hinf major virulence factors. In typeable Hinf the capsule is acknowledgedto mask other virulence factors, such as lipooligosaccharides (LOS) or integral outer membraneadhesins. However, huge surface appendages, including pili and autotransporters, traverse thecapsule.
General Introduction
6
i. The capsule
Encapsulated Hinf can readily lose their capsule. When present, the
polysaccharide capsule of Hinf, by protecting the bacterium from desiccation,
improves the survival of the bacterium in the environment and therefore increases
the propagation to new host (St. Geme III and Cutter, 1996). In the host, Hinf
capsule induces resistance to complement-mediated killing and phagocytosis,
both of which promote bacterial survival in human blood stream. However, the
capsule masks bacterial adhesins that are instrumental in the initial steps of the
bacterial colonization (e.g: adhesion to and invasion into epithelial cells).
Therefore, it has been proposed that, depending on the stage of the infection and
in response to environmental stimuli, encapsulated Hinf strains are able to
modulate the expression of their capsule (St. Geme III and Cutter, 1996).
ii. The pili
Selected typeable and non-typeable Hinf express pili, which are hairlike surface
appendages reaching up to 450 nm in length. Pili are used by Hinf to establish
long-range contact with epithelial cells, mucin or even extracellular matrix proteins
(St. Geme III, 2002). The adhesive pilus is encoded by a gene cluster made of 5
genes (hifA–hifE), with hifE coding for the pilus adhesin. Of note, these huge pili
play a role in the bacteria-induced agglutination of red blood cells and are
therefore called hemagglutinating pili.
Besides, the hemagglutinating pili not involved in bacterial motility, the relatively
small (5–7nm) type IV pilus (TfP) was initially characterized in the NTHi isolate
086-026NP and shown to be indispensable for the twitching motility of this strain
(Bakaletz et al., 2005). Since, the implication of TfP has been extended to:
adherence to epithelial cells, biofilm formation (in vitro and in vivo) and efficient
colonization of animal models (Carruthers et al., 2012).
iii. Fibrils
Fibrils are adhesive structures significantly thinner and smaller than the
hemagglutinating pilus and coded by hsf or hia gene in encapsulated or
nonencapsulated Hinf, respectively.
Hsf (Haemophilus surface fibril) is a huge trimeric autotransporter (100 nm) that
was proven, in Hinf type b, to promote adherence to respiratory epithelial cells
General Introduction
7
and, via binding to vitronectin, to contribute to serum resistance (Singh et al.,
2014).
Hia (Haemophilus influenzae adhesin), smaller than Hsf, is a high-molecular-
weight autotransporter (115 kDa) sharing significant homology with Hsf. Hia is also
an adhesin, whose adhesive properties are conserved when ectopically expressed
in Escherichia coli (Rao et al., 1999).
iv. HMW1 and HMW2 proteins
High-molecular-weight (HMW) adhesins HMW1 and HMW2 are autotransporters
present in up to 75% of all NTHi isolates and absent in typeable Hinf. The huge
majority of NTHi strains deficient in HMW1 and HMW2 expressed alternatively
Hia. HMW1 and HMW2 mediate binding to epithelial cells. Although very similar
(80% similarity), HMW1 and HMW2 differ in their binding specificity. For instance,
when expressed in E. coli, HMW1-mediated adherence to laryngeal Hep-2 cells
was 3-fold higher compared to HMW2 (Hultgren et al., 1993).
v. Hap
Hap (Haemophilus adhesion and penetration protein) is also an autotransporter
present in NTHi strains. The mature protein encompasses two domains: the
autotransporter β-barrel embedded in the outer membrane (designated Hapβ) and
the extracellular passenger domain (designated HapS). Once Hap is translocated
on the bacterial surface, an autoproteolytic cleavage event releases HapS from
Hapβ (St. Geme III, 2002). Interestingly, HapS mediates attachment to epithelial
cells and interaction with extracellular matrix proteins (fibronectin, laminin and
collagen IV). HapS also promotes bacterial aggregation and microcolony formation.
Most importantly, secretory leukocyte protease inhibitor, abundant in human upper
respiratory tract, inhibits Hap autoproteolytic cleavage and therefore potentiates its
adhesive functions (Rao et al., 1999). Moreover, when expressed in the otherwise
Hap-deficient Hinf Rd strain, Hap consistently increases the attachment of Rd to
epithelial cells (Euba et al., 2015).
vi. LOS
Lipooligosaccharide (LOS) represents a major component of the Hinf cell wall. The
role of LOS in Hinf virulence has been better characterized in NTHi isolates where
it is not masked by a capsule. LOS is involved in the early stages of Hinf
General Introduction
8
pathogenesis where it contributes to the adherence of the bacterium to epithelial
cells by interacting with the PAF (platelet-activating factor) receptor (Swords et al.,
2000). Importantly, LOS needs to incorporate phosphorylcholine (ChoP) in order to
bind the PAF receptor. The relevance of ChoP+ LOS is supported further by the
finding in NTHi that phase variable licA gene coding for the phosphorylcholine
kinase incorporating ChoP into LOS, was turned on 7 times more after 6 days of
human colonization compared to the initial inoculum (Poole et al., 2013).
When established in the mucosa, the bacteria use secreted form of LOS to induce
paralysis and loss of cilia by the epithelial cells (Rao et al., 1999). Importantly,
LOS plays a major role in the persistence of Hinf. In details, LOS contributes to the
bacterial resistance to complement killing. Also, by integrating host molecules,
such as ChoP or sialic acid, in its structure, LOS participates to bacterial host
mimicry and immune evasion (Bouchet et al., 2003).
vii. The outer membrane proteins
Outer membrane protein (OMP) P2. The OMP P2 is the most abundant protein
in the Hinf outer membrane. It is a trimeric, β-barreled protein with porin activity.
P2 binds to human nasopharyngeal mucin (Reddy et al., 1996). OMP P2 may also
play a role during systemic disease caused by Hinf type b as an otherwise
isogenic P2 mutant of a virulent type b Hinf strain was unable to provoke
bacteremia in infant rat (Cope et al., 1990). Besides, OMP P2 was shown to be
upregulated on bacteria within biofilm and also in the biofilm extracellular matrix
(Wu et al., 2014). It is however unclear, if P2 is necessary for biofilm formation in
Hinf.
OMP P4. The outer membrane protein P4 (or lipoprotein e) is primordial for the
transport hemin and nicotinamide nucleotides, across the outer membrane of Hinf
(Reidl and Mekalanos, 1996; Reidl et al., 2000). Interestingly a P4-deficient mutant
was proven to be less virulent than its isogenic wild-type Hib strain in rat model of
invasive disease (Morton et al., 2007). Lately, OMP P4 was shown to: (i) bind
extracellular matrix (ECM) proteins; (ii) promote bacterial attachment to epithelial
cells; (iii) improve serum resistance and (iv) be necessary for bacterial colonization
of mouse middle ear (Su et al., 2015).
OMP P5. The outer membrane P5 is a β-barrel transmembrane protein whose
porin activity has been not yet characterized. In clear contrast, its role as an
General Introduction
9
adhesin has been extensively investigated. For instance, OMP P5 associates with
human nasopharyngeal mucin (Reddy et al., 1996). P5 depletion decreases both
the adherence of Hib to human oropharyngeal cells and the severity of otitis media
in chinchilla model (Sirakova et al., 1994). The receptors of P5 on human epithelial
cells were suggested to be CEACAM1 (carcinoembryonic antigen-related cell
adhesion molecule 1) and ICAM1 (intercellular adhesion molecule 1) (Avadhanula
et al., 2006; Hill et al., 2001). OMP P5 is also implicated in the resistance of NTHi
to both classical and alternative complement pathways in typeable and non-
typeable Hinf strains (Rosadini et al., 2014). More recently, OMP P5 importance
for human epithelial cell invasion and bacterial persistence in mice lungs was
nicely demonstrated (Euba et al., 2015).
TbpA and TbpB. The transferrin binding protein (Tbp) A and B are outer
membrane proteins whose expression is iron-repressible. TbpA and TbpB interact
together with human transferrin, which is an iron transporter. Upon binding to
transferrin, TbpA releases the iron in the bacterial periplasm (Rao et al., 1999).
Tbps and iron uptake systems as a whole are essential for bacterial persistence as
the human host maintains the free iron at levels too low to sustain bacterial
survival and growth.
The other outer membrane proteins. As our understanding of Hinf pathogenesis
increases, so does the number of outer membrane proteins that are somehow
involved in Hinf virulence. For example, during the last decade two novel adhesins
were discovered in the Riesbeck lab, namely Protein E (PE) and Protein F (PE).
PE and PF are able to adhere to extracellular matrix proteins as well as to host
cells. Furthermore, PE and PF play in an important role in the subversion of host
innate immunity (Jalalvand and Riesbeck, 2014). In sum, many other outer
membrane proteins exploited by Hinf to colonize its human host may remain
unknown and the relevance of many other known OMPs in Hinf pathogenesis
certainly require further investigations.
viii. The genetic plasticity of Haemophilus influenzae
Hinf is naturally competent. In other words, Hinf is able to take up free DNA from
its environment. The bacterium is transformed, when this DNA uptake leads to
genotype change by either recombination or establishment of a plasmid. Given the
competence of many Hinf strains, it is not a surprise that the number of genes
General Introduction
10
within the Hinf species varies between 1765 and 2355, with the core genome
containing only 1485 genes (Jalalvand and Riesbeck, 2014).
Beside this genetic heterogeneity, all the genes possessed by Hinf are not evenly
expressed, even within a clonal population. Interestingly, some genes are phase
variable. These genes possess short nucleotide repeats in either their coding or
upstream promoter regions. Spontaneous gain or loss of these repeats results in
translational frame shifts, with genes being kept on or turned off. For a phase
variable gene coding for a simple molecule, this will result in expression or not of
the molecule. When the phase variable gene product is involved in the synthesis
of a complex macromolecule, such as LOS, phase variation can result in a vast
array of structures. As an example, the NTHi strain 9274 has as many as 25
different LOS forms (Rahman et al., 1999).
Once expressed, numerous Hinf surface molecules are different from one strain to
another within the immunogenic regions. These differences result from genetic
polymorphisms, including gene point mutations, insertions or deletions. The
ensuing antigenic variations prevent the antibody cross-reaction on heterologous
strains. The OMP P2, for instance, is able to modify its sequence within a clonal
population during the course of persisting infection (St. Geme III, 2002).
Virtually all the virulence factors that were discussed above (in the Subheading 4.)
and other surface-exposed Hinf are highly immunogenic. However, the genetic
and phenotypic variabilities of the bacterium maintain a highly heterogeneous
bacterial population that is able to evade host immune response and persist within
its preferred human niche.
5. CEACAM-recognition by Haemophilus influenzae
The carcinoembryonic antigen-related cell adhesion molecules (CEACAMs) are a
subgroup of heavily glycosylated proteins belonging to immunoglobulin
superfamily. CEACAMs are expressed by a wide range of cells, including
granulocytes, lymphocytes, endothelial and epithelial cells. In addition to their
physiological roles (extensively discussed in the Chapter I), CEACAMs are
targeted by selected bacteria in order to get a grip on the human mucosal surface.
Since CEACAMs were known to be expressed by epithelial cells of the respiratory
General Introduction
11
tract (Virji et al., 1996a) and Hinf attachment to epithelial cells acknowledged as an
irreplaceable step of the bacterial pathogenesis (Rao et al., 1999), Virji and
colleagues had the good idea to screen typeable and non-typeable Hinf isolates
for CEACAM-binding. Rewardingly, they showed that most of the strains tested
were able to interact with CEACAM1 (Virji et al., 2000). Additionally,
unencapsulated Hinf of the biogroup aegyptius was interacting also with CEA (the
protein product of the CEACAM5 gene). The Hinf binding site on the CEACAM1
side was defined to be the non-glycosylated CFG face of its amino-terminal Ig
variable-like domain (Virji et al., 2000). Additional experiments provided hints that
the Hinf CEACAM-binding adhesin was proteinaceous in nature. After deletion of
its outer membrane protein P5, the otherwise CEACAM-binding Hinf Rd strain
failed to recognize recombinant, soluble CEACAM1 but still interacted with
CEACAM1-expressing cells. Therefore, the authors suggested P5 as the
CEACAM1-binding adhesin of Hinf and speculated that an additional ligand may
intervene in cellular context (Hill et al., 2001).
Aiming to test the relevance of Hinf – CEACAM interaction in vivo, Bookwalker and
colleagues used the chinchilla model of Hinf nasopharyngeal colonization. Of note,
chinchilla express an orthologue of human CEACAM1, chinchilla CEACAM1
(cCEACAM1). cCEACAM1 cross-reacted with anti-human CEACAM1 antibody.
Strikingly, pretreatment of the animals with anti-human CEACAM1 antibody
severely impaired the colonization of chinchilla nasopharynx with Hinf, suggesting
the importance of cCEACAM1 for successful colonization (Bookwalter et al.,
2007). Since these authors had previously shown that the Hinf OMP P5 is required
for nasopharyngeal colonization of chinchilla (Sirakova et al., 1994), they deduced
that P5 was interacting with cCEACAM during chinchilla colonization. Therefore,
evidence of direct interaction between Hinf P5 and cCEACAM1 are missing.
Nevertheless, a recent study has demonstrated that CEACAMs are expressed in
non-cancer, human, lung tissues (Klaile et al., 2013). Most importantly, normal
human brochoepithelial cells upregulated their CEACAM1 expression after 24 h of
infection with NTHi isolates. (Klaile et al., 2013). It is unclear whether or not this
effect is restricted to bacteria exclusively colonizing human respiratory airways, as
no negative controls, such as commensal, gut bacteria were included by Klaile and
coworkers (Klaile et al., 2013).
General Introduction
12
6. CEACAMs – a common target for many bacteria
A quarter-century has passed since Leusch and colleagues first reported that
human E. coli isolates were able to bind members of the CEACAM family, namely
CEA and CEACAM6 (Leusch et al., 1990). At present, almost a dozen of bacterial
species have been characterized to target CEACAMs and it is very likely that more
CEACAM-binding bacteria will be uncovered in the future. For better or for worse,
much of the current research on the bacteria interacting with CEACAMs has
focused on the most pathogenic amongst them. Few of them will be briefly
presented in the following lines.
Neisseria gonorrhoeae. Probably the most studied of the CEACAM-binding
bacteria, Neisseria gonorrhoeae (Ngo), the causative agent of the sexually
transmitted disease gonorrhea, binds many members of the CEACAM family via
its “opacity-associated” (Opa) proteins (Virji et al., 1996a; Virji et al., 1996b). The
opa gene is phase variable owing to several pentameric repeats present in its 5’
coding region and each gonococcal strain expresses up to 12 distinct opa genes
(Stern et al., 1986). Noteworthy, each opa gene is independently turned off or on,
and each expressed Opa protein exhibits a unique CEACAM-binding profile
(ranging from no binding to CEACAMs, to multiple interactions with CEACAM1,
CEACAM3, CEA and CEACAM6) (Bos et al., 1997). Thus, the CEACAM-binding
capabilities of a clonal population integrate these genetic and phenotypic
variations. The outcome of the Ngo-CEACAM interaction depends on the
CEACAMs that are engaged. Targeting of epithelial CEACAMs (CEACAM1, CEA
and CEACAM6) provide the bacteria with a solid foothold on the mucosal surface.
Signaling cascades induced upon engagement of epithelial CEACAMs increase
integrin-mediated host cell adhesion to extracellular matrix and block exfoliation of
infected epithelial cells (Muenzner et al., 2010). Whereas the engagement of the
epithelial CEACAMs is beneficial to the bacteria, their recognition by granulocyte
receptor CEACAM3 leads to their opsonin-independent phagocytosis and rapid
elimination (Schmitter et al., 2004). Amazingly, Ngo isolates causing disseminating
diseases seem to have evolved to avoid interaction with CEACAM3 (Roth et al.,
2013). However, it remains puzzling which subtle structural features dictate the
binding of the β-barreled Opa protein to CEACAM1 and not to CEACAM3 (both
General Introduction
13
molecules sharing ~90% amino acid identity in their Opa-binding, N-terminal
domain).
Neisseria meningitidis. Living in the human nasopharynx, as a commensal,
Neisseria meningitidis (Nme) is able to cause life-threatening and fulminant
meningitis or septicaemia (Apicella, 2010). The interaction of Nme with CEACAMs
is also mediated by Opa proteins (Virji et al., 1996b). Similar to gonococcal Opa
proteins, menningococal Opa proteins utilize the surface-exposed hypervariable
(HV) regions 1 and 2 of their loops 2 and 3, respectively, to associate with
CEACAMs. Although highly variable, HV-2 regions of various Nme Opa proteins
have a conserved GxI/V/LxQ motif that is critical for CEACAM engagement (de
Jonge et al., 2003). In clear contrast, this motif is absent in Ngo Opa proteins.
Another remarkable difference between Ngo and Nme is the presence in the latter
of a polysaccharide capsule. The capsule of Nme, masking the CEACAM binding
OMP, has long cast doubt on the relevance of Opa – CEACAM interaction in vivo.
A recent study nicely demonstrated that capsulated Nme strain MC58 colonizes
better and persists longer in the nasopharynx of mice transgenic for human
CEACAM1 compared to wild-type mice (Johswich et al., 2013). Besides, the Nme
strain MC58 was also shown to keep binding to CEACAM1 even in absence of
Opa proteins, suggesting an additional CEACAM-targeting ligand in Nme
(Kuespert et al., 2011). Still, the identity of this adhesin has to be determined.
Moraxella catarrhalis. Another inhabitant of human respiratory tract, Moraxella
catarrhalis (Mca) is one the major causes of otitis media in children and respiratory
diseases in adults (Hill and Virji, 2003). Mca associates to CEACAMs via its
ubiquitous surface protein (Usp) A1, which is a “lollipop”-like autotransporter (Hill
and Virji, 2003). Of the ~90 kDa UspA1 protein, a ~17 kDa polypeptide named rD-
7 and representing 16% of UspA1 is able to interact with CEACAMs (Hill et al.,
2005). Multiple sequence alignment of this CEACAM-targeting portion of UspA1
from CEACAM-binding and non-CEACAM-binding strains of Mca, revealed a
putative CEACAM-binding motif (Brooks et al., 2008) that was refined to a 20-
amino acid stretch by structural studies (Conners et al., 2008). Thus, UspA1
exhibits a conserved CEACAM-binding motif within its coiled-coil stalk (Conners et
al., 2008). Of note, an additional non-CEACAM binding ubiquitous surface protein,
namely UspA2, is expressed by Mca. Upon recombination, UspA1/UspA2 chimeric
General Introduction
14
molecules can be generated. Intriguingly, Hill and colleagues demonstrated that
14% of the CEACAM-binding Mca clinical isolates use an UspA2 variant that has
incorporated the CEACAM-targeting motif of UspA1 (Hill et al., 2012). These
findings underline the flexibility of the bacteria interacting with CEACAMs, gaining
or losing the binding capabilities depending on their specific needs. An in vivo
study snapshotting the transfer of the CEACAM-binding motif from a Mca strain to
another in order to promote bacterial persistence on the mucosal surface is
warmly awaited.
Diffusely adhering Escherichia coli (DAEC). One group of the pathogenic
strains of E.coli interacting with CEACAMs is represented by the diffusely adhering
E. coli (DAEC). The DAEC are associated with both enteric and urinary tract
infections in humans and animals (Servin, 2005). The CEACAM-binding strains of
this group use their Afa/Dr adhesins to target CEACAMs. The Afa/Dr family of
adhesins includes both fimbrial and afimbrial adhesins. Afa adhesins, for instance,
have afimbrial adhesive sheaths encoded by an afa operon containing up to 6
genes (Servin, 2005). The products of these genes are transcriptional regulators,
periplasmic chaperon proteins, anchoring proteins or adhesin subunits. In most of
the cases the subunit AfaE and AfaD build up the adhesive sheath of the Afa
adhesins. However, AfaE adhesin was shown to be necessary and sufficient for
CEACAM-binding (Guignot et al., 2009). On the structural level, AfaE-III, AfaE
subunit of the uropathogenic E. coli strain A30, had its crystal structure solved a
decade ago. AfaE-III consists of an immunoglobulin-like fold missing a central
antiparallel β-strand and possessing an N-terminal extension. During the Afa
adhesin assembly, a periplasmic chaperone (AfaB) provides the missing strand in
a process that aids folding and targets the adhesin subunits to the outer
membrane usher protein for export (Anderson et al., 2004). The subsequent
incorporation of the adhesin subunit into the adhesive sheath on the bacterial
surface requires the N-terminal strand of AfaE-III to attach another adhesin subunit
by taking over the role previously performed by the chaperone in an anti-parallel
arrangement (Anderson et al., 2004). The CEACAM-binding interface, present on
side of AfaE-III, is a 1,446 Å-area encompassing portions of the strands A, B, E
and D (Korotkova et al., 2006). Even if some mutagenesis analyses have been
carried out in an attempt to pinpoint which AfaE-III amino residues are important
for CEACAM-targeting (Korotkova et al., 2006), molecular and structural features
General Introduction
15
explaining the failure of some Afa adhesins (e.g.: AfaE-VIII) to bind CEACAMs are
still a mystery.
In sum, the exploitation by distinct bacteria of structurally unrelated proteins to
target the common human CEACAM receptor is a clear indication of the pivotal
role of CEACAM-targeting in bacterial pathogenesis. The numerous unsolved
questions in this fascinating scientific field warrant further investigations.
Aims of the study
16
Aims of the study
Haemophilus influenzae (Hinf) exhibits an impressing genetic and phenotypic
amenability. However, most of its strains safeguard their interaction with the
human CEACAMs. Therefore, we embarked in this study with the enthusiastic
wish of contributing to the molecular understanding of Hinf pathogenesis,
especially with respect to the exquisite interaction between the postulated adhesin,
the antigenic variable OMP P5, and its human receptor CEACAM. We speculated
that the identification of a conserved epitope sustaining CEACAM-binding may
serve directly as a vaccine candidate or indirectly as rationale for the development
of adhesion-disrupting therapeutics. The fruit of our investigations is organized
here in 4 chapters, as follows:
1) Epithelial CEACAMs highly abundant on the mucosal surface are target of
choice for the incoming bacterial pathogens, including Haemophilus
influenzae. In the first chapter, we review the physiological roles of CEACAMs
before providing an overview on how the bacteria can hijack signaling induced
upon epithelial CEACAM engagement for their own profit. We postulate that
bacterial ligands of CEACAM may be used in the future in order to decrypt
physiological and pathologic processes involving epithelial CEACAMs.
2) Hinf was reported to use the outer membrane protein P5 to target CEACAM1.
The initial aim of this chapter was to determine the CEACAM-binding motif of
OMP P5. However, our initials results cast serious doubt on P5 as the Hinf
CEACAM-binding adhesin. Consequently, we screened a panel of wild-type
and mutants Hinf strains for CEACAM-recognition and identified the outer
membrane protein P1 as the Hinf ligand targeting human CEACAMs. One of
the remaining opened questions of this section was to provide an explanation,
as to why OMP P5 was wrongly considered as CEACAM-ligand. Further, by
expressing OMP P1 of Hinf strains with distinct CEACAM aptitudes in E. coli, it
should be tested whether the E. coli transformants recapitulate the profile of
the respective parent Hinf strain. Lastly, it should be analyzed if the OMP P1
targets exclusively human CEACAMs and if this species-specific recognition of
the receptor is also preserved when OMP 1 is ectopically expressed in E. coli.
Aims of the study
17
3) With the exception of its surface-exposed loops, OMP P1 is highly conserved
within the Pasteurellaceae family. Here, we sought to investigate if the
conserved structure corresponds to preserved CEACAM binding capability for
all the Pasteurellaceae family members. Mutagenesis analyses and molecular
modeling were used in order to pinpoint the molecular determinants of
CEACAM-targeting by OMP P1. Also, extracellular loops restriction
approaches were established in order to rigorously test the contribution of
loops flexibility in efficient interaction of OMP P1 with CEACAMs. Finally,
three-dimensional modeling of the bimolecular complex OMP P1 – CEACAM
should give visual appreciation of the extent to which extracellular loops of
OMP P1 are implicated in CEACAM-targeting.
4) The predicted structure of Hinf OMP P1 is mesmerizingly similar to the solved
structure of the E. coli fatty acid transporter FadL. In this chapter, we aimed at
analyzing the modulation of OMP P1 levels by fatty acids. The modified OMP
P1 levels were tested for any functional relevance with regard to bacterial
binding to recombinant CEACAMs, CEACAM-induced invasion of epithelial
cells and phagocytosis by human granulocytes.
18
CHAPTER I
Signaling by epithelial members of the
CEACAM family – mucosal docking sites
for pathogenic bacteria
Arnaud Kengmo Tchoupa1, Tamara Schumacher1 and Christof R. Hauck1,2
1Lehrstuhl für Zellbiologie, Universität Konstanz, Konstanz, Germany and 2Konstanz Research School Chemical Biology, Universität Konstanz, Konstanz,
Germany
Cell Communication and Signaling 2014: 12:27
Pathogen interaction and signaling by epithelial CEACAMs
19
I.1 Abstract
Carcinoembryonic antigen-related cell adhesion molecules (CEACAMs) comprise
a group of immunoglobulin-related vertebrate glycoproteins. Several family
members, including CEACAM1, CEA, and CEACAM6, are found on epithelial
tissues throughout the human body. As they modulate diverse cellular functions,
their signaling capacity is in the focus of current research. In this review we will
summarize the knowledge about common signaling processes initiated by
epithelial CEACAMs and suggest a model of signal transduction by CEACAM
family members lacking significant cytoplasmic domains. As pathogenic and non-
pathogenic bacteria exploit these receptors during mucosal colonization, we try to
highlight the connection between CEACAMs, microbes, and cellular responses.
Special emphasis in this context is placed on the functional interplay between
CEACAMs and integrins that influences matrix adhesion of epithelial cells. The
cooperation between these two receptor families provides an intriguing example of
the fine tuning of cellular responses and their manipulation by specialized
microorganisms.
I.2 Introduction
The carcinoembryonic antigen-related cell adhesion molecules (CEACAMs), a
subgroup of the CEA family of immunoglobulin-related proteins, are encoded in
the human genome by 12 genes (Kuespert et al., 2006; Zebhauser et al., 2005)
(Figure I-1). All 12 expressed CEACAM genes and a number of derived
pseudogenes cluster on chromosome 19q13 (Kammerer et al., 2007; Teglund et
al., 1994). CEACAMs show distinct expression patterns on different cell types
(Hammarstrom, 1999; Zebhauser et al., 2005). Whereas particular CEACAMs are
only expressed in certain epithelial or myeloid cells, others are found in various
tissues (Gray-Owen and Blumberg, 2006). Some family members play a precise
functional role in particular events such as hearing in the inner ear (CEACAM16)
or phagocytosis of specific bacterial pathogens (CEACAM3) (Schmitter et al.,
2004; Zheng et al., 2011). However, most CEACAMs can be seen as modulators
of general cellular processes such as cell adhesion, differentiation, proliferation,
and survival. To fulfill such diverse functions, CEACAMs have to intersect with
Pathogen interaction and signaling by epithelial CEACAMs
20
other cellular receptors and to transmit signals into the cell. Indeed, signal
transduction mediated by distinct CEACAM family members, which encompass a
cytoplasmic domain, such as CEACAM3 and a splice variant of CEACAM1 with
long cytoplasmic domain, has been studied in great detail (Buntru et al., 2012;
Gray-Owen and Blumberg, 2006). Given the fact that several CEACAMs are GPI-
anchored proteins or that they sustain functionality in the absence of a cytoplasmic
domain, the mechanistic details of signal transduction processes initiated by these
CEACAM family members are still widely unresolved. Interestingly, CEACAMs are
utilized by bacterial pathogens as host receptors on epithelial cells. Similar to
physiological stimulation of CEACAMs, bacteria-initiated clustering of CEACAMs
can induce robust cellular responses including activation of certain kinases,
stimulation of small G proteins, cytoskeletal rearrangements, induction of novel
gene expression events, enhanced cell adhesion, and receptor endocytosis. It has
become clear that CEACAM-binding bacterial pathogens exploit the signaling
capacity of these immunoglobulin superfamily receptors to enhance their chances
of colonizing the mucosal surface. As CEACAM family members without significant
cytoplasmic domains dominate on several epithelial surfaces such as breast, liver,
or prostate (Gaur et al., 2008), we will use this review, to summarize the current
knowledge about the signaling function of these epithelial CEACAMs. By
highlighting recent advances in the understanding of bacteria-induced CEACAM-
mediated processes, we provide a framework for further dissecting the molecular
signaling connections emanating from epithelial members of this family.
Pathogen interaction and signaling by epithelial CEACAMs
21
I.3 Physiological roles of epithelial CEACAMs
Since the discovery of carcinoembryonic antigen (CEA) some 50 years ago (Gold
and Freedman, 1965), and the subsequent appreciation of a family of CEA-related
cell adhesion molecules (Beauchemin et al., 1999) (Figure I-1), numerous
physiological and pathological processes have been associated with these
mammalian membrane glycoproteins. Historically, cancer is one of the disease
states linked to aberrant CEACAM function and the role of epithelial CEACAMs in
tumour progression and metastasis has been summarized in an excellent review
recently (Beauchemin and Arabzadeh, 2013). In particular, human CEACAM1,
CEA, and CEACAM6, which can be found on various epithelial cell types and
derived carcinomas, are thought to shape the interaction between tumour cells
and their stromal counterparts as well as immune cells. Apart from their potential
utilization as clinical biomarkers and promising therapeutic targets in melanoma,
lung, colorectal, and pancreatic cancers, these epithelial CEACAMs are also
implicated in morphogenesis (Kirshner et al., 2003; Yokoyama et al., 2007),
Figure I-1: The human CEACAM family. Schematic depiction of the twelve members of the human carcinoembryonic antigen-related cell adhesion molecules. The red spheres indicate IgV-like domains, the blue spheres indicate IgC2-like domains, which are stabilized by disulfide bonds (S-S). The green spirals indicate transmembrane helices. GPI-anchors are depicted in the form of a green arrow ending in the lipid bilayer. CEACAM20 encodes only a partial IgV-like domain (N*). Graph modified from http://www.carcinoembryonic-antigen.de/.
Pathogen interaction and signaling by epithelial CEACAMs
22
angiogenesis (Ergun et al., 2000; Horst et al., 2006), cell proliferation (Singer et
al., 2010), cell motility (Klaile et al., 2005; Muller et al., 2005), apoptosis (Singer et
al., 2005), regulation of cell matrix attachment (Muenzner et al., 2005; Wong et al.,
2009), as well as epithelial cell-cell interaction and cell polarisation (Benchimol et
al., 1989; Huang et al., 1999). Clearly, forward and reverse genetic approaches in
animal models have suggested that CEACAMs are not essential for all these
processes. For example, mice lacking CEACAM1 are viable and fertile and do not
show gross morphological alterations (Leung et al., 2006). Furthermore,
heterologous expression of human CEACAM1 in the mouse or expression of
additional human epithelial CEACAMs, which are not encoded in the murine
genome (such as CEA and CEACAM6), does not result in perturbation of tissue
architecture or normal tissue homeostasis (Chan and Stanners, 2004; Eades-
Perner et al., 1994; Gu et al., 2010). Therefore, epithelial CEACAMs seem to
contribute to the fine-tuning of cellular behaviour and their contribution might
become critical during stressful conditions, such as tissue damage and repair,
which are not readily obvious in laboratory kept animals.
Most studies of CEACAM-initiated signal transduction have focused on CEACAM1
in immune cells and transformed epithelial cells (nicely summarized in
(Beauchemin and Arabzadeh, 2013; Gray-Owen and Blumberg, 2006)).
Investigations into CEACAM1 structure and function have also profited from the
fact that this family member is expressed in different cell types and that CEACAM1
orthologs exist in other mammalian species (Kammerer and Zimmermann, 2010).
Due to differential splicing, human CEACAM1 occurs in 11 isoforms with the
number of extracellular Ig domains ranging from one to four (see the CEA
homepage at http://www.carcinoembryonic-antigen.de/index.html; (Barnett et al.,
1993)). The major isoforms in human cells are CEACAM1-4 and CEACAM1-3,
which possess an extracellular amino-terminal IgV-like domain, followed by three
(A1, B, A2) or two (A1, B) IgC2-like domains, respectively. Similarly, in other
epithelial CEACAMs, such as CEA or CEACAM6, up to six extracellular IgC2-like
domains follow the amino-terminal IgV-like domain (Figure I-1). Accordingly,
engagement of the extracellular domains of epithelial CEACAMs serves as the
primary stimulus for CEACAM-mediated transmembrane signaling. Under
physiologic conditions, homophilic interactions between CEACAMs on opposing
cells are thought to be the major trigger of CEACAM-initiated signaling processes,
Pathogen interaction and signaling by epithelial CEACAMs
23
although CEACAMs can also engage in heterophilic interactions, e.g. with
selectins (Obrink, 1997).
I.4 Role of CEACAM extracellular domains in mediating cis- and
trans-oligomerization
Trans-oligomerization resulting from homophilic interactions between the amino-
terminal IgV-like domains of CEACAMs on neighbouring epithelial cells is the basis
of CEACAM-mediated cell-cell adhesion (Teixeira et al., 1994; Watt et al., 2001;
Wikstrom et al., 1996; Zhou et al., 1993a). However, it has become clear that this
homophilic type of trans-oligomerization is further supported by the presence of
IgC2-like domains (Cheung et al., 1993; Zhou et al., 1993a). In a tissue context,
these additional extracellular Ig domains might allow these receptors to extend
farther from the membrane surface to facilitate binding, but they might also be
directly involved in homophilic trans-interactions (Zhou et al., 1993a; Zhou et al.,
1993b). Moreover, recent electron tomography studies of soluble and membrane-
attached CEACAM1 ectodomains have not only confirmed the critical role of the
IgV-like amino-terminal domain for trans-oligomerisation, but also pointed to
additional cis-interactions in the extracellular part of CEACAM1 (Klaile et al.,
2009). Indeed, the extracelluar chain of Ig domains in CEACAM1 appears to be
rather flexible, but can be stabilized by cis-interactions between either IgV-like
domains or IgC2-like domains of parallel CEACAM1 molecules in the same
membrane plane (Klaile et al., 2009). As a consequence, CEACAMs might occur
in different oligomerization states, partially dictated by the occurrence of trans- or
cis-interactions between their extracellular domains. At least in the case of
CEACAM1, these different oligomerization states clearly have an influence on its
signaling function (Muller et al., 2009). In one of the following sections, it will
become clear that the issue of CEACAM1 oligomerization is even more complex,
as the transmembrane domain of this receptor also sustains cis-interactions,
presumably depending on the lipid context.
Pathogen interaction and signaling by epithelial CEACAMs
24
I.5 Signaling by epithelial CEACAMs
As transmembrane signaling requires a connection to the cytosol, the
transmembrane domain containing CEACAM1 has been the focus of a multitude
of studies (Beauchemin and Arabzadeh, 2013; Gray-Owen and Blumberg, 2006).
Indeed, CEACAM1 harbors a cytoplasmic domain, which can either be long (L; 71
amino acids in humans) or short (S; 10 amino acids). The “L” isoforms encompass
a functional immunoreceptor tyrosine-based inhibitory motif (ITIM) and both
CEACAM1-L and CEACAM1-S isoforms are often co-expressed in the same cell,
with expression ratios varying between different cell types and between different
cellular states (Singer et al., 2010; Singer et al., 2000). In many cases, expression
of the short isoform interferes with CEACAM1-L generated signals (Muller et al.,
2009; Turbide et al., 1997). Therefore, the signal transduction role of CEACAM1
has been mostly attributed to the CEACAM1-L isoform and its cytoplasmic
domain. Indeed, CEACAM1-L can interact with cytoplasmic protein tyrosine
kinases and protein tyrosine phosphatases, as well as with calmodulin, β-catenin,
actin, filamin, shc, and tropomyosin (for review see (Beauchemin and Arabzadeh,
2013)). Only few of these interactions are sustained by the short cytoplasmic
domain of CEACAM1-4S. However, investigations of transformed mammary
epithelial cells (MCF7 cells) grown in a 3D-matrigel environment have suggested
that CEACAM1-4S can induce lumen formation in these carcinoma cells resulting
in acinar-like structures (Kirshner et al., 2003). In follow up studies, the effect of
CEACAM1-4S was pinpointed to binding interactions of the short cytoplasmic
domain. In particular, in CEACAM1-4S the membrane-proximal phenylalanine
F454 or lysine K456 residues (-HFGKTGSSGPLQ), respectively, interact with
cytoskeletal components and T457 (-HFGKTGSSGPLQ) is phosphorylated (Chen
et al., 2007). Furthermore, MCF7 cells injected together with human fibroblasts in
the fat pad of mice show a more normal phenotype (with lumen formation), when
CEACAM1 is stably expressed in these cells (Li et al., 2009). In this situation, both
CEACAM1-4S and CEACAM1-4L are able to induce lumen formation and gland
development in the xenograft (Samineni et al., 2011). Therefore, despite major
differences in their cytoplasmic sequences and their distinct profiles of protein-
protein interactions, both CEACAM1-4L as well as CEACAM1-4S appear to
modulate the growth behaviour of epithelial cells in a similar manner. These
findings imply that they can transmit at least some overlapping signals into the
Pathogen interaction and signaling by epithelial CEACAMs
25
cells. Indeed, phosphorylation of the membrane proximal threonine residue
(T457), present in the cytoplasmic domains of CEACAM1-4S and CEACAM1-4L,
by calmodulin kinase IID (CaMKIID) is the critical event required for CEACAM1-
driven lumen formation in transformed breast epithelial cells (Nguyen et al., 2014).
A similar contribution of CEACAM1 to morphogenesis has now been reported in
3D cultures of prostate cells (Zhang et al., 2013). The primary human prostate
cells formed organoids with a lumen and small tubular outgrowth, which was
inhibited, when anti-CEACAM1 antibodies were added to the cultures or when
CEACAM1 expression was reduced by about 50% with antisense oligonucleotides
(Zhang et al., 2013). As these cells express both CEACAM1 isoforms, with either
short or long cytoplasmic domain, it is unclear if one or both proteins are
responsible for the phenotype. Prostate epithelial cells express an additional
member of the CEACAM family, CEACAM20, which is found together with
CEACAM1 on the luminal surface of normal prostate glands. Again, antisense
oligonucleotides against CEACAM20 reduced tubule outgrowth (Zhang et al.,
2013). Clearly, CEACAM20 has a cytoplasmic domain sequence distinct from
CEACAM1. Even more striking, CEACAM20 lacks a complete IgV-like amino-
terminal domain, which is instrumental in CEACAM1 for homophilic interactions
between CEACAM1 on neighbouring cells. Together, these recent insights point to
functional commonalities between epithelial CEACAM family members, which
show striking sequence divergence in their amino-terminal IgV-like domain or their
cytoplasmic sequences.
One important implication arising from these results is the realization that signaling
by epithelial CEACAMs could involve parts of these receptors other than the
cytoplasmic domain or the amino-terminal IgV-like domain, such as the
transmembrane or additional extracellular domains. Indeed, recent experiments
employing either carcinoma cell lines or using bacterial pathogens as CEACAM
ligands have pointed into this unexpected direction.
Pathogen interaction and signaling by epithelial CEACAMs
26
I.6 CEACAM1 cis-oligomerization sustained by the
transmembrane domain
A long standing observation in the field is the reduced expression of CEACAM1
that accompanies transformation of epithelial cells from different tissues
(Beauchemin and Arabzadeh, 2013), including the transition from hepatocytes to
hepatoma cells. It is therefore not surprising, that re-expression of CEACAM1-4L
in rat hepatocellular carcinoma cells results in growth suppression in vitro and
reduced tumour formation in vivo (Laurie et al., 2005). In contrast, expression of
CEACAM1-4S in an anchorage-dependent hepatocellular carcinoma cell line
promoted robust growth of the cells in soft-agar, suggesting that CEACAM1-4S-
initiated signaling rendered the cells anchorage-independent (Lawson et al.,
2012). Strikingly, this effect could be abolished by mutations in the transmembrane
domain. In particular, point mutations disrupting a membrane-integral GxxxG motif
resulted in the loss of the anchorage-independent growth promoting properties of
CEACAM1-4S. As GxxxG motifs in α-helical domains are known to support helix-
helix interactions, it was proposed that such mutations might disrupt cis-dimer
formation of CEACAM1. Recent biochemical approaches based on chemical
crosslinking support the idea that CEACAM1 oligomerizes laterally via the
transmembrane domain to sustain downstream function (Patel et al., 2013).
Together, these results indicate that the transmembrane domain of CEACAM1
promotes clustering and oligomerization of the receptor as a pre-requisite for
signaling into the cell (Figure I-2).
Pathogen interaction and signaling by epithelial CEACAMs
27
I.7 CEACAM-binding bacteria reveal the lipid raft association of
their receptors
Further insight into CEACAM signaling connections has been gained by the use of
bacterial pathogens as selective and multivalent stimuli of these receptors. Over
the last two decades, diverse CEACAM-binding pathogens including pathogenic
Escherichia coli strains, Neisseria gonorrhoeae, Neisseria meningitidis,
Haemophilus influenzae, and Moraxella catarrhalis, have been found to bind to
CEACAM1 or other epithelial CEACAMs such as CEA and CEACAM6 (Chen and
Gotschlich, 1996; Chen et al., 1997; Hill and Virji, 2003; Leusch et al., 1991; Virji et
al., 2000; Virji et al., 1996a). In an intriguing example of convergent evolution,
these bacteria employ structurally distinct adhesive surface proteins (adhesins) to
Figure I-2: Signaling initiated by epithelial CEACAMs. Schematic summary of recent findingswith regard to CEACAM-initiated signaling events in epithelial cells. Upon ligand binding,CEACAM1 forms oligomers supported by cis-interactions between the extracellular and thetransmembrane domains (1) and is recruited to membrane microdomains (2). GPI-anchoredepithelial CEACAMs, such as CEA or CEACAM6, constitutively localize to membranemicrodomains (3). In membrane microdomains, epithelial CEACAMs connect to putative co-receptor(s) (black) via extracellular IgC2-like domains (4). Intracellular signaling triggered byepithelial CEACAMs either directly or indirectly via co-receptor(s) leads to phosphatidylinositol-3’-kinase dependent signaling processes connected to receptor-mediated endocytosis (5).Furthermore, stimulation of epithelial CEACAMs triggers novel gene expression events, e.g. denovo expression of CD105, which extracts zyxin from basal integrin-rich focal adhesion sites (6),resulting in increased integrin activity and enhanced binding to the basal extracellular matrix (ECM)(7).
Pathogen interaction and signaling by epithelial CEACAMs
28
connect to the same group of human receptors (Table 1). As CEACAM1, CEA,
and CEACAM6 are exposed on the apical membrane of mucosal cells, they
provide an accessible handle for incoming bacteria (for review see (Kuespert et
al., 2006)). Indeed, all CEACAM-binding pathogenic bacteria characterized so far
exploit the human mucosa as a platform for colonization, multiplication and further
spread (Virji, 2009) . In addition to mere binding to host cells, CEACAM
engagement triggers endocytosis of the bacteria into epithelial cells and
transcytosis of microorganisms through intact epithelial layers (Gray-Owen et al.,
1997; Virji et al., 1996a; Wang et al., 1998). In this respect, it has been reported
before that GPI-anchored CEA and CEACAM6 as well as CEACAM1 initiate a
characteristic uptake pathway that is distinct from phagocytosis mediated by the
granulocyte receptor CEACAM3 (McCaw et al., 2004; Schmitter et al., 2007). Due
to its exceptional phagocytosis promoting properties, CEACAM3-initiated signaling
has been studied in great detail (for review see (Buntru et al., 2012)). In contrast to
epithelial CEACAMs, CEACAM3-initiated uptake of bacteria critically relies on a
cytoplasmic sequence motif and involves extensive actin cytoskeleton
rearrangements orchestrated by the small GTPase Rac and its effector protein
WAVE2 (Pils et al., 2012; Schmitter et al., 2004). Importantly, CEACAM3-
mediated phagocytosis is independent of sphingolipid- and cholesterol-rich
membrane microdomains, as cholesterol chelators do not interfere with this
process (Guignot et al., 2009; Schmitter et al., 2007). This is strikingly different for
epithelial CEACAMs, where internalization of bacteria is sensitive against
cholesterol depletion (Korotkova et al., 2008; Schmitter et al., 2007). Therefore, in
addition to receptor dimerization and oligomerization, it appears that signaling
initiated by epithelial CEACAMs also requires the proper lipid environment in the
membrane. For GPI-linked CEA and CEACAM6 it is known for some time that
these glycoproteins localize to detergent-resistant membrane fractions (Screaton
et al., 2000). In this regard, the GPI anchor of CEA is sufficient to localize proteins
to membrane microdomains (Nicholson and Stanners, 2006). Transmembrane
CEACAM1 has also been found in detergent-resistant membrane microdomains in
epithelial and endothelial cells (Muenzner et al., 2008; Muller et al., 2005). In
contrast to GPI-anchored CEACAMs, which constitutively localize to the detergent-
resistant membrane fraction, CEACAM1 is only found in membrane microdomains
upon receptor clustering (Muenzner et al., 2008). This suggests an additional layer
Pathogen interaction and signaling by epithelial CEACAMs
29
of regulation, which drives this receptor into specific membrane regions upon
receptor engagement. As mutations in the transmembrane, but not the
cytoplasmic domain of CEACAM1 affect localization in detergent-resistant
membrane fractions (Muenzner et al., 2008), it is tempting to speculate that the
receptor oligomerization function of the CEACAM1 transmembrane domain directs
this receptor into membrane microdomains (Figure I-2). Together, epithelial
CEACAMs require a specific lipid environment in the plasma membrane for proper
function, where GPI-anchored CEACAMs constitutively localize and where
CEACAM1 can be recruited to upon receptor oligomerization.
Bacterial species Primary target tissue
Adhesin CEACAM specificity References
1 3 5 6
Adherent-invasive Escherichia coli (AIEC)
Digestive tract FimH - NC - + (Barnich et al., 2007)
Diffusely adhering Escherichia coli (DAEC)
Digestive tract Urogenital tract
Afa/Dr-I + - + + (Berger et al., 2004)
Haemophilus influenzae Nasopharynx OmpP5 + NC + NC (Hill et al., 2001; Virji et al., 2000)
Moraxella catarrhalis Nasopharynx UspA1 + + + NC (Brooks et al., 2008; Hill and Virji, 2003)
Neisseria gonorrhoeae Urogenital tract OpaCEA + + + + (Chen et al., 1997; Virji et al., 1996b)
Neisseria meningitidis Nasopharynx OpaCEA + + + + (Virji et al., 1996a)
commensal Neisseria Nasopharynx OpaCEA + NC NC NC (Toleman et al., 2001)
Salmonella sp Digestive tract Uncharacterized fimbrial adhesin
+ NC + + (Leusch et al., 1991)
NC: not yet characterized
I.8 CEACAM1 signaling initiated by the IgC2-like extracellular
domains
Though the localization in membrane microdomains is shared by epithelial
CEACAMs this does not provide a direct explanation for their signaling capacity.
Again, CEACAM-mediated internalization of bacterial pathogens provided novel
insight into how epithelial CEACAMs might be mechanistically connected to
intracellular signaling pathways. In numerous endocytic processes,
phosphatidylinositol phosphates (PIPs) play an important role (Araki et al., 1996;
Table 1: CEACAM-binding bacteria and their adhesive proteins
Pathogen interaction and signaling by epithelial CEACAMs
30
Indik et al., 1995). Therefore, the observation that CEACAM3–mediated
internalization is not blocked by wortmannin, an inhibitor of phosphatidylinositol-3’
kinase (PI3K), was particularly striking (Buntru et al., 2011). This surprising finding
with regard to CEACAM3 prompted an investigation of PI3K and PIPs in bacterial
internalization via epithelial CEACAMs. Interestingly, in CEACAM1-expressing
cells, a strong accumulation of phosphatidylinositol 3’,4’,5’-trisphosphate
(PI3,4,5P) was observed around bacterial uptake sites (Voges et al., 2012).
Furthermore, overexpression of class I PI3K increased bacterial uptake, whereas
wortmannin blocked CEACAM1-, CEA-, and CEACAM6-mediated internalisation.
Expression of the 5’-phosphate-directed PIP phosphatase SHIP (SH2 domain-
containing inositol phosphatase), which dephosphorylates PI3,4,5P, reduces
CEACAM1-mediated internalization. Intriguingly, PI3K-dependent endocytosis via
CEACAM1 was not linked to cytoplasmic determinants of the receptor, but rather
required the extracellular IgC2-like domains of CEACAM1 (Voges et al., 2012).
Accordingly, expression of CEACAM1 mutants lacking either one or all IgC2-like
domains resulted in lower numbers of endocytosed bacteria in comparison to
wildtype CEACAM1 despite a similar binding of the microorganisms to the
truncated receptor. It is interesting to note, that inhibition of PI3K by wortmannin
did not interfere with the re-location of CEACAM1 to membrane microdomains,
suggesting that PI3K signaling is downstream of receptor oligomerization and
membrane microdomain association of the receptor. A plausible explanation would
be that the IgC2 domains of CEACAM1 connect bacteria-bound CEACAM1,
presumably via the extracellular part of a membrane-microdomain located
receptor, with PI3K signaling inside cells (Figure I-2).
It is interesting to note that the IgC2 domains of CEACAM1 orthologs from human,
cattle, mouse and rat show a higher degree of sequence conservation than the
amino-terminal IgV-like domain (Kammerer et al., 2004; Kammerer and
Zimmermann, 2010). The lower sequence conservation in the amino-terminal IgV-
like domain compared to the IgC2-like domains has always been interpreted as a
sign of positive selection for sequence variation in the amino-terminal domain.
However, together with the loss of function upon deletion of the IgC2 domains, the
relative conservation of the IgC2 domains of epithelial CEACAMs may reflect
conserved functions and therefore evolutionary constraints on this region.
Importantly, whereas all isoforms of CEACAM1, CEA, and CEACAM6 encompass
Pathogen interaction and signaling by epithelial CEACAMs
31
at least one IgC2-like extracellular domain, CEACAM3 lacks such an extracellular
domain. The absence of an IgC2-like extracellular domain in CEACAM3 correlates
well with the mechanistically distinct endocytosis mediated by CEACAM3 in
comparison to epithelial CEACAMs. Altogether, it is very tempting to speculate
that engagement of epithelial CEACAMs will promote the association of their
extracelluar IgC2 domain(s) with not yet identified co-receptor(s), which in turn
transmit the PI3K activating signal into the cell (Figure I-2). This model would also
explain why CEACAMs with differences in the amino-terminal and the cytoplasmic
domain (such as CEACAM1 and CEACAM20) can promote similar cellular
responses as discussed above for prostate morphogenesis. Such a common co-
receptor for multiple CEACAMs might also be located in membrane microdomains,
where CEACAM1 re-locates upon oligomerization and where GPI-anchored
CEACAMs constitutively localize. The identification of this putative co-receptor
might be the turning point in the quest to completely understand the fascinating
physiology of epithelial CEACAMs.
I.9 CEACAM cooperation with integrins and other membrane
receptors
Several cellular receptors have already been proposed to act as co-receptors for
CEA or to co-operate with epithelial CEACAMs (Li et al., 2010; Samara et al.,
2007; Slevogt et al., 2008). For example, in lung epithelial cells, CEACAM1 has
been shown to co-immunoprecipitate with Toll-like receptor 2 (TLR2) and bacterial
engagement of CEACAM1 has been suggested to interfere with TLR2-induced
pro-inflammatory responses (Slevogt et al., 2008). However, as the Moraxella
catarrhalis strain O35E employed in these studies does not bind to any CEACAM
(Brooks et al., 2008), it is unclear, how CEACAM-initiated responses are triggered
in this context.
In several studies, it has been observed that CEACAM stimulation has a positive
effect on cell-matrix adhesion of epithelial cells as well as on integrin-mediated
cell-cell adhesion in leukocytes (Lobo et al., 2009; Muenzner et al., 2005; Muller et
al., 2005). In the case of CEACAM1, a phosphorylation-dependent interaction with
integrin β3 has been reported (Brummer et al., 2001) and CEACAM1 colocalizes
with integrin β1 in MCF7 cells grown in Matrigel (Kirshner et al., 2004) suggesting
Pathogen interaction and signaling by epithelial CEACAMs
32
that CEACAMs functionally interact with integrins. Since ligand-bound integrins
locally organize membrane microdomains, they could constitute a co-receptor for
epithelial CEACAMs (del Pozo et al., 2004; Hoffmann et al., 2010). Indeed, the
observed functional co-operation was suggested to result from co-clustering of
GPI-linked CEACAMs together with integrins in these membrane areas (Ordonez
et al., 2007). A co-operation between CEACAMs and integrins would nicely
explain the modulation of cellular functions such as cell adhesion and cell survival
in the absence of matrix-attachment (Camacho-Leal and Stanners, 2008;
Camacho-Leal et al., 2007). However, biochemical evidence for a close physical
interaction between CEA or CEACAM6 and integrins is lacking. Furthermore,
CEACAMs localize to lateral cell-cell contacts or the apical membrane
compartment in polarized cells, whereas ligand-bound integrins cluster at basal
cell-matrix adhesion sites. The seeming contradiction between functional
cooperation and distinct subcellular localization of epithelial CEACAMs and
integrins has been nicely resolved. Using CEACAM-binding bacteria as a naturally
occurring, highly selective and multivalent ligand for CEACAM1, CEA, and
CEACAM6, an unbiased gene-expression analysis revealed a number of genes,
which are specifically induced following CEACAM stimulation in epithelial cells
(Muenzner et al., 2005). Further analysis showed that upregulation of a member of
the TGF-β1 receptor family, termed endoglin or CD105, is observed upon
stimulation of GPI-anchored CEACAMs or stimulation of a CEACAM1 mutant
lacking the complete cytoplasmic domain (Muenzner et al., 2005). In all these
cases, CEACAM engagement by bacteria results in an elevated CD105 mRNA
level, which is observed within 1–3 hours after bacterial infection (Muenzner et al.,
2005). In a similar time frame, the infected epithelial cells display enhanced
integrin-mediated adhesion to the extracellular matrix and CD105 expression is
necessary and sufficient for this phenotype (Muenzner et al., 2005). CD105
expression in turn does not alter the amount of integrins on the cells, but initiates
the redistribution of the focal adhesion protein zyxin. Indeed, zyxin binds with high
affinity to the cytoplasmic domain of CD105, and disappears from integrin-rich
focal adhesion sites as soon as CD105 is expressed in epithelial cells (Muenzner
et al., 2010) (Figure I-2). Due to the lack of zyxin at focal adhesions, integrin
activity, and therefore extracellular matrix (ECM) binding of the infected cells,
increases over the course of several hours following contact with CEACAM-
Pathogen interaction and signaling by epithelial CEACAMs
33
binding bacteria. Increased integrin activity and strengthened ECM-binding is also
observed in zyxin-deficient or CD105-overexpressing cells, suggesting that
CEACAM-binding bacteria exploit physiological regulators of cell adhesion to
indirectly manipulate integrin activity (Muenzner et al., 2010) (Figure I-2). This
functional interplay between CEACAM stimulation, CD105 expression and its
effect on focal adhesion site composition provide a plausible scenario, how
CEACAMs can modulate integrin-mediated cell adhesion even without directly
associating with integrins. However, it should be noted that several CEACAM-
binding bacteria also possess surface adhesins, which associate with extracellular
matrix (ECM) proteins of their host, such as fibronectin or vitronectin (Brooks et al.,
2008; Hill and Virji, 2003; Virji, 2009). In this manner, ECM protein binding could
allow such bacteria to simultaneously engage integrins and CEACAMs once the
integrity of the epithelial barrier and the spatial separation of CEACAMs and
integrins might be compromised. If such a potential co-stimulation of integrins and
CEACAMs by pathogenic microbes has consequences for the outcome of
bacteria-host interaction has not been explored so far.
Nevertheless, already the indirect connection between CEACAMs and integrins
must be advantageous for bacteria trying to get a foothold on the mucosal surface,
given the fact that so many unrelated microbes target CEACAMs (Table 1).
Indeed, this functional connection allows bacteria to engage receptors on the
apical side of the epithelium, while ultimately impacting on the activity of integrins,
which are located on the basolateral side of polarized epithelial cells. In the case
of CEACAM-binding Neisseria gonorrhoeae, which infects the urogenital tract, it
has been observed that the enhanced matrix binding of the infected epithelial cells
strongly reduces the exfoliation of the superficial mucosal cell layer (Muenzner et
al., 2010). Suppressing CD105 upregulation or inhibiting the zyxin-CD105
interaction in the urogenital tract of CEA-transgenic mice allows exfoliation to
proceed despite the presence of CEACAM-binding bacteria, providing
experimental proof that CEACAM engagement is instrumental for successful
colonization of the mucosal surface (Muenzner et al., 2010). Further examples
have emerged, which demonstrate that colonization of the nasopharyngeal
mucosa by Neisseria meningitidis or Moraxella catarrhalis profits from the
presence of epithelial CEACAMs (Bookwalter et al., 2007; Johswich et al., 2013).
In the case of N. meningitidis, bacteria are not detected in wildtype mice three
Pathogen interaction and signaling by epithelial CEACAMs
34
days after inoculation, whereas the same bacterial strain is present for up to a
week in the nasopharynx of CEACAM1-transgenic mice (Johswich et al., 2013). It
is currently unclear, if suppression of epithelial exfoliation, CEACAM-integrin
cooperation, or other forms of CEACAM-initiated cellular responses are involved in
nasopharyngeal colonization. However, these examples again demonstrate that
epithelial CEACAMs, either with or without cytoplasmic domain, can orchestrate
signaling events in epithelial cells. Furthermore, they also showcase that much of
CEACAM functionality can be learned by using CEACAM-binding bacteria, such
as N. gonorrhoeae, as selective and potent stimuli.
I.10 Conclusions
Over the last decade, CEACAMs have emerged as important modulators of
signaling events in leukocytes, endothelial, and epithelial cells. The simultaneous
expression of multiple CEACAM family members by most human epithelial tissues,
including GPI-anchored and transmembrane forms in different splice variants, has
hampered progress in deciphering the molecular signaling connections initiated by
CEACAM-mediated cell-cell interactions. To understand the contribution of
CEACAMs in these processes, well characterized antibodies have been employed
to interfere with CEACAM-CEACAM interactions, but due to sterical hindrance
such approaches might also block a number of other cell-cell interactions. The use
of CEACAM-binding bacteria as multivalent, high affinity ligands for a number of
epithelial CEACAMs has provided an additional opportunity to selectively trigger
CEACAM signaling in vitro and in vivo. These natural probes allow the
visualization of local CEACAM-initiated signaling complexes as well as signaling
intermediates and, therefore, have provided novel insight. Combining these
different approaches will further help to refine our understanding of epithelial
CEACAM physiology and of the involved molecular and cellular processes.
35
CHAPTER II
Outer membrane protein P1 is the
CEACAM-binding adhesin of
Haemophilus influenzae
Arnaud Kengmo Tchoupa1, Sabine Lichtenegger2, Joachim Reidl2 and Christof
R. Hauck1,3
1Lehrstuhl für Zellbiologie, Universität Konstanz, Konstanz, Germany
2Institute of Molecular Biosciences, University of Graz, Graz, Austria
3Konstanz Research School Chemical Biology, Universität Konstanz, Konstanz,
Germany
Mol Microbiol. 2015 Jul 16. doi: 10.1111/mmi.13134. [Epub ahead of print]
Hinf binds human CEACAMs via P1 to colonize mucosa
36
II.1 Summary
Haemophilus influenzae is a Gram-negative pathogen colonizing the upper
respiratory tract mucosa. H. influenzae is one of several human-restricted bacteria,
which bind to carcinoembryonic antigen related cell adhesion molecules
(CEACAMs) on the epithelium leading to bacterial uptake by the eukaryotic cells.
Adhesion to CEACAMs is thought to be mediated by the H. influenzae outer
membrane protein (OMP) P5. However, CEACAMs still strongly bound to H.
influenzae lacking OMP P5 expression, and soluble CEACAM receptor
ectodomains failed to bind to OMP P5, when heterologously expressed in E. coli.
Screening of a panel of H. influenzae OMP mutants revealed that lack of OMP P1
completely abrogated CEACAM binding and supressed CEACAM-mediated
engulfment of H. influenzae by epithelial cells. Moreover, ectopic expression of
OMP P1 in E. coli was sufficient to induce CEACAM-binding and to promote
attachment to and internalization into CEACAM-expressing cells. Interestingly,
OMP P1 selectively recognizes human CEACAMs, but not homologs from other
mammals and this binding preference is preserved upon expression in E. coli.
Together, our data identify OMP P1 as the bona-fide CEACAM-binding invasin of
H. influenzae. This is the first report providing evidence for an involvement of the
major outer membrane protein P1 of H. influenzae in pathogenesis.
II.2 Introduction
Haemophilus influenzae is a human-restricted Gram-negative coccobacillus that
colonizes asymptomatically the human respiratory tract of healthy individuals. The
shift from commensal to pathogen happens when the bacterium reaches privileged
anatomical sites under various predisposing conditions, such as age, viral
infections, or a constant exposure to pollutants. Haemophilus infection can result
in otitis media, sinusitis, conjunctivitis, respiratory infections, septicaemia, or
meningitis (Foxwell et al., 1998; Rao et al., 1999). The capsulate members of this
species (typeable Hinf) are grouped into six serotypes (a-f) depending on the
composition of their capsular polysaccharides, with serotype b strains causing
invasive infections in infants and children (Agrawal and Murphy, 2011). These
strains may switch off capsule expression in vivo and in vitro (Hoiseth and
Hinf binds human CEACAMs via P1 to colonize mucosa
37
Gilsdorf, 1988), whereas the truly acapsulate strains lack the genetic information
for capsulation (non-typeable H. influenzae, NTHi). During the last 2 decades, the
widespread use of H. influenzae type b (Hib) conjugate vaccines has nearly
eradicated invasive disease by Hib in countries where the vaccines are used
widely (Agrawal and Murphy, 2011). In contrast, NTHi remains one of the most
common causes of otitis media, while also playing major roles in sinusitis and
conjunctivitis (Murphy et al., 2009). Furthermore, chronic obstructive pulmonary
disease (COPD), the fourth leading cause of death worldwide, can be exacerbated
by bacterial pathogens with NTHi being the most common (Roier et al., 2012).
Besides, the non-capsulate H. influenzae biogroup aegyptius (H. influenzae-aeg)
is an important cause of self-limiting conjunctivitis and Brazilian purpuric fever, a
fulminant systemic disease observed in children (Harrison et al., 2008).
The first step of in H. influenzae pathogenesis is the colonization of the
nasopharynx, which requires that the microbe overcomes host clearance
mechanisms, including the mucociliary escalator and local immunity (St. Geme III,
2002). Similar to other pathogenic bacteria, H. influenzae can express multiple
adhesins, which mediate attachment to host target tissues (Abraham et al., 1998;
Finlay and Falkow, 1997). One group of host receptors recognized by both typable
H. influenzae and NTHi belongs to the carcinoembryonic antigen-related cell
adhesion molecules (CEACAMs), a subgroup of the immunoglobulin superfamily
glycoproteins present on the apical side of nasopharyngeal epithelial cells (Virji et
al., 2000). It has been reported earlier, that the outer membrane protein (OMP) P5
of H. influenzae is the bacterial adhesin responsible for CEACAM binding (Hill et
al., 2001). However, P5 shows large antigenic variation between different strains
and it is unclear, which structural feature of this OMP mediates the interaction with
host CEACAMs.
In this study, we report that the OMP P5 of typeable H. influenzae or NTHi fails to
promote CEACAM binding. Instead, we identify H. influenzae OMP P1, a
homologue of E. coli FadL, as the CEACAM-binding adhesin. Deletion of P1 in H.
influenzae as well as heterologous expression in E. coli demonstrates that P1 is
necessary and sufficient to bind several human CEACAMs. Concordantly, when
expressed on the surface of E. coli, P1 mediates adhesion to and invasion into
epithelial cells. By identifying the bona fide CEACAM-binding adhesin of typeable
Hinf binds human CEACAMs via P1 to colonize mucosa
38
and non-typeable H. influenzae our study opens up novel avenues in designing
preventive strategies against these human pathogens.
II.3 Results
Haemophilus influenzae strains display distinct CEACAM binding profiles
More than a decade after the initial discovery, that typeable H. influenzae and
NTHi strains bind to the amino-terminal immunoglobulin variable (Igv)-like domain
of human CEACAM1 (Virji et al., 2000), the structural details of this interaction are
still unknown. Given the fact that the proposed CEACAM-binding adhesin, the
OMP P5, is highly variable, we wondered, whether this variation dictates binding to
selected human CEACAM family members. Accordingly, fifteen different H.
influenzae strains were employed in bacterial pull-down assays with soluble Igv-
like amino-terminal domains of human CEACAM1, CEACAM3, CEACAM4,
CEACAM5, and CEACAM6 expressed as GFP-fusion proteins (Figure II-1A and
B). Four different CEACAM-binding profiles were observed for the various strains:
14 out of 15 H. influenzae strains recognized CEACAM1, about a quarter of the
strains recognized one additional CEACAM family member, namely CEA, and one
strain of H. influenzae biogroup aegyptius (ATCC 11116) was able to bind
CEACAM1, CEA, and CEACAM3 (Figure II-1B and C). None of the H. influenzae
strains bound to CEACAM4, CEACAM6, or CEACAM8, which share a sequence
identity of 45, 90, or 72%, respectively, with human CEACAM1 in their N-terminal
domains (Figure II-1B and C).
Importantly, one NTHi isolate (strain 7502) completely lacked CEACAM binding
and even prolonged exposure did not reveal any positive signals (Figure II-1B).
Our recombinant CEACAM proteins comprised the Igv-like amino-terminal
domains only. However, the IgC2-like domains of CEACAM1 (A1, B or A2) were
suggested to contribute to the interaction of H. influenzae with CEACAM1 (Virji et
al., 2000). Therefore, we tested whether NTHi 7502 would be able to bind a
recombinant CEACAM1 protein comprising all extracellular domains (hCEA1-
NA1BA2; Figure II-1A).
Hinf binds human CEACAMs via P1 to colonize mucosa
39
Figure II-1. Differential recognition of human CEACAMs by various H. influenzae strains. A. Schematic drawing of human CEACAM1, the recombinant CEACAM1 amino-terminal Igv-like domain fused to GFP (hCEA1-N), the soluble CEACAM1 containing 4 extracellular domains (hCEA1-NA1BA2) and the recombinant chimera consisting of N-terminal Igv-like domain of CEACAM8 followed by 3 IgC2-like domains of CEACAM1 (hCEA8/1-NA1BA2). Soluble, GFP-fused amino-terminal Igv-like domains of different human and mammalian CEACAMs were used in this study. B. GFP-tagged N-terminal domains of human CEACAM1 (hCEA1-N), CEACAM3 (hCEA3-N), CEA (hCEA-N), CEACAM6 (hCEA6-N) and CEACAM8 (hCEA8-N) were generated as secreted proteins in 293 cells. Equivalent levels of the different soluble N-terminal domains were verified by probing the supernatants with anti-GFP antibodies (lowest panel). Supernatants containing the indicated GFP-fusion proteins were incubated with Hinf-aeg or NTHi strains 7502, KN4 or H86, respectively, and CEACAM-binding was determined upon bacterial pull-down and anti-GFP
Hinf binds human CEACAMs via P1 to colonize mucosa
40
Western Blot (upper panels). Pull-downs using NTHi strain 7502 (uppermost panel) were exposed 3-times longer than the other panels. C. Receptor binding assays as described in (B) were performed with eleven additional H. influenzae strains. The binding to CEACAMs is symbolized by “+” in case of binding or “─” in absence of binding. D. Cell culture supernatants (supe, left panel), enriched in equivalent amounts of GFP-tagged soluble proteins containing hCEA1-NA1BA2 or hCEA8/1-NA1BA2, respectively, were incubated with NTHi 7502 (middle panel) or H. influenzae Rd (right panel). The binding to CEACAMs was analyzed by flow cytometry. Bacteria incubated in plain medium (gray line) were included as negative controls. The number (6x) shows the -fold change in mean fluorescence intensity (MFI) of bacteria incubated with hCEA1-NA1BA2 in comparison to bacteria incubated with hCEA8/1-NA1BA2.
Similar to its inability to bind the CEACAM1 N-terminal domain (Figure II-1B), NTHi
7502 failed to recognize hCEA1-NA1BA2 (Figure II-1D). To investigate the
contribution of the Igv-like amino-terminal domain versus the IgC2-like domains of
CEACAM1 in H. influenzae binding we used hCEA1-NA1BA2 as well as a
chimeric protein, where the CEACAM1 N-terminal domain was exchanged for the
CEACAM8 N-terminal domain (hCEA8/1-NA1BA2; Figure II-1A). Clearly, H.
influenzae Rd associated with hCEA1-NA1BA2, but not with hCEA8/1-NA1BA2
highlighting the essential role of the Igv-like amino-terminal domain for the H.
influenzae-CEACAM1-interaction (Figure II-1D). Together, these data demonstrate
that the Igv-like domain of CEACAM1 is necessary and sufficient to mediate H.
influenzae binding, but that there is considerable heterogeneity in CEACAM-
recognition by H. influenzae strains that may be linked to variations in the
CEACAM-binding adhesin.
The OMP P5 protein of H. influenzae does not confer CEACAM binding
properties
The -barrel outer membrane protein (OMP) P5 was proposed to constitute the
CEACAM-binding adhesin of H. influenzae (Hill et al., 2001). To better define the
structure-activity relationship of this bacterial adhesin, we set out to compare P5
sequence variation to the observed CEACAM binding profiles. Therefore, we
cloned and sequenced the ompP5 gene from the 15 H. influenzae strains used in
the CEACAM-binding assay and aligned the protein sequences to delineate a
CEACAM-binding motif (Figure II-2A). By concentrating on surface exposed loops
3 and 4 of OMP P5, which form the putative binding interface (Bookwalter et al.,
2008), the amino acid sequence variation in these regions became obvious
(Figure II-2A). Interestingly, homologies in loop sequences did not co-segregate
with the observed pattern of CEACAM-binding. In particular, the OMP P5
Hinf binds human CEACAMs via P1 to colonize mucosa
41
sequence of strain 7502, which showed no CEACAM binding in our assays, was
not grossly different from the other strains (e.g. compare strain 7502 with strains
2019 or H51; Figure II-1C and Figure II-2A.). To directly link OMP P5 with the
ability of bacteria to bind CEACAM1, we expressed OMP P5 derived from one of
the CEACAM-binding H. influenzae strains, strain Rd, heterologously in E. coli
(Eco-P5-Rd). OMP P5 was not only expressed, but also integrated into the outer
membrane of E. coli as demonstrated by its accessibility to anti-P5 antibodies
(Figure II-2B and C) and its heat modifiable migration in gel electrophoresis
(Figure. S2). However, similar to the control E. coli strain harboring the empty
plasmid (Eco pET28), Eco-P5-Rd was not able to bind to the GFP-fused
CEACAM1 Igv-like amino-terminal domain, whereas Hinf Rd showed the expected
strong binding to CEACAM1-GFP in this assay (Fig. 2D). In addition, we
heterologously expressed two additional OmpP5 genes derived from strains H51
and H86, respectively, in E. coli (Eco-P5-H51; Eco-P5-H86). Similar to Eco-P5-Rd
we did not observe any binding to CEACAMs (data not shown). Together, these
results suggested that either OMP P5 alone is not sufficient to promote CEACAM1
binding or that this membrane protein is not the CEACAM-binding adhesin of H.
influenzae.
Hinf binds human CEACAMs via P1 to colonize mucosa
42
Figure II-2. The OMP P5 of H. influenzae is not the CEACAM-binding adhesin. A. 15 H.influenzae OMP P5 sequences spanning the predicted surface-exposed loops 3 and 4 werealigned with Clustal Omega. Spaces (-) are introduced to maximize homology. The dashed linesseparate H. influenzae strains exhibiting the same CEACAM-binding profile. B. Presence of H.influenzae OMP P5 in the outer membrane of recombinant E. coli (Eco-P5-Rd) was detected byWestern blotting with polyclonal anti-P5 mouse serum (-P5 blot; upper panel). E. coli harboringthe empty vector (Eco pET28) and outer membrane protein preparation of H. influenzae Rd servedas negative and positive controls, respectively. The arrow head points to Hinf OMP P5, whereasthe asterisk indicates a non-specific cross-reactive band from E. coli, presumably OMP A.Equivalent protein loading was demonstrated by Coomassie staining of the membrane (bottompanel). C. Surface location of OMP P5 in Eco-P5-Rd was confirmed by flow cytometry followingstaining with -P5 serum. Grey area indicates staining of the Eco-pET28 control strain. D. IPTG-induced Eco-pET28 and Eco-P5-Rd were incubated with cell culture supernatants containinghuman CEACAM1-N-GFP (hCEA1-N), human CEACAM8-N-GFP (hCEA8-N), or plain culturemedium. After washing, samples were analysed by flow cytometry to determine bacteria-associated GFP fluorescence. H. influenzae Rd was included as positive control for humanCEACAM1-binding. Grey area indicates staining of bacteria incubated in plain culture medium. Thenumber (5x) shows the -fold change in mean fluorescence intensity (MFI) of H. influenzae Rdincubated with hCEA1-N in comparison to hCEA8-N.
Hinf binds human CEACAMs via P1 to colonize mucosa
43
Genetic deletion of OMP P1 abrogates Hinf interaction with CEACAM1
The failure of heterologously expressed OMP P5 to engage CEACAMs prompted
us to search for alternative CEACAM-binding outer membrane proteins in H.
influenzae. Accordingly, we tested a panel of isogenic H. influenzae mutants with
deletions of defined outer membrane proteins for their ability to associate with the
recombinant CEACAM1-GFP fusion protein (see Table S1 for a list of strains and
mutants). Importantly, whereas the P5-deficient mutant of H. influenzae strain Rd
showed reduced CEACAM1 binding, the deletion of the OMP P1 of this strain
completely abrogated CEACAM1 binding (Figure II-3A). Deletion of genes
encoding OMP P2, OMP P4, or transferrin-binding protein did not alter CEACAM
recognition by Hinf strain Rd (Figure II-3A). To corroborate these findings in a
more physiological context, we analyzed adhesion of H. influenzae strain Rd to
human lung adenocarcinoma A549 cells. To this end, cells were infected with
either the wildtype bacteria (Hinf Rd), the OMP P5 mutant (Hinf Rd ∆P5), or the
OMP P1 mutant (Hinf Rd ∆P1), respectively. Two hours after infection, samples
were washed and processed for scanning electron microscopy. Clearly, whereas
numerous wildtype bacteria and similar numbers of the isogenic Hinf Rd ∆P5
mutant were found attached to the epithelial cells, cell adhesion by Hinf Rd ∆P1
was almost completely absent (Figure II-3B). Viability and growth of all three
strains in cell culture medium was not affected by the presence or absence of
OMP P1 or OMP P5, respectively (Suppl. Figure S1A). Together, these data
highlight the necessity of H. influenzae OMP P1 for CEACAM-targeting.
Hinf binds human CEACAMs via P1 to colonize mucosa
44
Figure II-3. Genetic deletion of OMP P1 abrogates the interaction of H. influenzae withhuman CEACAMs. A. H. influenzae Rd and its respective P5- (Hinf Rd ∆P5), P1- (Hinf Rd ∆P1),P2- (Hinf Rd ∆P2), P4- (Hinf Rd ∆P4) or TbpX- (Hinf Rd ∆TbpX) deficient mutants were incubatedwith the indicated human CEACAM-GFP fusion proteins or cell culture medium. Flow cytometrywas used to detect CEACAM binding. The fold change in MFI upon incubation with hCEA1-Ncompared to hCEA8-N is indicated for each sample. B. A549 cells were infected for 2 hours withwild-type H. influenzae Rd strain, its isogenic P5 or P1 mutant and processed for scanning electronmicroscopy upon fixation. Shown are pseudocolored bacteria (blue) attached to A549 cells (grey).
Hinf binds human CEACAMs via P1 to colonize mucosa
45
CEACAM1-mediated host cell invasion of H. influenzae is OMP P1-dependent
Although being considered a classical extracellular pathogen, H. influenzae has
been consistently reported within host cells, including epithelial cells (for a
comprehensive review, please refer to (Clementi and Murphy, 2011)). In order to
unambiguously identify the contribution of CEACAM1 in epithelial cell invasion by
H. influenzae, we used HeLa cells, which do not express CEACAMs
endogenously. To allow CEACAM-mediated interactions, HeLa cells were stably
transfected with GFP-tagged human CEACAM1 or with GFP alone (Suppl. Fig.
S1B). These stable cell lines were infected for 2 h with Hinf Rd, Hinf Rd ∆P1, or
Hinf Rd ∆P5, respectively, and internalized bacteria were evaluated by gentamicin
protection assays. Importantly, Hinf Rd invaded HeLa cells in a CEACAM1-
dependent manner and also Hinf Rd ∆P5 showed a CEACAM1-dependent
invasion, which was reduced to about 50% of the wildtype strain (Figure II-4A).
Consistent with its inability to bind CEACAM1, Hinf Rd ∆P1 completely failed to
invade HeLa CEACAM1 cells (Figure II-4A). Analysis of the outer membrane
protein profile of the three strains indicated that OMP P5 expression in Hinf Rd
∆P1 was comparable to the wildtype Rd strain (Figure II-4B). However, OMP P1
expression was clearly reduced in the Hinf Rd ∆P5 mutant indicating that deletion
of OMP P5 could affect surface expression of OMP P1 (Figure II-4B and C).
Alterations in outer membrane content of OMP P1 in Hinf Rd ∆P5 were not
correlated with gross morphological changes of the microorganisms as observed
by scanning electron microscopy (Figure II-4D). Together, our observation of
compromised OMP P1 levels in a Hinf Rd ∆P5 mutant might explain the reduction
in CEACAM binding and CEACAM-mediated invasion by P5-deficient H.
influenzae strains (Figure II-3A and Figure II-4A).
Hinf binds human CEACAMs via P1 to colonize mucosa
46
Figure II-4. H. influenzae invades CEACAM-expressing cells in an OMP P1-dependentmanner. A. HeLa cells stably expressing GFP alone or GFP-tagged human CEACAM1 wereinfected with H. influenzae Rd or the indicated mutants (Rd ∆P5 or Rd ∆P1). Two hours post-infection, extracellular bacteria were killed with gentamicin and the amount of viable, intracellularbacteria was determined via colony counts. Results represent means ± SD of three independentexperiments done in triplicate (n=3) *P < 0.05, ***P < 0.001). B. Wild-type H. influenzae Rd strain(WT), the OMP P5 mutant (∆P5) and the OMP P1 mutant (∆P1) outer membrane proteins wereisolated and immunoblotted with anti-P5 (upper panel) or anti-P1 (middle panel) mouse sera.Similar amounts of protein in the different samples were verified by Coomassie staining of themembrane (lower panel). C. Densitometry of intensities of OMP P1 band in ∆P5 relative to that ofthe wild-type Rd strain. Values are mean ± SD of the amount of OMP P1 as observed in threeindependent experiments. (*P < 0.05). D. Scanning electron microscopy pictures of H. influenzaeRd wildtype and Rd ∆P5, respectively. Scale bars 200 nm.
Hinf binds human CEACAMs via P1 to colonize mucosa
47
Complementation of OMP P1 in Hinf Rd ∆P1 restores CEACAM-binding
To firmly establish OMP P1 as a CEACAM-binding adhesin, we complemented the
P1-deficient Hinf Rd mutant (Hinf Rd ∆P1) by restoration of the chromosomal OMP
P1 locus. The resulting strain (Hinf Rd ∆P1 P1+) expressed P1 at levels similar to
the H. influenzae Rd parent strain (Figure II-5A) and bound to the soluble
CEACAM1-N-domain (Figure II-5B). To determine if re-expression of OMP P1 in
Hinf Rd ∆P1 leads to cellular uptake of the bacteria, flow cytometry-based invasion
assays were conducted with 293 cells and FITC-labeled bacteria. Cells were
transfected with a full length CEACAM1-4L-encoding expression plasmid or the
empty control plasmid. Cells were then infected with Hinf Rd, the OMP P5 mutant
(Hinf Rd ∆P5), the OMP P1 mutant (Hinf Rd ∆P1), or the P1 complemented strain
(Hinf Rd ∆P1 P+). After 2h of infection, the internalization of FITC-labeled microbes
was quantified by flow cytometry in the presence of trypan blue, which quenches
the fluorescence of extracellular bacteria. Consistent with the gentamicin
protection assays (Figure II-4A), Hinf Rd and, to a slightly lesser extent, Hinf Rd
∆P5 were internalized by CEACAM1-expressing 293 cells (Figure II-5C). However,
Hinf Rd ∆P1 showed no invasion and this non-invasive phenotype of Hinf Rd ∆P1
was completely reversed upon complementation with OMP P1 (Figure II-5C).
Together, these results confirmed the critical role of OMP P1 in CEACAM-
mediated interactions with host cells and suggested that OMP P1 can function as
a H. influenzae invasin for CEACAM-expressing human cells.
Hinf binds human CEACAMs via P1 to colonize mucosa
48
Figure II-5. Restoration of ompP1 in H. influenzae OMP P1 mutant reverses its CEACAM-binding phenotype. A. Outer membrane proteins of H. influenzae Rd (WT), its P1 mutant (∆P1)and the P1-restored strain (∆P1 P1+) were analyzed for P1 or P5 expression using anti-P1 or anti-P5 mouse sera, respectively. B. CEACAM-binding of the P1-complemented ∆P1 mutant Hinf Rd(Hinf Rd ∆P1 P1+) was determined by flow cytometry after incubation with the indicated GFP-fusedCEACAM soluble proteins. Grey area indicates staining of bacteria incubated in plain culturemedium. The fold change in MFI upon incubation with hCEA1-N compared to hCEA8-N isindicated. C. 293 cells were transiently transfected with the empty vector (pcDNA) or with a vectorencoding CEACAM1-4L-HA. Expression of CEACAM1-4L-HA in the transfected 293 cells wasdetected by Western Blotting of whole cell lysates (WCL) with monoclonal CEACAM-antibody(clone D14HD11; upper panel) and similar amounts of loaded WCL was verified by anti-tubulinantibodies (lower panel). Transfected cells were infected for 2 hours with FITC-labelled H.influenzae Rd wildtype, Rd ∆P5, Rd ∆P1, or the P1 complemented strain (Rd ∆P1 P1+),respectively. The infected cells were suspended and the fluorescence of extracellular bacteria wasquenched by trypan blue. Cell-associated fluorescence, which was proportional to the amount ofintracellular bacteria, was analyzed by flow cytometry. Uptake of H. influenzae Rd wildtype byCEACAM1-expressing cells was set to 100%. Bars show means ± SD of the intracellular bacteriaas observed in three independent experiments (*P < 0.05, ***P < 0.001).
Hinf binds human CEACAMs via P1 to colonize mucosa
49
Modelling of H. influenzae OMP P1 identifies surface-exposed structures
In an attempt to delineate CEACAM-binding structural features of OMP P1, we
combined three-dimensional modeling of OMP P1 with multiple protein sequence
alignment of OMP P1 from H. influenzae strains exhibiting distinct CEACAM-
recognition profiles. Interestingly, H. influenzae OMP P1 is a homologue of the E.
coli long-chain fatty acid transporter (FadL), which share around 39% amino acid
sequence identity. Using the atomic structure of FadL (van den Berg et al., 2004),
Hinf OMP P1 was modeled using SWISS-MODEL (http://swissmodel.expasy.org
/interactive). The resulting three-dimensional structure suggested that OMP P1
consists of a 14-stranded β barrel embedded in the Haemophilus outer membrane
Figure II-6A). The lumen of the barrel is occluded by the amino-terminal domain
and there are seven extracellular loops (L1 to L7), of which four, L1, L3, L4, and
L7, are prominent and easily accessible (Figure II-6A). Next, we extracted the
OMP P1 coding sequences of strains Hinf Rd KW20 (access. No. NC_000907
REGION: 422144..423523) and NTHi 86-028 NP (access. No. NC_007146
REGION: 491079..492455) from NCBI (http://www.ncbi.nlm.nih.gov). Furthermore,
ompP1 genes from thirteen additional isolates (see Table S1) were cloned and
sequenced. The derived protein sequences were aligned with Clustal Omega
(Sievers et al., 2011). The alignment highlighted conserved as well as the highly
variable regions of OMP P1, which coincide with the surface exposed loops 1, 3,
4, and 7 (Figure II-6B). Strikingly, the only strain in our collection, which had been
negative for CEACAM-binding, strain NTHi 7502, turned out to be a natural OMP
P1 mutant. Indeed, the ompP1 coding sequence of this strain contained a non-
sense mutation at position 875 creating a premature stop codon (Figure II-6B).
The disruption of expression was confirmed at the protein level, as no OMP P1
was detected in outer membrane preparations of strain NTHi 7502, while OMP P5
was expressed at normal levels (Figure II-6C). This result provides further genetic
support for the idea that OMP P1 is the CEACAM-binding adhesin. Further
inspection of the surface exposed loops of the CEACAM-binding OMP P1 variants
did not reveal a consistent sequence motif, which would explain binding to one or
more CEACAMs (Figure II-6B). Therefore, these findings suggest that a three-
dimensional structure rather than a conserved linear stretch of amino acids
determines CEACAM-targeting by H. influenzae OMP P1.
Hinf binds human CEACAMs via P1 to colonize mucosa
50
Figure II-6. H. influenzae OMP P1 is a β-barrel porin with highly variable extracellular loops.A. Putative Hinf-aeg OMP P1 three-dimensional structure generated by SWISS-Model. OMP P1 ispredicted to consist of a 14-stranded transmembrane β-barrel obstructed by an N-terminal hatchdomain. B. The deduced amino acid sequences of OMP P1 from 15 H. influenzae strains werealigned with Clustal Omega. Spaces [.] are introduced to allow maximal alignment. The colorshighlight amino acids, which are conserved in >90% of strains (dark grey), 60 – 90 % of strains(grey), or 45 – 60% (light grey) of analysed strains. The black arrows above the sequencesdelineate the surface-exposed loops (loop 1 — 7). C. Outer membrane proteins were isolated fromH. influenzae 7502 or H. influenzae Rd, respectively, and immunoblotted with anti-P5 (upper panel)or anti-P1 (lower panel) mouse sera.
Hinf binds human CEACAMs via P1 to colonize mucosa
51
Heterologous expression of H. influenzae OMP P1 transfers CEACAM-
binding capability to E. coli
To investigate if OMP P1 is sufficient to recognize CEACAMs, we cloned and
expressed the ompP1 gene of H. influenzae Rd in E. coli (Eco-P1-Rd).
Immunoblot analysis of outer membrane preparations as well as antibody staining
of intact bacteria indicated that OMP P1 is expressed and located on the surface
of Eco-P1-Rd (Figure II-7A and B). Importantly, expression of H. influenzae OMP
P1 converted E. coli into a CEACAM1-binding bacterium (Figure II-7C). To verify
OMP P1–CEACAM interaction in a cellular context, we infected HeLa cells
expressing GFP or GFP-tagged human CEACAM1 with Eco-P1-Rd, Eco-P5-Rd,
or E. coli harboring the empty vector (Eco-pET28), respectively. Neither Eco-
pET28, nor Hinf OMP P5-expressing E. coli (Eco-P5-Rd) was taken up by GFP- or
CEACAM1-GFP HeLa cells (Figure II-7D). In sharp contrast, Eco-P1-Rd was
readily internalized by CEACAM1-expressing cells (Figure II-7D). The results
obtained with antibiotic protection assays were corroborated with microscopic
analysis of infected samples. Differential staining of intracellular vs. extracellular
bacteria revealed that HeLa-CEACAM1-GFP cells contained numerous
intracellular Eco-P1-Rd (Figure II-7E). However, intracellular bacteria were not
observed in the case of HeLa-CEACAM1-GFP cells infected with Eco-P5-Rd or
Eco-pET28 (Figure II-7E). Taken together, these data demonstrate that OMP P1
of H. influenzae is not only necessary, but also sufficient to target CEACAMs
present on human cells and to trigger bacterial invasion in non-professional
phagocytic cells.
Hinf binds human CEACAMs via P1 to colonize mucosa
52
Figure II-7. . OMP P1 expression in E. coli promotes binding to recombinant CEACAMs and invasion into CEACAM-expressing cells. A. Outer membrane proteins (OMPs) were isolated from Eco pET28 or E. coli expressing the ompP1 gene of H. influenzae Rd (Eco P1-Rd) and immunoblotted with anti-P1 antibodies (upper panel). Similar loading of OMPs was verified by Coomassie staining. B. Surface location of P1 in Eco-P1-Rd was confirmed by flow cytometry following staining with -P1 serum. Black line indicates staining of the Eco pET28 control strain. C. GFP-tagged amino-terminal domains of CEACAM1 (hCEA1-N) or CEACAM8 (hCEA8-N) as well as plain cell culture medium were incubated with either Eco P1-Rd or Eco pET28. Flow cytometry analysis was used to determine bacteria-associated GFP fluorescence as a measure of binding to human CEACAMs. The fold change in MFI upon incubation with hCEA1-N compared to hCEA8-N is indicated. D. HeLa GFP and HeLa CEACAM1-GFP were infected with Eco pET28, Eco P5-Rd,
Hinf binds human CEACAMs via P1 to colonize mucosa
53
or Eco P1-Rd for 2 hours at MOI 100. The number of viable, internalized bacteria was subsequently determined by gentamicin protection assays. The data represent means ± standard deviation of three independent experiments done in triplicate (**P < 0.01). E. HeLa CEACAM1-GFP cells were infected as in (D). Samples were fixed and extracellular bacteria (red) were stained with polyclonal anti- E. coli LPS antibody in combination with a Cy3-conjugated secondary antibody. After cell permeabilization, both intra- and extracellular bacteria were stained again using polyclonal anti-E. coli lipopolysaccharide antibody in combination with a Cy5-conjugated secondary antibody (cyan). Arrowheads point to extracellular bacteria labelled with Cy3 and Cy5, while arrows indicate internalized bacteria selectively stained with Cy5. Scale bars 10 µM.
Heterologous expression of OMP P1 recapitulates the species-specific
CEACAM binding profile of H. influenzae strains
CEACAM-binding adhesins of several Gram-negative bacteria, including Neisseria
gonorrhoeae, N. meningitidis as well as Moraxella catarrhalis, selectively
associate with human CEACAM1, but not with CEACAM1 orthologues from other
mammals (Voges et al., 2010). The species–specificity of the H. influenzae
CEACAM-binding adhesin has not been evaluated. To first test, if H. influenzae
OMP P1 expressed in E. coli faithfully recapitulates the binding specificity of the
adhesin, we cloned and expressed in E. coli two additional OMP P1 genes derived
from NTHi KN4 or H. influenzae biogroup aegyptius (Eco-P1-KN4 and Eco-P1-
Hae), respectively. Similar to the CEACAM binding spectrum observed for the
parent H. influenzae strains, Eco-P1-KN4 recognized exclusively CEACAM1, Eco-
P1-Rd bound to CEACAM1 or CEA, whereas Eco-P1-Hae associated with human
CEACAM1, CEACAM3, and CEA (Figure II-8A). Most importantly, recombinant
OMP P1 levels in E. coli were similar to endogenous OMP P1 levels in the
respective H. influenzae parent strains (Figure II-8). Combined, these results
demonstrate that heterologous expression of OMP P1 in E. coli preserves the
binding specificity of the adhesin. Next, we used soluble GFP fusion proteins of
the IgV-like amino-terminal domains derived from human, bovine, murine, or
canine CEACAM1 orthologues (Figure II-8C) and employed them in binding
assays with H. influenzae strain Rd. Clearly, H. influenzae selectively bound to
human CEACAM1, but not to any other mammalian CEACAM1 orthologue (Figure
II-8C).
Hinf binds human CEACAMs via P1 to colonize mucosa
54
Furthermore, upon expression of OMP P1 in E. coli the identical binding specificity
was observed (Figure II-8D). Together, these results demonstrate that H.
influenzae OMP P1 has exquisite selectivity for human CEACAMs and indicate
that this outer membrane protein co-evolved with the human receptor(s). From
these detailed binding studies and their evolutionary implications, we speculate
that this molecular interaction plays a vital role in the microbe–host relationship.
Figure II-8. E. coli heterologously expressing H. influenzae OMP P1 mimic the species-specific CEACAM-recognition profile of parent H. influenzae strains. A. E. coli expressingOMP P1 of NTHi KN4 (Eco-P1-KN4), Hinf Rd (Eco-P1-Rd), or Hinf-aeg (Eco-P1-Hae) as well as E.coli pET28 (Eco-pET28) were incubated with equivalent levels of the indicated hCEA-N-GFPproteins. Presence of hCEA-N domains was verified by probing the supernatants with anti-GFPantibodies (supe; top panel). After incubation, bacteria-associated CEACAM-N domains weredetected by Western blotting with GFP antibodies (pull-down; bottom panels). B. The E. coli strainsexpressing different OMP P1 proteins from A) were compared side-by-side for OMP P1 expressionwith equal amount (10 µg) of outer membrane proteins derived from the parental H. influenzaestrains. C. The amino-terminal Igv-like domains (N) of human CEACAM8 (hCEA8), humanCEACAM1 (hCEA1), canine CEACAM1 (cCEA1), bovine CEACAM1b (bCEA1b), bovineCEACAM1a (bCEA1a) and murine CEACAM1 (mCEA1), were expressed as secreted GFP-fusionproteins in 293 cells. Cell culture supernatants containing the indicated GFP-fusion proteins wereincubated with H. influenzae Rd. After washing, bacteria-associated CEACAMs were detected byWestern blotting with anti-GFP antibodies (top panel). Equivalent levels of soluble N-terminaldomains in the cell culture supernatants were verified by probing the supernatants with anti-GFPantibodies (supe; bottom panel). D. Binding of Eco-P1-Rd to GFP-tagged amino-terminal domain ofvarious mammalian CEACAM1 of bovine, canine, human or mouse origin was analyzed as in C).
Hinf binds human CEACAMs via P1 to colonize mucosa
55
II.4 Discussion
Haemophilus influenzae is a clinically relevant pathogen inhabiting the human
nasopharynx. It has been previously reported that H. influenzae attaches to host
receptors of the CEACAM family via its antigenic variable outer membrane protein
P5 (Hill et al., 2001). In vivo, CEACAM-binding seems to be instrumental for H.
influenzae to successfully colonize model organisms (Bookwalter et al., 2007).
Here, we report that OMP P5 does not support receptor binding and OMP P5-
deficient H. influenzae have only a slightly reduced ability to engage CEACAMs.
Instead, a distinct membrane protein, the so-called OMP P1, is the critical
CEACAM-binding adhesin of H. influenzae. OMP P1 is one of the major surface-
exposed proteins and accounts for ~10% of H. influenzae OMP content (Bolduc et
al., 2000). Different binding assays and functional investigations with typeable and
non-typeable H. influenzae strains, deletion and complementation of OMP P1 as
well as heterologous expression of OMP P1 in E. coli unequivocally demonstrate
that this membrane protein is necessary and sufficient to engage CEACAMs. In
line with the species specificity of H. influenzae, OMP P1 shows high selectivity for
human, but not other mammalian CEACAMs. Furthermore, OMP P1 variants
derived from a diverse set of H. influenzae strains with defined CEACAM-binding
characteristics display an identical CEACAM-binding pattern when expressed in E.
coli strongly arguing that OMP P1 dictates CEACAM recognition by Haemophilus
influenzae.
Clearly, our results are on first sight in conflict with prior data, which suggested
that OMP P5 is the CEACAM-binding adhesin of H. influenzae (Hill et al., 2001).
However, already these investigators noted that deletion of the ompP5 gene only
reduces, but does not abrogate the CEACAM-binding capacity of the pathogens
(Hill et al., 2001). In light of our finding, that OMP P1 is the proper CEACAM-
binding surface protein of Hinf, an alternative explanation for the phenotype of Hinf
OMP P5 mutants is needed. For example, the effect of the ompP5 deletion on
CEACAM binding could be indirect via an influence on OMP P1 expression or
localization. Indeed, examination of the outer membrane profile from an OMP P5-
deficient strain revealed that in the absence of OMP P5 the abundance of OMP P1
is affected. In particular, we found lower levels of OMP P1 in outer membrane
preparations from an OMP P5-deficient strain compared to wildtype H. influenzae.
Hinf binds human CEACAMs via P1 to colonize mucosa
56
This observation provides an immediate explanation for the reduced CEACAM-
binding of an OMP P5 mutant, and unifies the prior reports with our novel findings.
The reduced presence of OMP P1 in H. influenzae ∆OMP P5 might be connected
to the known role of OMP P5 homologues from other Gram-negative bacteria,
such as E. coli OmpA, Porphyromonas gingivalis Pgm6, or Pseudmonas
aeruginosa OprF to regulate the stability of the bacterial outer membrane,
presumably via interactions with peptidoglycan (Confer and Ayalew, 2013; Gotoh
et al., 1989a; Gotoh et al., 1989b; Nagano et al., 2005; Rawling et al., 1998;
Wang, 2002).
Strong support to our conclusion that OMP P1 is the CEACAM-binding adhesin
comes from the chance observation of a Hinf isolate, NTHi strain 7502, which
expresses abundant OMP P5, but does not bind to any CEACAM. As it turns out,
NTHi strain 7502 is a natural mutant of OMP P1 harboring a missense mutation in
the OMP P1 reading frame. Interestingly, Swords et al. noted reduced epithelial
cell binding and invasion by this NTHi strain in an independent earlier study
(Swords et al., 2000). Together with the gain-of-function upon OMP P1 expression
in E. coli, OMP P1 is necessary and sufficient to engage human CEACAMs.
Interestingly, H. influenzae OMP P1 shares ~ 39% amino acid identity with the
FadL protein of E. coli, an outer membrane protein involved in the uptake of long
chain fatty acids (LCFA). FadL is characterized by a long -barrel consisting of 14
antiparallel -strands connected by several extended extracellular loops. Despite
the large diameter of the FadL monomer, the crystal structure reveals a ~40
residue N-terminal extension, which folds into a compact, plug-like structure
occupying the inner lumen of the -barrel. Therefore, FadL exists as a closed pore
and LCFA transport seems to require conformational rearrangements in both the
N-terminal plug and the -barrel wall (Hearn et al., 2009; van den Berg, 2005). It is
not known if H. influenzae OMP P1 is also involved in LCFA transport. In previous
studies, OMP P1-deficient strains were analyzed in animal models of invasive Hib
or Hinf-aeg disease, where no differences in the growth and pathogenic potential
compared to wildtype bacteria were observed, when these bacteria were injected
intraperitoneally (Hanson et al., 1989; Munson and Hunt, 1989; Segada et al.,
2000). This indicates that OMP P1 is not essential for bacterial survival and growth
under such in vivo conditions. However, given the fact the OMP P1 can serve as
Hinf binds human CEACAMs via P1 to colonize mucosa
57
an adhesin, which selectively binds to human, but not other mammalian
CEACAMs, it is in retrospect not surprising that an OMP P1-deficient strain might
not show a phenotype in animal models of H. influenzae infection.
In other -barrel shaped CEACAM-binding adhesins, such as the colony opacity-
associated (Opa) proteins of pathogenic Neisseriae, several extracellular loops are
highly variable between different isolates (Robertson and Meyer, 1992). Despite
their sequence variation, a short amino acid motif, GxI/V/LxQ, found in HV-2 of
N. meningitidis Opa proteins, appears to be critical for receptor engagement (de
Jonge et al., 2003). Interestingly, this motif can be found in OMP P1 loop 3 of
some H. influenzae strains (see e.g. strain H86; Figure II-6B). However, as only a
fraction of the CEACAM-binding H. influenzae strains harbors such a GxI/V/LxQ
motif, this sequence is clearly not necessary for engaging the host cell receptor.
Therefore, further refinement of the structure-activity relationship of OMP P1 and
ideally an atomic resolution of the ligand-receptor complex might allow
appreciation of this highly species-specific interaction between the bacterial
surface protein and the host receptor. Nevertheless, the unequivocal identification
of the bacterial adhesin and its functional determinants provides a solid ground to
devise strategies to interfere with Haemophilus – host tissue interaction.
II.5 Materials and methods
Haemophilus influenzae strains and growth conditions
The H. influenzae wild type strains and mutants used in this study and their
original sources are listed in Table S1. The four strains KN1–KN4 were originally
isolated from ear effusions of pediatric patients. Selected for their colony
morphology and their requirement of factors X and V, the identity of the strains
was confirmed by 16S rRNA gene sequencing. All H. influenzae strains were
grown on brain-heart infusion agar supplemented with Levinthal base (Hill et al.,
2001). Prior to infection or receptor binding assays, the bacteria were precultured
in brain-heart Levinthal broth until the mid-log phase (OD600 = 1).
Cell culture and generation of stable cell lines
Human embryonic kidney epithelial 293T cells (293 cells; ACC-635, German
collection of microorganisms and cell cultures, DSMZ, Braunschweig, Germany)
Hinf binds human CEACAMs via P1 to colonize mucosa
58
were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10%
calf serum. Transfection with pcDNA CEACAM1-4L HA or the empty control vector
(pcDNA) was done by calcium phosphate co-precipitation as described (Muenzner
et al., 2005). HeLa (human cervix) epithelial cells, a kind gift of T.F. Meyer (Max-
Planck Institut für Infektionsbiologie, Berlin, Germany), were stably transduced
with lentiviral particles encoding either CEACAM1 ∆CT-GFP or GFP alone as
previously described (Kuespert et al., 2011). HeLa as well as A549 epithelial cells
(human alveolar lung adenocarcinoma cells; ATCC CCL87) were cultured in
DMEM containing 10% fetal calf serum at 37 °C with 5% CO2 and subcultured
every 2–3 days (HeLa cells) or cultured until confluency was reached and seeded
in new culture vessels 16h prior to infection (A549 cells).
Binding assays with soluble CEACAMs
The production of recombinant GFP-tagged CEACAM proteins by 293 cells and
the subsequent bacterial binding assays were performed as described (Kuespert
and Hauck, 2009; Kuespert et al., 2007). Briefly, soluble recombinant GFP-tagged
CEACAM proteins, contained in cell culture supernatant, were clustered with
polyclonal anti-GFP antibodies (tag-tools GmbH, Konstanz, Germany) overnight at
4 °C. Then, 10 × 108 bacteria were added in 1 ml supernatant containing the
indicated GFP-fusion proteins and incubated for 2 h at 37 °C under gentle rotation.
Thereafter, bacteria were washed twice with PBS and samples were processed
either for Western blotting with GFP antibodies or flow cytometry to detect
bacteria-associated fluorescence.
Expression of outer membrane proteins in E. coli
Chromosomal DNA was used as template to amplify Hinf ompP1 or ompP5. The
respective genes were then cloned into the NcoI and XhoI site of pET28a
(Novagen, Madison, Wisconsin). The pET28a vectors encoding OMP P1 or OMP
P5 were transformed in E. coli Rosetta(DE3) (Novagen, Madison, Wisconsin),
which was induced for protein expression by IPTG for 3 hours. All E. coli strains
were cultured at 37 °C in Luria-Bertani (LB) supplemented with kanamycin.
ompP1 complementation
Chromosomal restoration of ompP1 was performed according to Lichtenegger et
al. (Lichtenegger et al., 2014). The ompP1 gene was linked to a kanamycin
resistance gene cassette via SOEing (Splicing by Overlap Extension) PCRs. After
Hinf binds human CEACAMs via P1 to colonize mucosa
59
transformation of P1-deficient Rd strain, the defective ompP1 gene (described in
(Kraiss et al., 1998)) was replaced by a functional one and the resistance to
kanamycin was used as positive selectable marker. Table S2 lists the primers
used.
Gentamicin protection assay
The amount of internalized bacteria after infection of epithelial cells was
determined by gentamicin protection assay (Roth et al., 2013). Concisely, 3 × 105
cells/well were seeded in 24-well plates coated with gelatin and were infected with
bacteria at a multiplicity of infection (MOI) = 100. Two hours later, extracellular
bacteria were killed by 1 h incubation with 100 µg/ml gentamicin in cell culture
medium. Internalized bacteria were freed during cell lysis by saponin treatment
and the viable amongst them were enumerated after plating, on appropriate agar,
the suitable dilutions.
Analysis of bacterial invasion by flow cytometry
Prior to infection, bacteria were labeled with 50 µg/ml fluorescein isothiocyanate
(FITC) in PBS (30 min at 37 °C). Labeled bacteria were used to infect cells (MOI =
100) for 2 h. Cells were processed and analyzed for intracellular bacteria by flow
cytometry as described earlier (Pils et al., 2006).
Scanning electron microscopy
A549 cells were infected with Hinf Rd wildtype strain, Hinf ∆P1-, or Hinf ∆P5
mutants for 2 h. Fixation and processing of the samples for scanning electron
microscopy was described (Pils et al., 2012). The micrographs were pseudo-
colored with Adobe Photoshop CS4 (Adobe Systems Gmbh, München, Germany).
Immunofluorescence staining
HeLa cells, seeded in 24-well plates, were infected for 2 h with different E. coli
strains at an MOI = 100. After their fixation, the samples were differentially stained
for extracellular and intracellular bacteria as described (Roth et al., 2013). As a
result, extracellular bacteria were marked by both Cy3 and Cy5, whereas
intracellular bacteria are only stained with Cy5.
Hinf binds human CEACAMs via P1 to colonize mucosa
60
Outer membrane proteins isolation
Proteins of the Hinf outer membrane were isolated as reported previously (Roier et
al., 2012). For E. coli, an identical method was used following IPTG induction.
Immunoblotting and used antibodies
Western blotting was performed as described earlier (Hauck et al., 2001) using a
monoclonal antibody (mAb) against GFP (clone JL-8; Clontech, Palo Alto,
California), mAb against tubulin (clone E7; DSHB, University of Iowa), or using
polyclonal mouse sera against OMP P1 (anti-P1) or against OMP P5 (anti-P5)
(Roier et al., 2014). CEACAM1-4L-HA expression in 293 cells was confirmed with
mouse monoclonal antibody against the HA-tag (clone TT7; tag-tools GmbH,
Konstanz, Germany) and with a mouse mAb against CEACAMs (CEACAM1,
CEACAM3, CEA, and CEACAM6) (clone D14HD11; Aldevron, Freiburg,
Germany).
Statistical analysis
Pairwise comparisons were made using Student’s t-test (SigmaStat software,
Systat Software Inc., San Jose, California). In all cases, a p value < 0.05 was
considered significant.
II.6 Acknowledgements
We are indebted to Stefan Schild and Sandro Roier (Universität Graz, Austria) for
providing us with the anti-P1 and anti-P5 serum. We are thankful to Monika Huff-
Nagel (practicing pediatrician; Konstanz, Germany) for contributing primary
isolates of NTHi (strains KN1 – KN4), to Susanne Feindler-Boeckh for technical
support and to Michael Laumann (EM Service, University of Konstanz, Germany)
for help with scanning electron microscopy. A.K.T. is recipient of a fellowship from
the Deutsche Akademische Austauschdienst (DAAD) and an associate fellow of
the RTG1331. S.L. was supported by the ERA-NET FWF I 662-B09. This work
was supported by DFG grant Ha 2856/6-2 to C.R.H.
Hinf binds human CEACAMs via P1 to colonize mucosa
61
II.7 Supporting information
Figure S1. A. Growth curves of Hinf Rd wildtype and its mutants in cell culture medium. The growth of Hinf Rdwildtype strain and its isogenic P5 or P1 mutant was monitored by reading the optical density at600 nm (OD600) in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum.Data shown are the mean of three independent experiments. B. CEACAM1-GFP or GFP expression in stable HeLa cell lines. HeLa cells were stably transducedto express either GFP alone or CEACAM1-GFP. Whole cell lysates (WCL) of virus-transducedHeLa cells were probed with GFP antibodies to confirm the expression of either GFP orCEACAM1-GFP (upper panel). The similar amount of cell lysate in the different samples wasdemonstrated by Western blotting with monoclonal anti-tubulin antibody (lower panel).
Figure S2. The outer membrane P5 is a heat-modifiable protein. 10 µg of outer membranepreparations of Hinf OMP P5-expressing Escherichia coli (Eco-P5-Rd) or Hinf Rd were boiled afteraddition of sample buffer (95ºC for 10 min) or left at room temperature. The samples, boiled (+) ornot (-), were immunoblotted with anti-P5 mouse serum. The arrow heads point to the ~27 kDa- and~35 kDa-bands characteristic of native (fast migrating) and denatured (slow migrating) OMP P5,respectively. The asterisk indicates a non-specific cross-reactive band from E. coli, presumablyOMPA
Hinf binds human CEACAMs via P1 to colonize mucosa
62
Hinf strain Source
Rd KW20 A. Wright Tuffs University
2019 3198 7502 9274
M. A. Apicella University of Iowa
KN1 KN2 KN3 KN4
This study
H46 H51 H86
R. Mutters Philipps-Universität Marburg
Hinf b Eagan
M. Herbert John Radcliffe Hospital, Oxford
86-028NP
R. S. Jr. Munson The Ohio State University
Hinf-aegyptius ATCC® 11116™
DSMZ-German Collection of Microorganisms and Cell Cultures
Hinf Rd mutant Description Source
∆P1 ∆P2 ∆P4 ∆P5 ∆Tbpx
ompP1::cat ompP2::cat ompP4::bla ompP5::cat HI1217::bla
(Kraiss et al., 1998) (Andersen et al., 2003) (Reidl and Mekalanos, 1996) (Lichtenegger et al., 2014) (Schlor et al., 2003)
Name Sequence (5’ - 3’) Product
Hinf P5-NcoI-forward ACGCCGCCATGGATGAAAAAAACTGCAATCGC OMP P5
Hinf P5-XhoI-reverse ACGCTGCTCGAGCGCGTTATTTAGTACCGTTTACCGCG
Hinf P1-NcoI-forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC OMP P1
Hinf P1-XhoI-reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG
P1-up-forward TTCCGCGTGGCTTGATCAACA
P1-KanR for Rd ∆P1 P1+
P1-down-Kan-reverse GTGTGCTTTATTATGCTAAATTTTTTCGTGAAGAAGGTGTTGCTGAC
Kan-P1-down-forward GTCAGCAACACCTTCTTCACGAAAAAATTTAGCATAATAAAGCACAC
Kan-down-reverse- TCTTGTGCAATGTAACATCAGAGAACGGATTAGAAACTATAATTTAAGT
Down-Kan-forward ACTTAAATTATAGTTTCTAATCCGTTCTCTGATGTTACATTGCACAAGA
Down-reverse AAAGGCAAGGTATCTATAATTTCG
Table S1. List of Hinf strains and mutants used in this study
Table S2 Primers used in this study
63
CHAPTER III
The outer membrane protein P1 of
Haemophilus influenzae utilizes its
extracellular loops to grasp human
CEACAM
Arnaud Kengmo Tchoupa1 and Christof R. Hauck1,2
1Lehrstuhl für Zellbiologie, Universität Konstanz, Konstanz, Germany
2Konstanz Research School Chemical Biology, Universität Konstanz, Konstanz,
Germany
In preparation
Multiple P1 loops form the CEACAM-binding interface
64
III.1 Abstract
Carcinoembryonic antigen-related cell adhesion molecules (CEACAMs) are a
group of immunoglobulin-related vertebrate glycoproteins expressed by various
cell types throughout the human body, including epithelial cells. Haemophilus
influenzae (Hinf), a Gram-negative colonizer of the nasopharyngeal mucosal
surface, exploits its outer membrane protein (OMP) P1 in order to intimately bind
CEACAMs present on the apical side of nasopharyngeal epithelial cells. Currently,
the molecular determinants in OMP P1, which are involved in providing the binding
interface to human CEACAMs, are largely unknown. Herein, we report that OMP
P1 is structurally conserved within the Pasteurellaceae family of Hinf. However,
Hinf OMP P1 has evolved unique features, absent in other Pasteurellaceae
species, enabling this adhesin to tightly bind CEACAMs. Moreover, the structural
homolog of Hinf OMP P1 in Escherichia coli (FadL) is unable to recognize
CEACAMs. Chimeric proteins between OMP P1 and FadL as well as deletion
mutants of OMP P1 indicated the involvement of a combination of several surface-
exposed loops of Hinf OMP P1 in CEACAM recognition. Most importantly,
restricting the mobility of OMP P1 extracellular loops abrogated the association of
OMP P1 expressing Hinf and E. coli with CEACAMs. Thus, our data strengthen
Hinf as exclusive CEACAM-binding Pasteurellaceae species and suggest a novel
CEACAM-binding clasp constituted by surface-located loops of OMP P1.
III.2 Introduction
Human mucosal surfaces are constantly exposed to incoming bacteria. To
establish themselves within the resident mucosal microbiota, the incoming bacteria
must withstand both mucosal immunity and mechanical cleansing (St. Geme III,
2002). Interestingly, these microorganisms possess specialized adhesive
structures, so-called adhesins, which recognize eukaryotic cell surface receptors
and thereby mediate the intimate attachment of the bacteria to the host tissues
(Abraham et al., 1998). The relevance of these adhesins is not limited to providing
the bacteria with a strong foothold on the mucosa. Indeed, the exquisite interaction
between the adhesins and the eukaryotic receptors can mediate signaling
cascades to ensure bacterial persistence (Tchoupa et al., 2014).
Multiple P1 loops form the CEACAM-binding interface
65
As a prominent example of convergent evolution, some human-specific
pathogens, such as Neisseria gonorrhoeae, Moraxella catarrhalis or Haemophilus
influenzae (Hinf), have evolved structurally unrelated adhesins in order to target
carcinoembryonic antigen-related cell adhesion molecules (CEACAMs), a
subgroup of the immunoglobulin superfamily glycoproteins expressed by a wide
range of cells, including epithelial cells (Kuespert et al., 2006). CEACAMs are
targeted by an extended “lollipop”-like trimeric autotransporter adhesin in
Moraxella catarrhalis: the ubiquitous surface protein (Usp) A1 (Hill and Virji, 2003).
In contrast, Neisseria gonorrhoeae (Ngo) and Neisseria meningitidis (Nme) bind to
human CEACAM receptors via 8-stranded β-barrel proteins with four extracellular
loops: the “opacity-associated” (Opa) proteins (Kuespert et al., 2011; Roth et al.,
2013).
The pivotal role of the Opa-mediated CEACAM-binding for successful colonization
has been documented by two outstanding studies with mice transgenic for human
CEACAM1 or human CEA, respectively (Johswich et al., 2013; Muenzner et al.,
2010). Muenzner and colleagues nicely demonstrated that the enhanced capability
of N. gonorrhoeae to colonize the urogenital tract of mice expressing human CEA
versus wild-type mice depended on the presence of the Opa adhesin on the side
of the microbes (Muenzner et al., 2010). Similarly, all the meningococci, that
persisted within the nasopharyngeal mucosa of mice transgenic for human
CEACAM1, expressed CEACAM1-specific Opa proteins (Johswich et al., 2013).
Moreover, the genetic depletion of all four opa genes in the meningococcal strain
H44/76 precluded this bacterium from colonizing the human CEACAM1-transgenic
mice (Johswich et al., 2013).
Hinf, a human-specific inhabitant of the respiratory tract, seems to rely, at least
partly, on CEACAM-binding in order to efficiently colonize the nasopharynx in
animal models (Bookwalter et al., 2007). However, that study failed to provide
unambiguous evidence that Hinf adhesins directly interact with CEACAMs present
in the nasopharynx of chinchilla (Bookwalter et al., 2007). Recently, we uncovered
the necessity and the sufficiency of the outer membrane protein (OMP) P1 in
CEACAM-binding by Haemophilus influenzae (Tchoupa et al., 2015). OMP P1 was
predicted to build a β-barrel structure with 14 strands and 7 surface-exposed
loops. The most prominent loops of OMP P1 corresponded to the highly variable
Multiple P1 loops form the CEACAM-binding interface
66
regions of this protein, as revealed by multiple sequence alignment and protein
modeling (Tchoupa et al., 2015). However, a detailed understanding of the
contribution of these extracellular loops to the interaction between OMP P1 and
CEACAM is lacking.
In this study, we report that, irrespective of the conserved structure of OMP P1
within the family Pasteurellaceae, the targeting of human CEACAM is restricted to
the species Haemophilus influenzae. Detailed investigations on structure-activity
relationships with OMP P1 mutants demonstrate that several flexible extracellular
loops allow OMP P1 to engage human CEACAMs. Furthermore, compromising
the mobility of OMP P1 extracellular loops by chemical crosslinking disrupted the
interaction of Haemophilus influenzae or OMP P1 expressing E. coli with
CEACAMs. Our findings provide the first insights on CEACAM-binding
determinants of OMP P1 and suggest a novel clamp mode of association between
the flexible, surface-exposed loops of the Hinf adhesin and host receptors of the
CEACAM family.
III.3 Results
CEACAM-binding is restricted to OMP P1 of Hinf within the Pasteurellaceae
family
Hinf OMP P1 targets members of the human CEACAM family (Tchoupa et al.,
2015). Several OMP P1 variants derived from Hinf strains have been analyzed in
detail, with OMP P1 derived from Hinf biogroup aegyptius (Hinf-aeg) (strain ATCC
11116) showing the most promiscuous binding to multiple CEACAMs, namely,
CEACAM1, CEA and CEACAM3 (Tchoupa et al., 2015). We wondered whether
OMP P1 and/or OMP P1-like proteins with potential CEACAM-binding properties
were present in related Gram-negative bacteria. Accordingly, the blastp tool was
used, under standard settings, to search for homologues / orthologues of Hinf-aeg
OMP P1 (P1-Hinf-aeg). Obviously, the best hits were from OMP P1 of other Hinf
strains (data not shown). However, several other members of the Pasteurellaceae
family exhibited OMP P1-like proteins highly similar to P1-Hinf-aeg, with OMP P1
of Haemophilus haemolyticus M21639 reaching ~75% amino acid identity (data
not shown). Therefore, we predicted a conserved structure of OMP P1 in
Multiple P1 loops form the CEACAM-binding interface
67
Pasteurellaceae. To confirm this assumption, we took advantage of the known
crystal structure of FadL, the long-chain fatty acid transporter (van den Berg et al.,
2004) and OMP P1 homolog in Escherichia coli. The crystal structure of FadL was
used as a scaffold to align amino acid sequences from OMP P1 of Haemophilus
haemolyticus (P1-Hhae), H. parainfluenzae (P1-Hpara), H. pittmaniae (P1-Hpitt),
Mannheimia heamolytica (P1-Mhae) and P1-Hinf-aeg (Figure III-1A). Of note, this
alignment was performed with P1 sequences lacking signal peptides, based on
mature P1 sequence published elsewhere (Munson and Grass, 1988). Obviously,
apart from the loops 1, 3, 4 and 7, which are highly variable within the Hinf species
with respect to the amino acid composition and length (Tchoupa et al., 2015),
Pasteurellaceae OMP P1 exhibited a striking conservation (Figure III-1A). This
conservation was supported by sequence similarity ranging from ~60% for P1-
Mhae to 74% for P1-Hhae, compared to P1-Hinf-aeg (Figure III-1B). Importantly,
the phylogenetic tree deduced from the multiple amino acid alignment of P1 and
FadL sequences revealed that P1-Mhae (the P1 protein with the lowest similarity
to P1-Hinf-aeg) rather clustered with FadL, whereas the other P1 proteins (P1-
Hinf-aeg, P1-Hhae, P1-Hpara and P1-Hpitt) clustered together (Figure III-1C).
Multiple P1 loops form the CEACAM-binding interface
68
Figure III-1. The structure of OMP P1 is highly conserved within the Pasteurellaceae family. A. OMP P1 amino acid sequences of Hinf-aeg (P1-Hinf-aeg), Haemophilus haemolyticus (P1-Hhae), H. parainfluenzae (P1-Hpara), H. pittmaniae (P1-Hpitt) and Mannheimia heamolytica (P1-Mhae) were aligned using the solved crystal structure of E. coli OMP FadL (FadL-Eco) and the program Expresso within T-Coffee web server. Dashes (-) are inserted to maximize the alignment. The colors indicate the accuracy of the alignment from indigo (<20%) to green (20 – 30%), yellow (30 – 40%) to light orange (40 – 60%) and dark orange (60 – 80 %) to red (>80%). Surface-exposed loops (loop 1 — 7) are delineated by black arrows above the sequences. B. The percent identity or similarity of the sequences described in (A) with P1-Hinf-aeg. C. The phylogenetic tree was constructed from the Clustal Omega alignment of the sequences described in (A) and edited using the Interactive Tree Of Life (iTOL) web tool. The branch length represents the mean number of amino acid substitution per site.
Multiple P1 loops form the CEACAM-binding interface
69
Urged by the similarity of their OMP P1 sequences with P1 Hinf-aeg, we tested a
couple of Pasteurellaceae family members for CEACAM-binding. GFP-tagged
amino-terminal domains of human CEACAM1 (hCEA1-N) or CEACAM8 (hCEA8-
N, negative control) were used in bacterial binding assays (Figure III-2A). Clearly,
no interaction with hCEA1-N could be detected for H. parainfluenzae, H.
haemolyticus or M. haemolytica (Figure III-2B). In clear contrast, Hinf-aeg
exhibited a strong interaction with hCEA1-N (Figure III-2B). Consistent with the
failure of the parent strain to target CEACAMs, E. coli ectopically expressing the
OMP P1 of H. parainfluenzae was not able to bind soluble CEACAMs nor it could
adhere to CEACAM-expressing epithelial cells (data not shown). Combined, these
data strongly suggest that P1 proteins of Hinf species have evolved distinct
features, absent in P1 from closely related species, enabling them to specifically
target human CEACAMs.
Figure III-2. CEACAM-binding properties are restricted to Haemophilus influenzae in thefamily Pasteurellaceae. A. Similar amounts of GFP-fused amino-terminal domain of humanCEACAM1 (hCEA1-N) or human CEACAM8 (hCE8-N) were harvested as cell culture supernatants(supe) of 293 cells and probed with anti-GFP antibodies B. hCEA1-N or hCEA8-N produced asdescribed in (A) was used in bacterial binding assays with H. parainfluenzae (upper left panel), H.haemolyticus (upper right panel), M. heamolytica (lower left panel) and Hinf-aeg (lower right panel).Bacteria incubated in plain cell culture medium were also included as negative controls (blacklines). CEACAM-binding was analyzed by flow cytometry.
Multiple P1 loops form the CEACAM-binding interface
70
Extracellular loops located at either end of OMP P1 are necessary for
CEACAM-binding
To understand the structural basis of OMP P1 CEACAM-binding, we modelled the
three-dimensional architecture of P1-Hinf-aeg according to the known crystal
structure of FadL (Figure III-3A). Overall, the 3-dimensional orientation of OMP P1
and FadL appeared to be highly similar (Figure III-3A). However the most
prominent extracellular loops of P1 (L1, L3, L4 and L7) were divergent (Figure
III-3B). Interestingly, non-CEACAM binding E. coli overexpressing FadL remains
unable to interact with CEACAMs (data not shown). Conversely, the heterologous
expression of P1-Hinf-aeg in E. coli (Eco P1-Hinf-aeg) confers CEACAM-binding
capabilities reminiscent of the parent Hinf-aeg strain (Tchoupa et al., 2015). These
strong discrepancies between CEACAM recognition by FadL or OMP P1 make
FadL suitable for gain-of-function approaches. Accordingly, to rapidly narrow down
the OMP P1 regions involved in CEACAM binding, we generated chimeras
between OMP P1 and FadL by either replacing the first 4 (P1-FadL) or the last 3
loops of FadL (FadL-P1), respectively, by the corresponding sequences from P1-
Hinf-aeg (Figure III-3C). P1-FadL or FadL-P1 where overexpressed in non-
CEACAM binding E. coli (Eco P1-FadL or Eco FadL-P1) and the bacteria were
subsequently incubated with equivalent amounts of GFP-tagged N-terminal
domain of human CEA (hCEA-N) or hCEA8-N (Figure III-3D). After washing, no
interaction with hCEA-N was detected for Eco P1-FadL or Eco FadL-P1 (Figure
III-3E). However, Eco P1-Hinf-aeg strongly bound hCEA-N (Figure III-3E).
Together, these results clearly indicate that the CEACAM-binding determinants of
OMP P1 are located on either moiety of the protein.
Multiple P1 loops form the CEACAM-binding interface
71
Figure III-3. A unique loop is not responsible of the targeting of CEACAMs by OMP P1. A. The solved structure of FadL (left panel) was retrieved from the database RCSB PDB (http://www.rcsb.org/pdb/home/home.do) and used to predict the three-dimensional architecture of Hinf-aeg OMP P1 (right panel). B. In the model of OMP P1, described in (A), highly variable portions of P1 are highlighted in red C. Schematic representations of FadL, P1 and the chimeras FadL-P1 and P1-FadL. Predicted extracellular loops are colored in black for FadL and in white for P1. D. Equivalent amounts of GFP-tagged N-terminal domain of human CEA (hCEA-N) or human
Multiple P1 loops form the CEACAM-binding interface
72
CEACAM8 (hCE8-N) in cell culture supernatants (supe) were verified by immunoblotting with anti-GFP antibodies. E. After overexpression of the indicated proteins in E. coli, the binding to hCEA-N or hCEA8-N was investigated by flow cytometry. The black lines indicate bacteria incubated with plain cell culture medium.
The highly variable loops of OMP P1 are instrumental for CEACAM-
recognition
The chimeras between OMP P1 and FadL have strongly suggested that a single
extracellular loop of OMP P1 may not be sufficient and that loops located at either
end of OMP P1 may be necessary for CEACAM binding. To determine individually,
which loop contributes to the binding interaction, we replaced single loops of OMP
P1 by a short antibody epitope PEYFK (Figure III-4A). Accordingly, we generated
seven OMP P1 mutants lacking distinct extracellular loops (OMP P1-∆L1 – OMP
P1-∆L7) and expressed them individually in E. coli. Similar to the wild-type OMP
P1, all the 7 mutants were detected in outer membrane preparations of E. coli
(Figure III-4B). CEACAM-binding assays revealed that, with the exception of
buried loops L2 and L5, all other loops of OMP P1 contributed to the interaction
with CEACAM (Figure III-4C).
Strikingly, the most prominent extracellular loop of P1-Hinf-aeg (L3) exhibited a
stretch, PEPQK, reminiscent of the antibody epitope PEYFK. Given the possible
practical advantages granted by the insertion of such a tag in a functional protein,
we substituted the proline-214 and the glutamine-215 residues of P1-Hinf-aeg by
tyrosine and phenylalanine, respectively. The mutant generated (P1-PQ-214-215-
YF) was expressed in E. coli outer membrane (Figure III-4D). Much to our
surprise, these double mutations, not predicted by protein modeling to change the
molecular architecture of P1-Hinf-aeg (data not shown), severely impaired its
binding to hCEA-N (Figure III-4E). Taken together, our data suggest that many
extracellular loops of OMP P1 are involved in the interaction with CEACAMs.
Multiple P1 loops form the CEACAM-binding interface
73
Figure III-4.The prominent extracellular loops of OMP P1 are necessary for CEACAM-binding. A. OMP P1 and the seven mutants carrying a specific loop deletion are schematically indicated. B. E. coli transformed with the empty pET28 vector (pET28) or pET28 carrying either OMP P1 (P1) or one of the 7 variants with loop substitutions (P1 ∆L1 — P1 ∆L7) were induced with IPTG. Their outer membrane proteins were subsequently immunoblotted with anti-P1 serum (upper panel). Similar amounts of protein in the different samples were verified by Coomassie staining of the membrane (lower panel). C. The ability of the wild-type P1 and its seven mutants to bind CEACAMs was estimated in pull-down assays with GFP-tagged soluble proteins. No binding is represented by “─”, weak binding by “+”, strong binding by “++”, and a very good by “+++”. D. OMP preparations of E. coli transformed with the empty pET28 vector (Eco pET28) and E. coli expressing P1 of Hinf-aeg (Eco P1-Hinf-aeg) or P1-Hinf-aeg mutant PQ-214-215-YF (Eco P1- PQ-214-215-YF) were probed with anti-P1 serum (upper panel). The membrane was also Coomassie stained to verify the equivalent amounts of protein in the different samples (lower panel). E. Flow cytometry was used to analyze the binding of Eco P1-Hinf-aeg or Eco P1- PQ-214-215-YF to the indicated CEACAMs. The fold change in MFI after hCEA-N–incubation compared to hCEA8-N is indicated for each sample.
Multiple P1 loops form the CEACAM-binding interface
74
Restriction of loop flexibility in OMP P1 abrogates CEACAM-binding
The structure of Neisseria gonorrhoeae Opa60 revealed that the loops of this OMP
are extraordinary flexible (Fox et al., 2014). In agreement with the assumptions of
Fox and colleagues, we speculated that extracellular loops of β-barreled
CEACAM-binding protein, including OpaCEA and OMP P1, function like a
mechanical clamp to tightly encircle their receptor from multiple angles. To
rigorously test this hypothesis, we took advantage of a 1.14 nm-wide crosslinking
agent bis-(sulfosuccinimidyl)-suberate (BS3) to constrain loop flexibility of OpaCEA
or OMP P1. Alternatively, the bacteria were also incubated with succinimidyl-2-
(biotinamido)ethyl-1,3-dithiopropionate (NHS-SS-Biotin). NHS-SS-Biotin, as well
as BS3, likely formed amide bonds with the primary amino groups (-NH2) of
multiple lysine residues present on loop 1, 3, 4, 5 and 7 of P1-Hinf-aeg or loops 1,
2 and 3 of OpaCEA. Interestingly, biotinylated OpaCEA-expressing Neisseria
gonorrhoea or Hinf-aeg was still able to bind GFP-tagged N-terminal domain of
human CEACAM3 (hCEA3-N) but not as good as the untreated bacteria (Figure
III-5A and B). Conversely, pre-treatment with BS3 completely blocked the
recognition of hCEA3-N by OpaCEA or P1-Hinf-aeg (Figure III-5A and B). Similar
amounts of hCEA3-N and hCEA8-N, used in the binding assays, were verified by
Western blotting (Figure III-5C).
Importantly, BS3 treatment did not abrogate CEACAM-binding of Moraxella
catarrhalis (Figure III-5D), a pathogen, which employs a linear amino acid motif in
the stalk of its UspA1 protein for receptor engagement (Hill et al., 2005).
Furthermore, BS3 consistently blocked the interaction between P1-Hinf-aeg and
CEACAM in the E. coli background (Figure III-5E); whereas, the uropathogenic E.
coli AfaE-III, targeting CEACAM via the multimeric immunoglobulin-like protein
AfaE-III (Anderson et al., 2004), still bound hCEA3-N after treatment with BS3
(Figure III-5E). Collectively, these results demonstrate that OMP P1 binding is
sensitive to agents compromising the flexibility of extracellular loops and suggest a
novel clamp mode of interaction between the flexible, extracellular loops of the
Hinf adhesin and host surface receptors of the CEACAM family.
Multiple P1 loops form the CEACAM-binding interface
75
Figure III-5. Loop flexibility in OMP P1 is essential to CEACAM recognition. A. OpaCEA expressing Neisseria gonorrhoeae (Ngo OpaCEA) was left untreated (left panel) or incubated with 2 mM of NHS-SS-Biotin (middle panel) or crosslinker (BS3) (right panel) prior binding assays with the indicated soluble GFP-fused N-terminal domains of human CEACAM3 (hCEA3-N) or CEACAM8 (hCEA8-N). Bacterium-associated fluorescence, as measured by flow cytometry, was used as readout to quantify the interaction of the bacteria with CEACAMs. The fold change in bacterial MFI upon incubation with hCEA3-N in comparison to hCEA8-N is indicated for each sample. The black lines indicate bacteria incubated with plain cell culture medium. B. Similar assays as described in (A) were performed with Hinf-aeg. C. Comparable levels of hCEA3-N or hCEA8-N, in cell culture supernatants (supe) used in A and B, were verified by immunoblotting with monoclonal anti-GFP antibody. D. Cell culture supernatants, containing similar amounts of GFP alone or CEACAM3 N-terminal domain (hCEA3-N) fused to GFP were incubated with bacteria pretreated or not with BS3. The CEACAM-binding of Hinf-aeg (top panel) or Moraxella catarrhalis (bottom panel) was detected with anti-GFP antibodies. E. Eco P1-Hinf-aeg (top panel) or Eco AfaE-III (bottom panel) was treated or not with BS3 prior binding assays with hCEA1-N or hCEA8-N. GFP signals characteristic of CEACAM-binding were detected with anti-GFP antibodies.
Multiple P1 loops form the CEACAM-binding interface
76
Modelling of the OMP P1-CEA complex
Fifteen years ago, Virji and colleagues, in a seminal study, have characterized one
of the two β-sheets of CEACAM1 N-terminal domain (the CFG-face) as the
primary target of CEACAM-binding Hinf (Virji et al., 2000). Furthermore, mutational
analyses identified amino acid residues Y-34 and I-91 as necessary for the
interaction of most Hinf strains with CEACAM1 (Virji et al., 2000). On the other
hand, our current findings underline the importance of P-214 and Q-215 residues
in Hinf OMP P1-recognition of CEACAMs (Figure III-4E). Therefore, to get an
impression of the bimolecular complex P1-Hinf-aeg – CEACAM1-N, the webserver
SwarmDock (http://bmm.cancerresearchuk.org/~SwarmDock/) was used. The
residues P-214 and Q-215 or Y-34 and I-91 were set as active sites in P1-Hinf-aeg
or CEACAM1-N, respectively. The predicted three-dimensional atomic architecture
of the bimolecular complex P1-Hinf-aeg – CEACAM1-N (Figure III-6) makes highly
plausible that many flexible loops of OMP P1 are engaged in the interaction with
CEACAM1-N and warrants further studies to co-crystalize OMP P1 and
CEACAM1-N.
Figure III-6. A putative model of the bimolecular complex CEACAM1-N – P1-Hinf-aeg.The interaction between the amino-terminal domain of human CEACAM1 (red) and the outermembrane protein P1 of Hinf-aeg (blue) was modeled using the docking server SwarmDock.
Multiple P1 loops form the CEACAM-binding interface
77
III.4 Discussion
An unbiased study has identified the outer membrane protein (OMP) P1 as the
CEACAM-binding adhesin of Hinf (Tchoupa et al., 2015). OMP P1 is predicted to
build a 14-stranded -barrel with 7 extracellular loops. The most prominent
surface-exposed loops of OMP P1 (L1, L3, L4 and L7) were also the most variable
amongst Hinf strains. Although this high versatility of their CEACAM-binding
adhesin, all the OMP P1-proficient Hinf strains analyzed to date were able to
interact with CEACAMs (Tchoupa et al., 2015). Therefore, we hypothesized that
OMP P1 might have a conserved function in all the members of the
Pasteurellaceae family.
In this study, by aligning the sequences of OMP P1 from several Pasteurellaceae
family members against the crystal structure of E. coli FadL, we have
demonstrated that the OMP P1 is structurally conserved in Pasteurellaceae.
However, except Hinf, no additional Pasteurellaceae species was found to bind
several human CEACAMs tested. Even though the unique features that confer a
CEACAM-binding to Hinf OMP P1 are still to be identified, structure-activity
investigations undertaken here highlight the relevance of several flexible loops for
the recognition of CEACAM by OMP P1.
Neisseria meningitidis is another inhabitant of the human nasopharynx interacting
with CEACAMs via the opacity-associated (Opa) proteins (Kuespert et al., 2011).
Within the Neisseria genus, the CEACAM-binding capability is not restricted to
species with a high pathogenic potential, such as N. meningitidis and N.
gonorrhoeae, but also shared by strictly commensal species, including N.
lactamica and N. subflava (Toleman et al., 2001). Contrastingly, Haemophilus
influenzae is the unique species of the genus Haemophilus that is able to interact
with human CEACAMs. Indeed, our current data reveal that less pathogenic
Haemophilus species (H. parinfluenzae and H. haemolyticus) are unable to bind
CEACAMs. Therefore, it is enticing to think that the recognition of CEACAMs by
Hinf may provide the bacterium with a selective advantage to outcompete other
Haemophilus species of the same ecological niche. Further studies with Hinf OMP
P1 as well as P1 of commensal species of Haemophilus, all expressed in a
heterologous background (e.g.: Escherichia coli), would elucidate in suitable
Multiple P1 loops form the CEACAM-binding interface
78
animal models the relevance of P1 – CEACAM interaction for a successful
colonization.
An additional human-restricted colonizer of the respiratory tract, which tightly
interacts with CEACAMs, is Moraxella catharralis (Hill and Virji, 2003). Moraxella
catharralis utilizes a stretch of 20 amino acids, within the stalk of its UspA1
adhesin, to bind CEACAMs with a high affinity (Conners et al., 2008). Consistently,
Moraxella catharralis strains deprived of this CEACAM-binding motif in their
UspA1 proteins (e.g.: strain O35E) are unable to interact with CEACAMs (Brooks
et al., 2008). However, such a linear motif is so far unknown in Hinf OMP P1. The
only non-CEACAM binding Hinf strain we have identified so far (strain 7502) is
totally deficient in OMP P1 (Tchoupa et al., 2015). Furthermore, our previous work
demonstrated that Hinf strains targeting one, two or three members of the
CEACAM family exhibited highly variable extracellular loops irrespective of their
CEACAM-binding profiles (Tchoupa et al., 2015).
Non-CEACAM binding OMP P1 of H. parainfluenzae (P1-Hpara) differs
significantly from P1-Hinf-aeg within the most prominent loops (L1, L3, L4 and L7).
Importantly, the deletion of each of these loops disrupts the interaction of P1-Hinf-
aeg with CEACAMs. The necessity of multiple loops to target the same receptor
may be puzzling at the first sight; however, other -barrel shaped CEACAM-
binding adhesins, such as the neisserial Opa proteins, combined the hypervariable
regions of their loops 2 and 3 (HV-1 and HV-2, respectively) to build the CEACAM-
binding interface (de Jonge et al., 2003). Intriguingly, the combination HV-1 — HV-
2 is so specific that a chimeric protein, made of HV-1 and HV-2 originated from
two different CEACAM-binding Opa proteins, is unable to interact with CEACAMs
(de Jonge et al., 2003). In a similar fashion, several variable loops of P1-Hinf-aeg
may cooperate in order to bind CEACAMs. Nevertheless, it is plausible that each
loop does not contribute to a similar extend to the CEACAM-binding. The loss of
the CEACAM-recognition after the substitution of the residues Pro-214 and Gln-
215 in the loop 3 of P1-Hinf-aeg supports this hypothesis. Furthermore, some of
the variations in the extracellular loops appears to be under positive selection
(Mes and van Putten, 2007). Especially in OMP P1 loop 1 and loop 3, both of
which are required for CEACAM binding, particular amino acids seem to be
Multiple P1 loops form the CEACAM-binding interface
79
preferred (Mes and van Putten, 2007), which could be due to their involvement in
receptor recognition.
Considering their size, 5 extracellular loops of Hinf OMP P1 (L2, L3, L4, L5 and
L6) are reminiscent of E. coli homolog FadL loops (van den Berg et al., 2004)
whereas two loops (L1 and L7) are relatively huge in P1. Since the flexibility of a β-
barrel outer membrane protein’s loop is often linked to its length (Fox et al., 2014),
the diverse conformations sampled by the prominent loops 1 and 7 may be
instrumental in stabilizing the complex OMP P1-CEACAM. Accordingly, chimeras
lacking L1 or L7 of P1, or P1 mutants depleted in L1 or L7 are unable to target
CEACAMs. More than the (variable) sequences of the huge loops of P1, the
various conformations conferred by the loop flexibility may promote the recognition
of the CEACAM receptor, as it was speculated for Opa proteins (Fox et al., 2014).
By restricting the flexibility of OpaCEA and P1-Hinf-aeg with chemical cross-linkers,
we abrogated their interaction with CEACAMs. Interestingly, biotinylated OpaCEA
and P1-Hinf-aeg exhibited a reduced interaction with CEACAMs, probably due to
the steric hindrance generated by the 2.34 nm-spacer arm of the biotin used.
Haemophilus influenzae constantly faces the challenge of (i) varying the surface-
exposed regions of its outer membrane protein P1 to escape the immune
response of its human host while (ii) preserving the advantage inherent in the
interaction of P1 with its human receptor CEACAM. This constraint may build the
framework within which the amino acid sequences of OMP P1 extracellular loops
are evolving. Therefore, co-crystal structures of Hinf OMP P1 variants in
bimolecular complexes with an identical CEACAM family member may represent
the turning point in the development of innovative, adhesion disrupting, therapeutic
strategies to combat Haemophilus influenzae.
III.5 Materials and methods
Multiple sequence alignment
OMP P1 sequences of Hinf-aeg ATCC 11116, Haemophilus parainfluenzae ATCC
33392, H. haemolyticus M19501, H. pittmaniae HK85, Mannheimia heamolytica
PKL10, as well as the FadL sequence of E. coli K12 were retrieved from Uniprot
(http://www.uniprot.org/). The PDB file of the crystal structure of E. coli FadL was
Multiple P1 loops form the CEACAM-binding interface
80
downloaded from the protein data bank RCSB (http://www.rcsb.org/pdb/home
/home.do). The different P1 and FadL sequences were aligned using the Expresso
application of the T-Coffee web server (http://tcoffee.crg.cat/apps/tcoffee/do:
expresso) and the solved structure of FadL. The phylogenic tree with the different
P1 or FadL OMPs was edited with iTOL (http://itol.embl.de/).
Bacterial strains and growth conditions
Hinf-aeg ATCC 11116, H. parainfluenzae ATCC 33392, Mannheimia haemolytica
ATCC 33396 and Moraxella catarrhalis ATCC 25238 were obtained from German
Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany).
H. haemolyticus NCTC 10659 was purchased from PHE (Public Health England,
Salisbury, UK). OpaCEA-expressing, non-piliated Neisseria gonorrhoeae MS11-
B2.1 (strain N309) is a generous gift of Prof. Thomas F. Meyer (Max-Planck
Institut für Infektionsbiologie, Berlin, Germany). Escherichia coli AfaE-III is the
courtesy of Prof. Ulrich Dobrindt (University of Münster, Münster, Germany). With
the exception of Neisseria gonorrhoeae and Escherichia coli AfaE-III, all the
bacteria were grown on brain-heart infusion agar supplemented with 10%
Levinthal base. Neisseria gonorrhoeae was grown on GC agar plates
supplemented with vitamins while Escherichia coli AfaE-III grew on Luria-Bertani
(LB) agar plates. All the bacteria were grown at 37°C in 5% CO2.
Outer membrane proteins isolation
After IPTG induction, proteins of the E. coli outer membrane were isolated as
described elsewhere (Roier et al., 2012).
Expression of outer membrane proteins in E. coli.
Chromosomal DNA of Hinf-aeg or E. coli K12 was used as template to amplify
ompP1 or fadL, respectively. The respective genes were then cloned into the NcoI
and XhoI site of pET28a (Novagen). E. coli Rosetta (DE3) (Novagen) was
subsequently transformed with the pET28a vectors encoding FadL or OMP P1,
and induced for protein expression by IPTG for 3 hours. LB medium supplemented
with kanamycin was used to grow all the transformed E. coli strains at 37°C.
ompP1 mutagenesis
OMP P1 chimeras and mutants were generated via SOEing PCRs (Horton et al.,
1990). For the chimeras, the primers were designed to fuse in-frame the N-
Multiple P1 loops form the CEACAM-binding interface
81
terminal moiety of P1 with the C-terminal moiety of FadL (P1-FadL) or inversely
(FadL-P1). In loop-depleted P1 mutants, the FIVE-epitope tag (PEYFK) coding
sequence was used as overlap extension to replace every single loop. In the P1-
PQ-214-215-YF mutant, SOEing PCRs were also used to replace the Pro-214 and
Gln-215 residues in P1-Hinf-aeg by Tyr and Phe, respectively. The Table 2 lists
the primers used.
Name Sequence (5’ - 3’) Product
P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1-FadL P1-FadL-reverse CATTTCAGGCAGGTTCAGCGTCAGATTACCTTTTTTACCTTCTTTGATG P1-FadL-forward CTGACGCTGAACCTGCCTGAAATG FadL- XhoI-reverse GCGCGGCTCGAGTCAGAACGCGTAGTTAAAGTTAGTACC FadL- NcoI-forward GGACGCCATGGATGAGCCAGAAAACCCTGTTTAC
FadL-P1 FadL-P1-reverse
GTAATCTGGCAATTTAAGGGTTAAATAACCCGATTGCGTTGCGCCACCTGTC
FadL-P1-forward TTAACCCTTAAATTGCCAGATTAC P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop1
P1 ∆Loop1-reverse GCCCTCTATCTTGAAGTACTCGGGTACATCACCATTCATATTAATTC P1 ∆Loop1-forward CCCGAGTACTTCAAGATAGAGGGCGGCTCAGCTTC P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop2
P1 ∆Loop2-reverse CAAGTCAGTTTTGAAGTACTCGGGTTCACTTTTTAGACCGAAATTGACATTCATTCC
P1 ∆Loop2-forward CCCGAGTACTTCAAAACTGACTTGAGTGCTATC P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop3
P1 ∆Loop3-reverse GGGCAAATACTTGAAGTACTCGGGAACACTATCCGCAATAATACC P1 ∆Loop3-forward CCCGAGTACTTCAAGTATTTGCCCTCTAAAGACACATCTG P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop4
P1 ∆Loop4-reverse ACCTTTTTTGAAGTACTCGGGGGCAGTGCGGTCAGTAAAATCAATGTCC P1 ∆Loop4-forward CCCGAGTACTTCAAAAAAGGTGATTTAACCCTTACATTGCC P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop5
P1 ∆Loop5-reverse TTCTTTATCCTTGAAGTACTCGGGGCTGGCATGTAATTTTGTTAAACG P1 ∆Loop5-forward CCCGAGTACTTCAAGGATAAAGAATTGCAATACAATAATAAC P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop6
P1 ∆Loop6-reverse GGTATCTGGCTTGAAGTACTCGGGTTGATCGTAAGCAATACCCGCGCGTAAGG
P1 ∆Loop6-forward CCCGAGTACTTCAAGCCAGATACCGATCGCACTTGGTATAG P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1 ∆Loop7
P1 ∆Loop7-reverse TGCATTTGTCTTGAAGTACTCGGGTACTTCTTTAAAGTGAACTTTTTTGCC
P1 ∆Loop7-forward CCCGAGTACTTCAAGACAAATGCAAATTATACTTCTCAAGCACATGCAAATC
P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG P1-NcoI_forward GCGCCGCCATGGATGAAAAAATTTAATCAATCTC
P1-PQ-214-215-
YF
P1-PQ-YF-reverse ATACTCAGCAATTGTTTTGAAGTATTCTGGCAGTACTGTGGATAC P1-PQ-YF-forward TACTTCAAAACAATTGCTGAGTATTTGAC P1-XhoI_reverse GCGCGGCTCGAGCGCGTTAGAAACTATAATTTAAGTTTAAACCGTAAAG
Table 2. Primers used in this study
Multiple P1 loops form the CEACAM-binding interface
82
Binding assays with soluble CEACAMs
Soluble GFP-tagged CEACAM ectodomains were produced by 293 cells and
subsequently used in bacterial binding assays as described elsewhere (Kuespert
and Hauck, 2009; Kuespert et al., 2007). Shortly, soluble CEACAM-N-GFP
proteins, contained in cell culture supernatant, were clustered overnight with
polyclonal anti-GFP antibodies at 4°C. Afterwards, 10 × 108 bacteria were added
in 1 ml supernatant containing the indicated GFP-fusion proteins and incubated for
2 h at 37 °C under gentle rotation. After extensive washing with PBS, the samples
were either immunoblotted with a monoclonal GFP antibody or processed for flow
cytometry to detect bacteria-associated fluorescence. In some cases, prior to
bacterial pull-down with CEACAMs, the bacteria were treated with either 2 mM
bis(sulfosuccinimidyl) suberate (BS3) (Thermo Scientific) or succinimidyl-2-
(biotinamido)ethyl-1,3-dithiopropionate (NHS-SS-Biotin) (Thermo Scientific)
according to the manufacturer’s instructions.
Western blotting and used antibodies
Immunoblotting was performed as reported earlier (Hauck et al., 2001) using a
monoclonal antibody (mAb) against GFP (clone JL-8; Clonetech) or polyclonal
mouse sera against OMP P1 (anti-P1) (Roier et al., 2014).
83
CHAPTER IV
The Haemophilus influenzae adhesin
OMP P1 is regulated by host-derived
fatty acids
Arnaud Kengmo Tchoupa1 and Christof R. Hauck1,2
1Lehrstuhl für Zellbiologie, Universität Konstanz, Konstanz, Germany, and 2Konstanz Research School Chemical Biology, Universität Konstanz, Konstanz,
Germany
In preparation.
Host-derived fatty acids modulate P1 levels
84
IV.1 Abstract
Haemophilus influenzae (Hinf) is a human-restricted colonizer of the upper
respiratory tract. Hinf outer membrane protein (OMP) P1 specifically binds to
carcinoembryonic antigen-related cell adhesion molecules (CEACAMs) present on
the apical surface of human epithelial cells. Strikingly, OMP P1 is a homolog of
long-chain fatty acid (LCFA) transporter FadL of Escherichia coli, which forms a
14-stranded β-barrel in the outer membrane. Whereas the role of FadL as LCFA
transporter and the regulation of FadL expression by environmental LCFAs have
been comprehensively deciphered during the last decades, a role of OMP P1 in
LCFA transport by Hinf is unknown. By employing an isogenic OMP P1 mutant, we
report that LCFAs drive Hinf growth in an OMP P1-independent fashion. However,
LCFAs appear to be potent stimulators of OMP P1 expression. Increased levels of
OMP P1 were functionally relevant as LCFA-treated bacteria showed improved
CEACAM-binding in vitro. As a result, Hinf invasion into CEACAM-expressing
epithelial cells was strongly increased upon exposure to host-derived lipids.
Interestingly, Hinf were found to be taken up by primary human granulocytes in an
opsonin-independent manner. Most importantly, the bacteria fed with arachidonic
acid and recognizing granulocyte CEACAM3 were clearly more phagocytosed.
Thus, our data provide the first evidence of the regulation of a bacterial adhesin by
host-derived lipid signals and provide impetus to study the effects of free fatty
acids on Hinf pathogenesis.
IV.2 Introduction
In Gram-negative bacteria, the outer membrane (OM) is the outermost layer of the
bacterial cell and the dynamic interface between the bacterium and its surrounding
environment, including higher animals and plants. The OM is of prime importance
for the bacterial survival as it has been linked to diverse functions such as the
uptake of nutrients (Nikaido, 2003), the adaptation in host niches (Lin et al., 2002)
and the resistance to antibacterial molecules (Pages et al., 2008). Our improved
understanding of OM importance is fuelled by detailed insights into its structure
and composition. The OM is an asymmetric bilayer with phospholipids in its inner
leaflet and lipopolysaccharides (LPS) in its outer leaflet. Whereas lipoproteins are
Host-derived fatty acids modulate P1 levels
85
anchored to the inner leaflet of the OM, β-barrel proteins are integral constituents
of the lipid bilayer (Nikaido, 2003).
β-barrel OM proteins (OMPs) build channels, through which hydrophilic molecules,
including most nutrients, move from the extracellular milieu into the periplasm.
OMPs allowing influx of structurally unrelated solutes are referred to as non-
specific channels or porins. On the contrary, for the specific channels, the
presence of binding sites for incoming molecules within the β-barrel facilitates the
diffusion of defined ligands (Nikaido, 2003). Specific channels are involved in the
ligand gated transport of various molecules, including small metal ions (Hohle et
al., 2011), a large variety of sugars (Nikaido, 2003) and fatty acids (van den Berg,
2005).
In Escherichia coli, more than four decades ago, Hill and Angelmaier
characterized a mutant of E. coli that was proficient in the degradation system of
long-chain fatty acids (LCFAs), but unable to metabolize the octadecanoic, oleic
acid. These authors suggested therefore the existence of a gene product, which
facilitates the diffusion of fatty acids through the outer membrane. (Hill and
Angelmaier, 1972). Six years later, this gene was identified and named fadL (Nunn
and Simons, 1978). The fadL gene product (FadL), a protein that locates within the
outer membrane of E. coli, was shown to be essential for long-chain fatty acid
transport (Black et al., 1987).
FadL belongs to the fatty acid degradation (fad) regulon genes, which are induced
when LCFAs (with a chain >C12) are present in the bacterial growth medium.
Explicitly, after their transport through the OM by FadL, the LCFAs cross the inner
membrane via a hitherto unknown mechanism (presumably by free diffusion). The
cytosolic LCFAs are activated into their acyl-CoA thioesters by the acyl-CoA
synthetase FadD, which is loosely associated with the inner leaflet of the inner
membrane (Black et al., 1992). The acyl-CoA thioesters bind to FadR, a repressor
of the entire fad regulon. The ensuing structural changes in FadR, upon acyl-CoA
binding, dissociate the protein from its operator sites and increase the fad regulon
expression (Feng and Cronan, 2012). Besides, the fad regulon is also modulated
by the global transcription factor ArcA (Cho et al., 2006) and the global regulator
CRC-cAMP complex (Feng and Cronan, 2012).
Host-derived fatty acids modulate P1 levels
86
Undeniably, the major breakthrough in the study of FadL was the elucidation of its
three-dimensional architecture (van den Berg et al., 2004). FadL is a 14-stranded
β-barrel protein, whose lumen is occluded by an amino-terminal hatch domain
functioning as a plug (van den Berg et al., 2004). A detailed understanding of the
FadL structure suggested that the LCFAs are transported by FadL via a so-called
lateral diffusion (Hearn et al., 2009). During lateral diffusion, LCFAs utilize the
extracellular part of FadL to circumvent the LPS polar barrier. Once inside the β-
barrel, LCFAs do not diffuse across the channel, but rather use a lateral opening in
the wall of FadL to get into the lipid bilayer from where they diffuse freely into the
periplasm (Hearn et al., 2009).
Almost a quarter century ago, Paul Black, struck by the similarity between FadL
and Haemophilus influenzae (Hinf) outer membrane protein (OMP) P1, has
hypothesized that OMP P1 may be involved in LCFAs transport (Black, 1991). He
speculated that, similar to FadL, OMP P1 expression may be induced by LCFAs or
some other nutrients (Black, 1991). However, no experimental evidences
supported his assumptions. Recently, OMP P1 was identified to play a role in Hinf
pathogenesis (Tchoupa et al., 2015). By binding to the carcinoembryonic antigen-
related cell adhesion molecules (CEACAMs), which are immunoglobulin-like
glycoproteins expressed by epithelial cells in the human respiratory tract and on
other mucosal surfaces, OMP P1 promotes the intimate attachment of Hinf to its
host cells (Tchoupa et al., 2015).
These novel insights into the pathophysiological importance of OMP P1 have
revived the need to understand, whether OMP P1 is involved in uptake of LCFAs
by Hinf and to what extent environmental LCFAs regulate OMP P1 levels. Here we
report that enrichment of Hinf growth medium with LCFAs accelerated not only
bacterial growth, but also upregulated the OMP P1 expression. The elevated
levels of OMP1 in the outer membrane were functionally relevant, as the bacteria
grown in LCFA-supplemented media exhibited an increased interaction with their
cognate receptors and showed increase invasiveness into CEACAM-expressing
epithelial cells. Our investigations suggest that Haemophilus pathogenicity is
guided by environmental cues present inside the host.
Host-derived fatty acids modulate P1 levels
87
IV.3 Results
Haemophilus influenzae growth is improved by long-chain fatty acids
The Rd strain of Hinf was previously shown to grow well in a lipid-depleted version
of the otherwise lipid-rich BHI (brain-heart infusion) medium (Wang and Cronan,
2003). Furthermore, the fatty acid composition of the bacteria remained identical
irrespective of the medium where they were grown (lipid-depleted or not) (Wang
and Cronan, 2003). In that context, we wondered whether Hinf modifies its growth
speed in response to increased concentration of fatty acids in the medium.
Therefore, Hinf strain Rd was grown in BHL broth (BHI + levinthal base)
supplemented with oleic acid, a monounsaturated octadecanoic acid (18:1, cis-9).
Clearly, Rd division rate was accelerated in the presence of oleic acid (Figure
IV-1A). Furthermore, this effect was concentration-dependent as the multiplication
of the bacteria increased with the amount of oleic acid added in the medium
(Figure IV-1A). The obvious stimulation of Hinf growth by oleic acid strongly
suggests that this LCFA is crossing the outer membrane and contributes to
bacterial metabolism. In E. coli, FadL is the principal LCFA transporter (Lepore et
al., 2011). Given the striking similarity between E. coli FadL and Hinf OMP P1, we
analyzed growth of the OMP P1-deficient Rd mutant (Hinf Rd ∆P1) in BHL broth
enriched with oleic acid. Similar to the wild-type Rd strain, the OMP P1 mutant
grew faster upon supplementation of medium with LCFA (Figure IV-1B). Taken
together these data demonstrate that LCFAs are able to boost the growth of Hinf
in an OMP1-independent manner.
Host-derived fatty acids modulate P1 levels
88
OMP P1 is upregulated by long-chain fatty acids
Apart from the stimulation of bacterial growth, addition of LCFAs in the medium
has been consistently shown to induce expression for most genes involved in fatty
acid degradation, including fadL (Dirusso and Black, 2004; Overath and Raufuss,
1967). Consequently, we compared the expression levels of the FadL homolog in
Hinf (OMP P1) upon bacterial growth in BHL broth enriched with oleic acid.
Interestingly, OMP P1 was more abundant on the surface of Hinf Rd once the
medium was supplemented with oleic acid (Figure IV-2A). Next, we sought to
determine if the upregulation of OMP P1 was a general feature shared by all the
LCFAs. Accordingly, arachidonic acid, a polyunsaturated eicosanoic acid (20:4,
all-cis-5, 8, 11, 14), was added in the BHL broth. In accordance with OMP P1
levels after oleic acid-treatment, incubation of the bacteria in arachidonic acid
tremendously augmented their OMP P1 expression (Figure IV-2B). These
observations, made with the Hinf lab strain Rd, were confirmed using a non-
typeable Hinf (NTHi) clinical isolate (NTHi H86) which showed increased OMP P1
levels with arachidonic acid (Figure IV-2B). Combined, these results strongly
suggest that many LCFAs, if not all, induce OMP P1 expression in Hinf. Moreover,
typeable and non-typeable Hinf are responsive to exogenous LCFAs.
Figure IV-1. Oleic acid stimulates Hinf growth in an OMP P1-independent manner.A. Hinf Rd growth was monitored by reading the optical density at 600 nm (OD600) in BHL brothsupplemented or not with oleic acid. B. As in (A), the growth of Hinf Rd OMP P1-deficient mutant(Hinf Rd ∆P1) was also monitored in the presence or not of oleic acid.
Host-derived fatty acids modulate P1 levels
89
Lipid-induced expression of OMP P1 enhances the binding of Hinf to human
CEACAMs
To date, the only experimentally validated role of Hinf OMP P1 is the binding to
human CEACAM family members in order to promote tight adhesion to host cells
(Tchoupa et al., 2015). In contrast to previous assumptions by Hill and colleagues
(Hill et al., 2001), we could not detect a direct CEACAM-binding function for OMP
P5, but found that OMP P5 facilitates the proper surface display of OMP P1 within
the outer membrane (Tchoupa et al., 2015). LCFAs and their effect on OMP P1
levels provided us with new tools to critically test our previous findings. After
arachidonic acid-treatment and intensive washing, Hinf Rd, its OMP P5 mutant or
its OMP P1 mutant was tested for CEACAM-binding with cell culture supernatants
enriched with similar amount of soluble GFP-tagged amino-terminal domain of
human CEACAM1 (hCEA1-N) or human CEACAM8 (hCEA8-N, negative control)
(Figure IV-3A). OMP P1 mutant (∆P1) failed to recognize hCEA1-N whereas Rd
wild-type strain (WT) and OMP P5-deficient mutant (∆P5) clearly bound hCEA1-N
Figure IV-2. Long-chain fatty acids increase OMP P1 levels of typeable and non-typeableHinf. A. Hinf Rd treated or not with 3.6 mM oleic acid were analyzed by flow cytometry followingstaining with -P1 serum in combination with a Dylight® 488-conjugated secondary antibody. B.After incubation with 1mM arachidonic acid, the OMP P1 levels in Hinf Rd (left panel) or NTHi H86(right panel) were analyzed as in (A). The DMSO-treated bacteria (black lines) were included ascontrols.
Host-derived fatty acids modulate P1 levels
90
(Figure IV-3B). Interestingly, upon arachidonic acid-treatment, the wild-type strain
engaged more hCEA1-N while the interaction of OMP P5 mutant with hCEA1
remained unchanged (Figure IV-3B). Consistent with their respective CEACAM-
binding capabilities, OMP P5-deficient mutant only showed a modest OMP P1
induction by LCFAs, whereas the wild-type Rd strain massively responded to
LCFAs by increasing its OMP P1 levels (Figure IV-3C). Taken together, these
results indicate that: (i) the enhanced CEACAM-binding of Hinf in response to
LCFA-treatment is due to increased levels of OMP P1 as a mutant genetically
deficient in OMP P1 is unable to interact with CEACAMs before and after LCFA
stimulation, and (ii) OMP P5 seems to play a role in the regulation of OMP P1
levels in the outer membrane since the OMP P5 mutant failed to upregulate its
OMP P1 levels in a manner comparable to the wild-type strain.
Figure IV-3. LCFA-treatment improves the CEACAM-binding in wild-type Hinf but not in the OMP P5-deficient mutant. A. Soluble GFP-tagged N-terminal domain of human CEACAM1 (hCEA1-N) or CEACAM8 (hCEA8-N) were harvested in cell culture supernatants of 293 cells. Similar amounts were verified by immunoblotting with anti-GFP antibodies. B. hCEA1-N and hCEA8-N, produced as described in (A) were incubated with Hinf Rd wild-type (WT), its OMP P5- (∆P5) or OMP P1-deficient mutant (∆P1) treated or not with arachidonic acid beforehand. After washing, the interaction with CEACAMs was determined upon anti-GFP Western Blot. C. OMP P1 levels in P5 mutant (Hinf Rd ∆P5, left panel) and wild-type (Hinf Rd, right panel), after incubation with arachidonic acid and binding assays with CEACAMs, were analyzed by flow cytometry upon staining with anti-P1 antibodies. The black lines indicate the staining of the DMSO-treated bacteria. The fold change in mean fluorescence (MFI) of LCFA-treated bacteria in comparison to DMSO is indicated for each sample.
Host-derived fatty acids modulate P1 levels
91
LCFA-treated Haemophilus influenzae invade epithelial cells more efficiently
Hinf adhesion to and invasion into epithelial cells were shown to correlate with
OMP P1 expression levels and CEACAM-binding (Tchoupa et al., 2015). By using
the clinical isolate NTHi H86, we wanted to determine if the increased OMP P1
levels upon incubation with LCFAs induce a better internalization of the bacterium
in a CEACAM-dependent manner. First, we tested the effect of LCFAs on NTHi
H86 recognition by soluble CEACAMs. In a gratifying similarity with the Rd strain,
NTHi H86 showcased an improved interaction with hCEA1-N upon treatment with
LCFAs (Figure IV-4A). To confirm the relevance of increased binding to soluble
CEACAMs in a cellular context, NTHi H86 treated or not with LCFAs was stained
with FITC. After FITC-labelling, the bacteria were used to infect 293 cells
transiently expressing HA-tagged full-length CEACAM1-4L or 293 cells transfected
with an empty control plasmid. Four hours post-infection, the fluorescence emitted
by the extracellular bacteria was quenched with trypan blue and the amount of
internalized FITC-labeled bacteria estimated by flow cytometry. Obviously, NTHi
H86 was unable to invade 293 control cells even after incubation with LCFAs
(Figure IV-4B). In clear contrast, the bacteria readily invaded CEACAM1-
expressing cells. However, the LCFA-treatment prior to infection significantly
improved invasiveness of NTHi H86 (Figure IV-4B). Collectively, these data
provide strong evidence that new OMP P1 molecules expressed upon addition of
LCFAs into the bacterial medium are functionally relevant, and promote a stronger
binding of the bacteria with their human receptor CEACAM and a better invasion
of epithelial cells in a CEACAM-dependent manner.
Host-derived fatty acids modulate P1 levels
92
Figure IV-4. Long-chain fatty acids increase the CEACAM-dependent invasiveness of Hinf. A. The CEACAM-binding of NTHi H86 was analyzed by flow cytometry after the treatment of the bacterium with DMSO (left panel) or arachidonic acid (right panel). The fold change in MFI of the bacteria binding to hCEA1-N (red line) is compared to hCEA8 (blue line, negative control). As an additional control, bacteria incubated with the plain cell culture medium were also included (black line). B. Upon incubation with either DMSO or arachidonic acid, NTHi H86 was FITC-labelled and used to infect 293 cells transiently expressing the full-length HA-tagged CEACAM1 (CEACAM1) or control 293 cells transfected with an empty vector (pcDNA3.1). Four hours post-infection, the infected cells were taken into suspension and trypan blue was utilized to quench the fluorescence of extracellular bacteria. 293 cells fluorescence, proportional to the quantity of internalized bacteria, was determined by flow cytometry. Relative values of the bacterial uptake are compared to DMSO-treated bacteria engulfed by CEACAM1-expressing cells. The data represent means ± standard deviation of three independent experiments (**P < 0.01).
Increased OMP P1 levels elicit a stronger innate immune response against
the bacteria
The airway epithelium expresses numerous inflammatory mediators in response to
colonizing Hinf. These cytokines induce the influx of various white blood cells,
including granulocytes (Rao et al., 1999). Human granulocytes possess a
Host-derived fatty acids modulate P1 levels
93
receptor, namely CEACAM3, which is able to induce the opsonin-independent
phagocytosis of bound bacteria (Roth et al., 2013). Recently, we reported that
Haemophilus influenzae biogroup aegyptius (Hinf-aeg) was binding to granulocyte
CEACAM3 (Tchoupa et al., 2015). We wondered whether CEACAM3-recognition
was contributing to efficient engulfment of Hinf by granulocytes. First, we
confirmed the interaction of Hinf-aeg with epithelial CEACAM1 and CEA, and with
granulocyte CEACAM3 ((Tchoupa et al., 2015); Figure IV-5A). Consistent with our
results with other Hinf strains, upon incubation with arachidonic acid, Hinf-aeg
clearly exhibited an improved interaction with the GFP-fused N-terminal domains
of CEACAM1, CEACAM3 and CEA (Figure IV-5B).
Figure IV-5. Hinf-aeg fed with LCFA augments its interaction with epithelial and granulocyte CEACAMs. A. Hinf-aeg was incubated with the indicated GFP-fused amino-terminal CEACAM domains (hCEAX-N). After extensive washing, the bacterial pellets were immunoblotted with anti-GFP antibodies (upper panel). Equivalent amounts of different hCEAX-N in cell culture supernatants (supe) were confirmed by western blotting (lower panel). B. Identical cell culture
Host-derived fatty acids modulate P1 levels
94
supernatants as in (A) were used to perform binding assays with Hinf-aeg that was pre-treated with DMSO or arachidonic acid. The interactions of bacteria with hCEAX-GFP were analyzed by flow cytometry. The fold change in MFI is indicated for each GFP-tagged N-terminal CEACAM domain and compares the CEACAM-binding of the LCFA-treated bacteria and the DMSO-treated bacteria
Next, FITC-labelled Hinf-aeg and the non-CEACAM3-binding NTHi H86 were used
to infect isolated human granulocytes. After 30 min of infection, the fluorescence
emitted by extracellular bacteria were quenched with trypan blue and the
fluorescence of the granulocytes upon infection was used as read-out to estimate
the amount of phagocytosed bacteria. Strikingly, the CEACAM3-binding Hinf-aeg
was more engulfed by the granulocytes in comparison to NTHi H86 (Figure IV-6A).
Moreover, in strict accordance with its better interaction with the soluble
CEACAMs, LCFA-treated Hinf-aeg was more phagocytosed by isolated human
granulocytes (Figure IV-6A and Figure IV-6B). Together, these data indicate that
CEACAM3 may play a major role in the opsonin-independent phagocytosis of
Hinf-aeg by granulocytes. Thus, increased OMP P1 levels upon LCFA-treatment
can be both beneficial (better binding to epithelial CEACAMs) and detrimental
(improved CEACAM3-recognition and faster engulfment by granulocytes).
Figure IV-6. Non-opsonic engulfment of Hinf-aeg by human granulocytes is potentiated bybacterial treatment with LCFA. A. Primary human granulocytes were infected for 30 min witheither Hinf-aeg or NTHi H86 at an MOI of 200. Prior to infection, the bacteria were incubated withDMSO or arachidonic acid and subsequently stained with FITC. After the infection, thefluorescence of extracellular and granulocyte-attached bacteria was quenched with trypan blue.Cell-associated fluorescence, correlating with the amount of taken up bacteria, was analyzed byflow cytometry. Uptake of DMSO-treated Hinf-aeg by human granulocytes was set to 100%. B.Phagocytosis assays, performed as described in (A), were extended to primary humangranulocytes from two additional donors. The isolated granulocytes were infected with DMSO- orLCFA-treated Hinf-aeg.
Host-derived fatty acids modulate P1 levels
95
IV.4 Discussion
Gram-negative bacteria, such as E. coli, are able to use fatty acids as a sole
source of carbon and as an energy source (Nunn and Simons, 1978). The
incoming fatty acids are either catabolized by the bacteria via the β-oxidation
pathway, or incorporated into complex membrane lipids (Overath and Raufuss,
1967). Whenever needed, the bacteria are also able to synthesize unsaturated
fatty acids. Given the high energy cost inherent in such a biosynthesis, the
bacteria must have evolved a finely tuned system to sense available exogenous
fatty acids and stop neosynthesis of fatty acids. In E. coli, the FadR protein
represses the entire set of fatty acid degradation (fad) regulon genes and activates
the fatty acid biosynthetic pathway (Feng and Cronan, 2012). Once the incoming
LCFAs are activated to acyl-CoAs, they bind FadR and abrogate its dual function
as repressor and activator (Iram and Cronan, 2005). Noticeably, the Hinf FadR
protein exhibits only poor acyl-CoA binding affinities (Iram and Cronan, 2005).
These findings question if the regulation of fatty acid degradation genes in Hinf in
response to exogenous fatty acids mimics the situation in E. coli.
In this study, we report that the E. coli FadL homolog, the Hinf outer membrane
protein P1 is induced by addition of LCFAs to the bacterial medium. Though OMP
P1 is not necessary for the increased growth of the bacteria in response to LCFA
supplementation, it nevertheless appears to be part of a fatty acid regulon.
Increased OMP P1 levels improve the exquisite interaction of Hinf with human
CEACAMs. Noteworthy, the genetic depletion of OMP P5 compromises the
increase of OMP P1 surface expression in response to LCFAs, strengthening our
previous hypothesis that OMP P5 may play a role in outer membrane integration
of OMP P1 (Tchoupa et al., 2015). As a consequence of LCFA-induced OMP P1
expression, Hinf invades epithelial cells more efficiently. At the same time, for Hinf
strains expressing a CEACAM3-binding OMP P1, the upregulation of OMP P1 by
LCFAs results in a more efficient phagocytosis by human granulocytes.
Strikingly, Hinf OMP P1 shares a high similarity with E. coli FadL (~39 % amino
acid identity). Three-dimensional modeling of OMP P1, based on the crystal
structure of FadL, predicts that OMP P1 possesses an opening within its β-barrel
(Tchoupa et al., unpublished observations). This opening was shown to be critical
for the lateral transport of LCFAs by FadL in E. coli (Lepore et al., 2011). Although
Host-derived fatty acids modulate P1 levels
96
our data demonstrate that the growth of OMP P1-deficient Hinf Rd mutant is
enhanced upon addition of oleic acid in the bacterial medium, we cannot exclude
that OMP P1 is involved in LCFA transport. Indeed, thanks to their amphipathic
character, LCFAs can freely diffuse across membranes without the need of a
protein-mediated transport (van den Berg et al., 2004). Once the bacteria, in
experimental settings are flooded with high amount of LCFAs, the free diffusion
may become the major mechanism of fatty acid uptake (Brash, 2001). In our
experimental settings, we supplemented the lipid-rich BHL broth with 3.6 mM oleic
acid. Further investigations should take advantage of the lipid-depleted variant of
Hinf growth medium (as described by (Wang and Cronan, 2003)) to test the effect
of increasing amounts of LCFAs on the growth of the P1-deficient Hinf Rd mutant
in comparison with the wild-type strain. Furthermore, physiological concentrations
of LCFAs (e.g.: within the micromolar range for arachidonic acid (Brash, 2001))
should be taken into account when deciphering the role of OMP P1 as a fatty acid
transporter in Hinf.
Whereas the physiological role of OMP P1 as LCFA transporter is still unclear,
recent findings have provided strong evidence that OMP P1 specifically interact
with human CEACAMs (Tchoupa et al., 2015). The relevance of OMP P1 in the
pathogenesis of Haemophilus influenzae is strengthened by a previous report
demonstrating that a natural OMP P1 deficient mutant, the NTHi strain 7502,
exhibits a compromised adherence to and invasion of bronchial epithelial cells
(Swords et al., 2000). Obviously, OMP P1-proficient strains of Hinf strongly attach
to CEACAM-expressing epithelial cells (Tchoupa et al., 2015; Virji et al., 2000).
Taking into account the findings presented in this manuscript, it is interesting to
mention that bacterial or viral infections of the mucosal surface are able to
stimulate the release of various pro-inflammatory cytokines initiating local
inflammations (Rao et al., 1999). It is a long standing observation that
inflammatory processes can trigger massive release of a LCFA from mammalian
cells, the arachidonic acid (Hammarstrom et al., 1975). Free arachidonic acid
molecules are converted in various bioactive eicosanoids, including
prostaglandins, leukotrienes and lipoxins (Harizi et al., 2008). Increased amounts
of eicosanoids during inflammation are, depending on the context of their
production, either pro-inflammatory or anti-inflammatory (Harizi et al., 2008).
Furthermore, a pro-inflammatory milieu also leads to a strong upregulation of
Host-derived fatty acids modulate P1 levels
97
CEACAMs, in particular CEACAM1 on epithelial and endothelial cells (Muenzner
et al., 2001, Muenzner et al., 2002). Therefore, inflammation might simultaneously
increase the presence of the Hinf adhesin as well as the abundance of the
pathogen’s receptor on the host side, which might tip the balance from harmless
colonization of a spatially restricted niche on the mucosa to invasive behavior of
the pathogen and systemic disease.
Interestingly, OMP P1 is not the only outer membrane protein of Haemophilus
influenzae, whose expression can be modulated by environmental cues. Classic
examples are provided by the transferrin binding protein (Tbp) A and B. The OMPs
TbpA and TbpB coordinately bind iron-saturated human transferrin in order to
extract and translocate iron into the bacterial periplasm (Gray-Owen and
Schryvers, 1993). The genes tbpA and tbpB are expressed as a single
transcriptional unit under the control of a ferric uptake regulatory (Fur-like)
repressor protein and their expression is suppressed by environmental iron (Rao
et al., 1999).
Importantly, arachidonic acid-derived leukotrienes, also produced by epithelial
bronchial cells (Jame et al., 2007), are powerful chemotactic agents of
granulocytes (Murray et al., 2003). Ex vivo studies with isolated human
granulocytes shed light on the interaction of infiltrating granulocytes with invading
bacteria. Clearly, the opsonization of the bacteria with antibodies and complement
components is essential for the granulocyte-mediated killing of NTHi (Langereis
and Weiser, 2014). However, our current data demonstrate that various NTHi
strains are readily phagocytosed by granulocytes in an opsonin-independent
manner. It is still unclear if such engulfment of NTHi will lead to their destruction.
Importantly, opsonin-independent recognition of Neisseria gonorrhoeae via the
granulocyte-specific CEACAM3 receptor was consistently reported to lead to
uptake and elimination of gonococci by human granulocytes (Buntru et al., 2011;
Roth et al., 2013; Schmitter et al., 2004). Concordantly, the results presented in
this manuscript speak for a better recognition of CEACAM3-binding Hinf-aeg by
the granulocytes in comparison to the non-CEACAM3-binding NTHi H86 strain.
However, due to the presence of multiple CEACAMs in human granulocytes, the
differential uptake might also reflect differences in the binding affinities to human
CEACAMs by the used Hinf strains. Therefore, experiments with human
Host-derived fatty acids modulate P1 levels
98
granulocytes in the presence of blocking antibodies against CEACAM3 or with
myeloid cell lines with genetic disruption of the CEACAM3 gene might help to
clarify this issue.
Nevertheless, the observed increased uptake of LCFA-fed Hinf-aeg by
granulocytes fits into a plausible scenario where the arachidonic acid released
upon bacterial attachment to the epithelium (via CECAM1 or CEA) would
simultaneously increase OMP P1 expression by the bacteria while alarming the
granulocytes. The CEACAM3-mediated bacterial recognition could lead to the
elimination of invading bacteria and tip the balance in favor of a local and
contained infection. However, some strains might escape this defence mechanism
by expressing non-CEACAM3-binding adhesin, increasing the chances of
disseminated disease. This concept is in accordance with a recent report where N.
gonorrohoeae strains isolated from patients with disseminated disease harbored
CEACAM-binding adhesins, which lacked CEACAM3-recognition (Roth et al.,
2013). As some strains of the biogroup aegyptius of Hinf are able to cause life-
threatening systemic diseases (Harrison et al., 2008), it would be of prime
importance to investigate whether these strains interact or not with granulocyte
CEACAM3.
In conclusion, we have shown that Hinf increases its OMP P1 levels in response to
fatty acids, such as arachidonic acid. Whereas the molecular mechanisms
underpinning this OMP P1 upregulation are still unclear, our findings suggest that
the Hinf can sense lipid environmental cues to regulate its adhesin. Given the
presence of arachidonic acid at virtually all the mucosal surfaces of human body
and the high structural conservation of OMP P1 / FadL in Gram-negative bacteria,
sensing and responding to lipid signals might represent a widespread mechanism
contributing to the efficient colonization of highly specific niches by the bacteria.
IV.5 Materials and methods
Haemophilus influenzae strains and growth conditions
Hinf Rd and its mutants used in this study were described previously (Tchoupa et
al., 2015). Hinf-aeg ATCC® 11116 was purchased from German Collection of
Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). NTHi H86
Host-derived fatty acids modulate P1 levels
99
was generously provided by Prof. Reidl (University of Graz, Graz, Austria). The
bacteria were all grown on brain-heart infusion agar enriched with 10% Levinthal
base (BHL).
Growth curves
The growth kinetics were determined for the bacteria grown in BHL broth
supplemented or not with 3.6 or 7.2 mM oleic acid (Sigma-Aldrich). 10 × 108
bacteria were added to 8 mL of BHL broth. Cultures were grown at 37°C with
constant shaking (200 rpm) and the absorbance at 600 nm was measured every
30–60 minutes.
Analysis of surface-exposed OMP P1 levels by flow cytometry
The bacteria were incubated for 3 h at 37 °C with 3.6 mM oleic acid (Sigma-
Aldrich), 1 mM arachidonic acid (Sigma-Aldrich) or the solvent DMSO (Roth,
Karlsruhe, Germany). PBS was then used to wash the bacteria and anti-OMP P1
mouse serum (diluted 1:200) was directly added to intact bacteria. The samples
were incubated for 90 min at 4 °C under gentle rotation. After staining with Dylight®
488-conjugated anti-mouse IgG antibody (Jackson Immuno-Research, West
Grove, PA), the bacteria were washed with PBS and analyzed with the BD™ LSR
II flow cytometer.
CEACAM-binding assays
Soluble GFP-fused amino-terminal CEACAM domains were produced by 293 cells
and used for bacterial pull-down assays as previously described (Tchoupa et al.,
2015). In short, soluble recombinant GFP-tagged CEACAM N-terminal domains,
enriched in cell culture supernatants, were clustered with polyclonal anti-GFP
antibodies (overnight, at 4 °C). 10 × 108 bacteria were afterwards added in 1 ml
supernatant containing the indicated GFP-fusion proteins and incubated for 2 h at
37 °C under constant rotation. After washing with PBS, the bacteria were either
immunoblotted with a monoclonal antibody against GFP or bacteria-associated
fluorescence was detected by flow cytometry. In some cases, prior to bacterial
pull-down with CEACAMs, 6 × 109 bacteria were incubated in BHL broth
supplemented with 100 µM of arachidonic acid (Sigma-Aldrich) for 35 min at 37 °C
and washed twice with PBS.
Host-derived fatty acids modulate P1 levels
100
Western blotting
Immunoblotting was performed as described earlier (Hauck et al., 2001) using a
monoclonal antibody (mAb) against GFP (clone JL-8; Clonetech).
Cell culture
Cells of the HEK (Human Embryonic Kidney) 293T cell line (293 cells; ACC-635,
German Collection of Microorganisms and Cell Cultures, DSMZ, Braunschweig,
Germany) were grown in Dulbecco’s modified Eagle’s medium (DMEM) enriched
with 10% calf serum at 37° C with 5% CO2 and subcultured every 2–3 days. 293
cells were transfected calcium phosphate co-precipitation with 5–8 µg of plasmid
DNA for each 10-cm culture dish.
Quantification of bacterial internalization by flow cytometry
DMSO- or LCFA-treated bacteria were labeled with fluorescein isothiocyanate
(FITC) as described (Tchoupa et al., 2015) and used to infect cells (MOI = 200) for
2 h. The amount of intracellular bacteria was estimated by flow cytometry as
described earlier (Pils et al., 2006).
Statistical analysis
Pairwise comparisons were conducted using Student’s t-test (SigmaStat software,
Systat Software Inc., San Jose, California). P value < 0.05 was considered
significant.
General Discussion
101
General Discussion
Haemophilus influenzae type b (Hib) conjugated vaccine has nearly eradicated Hib
invasive infections in countries where its use is widespread (Agrawal and Murphy,
2011). Due to its components, the Hib capsular polysaccharide PRP (poly-ribose
ribitol phosphate) attached to a carrier protein, the Hib conjugate vaccine is
ineffective against unencapsulated Haemophilus influenzae (NTHi), representing
most of the circulating strains in the human population. The enormous genetic and
phenotypic diversity of NTHi strains is the major obstacle for vaccine development
against unencapsulated Hinf. Interestingly, notwithstanding their heterogeneity,
most of the NTHi isolates, as well as encapsulated strains, are able to bind the
amino-terminal domain of the human glycoprotein CEACAM1 (Virji et al., 2000).
An antigenic variable outer membrane protein, the OMP P5, was suggested to be
the bacterial ligand used by Hinf to target CEACAMs (Hill et al., 2001). Importantly,
the ompP5 gene is present in all Hinf isolates whose genome sequences are
available (Euba et al., 2015). Therefore, we hypothesized that identification of a
conserved CEACAM-binding motif within the OMP P5 may represent a major
breakthrough to allow novel preventive and curative measures against Hinf.
In this study, multiple sequence alignment of Hinf OMP P5 from strains with
different CEACAM-binding profiles did not provide any clue as to why NTHi strain
7502 is unable to target CEACAMs. Moreover, OMP P5 of various CEACAM-
binding Hinf strains, when expressed in non-CEACAM binding E. coli, was not
sufficient to promote the recognition of human CEACAMs. These initials
observations raised uncertainty about OMP P5 as the Hinf CEACAM-binding
ligand. Therefore, we screened mutant and wild-type Hinf strains to finally identify
OMP P1 as the bona fide adhesin recognizing human CEACAMs. Our results on
first sight were in conflict with former work, which suggested a direct interaction
between OMP P5 and CEACAMs (Hill et al., 2001). However, this previous study
already conceded that: (i) the non-CEACAM binding strain NTHi H305 expresses
the OMP P5; (ii) genetic depletion of OMP P5 weakens the interaction of
CEACAM-binding Hinf with soluble CEACAMs but not with CEACAM-expressing
cells and (iii) the adhesion of OMP P5-deficient Hinf to CEACAM-expressing cells
was inhibited by an anti-CEACAM antibody (Hill et al., 2001). Combined, these
observations strongly suggested that another Hinf factor, whose surface levels
General Discussion
102
might be affected by OMP P5-depletion, was responsible for CEACAM-targeting
by the bacteria. Consistent with this plausible hypothesis, our data demonstrate
reduced levels of OMP P1 in OMP P5 mutant, providing a reasonable explanation
for the altered CEACAM-binding phenotype upon OMP P5 deletion. Fluctuations in
OMP P1 levels upon knock out of another outer membrane protein are not
unusual. For instance, Cope and co-workers noticed increased production of OMP
P1 in a OMP P2-depleted Hib mutant. Importantly, the complementation of OMP
P2 in the OMP P2-deficient mutant brings back the OMP P1 synthesis to the wild-
type levels (Cope et al., 1990). Conversely, we did not observe major changes in
OMP P2 levels upon OMP P1 deletion. However, a Hinf strain engineered to
express more OMP P1 in comparison to the otherwise isogenic wild-type strain,
exhibits less OMP P2 compared to the wild-type strain (Hanson et al., 1989). It is
likely that OMP P1 together with other outer membrane proteins builds the so-
called OMP islands where the absence or the overabundance of one OMP may
results in remodeling of OMP islands’ composition in order to preserve their
stability.
Lately, the strain NTHi strain 375 was characterized as a non-CEACAM-binding
strain notwithstanding its well-expressed OMP P5 protein sharing 95% amino acid
identity with OMP P5 of Hinf Rd strain (Euba et al., 2015). Furthermore, the
massive attachment of the Rd strain to CEACAM1-expressing cells was reduced,
not abrogated, upon deletion of OMP P5 (Euba et al., 2015). Whereas the residual
expression of OMP P1 in Rd ∆P5 mutant explains why this mutant is able to
adhere to cells in a CEACAM-dependent manner, it is unclear why NTHi 375 does
not bind CEACAMs. NTHi 7502, the only strain of our Hinf collection unable to
recognize CEACAMs, does not express any OMP P1 because of a mutation in its
ompP1 gene creating a premature stop codon (Tchoupa et al., 2015). In contrast,
NTHi 375 ompP1 gene is intact. It would be interesting to analyze the amount of
OMP P1 expressed by NTHi 375. High expression levels of OMP P1 in NTHi 375
would establish NTHi 375 as the first OMP P1-proficient Hinf strain unable to bind
CEACAMs, a magnificent tool in order to decrypt the molecular requirements for
OMP P1 – CEACAM interaction.
There are numerous examples, where the deletion of a particular OMP results in
readjusted levels of a different OMP. Therefore, it is essential to combine loss-of-
General Discussion
103
function assays upon OMP deletion with gain-of-function experiments after OMP
overexpression. Accordingly, our findings demonstrate that non-CEACAM-binding
E. coli is converted to a CEACAM-binding strain upon OMP P1 overexpression.
Thus, Hinf OMP P1 is not only necessary but also sufficient for CEACAM-
targeting. Of note, OMP P1 is only the second CEACAM-binding adhesin, after the
neisserial Opa proteins, to have been rigorously tested in a heterologous
background (Gray-Owen et al., 1997; Roth et al., 2013). The misidentification of
OMP P5 as the Hinf CEACAN-binding ligand (Hill et al., 2001) underlines the
importance of unbiased assays evaluating the sufficiency of an OMP to target its
postulated receptor.
The OMP P1, one of the major Hinf outer membrane proteins, is highly
immunogenic. Its use as a vaccine has been hampered by the lack of cross-
protection provided against heterologous strains. Indeed, chinchilla immunized
with recombinant OMP P1 from a given strain are protected against this strain but,
not against a strain with a different P1 (91% amino acid identity) (Bolduc et al.,
2000). This failure is due to the variability of the immunogenic and surface-
exposed regions of OMP P1 amongst Hinf strains (Bolduc et al., 2000).
Importantly, the atomic resolution structure of E. coli FadL (van den Berg et al.,
2004), the homolog of OMP P1, has greatly contributed to the understanding of
OMP P1 molecular architecture as FadL can be used to model OMP P1 (Mes and
van Putten, 2007). OMP P1 is predicted to build a 14-stranded β-barrel with 7
extracellular loops. By combining protein modeling and multiple sequence
alignment, we assigned the P1 immunogenic regions, previously known as
variable regions I, II, III and IV, to the loops 1, 3, 4 and 7, respectively. Although
these loops significantly differ from one Hinf strain to another, they are essential
for CEACAM-binding as their deletion result in loss of interaction. Interestingly, the
Opa protein, another β-barreled CEACAM-targeting adhesin, also uses two
hypervariable regions (HV1 and HV2) located on its loops 2 and 3 in order to bind
CEACAMs (Bos et al., 2002). Notwithstanding its variability, HV2 of
meningococcal Opa proteins harbors a conserved amino acid motif, GxI/V/LxQ,
that is critical for CEACAM recognition (de Jonge et al., 2003). This motif is
present in OMP P1 loop3 of only a small fraction of Hinf isolates, suggesting its
dispensability in CEACAM-binding by Hinf.
General Discussion
104
The relevance of Hinf attachment to epithelial CEACAMs in vivo was previously
investigated using the chinchilla as a model (Bookwalter et al., 2007). Bookwalter
and colleagues showed that animals pretreated with anti-human CEACAM
antibodies were less colonized by Hinf and suggested therefore a direct interaction
between Hinf and chinchilla CEACAM1 (cCEACAM1) for nasopharyngeal
colonization (Bookwalter et al., 2007). Given the inexistence of direct evidence for
the binding of Hinf to cCEACAM1 and the relatively poor similarity (56% identity)
between chinchilla and human CEACAM1 amino-terminal domains, it is necessary
to rigorously test Hinf for cCEACAM1-binding. Our experimental setup, with GFP-
tagged CEACAM N-terminal domains, may be very useful for such a purpose and
has enabled us already to demonstrate that Hinf interacts with human CEACAM1,
but not with CEACAM1 from murine, bovine or canine origin. This apparent
species-specificity of CEACAM-targeting was also observed in other human-
restricted, CEACAM-binding pathogens (Neisseria, gonorrhoeae, Neisseria
meningitidis and Moraxella catarrhalis) that also selectively target human, but not
other mammalian CEACAM1 orthologues (Voges et al., 2010).
CEACAM homologs are present in many vertebrates, including cartilaginous fish,
frogs, lizards and humans, but absent in avian species (Chang et al., 2013). In
each species where they are expressed, CEACAMs have undergone a rapid
evolution to acquire or lose functions in response to the environmental selection
pressure and their recognition by pathogens is thought to constitute a driving force
for their evolution (Kammerer and Zimmermann, 2010). Our current results speak
for a co-evolution of Hinf OMP P1 with its human receptor(s). Although OMP P1 is
well-conserved in the Pasteurellaceae family, it is unclear which genetic alteration
has enabled only Hinf to recognize and exploit CEACAMs. It is plausible that
antigenic variations occurring on the prominent extracellular loops of OMP P1 may
have, by chance, rendered a Hinf strain able to interact with CEACAMs. The
ensuing advantages in terms of colonization and survival within the host might
have enabled the CEACAM-binding Hinf strains to gradually outcompete the non-
CEACAM-binding isolates. However, the importance of CEACAM-targeting for Hinf
colonization has to be explored in vivo with suitable animal models (Johswich et
al., 2013; Muenzner et al., 2010) or with human volunteers (Poole et al., 2013)
taking in account the species-specificity of OMP P1 – CEACAM interaction. The
only non-CEACAM binding Hinf strain we have characterized to date, NTHi strain
General Discussion
105
7502, was isolated from a patient with chronic obstructive pulmonary disease
(COPD) (Swords et al., 2000). Owing to the complex pathophysiology of COPD, it
is a mystery if NTHi 7502 was responsible of one exacerbation episode of the
disease or if this isolate is able to colonize a healthy, human airway.
Modelling of the OMP P1 structure has revealed that it might exhibit a large
opening in its putative outer membrane embedded β-barrel, reminiscent of the
lateral opening through which E. coli FadL transports long-chain fatty acids
(LCFAs) (Hearn et al., 2009). Our preliminary attempts failed to demonstrate an
essential role of Hinf OMP P1 in fatty acid uptake. However, a strong upregulation
of OMP P1 is noticeable upon treatment of inf with fatty acids. Similarly, FadL is
acknowledged to be induced by LCFAs (Feng and Cronan, 2012). Furthermore, E.
coli can: (i) sense the presence of LCFA in its environment; (ii) induce the
expression of its LCFA transporter (FadL) and (iii) shut down its own fatty acid
biosynthesis machinery (Feng and Cronan, 2012). It is currently unclear if this
intricate regulation of fatty acids degradation or neosynthesis is present in Hinf.
Nevertheless, the increased levels of OMP P1 upon addition of LCFAs enable Hinf
to better bind to soluble CEACAMs and to invade epithelial cells more efficiently in
a CEACAM-dependent manner. Consistently, the Hinf P1-deficient mutant is
unable to interact with CEACAMs even after LCFA-treatment. Thus, our data
underline the role of OMP P1 as the Hinf CEACAM-binding ligand while
suggesting that OMP P1 levels can be controlled by host-derived environmental
cues. This is the first example of a bacterial adhesin, whose expression is
modulated by lipid messengers that can be delivered by the host.
In conclusion, our current work provides strong evidence on the necessity and the
sufficiency of OMP P1 in the recognition of human CEACAMs by Haemophilus
influenzae and demonstrates that OMP P5 does not function as a CEACAM-
binding adhesin. OMP P1 utilizes its immunogenic, flexible and surface-exposed
extracellular loops to build the binding interface towards CEACAMs. Besides, the
induction of OMP P1 expression upon supplementation of fatty acids strongly
enhances the interaction of Hinf with CEACAMs and strengthens the finding that
OMP P1 is the bona fide CEACAM-binding adhesin of Hinf. Combined, these data
shed new light on the critical role played by OMP P1 in the crosstalk between
Haemophilus influenzae and its human host. The knowledge gained here is a solid
General Discussion
106
starting point for further studies pertaining to the relevance of OMP P1 in Hinf
pathogenesis, while paving the way for OMP P1-based preventive and therapeutic
approaches against Haemophilus influenzae.
Declaration of author’s contributions
107
Declaration of author’s contributions
Chapter I: “Signaling by epithelial members of the CEACAM family – mucosal
docking sites for pathogenic bacteria”. Cell Communication and Signaling 2014:
12:27
Arnaud Kengmo Tchoupa, Tamara Schumacher and Christof R. Hauck.
The manuscript was written by AKT and CRH. AKT, TS and CRH revised the
manuscript.
Chapter II: “Outer membrane protein P1 is the CEACAM-binding adhesin of
Haemophilus influenzae”. Mol Microbiol. 2015 Jul 16. doi: 10.1111/mmi.13134.
Arnaud Kengmo Tchoupa, Sabine Lichtenegger, Joachim Reidl and Christof R.
Hauck.
CRH and JR conceived and supervised the project. AKT and SL performed the
experiments. AKT and CRH wrote the paper.
Chapter III: “The outer membrane protein P1 of Haemophilus influenzae utilizes
its extracellular loops to grasp human CEACAM.” In preparation.
Arnaud Kengmo Tchoupa and Christof R. Hauck.
CRH and AKT designed the study. Experiments were conducted by AKT. AKT
wrote the manuscript.
Chapter IV: “Haemophilus influenzae upregulates its outer membrane protein P1
in response to fatty acids”. In preparation.
Arnaud Kengmo Tchoupa and Christof R. Hauck.
CRH and AKT designed the study. AKT performed the experiments. The
manuscript was written by AKT.
List of publications
108
List of publications
1. Publications that are included in this thesis
Tchoupa, A.K., T. Schuhmacher, and C.R. Hauck. 2014. Signaling by
epithelial members of the CEACAM family - mucosal docking sites for
pathogenic bacteria. Cell Commun Signal. 12:27.
Tchoupa, A.K., S. Lichtenegger, J. Reidl, and C.R. Hauck. 2015. Outer
membrane protein P1 is the CEACAM-binding adhesin of Haemophilus
influenzae. Molecular microbiology.
2. Publications not part of this thesis
Roth, A., C. Mattheis, A.K. Tchoupa, M. Hupfeld, N. Hartmann, M. Unemo,
C.R. Hauck. Induction of the opsonophagoctosis by CEA-Fc fusion proteins.
In preparation.
Kuespert, K., A. Roth, J. Adrian, A.K. Tchoupa, C. Sprissler, T. Güntert, G.
O. Cory, C.R. Hauck. JNKs are involved in CEACAM3-mediated opsonin-
independent phagocytosis of Neisseria gonorrhoeae. The Journal of
biological chemistry. In revision.
Muenzner, P., A.K. Tchoupa, B. Klauser, T. Brunner, U. Dobrindt, C.R.
Hauck. Uropathogenic E. coli exploit CEA to promote colonization of the
urogenital tract mucosa. Plos Pathogens. In revision.
References
109
References
Abraham, S.N., A.B. Jonsson, and S. Normark. 1998. Fimbriae-mediated host-pathogen cross-talk. Curr Opin Microbiol. 1:75-81.
Agrawal, A., and T.F. Murphy. 2011. Haemophilus influenzae infections in the H. influenzae type b conjugate vaccine era. Journal of clinical microbiology. 49:3728-3732.
Andersen, C., E. Maier, G. Kemmer, J. Blass, A.K. Hilpert, R. Benz, and J. Reidl. 2003. Porin OmpP2 of Haemophilus influenzae shows specificity for nicotinamide-derived nucleotide substrates. The Journal of biological chemistry. 278:24269-24276.
Anderson, K.L., J. Billington, D. Pettigrew, E. Cota, P. Simpson, P. Roversi, H.A. Chen, P. Urvil, L. du Merle, P.N. Barlow, M.E. Medof, R.A. Smith, B. Nowicki, C. Le Bouguenec, S.M. Lea, and S. Matthews. 2004. An atomic resolution model for assembly, architecture, and function of the Dr adhesins. Molecular cell. 15:647-657.
Apicella, M.A. 2010. Neisseria meningitidis. Churchill Livingstone, Philadelphia. 2737-2752 pp.
Araki, N., M.T. Johnson, and J.A. Swanson. 1996. A role for phosphoinositide 3-kinase in the completion of macropinocytosis and phagocytosis by macrophages. The Journal of cell biology. 135:1249-1260.
Avadhanula, V., C.A. Rodriguez, G.C. Ulett, L.O. Bakaletz, and E.E. Adderson. 2006. Nontypeable Haemophilus influenzae adheres to intercellular adhesion molecule 1 (ICAM-1) on respiratory epithelial cells and upregulates ICAM-1 expression. Infection and immunity. 74:830-838.
Bakaletz, L.O., B.D. Baker, J.A. Jurcisek, A. Harrison, L.A. Novotny, J.E. Bookwalter, R. Mungur, and R.S. Munson, Jr. 2005. Demonstration of Type IV pilus expression and a twitching phenotype by Haemophilus influenzae. Infection and immunity. 73:1635-1643.
Barnett, T.R., L. Drake, and W. Pickle, 2nd. 1993. Human biliary glycoprotein gene: characterization of a family of novel alternatively spliced RNAs and their expressed proteins. Molecular and cellular biology. 13:1273-1282.
Barnich, N., F.A. Carvalho, A.L. Glasser, C. Darcha, P. Jantscheff, M. Allez, H. Peeters, G. Bommelaer, P. Desreumaux, J.F. Colombel, and A. Darfeuille-Michaud. 2007. CEACAM6 acts as a receptor for adherent-invasive E. coli, supporting ileal mucosa colonization in Crohn disease. The Journal of clinical investigation. 117:1566-1574.
Beauchemin, N., and A. Arabzadeh. 2013. Carcinoembryonic antigen-related cell adhesion molecules (CEACAMs) in cancer progression and metastasis. Cancer Metastasis Rev.
Beauchemin, N., P. Draber, G. Dveksler, P. Gold, S. Gray-Owen, F. Grunert, S. Hammarstrom, K.V. Holmes, A. Karlsson, M. Kuroki, S.H. Lin, L. Lucka, S.M. Najjar, M. Neumaier, B. Obrink, J.E. Shively, K.M. Skubitz, C.P. Stanners, P. Thomas, J.A. Thompson, M. Virji, S. von Kleist, C. Wagener, S. Watt, and W. Zimmermann. 1999. Redefined nomenclature for members of the carcinoembryonic antigen family. Experimental cell research. 252:243-249.
Benchimol, S., A. Fuks, S. Jothy, N. Beauchemin, K. Shirota, and C.P. Stanners. 1989. Carcinoembryonic antigen, a human tumor marker, functions as an intercellular adhesion molecule. Cell. 57:327-334.
References
110
Berger, C.N., O. Billker, T.F. Meyer, A.L. Servin, and I. Kansau. 2004. Differential recognition of members of the carcinoembryonic antigen family by Afa/Dr adhesins of diffusely adhering Escherichia coli (Afa/Dr DAEC). Molecular microbiology. 52:963-983.
Black, P.N. 1991. Primary sequence of the Escherichia coli fadL gene encoding an outer membrane protein required for long-chain fatty acid transport. Journal of bacteriology. 173:435-442.
Black, P.N., C.C. DiRusso, A.K. Metzger, and T.L. Heimert. 1992. Cloning, sequencing, and expression of the fadD gene of Escherichia coli encoding acyl coenzyme A synthetase. The Journal of biological chemistry. 267:25513-25520.
Black, P.N., B. Said, C.R. Ghosn, J.V. Beach, and W.D. Nunn. 1987. Purification and characterization of an outer membrane-bound protein involved in long-chain fatty acid transport in Escherichia coli. The Journal of biological chemistry. 262:1412-1419.
Bolduc, G.R., V. Bouchet, R.Z. Jiang, J. Geisselsoder, Q.C. Truong-Bolduc, P.A. Rice, S.I. Pelton, and R. Goldstein. 2000. Variability of outer membrane protein P1 and its evaluation as a vaccine candidate against experimental otitis media due to nontypeable Haemophilus influenzae: an unambiguous, multifaceted approach. Infection and immunity. 68:4505-4517.
Bookwalter, J.E., J.A. Jurcisek, S.D. Gray-Owen, S. Fernandez, G. McGillivary, and L.O. Bakaletz. 2007. A CEACAM1 homologue plays a pivotal role in nontypeable Haemophilus influenzae colonization of the chinchilla nasopharynx via the OMP P5-homologous adhesin. Infection and immunity.
Bookwalter, J.E., J.A. Jurcisek, S.D. Gray-Owen, S. Fernandez, G. McGillivary, and L.O. Bakaletz. 2008. A carcinoembryonic antigen-related cell adhesion molecule 1 homologue plays a pivotal role in nontypeable Haemophilus influenzae colonization of the chinchilla nasopharynx via the outer membrane protein P5-homologous adhesin. Infection and immunity. 76:48-55.
Bos, M.P., F. Grunert, and R.J. Belland. 1997. Differential recognition of members of the carcinoembryonic antigen family by Opa variants of Neisseria gonorrhoeae. Infection & Immunity. 65:2353-2361.
Bos, M.P., D. Kao, D.M. Hogan, C.C. Grant, and R.J. Belland. 2002. Carcinoembryonic antigen family receptor recognition by gonococcal Opa proteins requires distinct combinations of hypervariable Opa protein domains. Infection and immunity. 70:1715-1723.
Bouchet, V., D.W. Hood, J. Li, J.R. Brisson, G.A. Randle, A. Martin, Z. Li, R. Goldstein, E.K. Schweda, S.I. Pelton, J.C. Richards, and E.R. Moxon. 2003. Host-derived sialic acid is incorporated into Haemophilus influenzae lipopolysaccharide and is a major virulence factor in experimental otitis media. Proceedings of the National Academy of Sciences of the United States of America. 100:8898-8903.
Brash, A.R. 2001. Arachidonic acid as a bioactive molecule. The Journal of clinical investigation. 107:1339-1345.
Brooks, M.J., J.L. Sedillo, N. Wagner, W. Wang, A.S. Attia, H. Wong, C.A. Laurence, E.J. Hansen, and S.D. Gray-Owen. 2008. Moraxella catarrhalis binding to host cellular receptors is mediated by sequence-specific determinants not conserved among all UspA1 protein variants. Infection and immunity. 76:5322-5329.
References
111
Brummer, J., A. Ebrahimnejad, R. Flayeh, U. Schumacher, T. Loning, A.M. Bamberger, and C. Wagener. 2001. cis Interaction of the cell adhesion molecule CEACAM1 with integrin beta(3). Am J Pathol. 159:537-546.
Buntru, A., K. Kopp, M. Voges, R. Frank, V. Bachmann, and C.R. Hauck. 2011. Phosphatidylinositol-3' kinase activity is critical for initiating the oxidative burst and bacterial destruction during CEACAM3-mediated phagocytosis. The Journal of biological chemistry. 286:9555-9566.
Buntru, A., A. Roth, N.J. Nyffenegger-Jann, and C.R. Hauck. 2012. HemITAM signaling by CEACAM3, a human granulocyte receptor recognizing bacterial pathogens. Archives of biochemistry and biophysics. 524:77-83.
Camacho-Leal, P., and C.P. Stanners. 2008. The human carcinoembryonic antigen (CEA) GPI anchor mediates anoikis inhibition by inactivation of the intrinsic death pathway. Oncogene. 27:1545-1553.
Camacho-Leal, P., A.B. Zhai, and C.P. Stanners. 2007. A co-clustering model involving alpha5beta1 integrin for the biological effects of GPI-anchored human carcinoembryonic antigen (CEA). Journal of cellular physiology. 211:791-802.
Carruthers, M.D., E.N. Tracy, A.C. Dickson, K.B. Ganser, R.S. Munson, Jr., and L.O. Bakaletz. 2012. Biological roles of nontypeable Haemophilus influenzae type IV pilus proteins encoded by the pil and com operons. Journal of bacteriology. 194:1927-1933.
Chan, C.H., and C.P. Stanners. 2004. Novel mouse model for carcinoembryonic antigen-based therapy. Molecular therapy : the journal of the American Society of Gene Therapy. 9:775-785.
Chang, C.L., J. Semyonov, P.J. Cheng, S.Y. Huang, J.I. Park, H.J. Tsai, C.Y. Lin, F. Grutzner, Y.K. Soong, J.J. Cai, and S.Y. Hsu. 2013. Widespread divergence of the CEACAM/PSG genes in vertebrates and humans suggests sensitivity to selection. PloS one. 8:e61701.
Chen, C.J., J. Kirshner, M.A. Sherman, W. Hu, T. Nguyen, and J.E. Shively. 2007. Mutation analysis of the short cytoplasmic domain of the cell-cell adhesion molecule CEACAM1 identifies residues that orchestrate actin binding and lumen formation. The Journal of biological chemistry. 282:5749-5760.
Chen, T., and E.C. Gotschlich. 1996. CGM1a antigen of neutrophils, a receptor of gonococcal opacity proteins. Proceedings of the National Academy of Sciences of the United States of America. 93:14851-14856.
Chen, T., F. Grunert, A. Medina-Marino, and E.C. Gotschlich. 1997. Several carcinoembryonic antigens (CD66) serve as receptors for gonococcal opacity proteins. Journal of Experimental Medicine. 185:1557-1564.
Cheung, P.H., W. Luo, Y. Qiu, X. Zhang, K. Earley, P. Millirons, and S.H. Lin. 1993. Structure and function of C-CAM1. The first immunoglobulin domain is required for intercellular adhesion. The Journal of biological chemistry. 268:24303-24310.
Cho, B.K., E.M. Knight, and B.O. Palsson. 2006. Transcriptional regulation of the fad regulon genes of Escherichia coli by ArcA. Microbiology. 152:2207-2219.
Clementi, C.F., and T.F. Murphy. 2011. Non-typeable Haemophilus influenzae invasion and persistence in the human respiratory tract. Front Cell Infect Microbiol. 1:1.
Confer, A.W., and S. Ayalew. 2013. The OmpA family of proteins: roles in bacterial pathogenesis and immunity. Veterinary microbiology. 163:207-222.
References
112
Conners, R., D.J. Hill, E. Borodina, C. Agnew, S.J. Daniell, N.M. Burton, R.B. Sessions, A.R. Clarke, L.E. Catto, D. Lammie, T. Wess, R.L. Brady, and M. Virji. 2008. The Moraxella adhesin UspA1 binds to its human CEACAM1 receptor by a deformable trimeric coiled-coil. The EMBO journal. 27:1779-1789.
Cope, L.D., S.E. Pelzel, J.L. Latimer, and E.J. Hansen. 1990. Characterization of a mutant of Haemophilus influenzae type b lacking the P2 major outer membrane protein. Infection and immunity. 58:3312-3318.
de Jonge, M.I., H.J. Hamstra, L. van Alphen, J. Dankert, and P. van der Ley. 2003. Mapping the binding domains on meningococcal Opa proteins for CEACAM1 and CEA receptors. Molecular microbiology. 50:1005-1015.
del Pozo, M.A., N.B. Alderson, W.B. Kiosses, H.H. Chiang, R.G. Anderson, and M.A. Schwartz. 2004. Integrins regulate Rac targeting by internalization of membrane domains. Science. 303:839-842.
Dirusso, C.C., and P.N. Black. 2004. Bacterial long chain fatty acid transport: gateway to a fatty acid-responsive signaling system. The Journal of biological chemistry. 279:49563-49566.
Eades-Perner, A.M., H. van der Putten, A. Hirth, J. Thompson, M. Neumaier, S. von Kleist, and W. Zimmermann. 1994. Mice transgenic for the human carcinoembryonic antigen gene maintain its spatiotemporal expression pattern. Cancer research. 54:4169-4176.
Ergun, S., N. Kilik, G. Ziegeler, A. Hansen, P. Nollau, J. Gotze, J.H. Wurmbach, A. Horst, J. Weil, M. Fernando, and C. Wagener. 2000. CEA-related cell adhesion molecule 1: a potent angiogenic factor and a major effector of vascular endothelial growth factor. Molecular cell. 5:311-320.
Euba, B., J. Moleres, C. Viadas, I. Ruiz de los Mozos, J. Valle, J.A. Bengoechea, and J. Garmendia. 2015. Relative Contribution of P5 and Hap Surface Proteins to Nontypable Haemophilus influenzae Interplay with the Host Upper and Lower Airways. PloS one. 10:e0123154.
Feng, Y., and J.E. Cronan. 2012. Crosstalk of Escherichia coli FadR with global regulators in expression of fatty acid transport genes. PloS one. 7:e46275.
Finlay, B.B., and S. Falkow. 1997. Common themes in microbial pathogenicity revisited. Microbiology and Molecular Biology Reviews. 61:136-169.
Fox, D.A., P. Larsson, R.H. Lo, B.M. Kroncke, P.M. Kasson, and L. Columbus. 2014. Structure of the neisserial outer membrane protein opa60: loop flexibility essential to receptor recognition and bacterial engulfment. J Am Chem Soc. 136:9938-9946.
Foxwell, A.R., J.M. Kyd, and A.W. Cripps. 1998. Nontypeable Haemophilus influenzae: pathogenesis and prevention. Microbiology and molecular biology reviews : MMBR. 62:294-308.
Gaur, S., J.E. Shively, Y. Yen, and R.K. Gaur. 2008. Altered splicing of CEACAM1 in breast cancer: identification of regulatory sequences that control splicing of CEACAM1 into long or short cytoplasmic domain isoforms. Molecular cancer. 7:46.
Gold, P., and S.O. Freedman. 1965. Specific carcinoembryonic antigens of the human digestive system. The Journal of experimental medicine. 122:467-481.
Gotoh, N., H. Wakebe, and T. Nishino. 1989a. Ultrastructural aspects of fragility of Pseudomonas aeruginosa outer membrane devoid of protein F. FEMS microbiology letters. 50:51-53.
References
113
Gotoh, N., H. Wakebe, E. Yoshihara, T. Nakae, and T. Nishino. 1989b. Role of protein F in maintaining structural integrity of the Pseudomonas aeruginosa outer membrane. Journal of bacteriology. 171:983-990.
Gray-Owen, S.D., and R.S. Blumberg. 2006. CEACAM1: contact-dependent control of immunity. Nat Rev Immunol. 6:433-446.
Gray-Owen, S.D., D.R. Lorenzen, A. Haude, T.F. Meyer, and C. Dehio. 1997. Differential Opa specificities for CD66 receptors influence tissue interactions and cellular response to Neisseria gonorrhoeae. Molecular microbiology. 26:971-980.
Gray-Owen, S.D., and A.B. Schryvers. 1993. The interaction of primate transferrins with receptors on bacteria pathogenic to humans. Microbial pathogenesis. 14:389-398.
Gu, A., Z. Zhang, N. Zhang, W. Tsark, and J.E. Shively. 2010. Generation of human CEACAM1 transgenic mice and binding of Neisseria Opa protein to their neutrophils. PloS one. 5:e10067.
Guignot, J., S. Hudault, I. Kansau, I. Chau, and A.L. Servin. 2009. Human decay-accelerating factor and CEACAM receptor-mediated internalization and intracellular lifestyle of Afa/Dr diffusely adhering Escherichia coli in epithelial cells. Infection and immunity. 77:517-531.
Hammarstrom, S. 1999. The carcinoembryonic antigen (CEA) family: structures, suggested functions and expression in normal and malignant tissues. Semin Cancer Biol. 9:67-81.
Hammarstrom, S., M. Hamberg, B. Samuelsson, E.A. Duell, M. Stawiski, and J.J. Voorhees. 1975. Increased concentrations of nonesterified arachidonic acid, 12L-hydroxy-5,8,10,14-eicosatetraenoic acid, prostaglandin E2, and prostaglandin F2alpha in epidermis of psoriasis. Proceedings of the National Academy of Sciences of the United States of America. 72:5130-5134.
Hanson, M.S., L.D. Cope, and E.J. Hansen. 1989. Expression of the heat-modifiable major outer membrane protein of Haemophilus influenzae type b is unrelated to virulence. Infection and immunity. 57:1639-1646.
Harizi, H., J.B. Corcuff, and N. Gualde. 2008. Arachidonic-acid-derived eicosanoids: roles in biology and immunopathology. Trends in molecular medicine. 14:461-469.
Harrison, L.H., V. Simonsen, and E.A. Waldman. 2008. Emergence and disappearance of a virulent clone of Haemophilus influenzae biogroup aegyptius, cause of Brazilian purpuric fever. Clinical microbiology reviews. 21:594-605.
Hauck, C.R., T. Hunter, and D.D. Schlaepfer. 2001. The v-Src SH3 domain facilitates a cell adhesion-independent association with FAK. Journal of Biological Chemistry. 276:17653-17662.
Hearn, E.M., D.R. Patel, B.W. Lepore, M. Indic, and B. van den Berg. 2009. Transmembrane passage of hydrophobic compounds through a protein channel wall. Nature. 458:367-370.
Hill, D.J., A.M. Edwards, H.A. Rowe, and M. Virji. 2005. Carcinoembryonic antigen-related cell adhesion molecule (CEACAM)-binding recombinant polypeptide confers protection against infection by respiratory and urogenital pathogens. Molecular microbiology. 55:1515-1527.
Hill, D.J., M.A. Toleman, D.J. Evans, S. Villullas, L. Van Alphen, and M. Virji. 2001. The variable P5 proteins of typeable and non-typeable Haemophilus influenzae target human CEACAM1. Molecular microbiology. 39:850-862.
References
114
Hill, D.J., and M. Virji. 2003. A novel cell-binding mechanism of Moraxella catarrhalis ubiquitous surface protein UspA: specific targeting of the N-domain of carcinoembryonic antigen-related cell adhesion molecules by UspA1. Molecular microbiology. 48:117-129.
Hill, D.J., C. Whittles, and M. Virji. 2012. A novel group of Moraxella catarrhalis UspA proteins mediates cellular adhesion via CEACAMs and vitronectin. PloS one. 7:e45452.
Hill, F.F., and D. Angelmaier. 1972. Specific enrichment of mutants of Escherichia coli with an altered acyl CoA synthetase by tritium suicide. Molecular & general genetics : MGG. 117:143-152.
Hoffmann, C., A. Berking, F. Agerer, A. Buntru, F. Neske, G.S. Chhatwal, K. Ohlsen, and C.R. Hauck. 2010. Caveolin limits membrane microdomain mobility and integrin-mediated uptake of fibronectin-binding pathogens. Journal of cell science. 123:4280-4291.
Hohle, T.H., W.L. Franck, G. Stacey, and M.R. O'Brian. 2011. Bacterial outer membrane channel for divalent metal ion acquisition. Proceedings of the National Academy of Sciences of the United States of America. 108:15390-15395.
Hoiseth, S.K., and J.R. Gilsdorf. 1988. The relationship between type b and nontypable Haemophilus influenzae isolated from the same patient. The Journal of infectious diseases. 158:643-645.
Horst, A.K., W.D. Ito, J. Dabelstein, U. Schumacher, H. Sander, C. Turbide, J. Brummer, T. Meinertz, N. Beauchemin, and C. Wagener. 2006. Carcinoembryonic antigen-related cell adhesion molecule 1 modulates vascular remodeling in vitro and in vivo. The Journal of clinical investigation. 116:1596-1605.
Horton, R.M., Z.L. Cai, S.N. Ho, and L.R. Pease. 1990. Gene splicing by overlap extension: tailor-made genes using the polymerase chain reaction. BioTechniques. 8:528-535.
Huang, J., J.D. Hardy, Y. Sun, and J.E. Shively. 1999. Essential role of biliary glycoprotein (CD66a) in morphogenesis of the human mammary epithelial cell line MCF10F. Journal of cell science. 112 ( Pt 23):4193-4205.
Hultgren, S.J., S. Abraham, M. Caparon, P. Falk, J.W. St Geme, 3rd, and S. Normark. 1993. Pilus and nonpilus bacterial adhesins: assembly and function in cell recognition. Cell. 73:887-901.
Indik, Z.K., J.G. Park, S. Hunter, and A.D. Schreiber. 1995. The molecular dissection of Fc gamma receptor mediated phagocytosis. Blood. 86:4389-4399.
Iram, S.H., and J.E. Cronan. 2005. Unexpected functional diversity among FadR fatty acid transcriptional regulatory proteins. The Journal of biological chemistry. 280:32148-32156.
Jalalvand, F., and K. Riesbeck. 2014. Haemophilus influenzae: recent advances in the understanding of molecular pathogenesis and polymicrobial infections. Current opinion in infectious diseases. 27:268-274.
Jame, A.J., P.M. Lackie, A.M. Cazaly, I. Sayers, J.F. Penrose, S.T. Holgate, and A.P. Sampson. 2007. Human bronchial epithelial cells express an active and inducible biosynthetic pathway for leukotrienes B4 and C4. Clinical and experimental allergy : journal of the British Society for Allergy and Clinical Immunology. 37:880-892.
Johswich, K.O., S.E. McCaw, E. Islam, A. Sintsova, A. Gu, J.E. Shively, and S.D. Gray-Owen. 2013. In Vivo Adaptation and Persistence of Neisseria
References
115
meningitidis within the Nasopharyngeal Mucosa. PLoS pathogens. 9:e1003509.
Kammerer, R., T. Popp, S. Hartle, B.B. Singer, and W. Zimmermann. 2007. Species-specific evolution of immune receptor tyrosine based activation motif-containing CEACAM1-related immune receptors in the dog. BMC Evol Biol. 7:196.
Kammerer, R., T. Popp, B.B. Singer, J. Schlender, and W. Zimmermann. 2004. Identification of allelic variants of the bovine immune regulatory molecule CEACAM1 implies a pathogen-driven evolution. Gene. 339:99-109.
Kammerer, R., and W. Zimmermann. 2010. Coevolution of activating and inhibitory receptors within mammalian carcinoembryonic antigen (CEA) families. BMC Biology. 8:12.
Kirshner, J., C.J. Chen, P. Liu, J. Huang, and J.E. Shively. 2003. CEACAM1-4S, a cell-cell adhesion molecule, mediates apoptosis and reverts mammary carcinoma cells to a normal morphogenic phenotype in a 3D culture. Proceedings of the National Academy of Sciences of the United States of America. 100:521-526.
Kirshner, J., J. Hardy, S. Wilczynski, and J.E. Shively. 2004. Cell-cell adhesion molecule CEACAM1 is expressed in normal breast and milk and associates with beta1 integrin in a 3D model of morphogenesis. J Mol Histol. 35:287-299.
Klaile, E., T.E. Klassert, I. Scheffrahn, M.M. Muller, A. Heinrich, K.A. Heyl, H. Dienemann, C. Grunewald, R. Bals, B.B. Singer, and H. Slevogt. 2013. Carcinoembryonic antigen (CEA)-related cell adhesion molecules are co-expressed in the human lung and their expression can be modulated in bronchial epithelial cells by non-typable Haemophilus influenzae, Moraxella catarrhalis, TLR3, and type I and II interferons. Respiratory research. 14:85.
Klaile, E., M.M. Muller, C. Kannicht, B.B. Singer, and L. Lucka. 2005. CEACAM1 functionally interacts with filamin A and exerts a dual role in the regulation of cell migration. Journal of cell science. 118:5513-5524.
Klaile, E., O. Vorontsova, K. Sigmundsson, M.M. Muller, B.B. Singer, L.G. Ofverstedt, S. Svensson, U. Skoglund, and B. Obrink. 2009. The CEACAM1 N-terminal Ig domain mediates cis- and trans-binding and is essential for allosteric rearrangements of CEACAM1 microclusters. The Journal of cell biology. 187:553-567.
Korotkova, N., E. Cota, Y. Lebedin, S. Monpouet, J. Guignot, A.L. Servin, S. Matthews, and S.L. Moseley. 2006. A subfamily of Dr adhesins of Escherichia coli bind independently to decay-accelerating factor and the N-domain of carcinoembryonic antigen. The Journal of biological chemistry. 281:29120-29130.
Korotkova, N., Y. Yarova-Yarovaya, V. Tchesnokova, N. Yazvenko, M.A. Carl, A.E. Stapleton, and S.L. Moseley. 2008. Escherichia coli DraE adhesin-associated bacterial internalization by epithelial cells is promoted independently by decay-accelerating factor and carcinoembryonic antigen-related cell adhesion molecule binding and does not require the DraD invasin. Infection and immunity. 76:3869-3880.
Kraiss, A., S. Schlor, and J. Reidl. 1998. In vivo transposon mutagenesis in Haemophilus influenzae. Appl Environ Microbiol. 64:4697-4702.
Kuespert, K., and C.R. Hauck. 2009. Characterizing host receptor recognition by individual bacterial pathogens. In Methods in Molecular Biology. Vol. 470. S. Rupp, K. Sohn, and N. Hauser, editors. Humana Press, Totowa. 57-65.
References
116
Kuespert, K., S. Pils, and C.R. Hauck. 2006. CEACAMs - their role in physiology and pathophysiology. Current opinion in cell biology. 18:565-571.
Kuespert, K., A. Roth, and C.R. Hauck. 2011. Neisseria meningitidis has two independent modes of recognizing its human receptor CEACAM1. PloS one. 6:e14609.
Kuespert, K., S. Weibel, and C.R. Hauck. 2007. Profiling of bacterial adhesin – host receptor recognition by soluble immunoglobulin superfamily domains. Journal of Microbiological Methods. 68:478-485.
Langereis, J.D., and J.N. Weiser. 2014. Shielding of a lipooligosaccharide IgM epitope allows evasion of neutrophil-mediated killing of an invasive strain of nontypeable Haemophilus influenzae. mBio. 5:e01478-01414.
Laurie, N.A., M.M. Comegys, M.P. Carreiro, J.F. Brown, D.L. Flanagan, K.E. Brilliant, and D.C. Hixson. 2005. Carcinoembryonic antigen-related cell adhesion molecule 1a-4L suppression of rat hepatocellular carcinomas. Cancer research. 65:11010-11017.
Lawson, E.L., D.R. Mills, K.E. Brilliant, and D.C. Hixson. 2012. The transmembrane domain of CEACAM1-4S is a determinant of anchorage independent growth and tumorigenicity. PloS one. 7:e29606.
Lepore, B.W., M. Indic, H. Pham, E.M. Hearn, D.R. Patel, and B. van den Berg. 2011. Ligand-gated diffusion across the bacterial outer membrane. Proceedings of the National Academy of Sciences of the United States of America. 108:10121-10126.
Leung, N., C. Turbide, M. Olson, V. Marcus, S. Jothy, and N. Beauchemin. 2006. Deletion of the carcinoembryonic antigen-related cell adhesion molecule 1 (Ceacam1) gene contributes to colon tumor progression in a murine model of carcinogenesis. Oncogene. 25:5527-5536.
Leusch, H.G., Z. Drzeniek, Z. Markos-Puztai, and C. Wagener. 1991. Binding of Escherichia coli and Salmonella strains to members of the carcinoembryonic antigen family: differential binding inhibition by aromatic glycosides of mannose. Infection & Immunity. 59:2051-2057.
Leusch, H.G., S.A. Hefta, Z. Drzeniek, K. Hummel, Z. Markos-Pusztai, and C. Wagener. 1990. Escherichia coli of human origin binds to carcinoembryonic antigen (CEA) and non-specific crossreacting antigen (NCA). FEBS letters. 261:405-409.
Li, C., C.J. Chen, and J.E. Shively. 2009. Mutational analysis of the cytoplasmic domain of CEACAM1-4L in humanized mammary glands reveals key residues involved in lumen formation: stimulation by Thr-457 and inhibition by Ser-461. Experimental cell research. 315:1225-1233.
Li, Y., H. Cao, Z. Jiao, S.B. Pakala, D.N. Sirigiri, W. Li, R. Kumar, and L. Mishra. 2010. Carcinoembryonic antigen interacts with TGF-{beta} receptor and inhibits TGF-{beta} signaling in colorectal cancers. Cancer research. 70:8159-8168.
Lichtenegger, S., I. Bina, S. Roier, S. Bauernfeind, K. Keidel, S. Schild, M. Anthony, and J. Reidl. 2014. Characterization of lactate utilization and its implication on the physiology of Haemophilus influenzae. International journal of medical microbiology : IJMM. 304:490-498.
Lin, J., S. Huang, and Q. Zhang. 2002. Outer membrane proteins: key players for bacterial adaptation in host niches. Microbes and infection / Institut Pasteur. 4:325-331.
Lobo, E.O., Z. Zhang, and J.E. Shively. 2009. Pivotal advance: CEACAM1 is a negative coreceptor for the B cell receptor and promotes CD19-mediated
References
117
adhesion of B cells in a PI3K-dependent manner. Journal of leukocyte biology. 86:205-218.
McCaw, S.E., E.H. Liao, and S.D. Gray-Owen. 2004. Engulfment of Neisseria gonorrhoeae: revealing distinct processes of bacterial entry by individual carcinoembryonic antigen-related cellular adhesion molecule family receptors. Infection and immunity. 72:2742-2752.
Mes, T.H., and J.P. van Putten. 2007. Positively selected codons in immune-exposed loops of the vaccine candidate OMP-P1 of Haemophilus influenzae. Journal of molecular evolution. 64:411-422.
Morton, D.J., A. Smith, T.M. VanWagoner, T.W. Seale, P.W. Whitby, and T.L. Stull. 2007. Lipoprotein e (P4) of Haemophilus influenzae: role in heme utilization and pathogenesis. Microbes and infection / Institut Pasteur. 9:932-939.
Muenzner, P., V. Bachmann, J. Hentschel, W. Zimmermann, and C.R. Hauck. 2010. Human-specific bacterial pathogens block shedding of epithelial cells by stimulating integrin activation. Science. 329:1197-1201.
Muenzner, P., V. Bachmann, K. Kuespert, and C.R. Hauck. 2008. The CEACAM1 transmembrane domain, but not the cytoplasmic domain, directs internalization of human pathogens via membrane-microdomains. Cellular microbiology. 10:1074-1092.
Muenzner, P., O. Billker, T.F. Meyer, and M. Naumann. 2002. Nuclear factor-kB directs CEACAM1 receptor expression in Neisseria gonorrhoeae-infected epithelial cells. Journal of Biological Chemistry. 277:7438-7446.
Muenzner, P., M. Naumann, T.F. Meyer, and S.D. Gray-Owen. 2001. Pathogenic Neisseria Trigger Expression of Their Carcinoembryonic Antigen-related Cellular Adhesion Molecule 1 (CEACAM1; Previously CD66a) Receptor on Primary Endothelial Cells by Activating the Immediate Early Response Transcription Factor, Nuclear Factor-kappa B. Journal of Biological Chemistry. 276:24331-24340.
Muenzner, P., M. Rohde, S. Kneitz, and C.R. Hauck. 2005. CEACAM engagement by human pathogens enhances cell adhesion and counteracts bacteria-induced detachment of epithelial cells. Journal of Cell Biology. 170:825-836.
Muller, M.M., E. Klaile, O. Vorontsova, B.B. Singer, and B. Obrink. 2009. Homophilic adhesion and CEACAM1-S regulate dimerization of CEACAM1-L and recruitment of SHP-2 and c-Src. The Journal of cell biology. 187:569-581.
Muller, M.M., B.B. Singer, E. Klaile, B. Obrink, and L. Lucka. 2005. Transmembrane CEACAM1 affects integrin-dependent signaling and regulates extracellular matrix protein-specific morphology and migration of endothelial cells. Blood. 105:3925-3934.
Munson, R., Jr., and S. Grass. 1988. Purification, cloning, and sequence of outer membrane protein P1 of Haemophilus influenzae type b. Infection and immunity. 56:2235-2242.
Munson, R., Jr., and A. Hunt. 1989. Isolation and characterization of a mutant of Haemophilus influenzae type b deficient in outer membrane protein P1. Infection and immunity. 57:1002-1004.
Murphy, T.F., H. Faden, L.O. Bakaletz, J.M. Kyd, A. Forsgren, J. Campos, M. Virji, and S.I. Pelton. 2009. Nontypeable Haemophilus influenzae as a pathogen in children. Pediatr Infect Dis J. 28:43-48.
Murray, J., C. Ward, J.T. O'Flaherty, I. Dransfield, C. Haslett, E.R. Chilvers, and A.G. Rossi. 2003. Role of leukotrienes in the regulation of human
References
118
granulocyte behaviour: dissociation between agonist-induced activation and retardation of apoptosis. British journal of pharmacology. 139:388-398.
Nagano, K., E.K. Read, Y. Murakami, T. Masuda, T. Noguchi, and F. Yoshimura. 2005. Trimeric structure of major outer membrane proteins homologous to OmpA in Porphyromonas gingivalis. Journal of bacteriology. 187:902-911.
Nguyen, T., C.J. Chen, and J.E. Shively. 2014. Phosphorylation of CEACAM1 Molecule by Calmodulin Kinase IID in a Three-dimensional Model of Mammary Gland Lumen Formation. The Journal of biological chemistry. 289:2934-2945.
Nicholson, T.B., and C.P. Stanners. 2006. Specific inhibition of GPI-anchored protein function by homing and self-association of specific GPI anchors. The Journal of cell biology. 175:647-659.
Nikaido, H. 2003. Molecular basis of bacterial outer membrane permeability revisited. Microbiology and molecular biology reviews : MMBR. 67:593-656.
Norskov-Lauritsen, N. 2014. Classification, identification, and clinical significance of haemophilus and aggregatibacter species with host specificity for humans. Clinical microbiology reviews. 27:214-240.
Nunn, W.D., and R.W. Simons. 1978. Transport of long-chain fatty acids by Escherichia coli: mapping and characterization of mutants in the fadL gene. Proceedings of the National Academy of Sciences of the United States of America. 75:3377-3381.
Obrink, B. 1997. CEA adhesion molecules: multifunctional proteins with signal-regulatory properties. Current opinion in cell biology. 9:616-626.
Ordonez, C., A.B. Zhai, P. Camacho-Leal, L. Demarte, M.M. Fan, and C.P. Stanners. 2007. GPI-anchored CEA family glycoproteins CEA and CEACAM6 mediate their biological effects through enhanced integrin alpha5beta1-fibronectin interaction. Journal of cellular physiology. 210:757-765.
Overath, P., and E.M. Raufuss. 1967. The induction of the enzymes of fatty acid degradation in Escherichia coli. Biochemical and biophysical research communications. 29:28-33.
Pages, J.M., C.E. James, and M. Winterhalter. 2008. The porin and the permeating antibiotic: a selective diffusion barrier in Gram-negative bacteria. Nature reviews. Microbiology. 6:893-903.
Patel, P.C., H.S. Lee, A.Y. Ming, A. Rath, C.M. Deber, C.M. Yip, J.V. Rocheleau, and S.D. Gray-Owen. 2013. Inside-out signaling promotes dynamic changes in the CEACAM1 oligomeric state to control its cell adhesion properties. The Journal of biological chemistry. 288(41):29654-29669.
Pils, S., K. Kopp, L. Peterson, J. Delgado-Tascon, N.J. Nyffenegger-Jann, and C.R. Hauck. 2012. The adaptor molecule Nck localizes the WAVE complex to promote actin polymerization during CEACAM3-mediated phagocytosis of bacteria. PloS one. 7:e32808.
Pils, S., T. Schmitter, F. Neske, and C.R. Hauck. 2006. Quantification of bacterial invasion into adherent cells by flow cytometry. Journal of Microbiological Methods. 65:301-310.
Poole, J., E. Foster, K. Chaloner, J. Hunt, M.P. Jennings, T. Bair, K. Knudtson, E. Christensen, R.S. Munson, Jr., P.L. Winokur, and M.A. Apicella. 2013. Analysis of nontypeable haemophilus influenzae phase-variable genes during experimental human nasopharyngeal colonization. The Journal of infectious diseases. 208:720-727.
References
119
Rahman, M.M., X.X. Gu, C.M. Tsai, V.S. Kolli, and R.W. Carlson. 1999. The structural heterogeneity of the lipooligosaccharide (LOS) expressed by pathogenic non-typeable Haemophilus influenzae strain NTHi 9274. Glycobiology. 9:1371-1380.
Rao, V.K., G.P. Krasan, D.R. Hendrixson, S. Dawid, and J.W. St Geme, 3rd. 1999. Molecular determinants of the pathogenesis of disease due to non-typable Haemophilus influenzae. FEMS Microbiol Rev. 23:99-129.
Rawling, E.G., F.S. Brinkman, and R.E. Hancock. 1998. Roles of the carboxy-terminal half of Pseudomonas aeruginosa major outer membrane protein OprF in cell shape, growth in low-osmolarity medium, and peptidoglycan association. Journal of bacteriology. 180:3556-3562.
Reddy, M.S., J.M. Bernstein, T.F. Murphy, and H.S. Faden. 1996. Binding between outer membrane proteins of nontypeable Haemophilus influenzae and human nasopharyngeal mucin. Infection and immunity. 64:1477-1479.
Reidl, J., and J.J. Mekalanos. 1996. Lipoprotein e(P4) is essential for hemin uptake by Haemophilus influenzae. The Journal of experimental medicine. 183:621-629.
Reidl, J., S. Schlor, A. Kraiss, J. Schmidt-Brauns, G. Kemmer, and E. Soleva. 2000. NADP and NAD utilization in Haemophilus influenzae. Molecular microbiology. 35:1573-1581.
Robertson, B.D., and T.F. Meyer. 1992. Genetic variation in pathogenic bacteria. Trends in Genetics. 8 (12):422-427.
Roier, S., T. Blume, L. Klug, G.E. Wagner, W. Elhenawy, K. Zangger, R. Prassl, J. Reidl, G. Daum, M.F. Feldman, and S. Schild. 2014. A basis for vaccine development: Comparative characterization of Haemophilus influenzae outer membrane vesicles. International journal of medical microbiology : IJMM:10.1016/j.ijmm.2014.1012.1005.
Roier, S., D.R. Leitner, J. Iwashkiw, K. Schild-Prufert, M.F. Feldman, G. Krohne, J. Reidl, and S. Schild. 2012. Intranasal immunization with nontypeable Haemophilus influenzae outer membrane vesicles induces cross-protective immunity in mice. PloS one. 7:e42664.
Rosadini, C.V., S. Ram, and B.J. Akerley. 2014. Outer membrane protein P5 is required for resistance of nontypeable Haemophilus influenzae to both the classical and alternative complement pathways. Infection and immunity. 82:640-649.
Roth, A., C. Mattheis, P. Münzner, M. Unemo, and C.R. Hauck. 2013. Innate recognition by neutrophil granulocytes differs between Neisseria gonorrhoeae strains causing local or disseminating infections. Infection and immunity. 81(7):2358-2370.
Samara, R.N., L.M. Laguinge, and J.M. Jessup. 2007. Carcinoembryonic antigen inhibits anoikis in colorectal carcinoma cells by interfering with TRAIL-R2 (DR5) signaling. Cancer research. 67:4774-4782.
Samineni, S., C. Glackin, and J.E. Shively. 2011. Role of CEACAM1, ECM, and mesenchymal stem cells in an orthotopic model of human breast cancer. Int J Breast Cancer. 2011:381080.
Schlor, S., G. Kemmer, and J. Reidl. 2003. Transposon Tn10. Methods in molecular medicine. 71:211-224.
Schmitter, T., F. Agerer, L. Peterson, P. Muenzner, and C.R. Hauck. 2004. Granulocyte CEACAM3 is a phagocytic receptor of the innate immune system that mediates recognition and elimination of human-specific pathogens. Journal of Experimental Medicine. 199:35-46.
References
120
Schmitter, T., S. Pils, S. Weibel, F. Agerer, A. Buntru, K. Kopp, and C.R. Hauck. 2007. Opa proteins of pathogenic Neisseriae initiate Src-kinase-dependent or lipid raft-mediated uptake via distinct human CEACAM isoforms. Infection & Immunity. 75:4116-4126.
Screaton, R.A., L. DeMarte, P. Draber, and C.P. Stanners. 2000. The specificity for the differentiation blocking activity of carcinoembryonic antigen resides in its glycophosphatidyl-inositol anchor. The Journal of cell biology. 150:613-626.
Segada, L.M., G.M. Carlone, L.L. Gheesling, and A.J. Lesse. 2000. Characterization of P1-deficient isogenic mutant of Haemophilus influenzae biogroup aegyptius associated with Brazilian purpuric fever. Microbial pathogenesis. 28:145-155.
Servin, A.L. 2005. Pathogenesis of Afa/Dr diffusely adhering Escherichia coli. Clinical microbiology reviews. 18:264-292.
Sievers, F., A. Wilm, D. Dineen, T.J. Gibson, K. Karplus, W. Li, R. Lopez, H. McWilliam, M. Remmert, J. Soding, J.D. Thompson, and D.G. Higgins. 2011. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol. 7:539.
Singer, B.B., E. Klaile, I. Scheffrahn, M.M. Muller, R. Kammerer, W. Reutter, B. Obrink, and L. Lucka. 2005. CEACAM1 (CD66a) mediates delay of spontaneous and Fas ligand-induced apoptosis in granulocytes. European journal of immunology. 35:1949-1959.
Singer, B.B., I. Scheffrahn, R. Kammerer, N. Suttorp, S. Ergun, and H. Slevogt. 2010. Deregulation of the CEACAM expression pattern causes undifferentiated cell growth in human lung adenocarcinoma cells. PloS one. 5:e8747.
Singer, B.B., I. Scheffrahn, and B. Obrink. 2000. The tumor growth-inhibiting cell adhesion molecule CEACAM1 (C-CAM) is differently expressed in proliferating and quiescent epithelial cells and regulates cell proliferation. Cancer research. 60:1236-1244.
Singh, B., Y.C. Su, T. Al-Jubair, O. Mukherjee, T. Hallstrom, M. Morgelin, A.M. Blom, and K. Riesbeck. 2014. A fine-tuned interaction between trimeric autotransporter haemophilus surface fibrils and vitronectin leads to serum resistance and adherence to respiratory epithelial cells. Infection and immunity. 82:2378-2389.
Sirakova, T., P.E. Kolattukudy, D. Murwin, J. Billy, E. Leake, D. Lim, T. DeMaria, and L. Bakaletz. 1994. Role of fimbriae expressed by nontypeable Haemophilus influenzae in pathogenesis of and protection against otitis media and relatedness of the fimbrin subunit to outer membrane protein A. Infection and immunity. 62:2002-2020.
Slevogt, H., S. Zabel, B. Opitz, A. Hocke, J. Eitel, D. N'Guessan P, L. Lucka, K. Riesbeck, W. Zimmermann, J. Zweigner, B. Temmesfeld-Wollbrueck, N. Suttorp, and B.B. Singer. 2008. CEACAM1 inhibits Toll-like receptor 2-triggered antibacterial responses of human pulmonary epithelial cells. Nature immunology. 9:1270-1278.
St. Geme III, J.W. 2002. Molecular and cellular determinants of non-typeable Haemophilus influenzae adherence and invasion. Cellular microbiology. 4:191 - 200.
St. Geme III, J.W., and D. Cutter. 1996. Influence of pili, fibrils, and capsule on in vitro adherence by Haemophilus influenzae type b. Molecular microbiology. 21.
References
121
Stern, A., M. Brown, P. Nickel, and T.F. Meyer. 1986. Opacity genes in Neisseria gonorrhoeae: Control of phase and antigenic variation. Cell. 47:61-71.
Su, Y.C., O. Mukherjee, B. Singh, O. Hallgren, G. Westergren-Thorsson, D. Hood, and K. Riesbeck. 2015. Haemophilus influenzae P4 Interacts With Extracellular Matrix Proteins Promoting Adhesion and Serum Resistance. The Journal of infectious diseases.
Swords, W.E., B.A. Buscher, K. Ver Steeg Ii, A. Preston, W.A. Nichols, J.N. Weiser, B.W. Gibson, and M.A. Apicella. 2000. Non-typeable Haemophilus influenzae adhere to and invade human bronchial epithelial cells via an interaction of lipooligosaccharide with the PAF receptor. Molecular microbiology. 37:13-27.
Tchoupa, A.K., S. Lichtenegger, J. Reidl, and C.R. Hauck. 2015. Outer membrane protein P1 is the CEACAM-binding adhesin of Haemophilus influenzae. Molecular microbiology.
Tchoupa, A.K., T. Schuhmacher, and C.R. Hauck. 2014. Signaling by epithelial members of the CEACAM family - mucosal docking sites for pathogenic bacteria. Cell Commun Signal. 12:27.
Teglund, S., A. Olsen, W.N. Khan, L. Frangsmyr, and S. Hammarstrom. 1994. The pregnancy-specific glycoprotein (PSG) gene cluster on human chromosome 19: fine structure of the 11 PSG genes and identification of 6 new genes forming a third subgroup within the carcinoembryonic antigen (CEA) family. Genomics. 23:669-684.
Teixeira, A.M., J. Fawcett, D.L. Simmons, and S.M. Watt. 1994. The N-domain of the biliary glycoprotein (BGP) adhesion molecule mediates homotypic binding: domain interactions and epitope analysis of BGPc. Blood. 84:211-219.
Toleman, M., E. Aho, and M. Virji. 2001. Expression of pathogen-like Opa adhesins in commensal Neisseria: genetic and functional analysis. Cellular microbiology. 3:33-44.
Turbide, C., T. Kunath, E. Daniels, and N. Beauchemin. 1997. Optimal ratios of biliary glycoprotein isoforms required for inhibition of colonic tumor cell growth. Cancer research. 57:2781-2788.
van den Berg, B. 2005. The FadL family: unusual transporters for unusual substrates. Current opinion in structural biology. 15:401-407.
van den Berg, B., P.N. Black, W.M. Clemons, Jr., and T.A. Rapoport. 2004. Crystal structure of the long-chain fatty acid transporter FadL. Science. 304:1506-1509.
Van Eldere, J., M.P. Slack, S. Ladhani, and A.W. Cripps. 2014. Non-typeable Haemophilus influenzae, an under-recognised pathogen. The Lancet. Infectious diseases. 14:1281-1292.
Virji, M. 2009. Pathogenic Neisseriae: surface modulation, pathogenesis and infection control. Nature reviews. Microbiology. 7:274-286.
Virji, M., D. Evans, J. Griffith, D. Hill, L. Serino, A. Hadfield, and S.M. Watt. 2000. Carcinoembryonic antigens are targeted by diverse strains of typable and non-typable Haemophilus influenzae. Molecular microbiology. 36:784-795.
Virji, M., K. Makepeace, D.J.P. Ferguson, and S.M. Watt. 1996a. Carcinoembryonic antigens (CD66) on epithelial cells and neutrophils are receptors for Opa proteins of pathogenic Neisseriae. Molecular microbiology. 22:941-950.
Virji, M., S.M. Watt, S. Barker, K. Makepeace, and R. Doyonnas. 1996b. The N-domain of the human CD66a adhesion molecule is a target for Opa proteins
References
122
of Neisseria meningitidis and Neisseria gonorrhoeae. Molecular microbiology. 22:929-939.
Voges, M., V. Bachmann, R. Kammerer, U. Gophna, and C.R. Hauck. 2010. CEACAM1 recognition by bacterial pathogens is species-specific. BMC Microbiology. 10:117.
Voges, M., V. Bachmann, J. Naujoks, K. Kopp, and C.R. Hauck. 2012. Extracellular IgC2-like domains of CEACAMS mediate PI3K sensitivity during uptake of pathogens. PloS One accepted. 7 e39908.
Wang, H., and J.E. Cronan. 2003. Haemophilus influenzae Rd lacks a stringently conserved fatty acid biosynthetic enzyme and thermal control of membrane lipid composition. Journal of bacteriology. 185:4930-4937.
Wang, J., S.D. Gray-Owen, A. Knorre, T.F. Meyer, and C. Dehio. 1998. Opa binding to cellular CD66 receptors mediates the transcellular traversal of Neisseria gonorrhoeae across polarized T84 epithelial cell monolayers. Molecular microbiology. 30:657-671.
Wang, Y. 2002. The function of OmpA in Escherichia coli. Biochemical and biophysical research communications. 292:396-401.
Watt, S.M., A.M. Teixeira, G.Q. Zhou, R. Doyonnas, Y. Zhang, F. Grunert, R.S. Blumberg, M. Kuroki, K.M. Skubitz, and P.A. Bates. 2001. Homophilic adhesion of human CEACAM1 involves N-terminal domain interactions: structural analysis of the binding site. Blood. 98:1469-1479.
Wikstrom, K., G. Kjellstrom, and B. Obrink. 1996. Homophilic intercellular adhesion mediated by C-CAM is due to a domain 1-domain 1 reciprocal binding. Experimental cell research. 227:360-366.
Wong, C., Y. Liu, J. Yip, R. Chand, J.L. Wee, L. Oates, B. Nieswandt, A. Reheman, H. Ni, N. Beauchemin, and D.E. Jackson. 2009. CEACAM1 negatively regulates platelet-collagen interactions and thrombus growth in vitro and in vivo. Blood. 113:1818-1828.
Wu, S., M.M. Baum, J. Kerwin, D. Guerrero, S. Webster, C. Schaudinn, D. VanderVelde, and P. Webster. 2014. Biofilm-specific extracellular matrix proteins of nontypeable Haemophilus influenzae. Pathogens and disease. 72:143-160.
Yokoyama, S., C.J. Chen, T. Nguyen, and J.E. Shively. 2007. Role of CEACAM1 isoforms in an in vivo model of mammary morphogenesis: mutational analysis of the cytoplasmic domain of CEACAM1-4S reveals key residues involved in lumen formation. Oncogene. 26:7637-7646.
Zebhauser, R., R. Kammerer, A. Eisenried, A. McLellan, T. Moore, and W. Zimmermann. 2005. Identification of a novel group of evolutionarily conserved members within the rapidly diverging murine Cea family. Genomics. 86:566-580.
Zhang, H., A. Eisenried, W. Zimmermann, and J.E. Shively. 2013. Role of CEACAM1 and CEACAM20 in an in vitro model of prostate morphogenesis. PloS one. 8:e53359.
Zheng, J., K.K. Miller, T. Yang, M.S. Hildebrand, A.E. Shearer, A.P. DeLuca, T.E. Scheetz, J. Drummond, S.E. Scherer, P.K. Legan, R.J. Goodyear, G.P. Richardson, M.A. Cheatham, R.J. Smith, and P. Dallos. 2011. Carcinoembryonic antigen-related cell adhesion molecule 16 interacts with alpha-tectorin and is mutated in autosomal dominant hearing loss (DFNA4). Proceedings of the National Academy of Sciences of the United States of America. 108:4218-4223.
References
123
Zhou, H., A. Fuks, G. Alcaraz, T.J. Bolling, and C.P. Stanners. 1993a. Homophilic adhesion between Ig superfamily carcinoembryonic antigen molecules involves double reciprocal bonds. The Journal of cell biology. 122:951-960.
Zhou, H., C.P. Stanners, and A. Fuks. 1993b. Specificity of anti-carcinoembryonic antigen monoclonal antibodies and their effects on CEA-mediated adhesion. Cancer research. 53:3817-3822.