UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Pros en cons of natural and recombinant allergens

van Oort, E.

Link to publication

Citation for published version (APA):van Oort, E. (2003). Pros en cons of natural and recombinant allergens

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 19 Jun 2018

ChapterChapter 6

Substitutio nn of Pichia pastoris-derived recombinan t protein ss wit h mannos e containin g O- and N-linke d glycan ss decrease s specificit y of diagnosti c tests .

Ericaa van Oort1, Patrice Lerouge2, Pleuni G. de Heer1, Martial Séveno2, Laurent

Coquet3,, Piet Modderman1, Loïc Faye2, Rob C. Aalberse1, Ronald van Ree1.

'Sanquin,, Department of Immunopathology and Laboratory for Experimental and Clinical

Immunology,, Academic Medical Center, University of Amsterdam, 1066 CX Amsterdam, The

Netherlands.. 2UMR CNRS 6037 and 3UMR CNRS 6522, IFRMP 23, Faculté des Sciences,

Universitéé de Rouen 76821 Mont Saint Aignan, France.

(submittedfor(submittedfor publication)

137 7

SubstitutionSubstitution of Picbia-derived recombinant proteins with O-and K-linked mannose

Abstract t

Background:: Recombinant proteins from Pichia pastoris are frequently

evaluatedd as diagnostic tools. Objective: The objective of this study was to investigate

whetherr glycosylation by Pichia pastoris interferes with the specificity of diagnostic

tests.. Methods: An auto-antigen involved in Wegener's disease (protease 3) and two

majorr inhalant allergens from grass pollen (Dae g 5) and house dust mite (Der p 1)

weree produced as recombinant molecules in Pichia pastoris. O-linked glycans on Dae g

55 were characterized bv MALDI-TO F MS. Immune-reactivity of the recombinant

proteinss was compared to that of their natural counterparts by ELISA and RAST as

welll as bv ELISA- and RAST-inhibition. Results: In contrast to the non-glycosylated

naturall allergen, recombinant Dae g 5 was shown to carry at least 2 small mannose

containingg O-glycans. Both these O-glycans and the N-linked glycans on recombinant

proteasee 3 and recombinant Der p 1 were shown to be recognized in ELISA by IgG

antibodiess in sera of 49%-87% of healthy individuals. These IgG responses were

closelyy correlated. The natural autoantigen and allergens were not recognized by IgG

antibodiess from healthy subjects. The carbohydrate nature of the epitopes recognized

byy IgG on the recombinant proteins was confirmed by inhibition studies with

mannosee and yeast mannan. IgE recognition of yeast glycans was observed in 2 out of

99 ABPA positive sera. Conclusion: Production of recombinant molecules in yeast (or

moulds)) can introduce IgG-binding glycans that negatively affect the specificity of

diagnosticc tests.

138 8

ChapterChapter 6

Introductio n n

Recombinantt technology has revolutionized the development of diagnostic

testss and the production of therapeutic proteins. It allows the production of disease-

relatedd molecules in reproducible fashion and large quantities. In the early years,

productionn of recombinant proteins was largely limited to prokaryotic expression

systemss like Escherichia coli. Since the machinery of prokaryotic organisms for post-

translationall modifications is limited, eukaryotic expression systems have rapidly been

introducedd next to E.coli. These include yeasts and moulds, insect cells, mammalian

cellss and plants. Yeast expression systems like Saccharomyces cerevisiae and Pichia pastoris

combinee the advantages of rapid and easy molecular manipulation with the ability to

introducee post-translational modifications, such as glycosylation and efficient

formationn of disulfide bonds. Glycosylation is often essential for conformation and

stabilityy of the proteins, but at the same time yeast glycans might have antigenic,

immunogenicc and immune-modulating properties for mammals (V4\ Antigenicity of

recombinantt yeast glycoproteins might affect the specificity of diagnostic tests,

whereass immunogenicity and immune-modulating properties may be relevant in

relationn to their behavior as therapeutic molecules. Yeast-derived mannans have been

shownn to bind to serum proteins, like mannan-binding protein ®, upon binding the

glycann structures may be able to activate the innate immune system via the mannan-

bindingg pathway.

S.cerevisiaeS.cerevisiae was the first yeast used for production of recombinant proteins,

butt at present P. pastoris is far more frequently used. One of the reasons why interest

inn S.cerevisiae has diminished, is its tendency to hyper-glycosylate. Proteins expressed in

S.cerevisiaeS.cerevisiae are often modified by substitution with N-linked glycans with a core

structuree harboring at least 9 to 13 mannose residues (Mang.nGlclSLA^), so-called

high-mannosee glycans. However, extensions of outer chains with up to 200 mannose

residuess may occur (f)). Usually these chains are linked to the a-l,3-branch on the core,

butt in some cases short a-l,2-linked mannose chains are attached. P. pastoris is

generallyy regarded to produce smaller high-mannose N-glycans. In both S\ cerevisiae

andd P. pastoris, charged groups can be linked to N-glycans in the form of mannose

phosphatee diesters C7-9). Yeast can also substitute serines and threonines with O-linked

glycans.. O-glycosylation begins in the endoplasmic reticulum by addition of a single

mannose.. After transportation to the Golgi, several a-(l,2)-linked mannoses can be

attachedd that may finally be capped with one or two a-(l,3)-linked mannoses (,0).

139 9

SubstitutionSubstitution of Pichia-derived recombinant proteins with O-and N-linked mannose

Specificc antibodies to yeast glycans have been shown to be directed to side-

chainss of yeast carbohydrate moieties. Such antibodies were for example described

againstt tri-and tetrasaccharide side chains of glycans found in the yeast Saccharomyces

lactislactis m) . In Saccharomyces cerevisiae a three-unit structure containing (1,3),(1,2),(1,2)-

linkedd Ot-D-mannopyranose was shown to be an antigenic structure in'K Recendy, we

producedd an auto-antigen and two inhalant allergens in P. pastoris for application in

diagnosticc tests. The auto-antigen, protease 3 (Pr3), involved in Wegener's disease (13>,

hass a putative N-glycosylation site and was found to be glycosylated by the yeast. The

recombinantt molecule binds Pr3-specific IgG in a similar way as the natural protein.

Thee major house dust mite allergen from Dermatophagoides pteronyssinus, Der p 1, is a

cysteinee protease with two consensus sequences for N-glycosylation, one in the pro-

sequencee of the pro-enzyme and one in the mature enzyme (14>. Some controversy

existss on the question whether, and if so at which position, the natural allergen is

glycosylatedd (15-16). The allergen was shown to be glycosylated upon expression in P.

pastoris,pastoris, and was reported to have similar IgE reactivity as its natural counterpart when

testingg sera of house dust mite allergic patients 'X1\ Unexpectedly, also the major

DactylisDactylis glomerata grass pollen allergen Dae g 5 was glycosylated by this yeast. The

primaryy sequence of Dae g 5 does not harbor any consensus sequences for N-

glycosylationn and the natural allergen is not glycosylated. Most likely, the allergen is

substitutedd with O-glycans during expression in P. pastoris. Despite these differences

betweenn natural and recombinant Dae g 5, both molecules displayed very similar IgE

bindingg characteristics when sera from grass pollen allergic patients were tested <18).

Thee aim of this study, was to investigate the structure of the O-linked

glycan(s)) on rDac g 5, and to examine whether yeast glycans influence the

performancee of immunoassays for the diagnosis of autoimmune and allergic diseases.

Materia ll & methods

ProductionProduction of recombinant proteins

Recombinantt Dae g 5 (rDac g 5) and proDer p 1 (rproDer p 1) were

expressedd in Pichia pastoris as described before (17;18). For the expression of

enzymaticallyy inactive recombinant Pr3 (rPr3), its active site serine (Ser 176) was

changedd to an alanine using the overlap extension method (19). The resulting cDNA

wass sequenced and subcloned into the P. pastoris expression vector pPicZaA, which

140 0

ChapterChapter 6

wass subsequently transformed to the P. pastoris strain KM71, according to the

manufacturerss instructions (Invitrogen, San Diego, Calif., USA). A Pr3-secreting clone

wass cultured at pH 5.0 in a Bioflo 3000 fermentor (New Brunswick Scientific,

Nijmegen,, The Netherlands), and synthesis of Pr3 was induced by addition of

methanol.. After 7 days of culture, the yeast supernatant was harvested, concentrated

andd dialyzed against PBS (pH 7.4) using a capillary dialyzer (Hemoflo F5, Fresenius,

Badd Homburg, Germany).

PurificationPurification of recombinant and natural proteins

Naturall Dae g 5 (nDac g 5) was purified from a Dactylis glomerata grass pollen

extractt by affinity-chromatography with a monoclonal antibody (4B1) coupled to

CNBr-activatedd Sepharose 4B (Amersham Pharmacia Biotech, Uppsala, Sweden) (2°).

Afterr the protein was allowed to bind, the column was washed with PBS, and

subsequendyy eluted with 0.1 mol/L glycine (pH 2.5). After dialysis against PBS the

sampless were stored at -20°C till further use. Natural Der p 1 (nDer p i ) was purified

fromm a spent culture medium extract (CSL, Melbourne, Australia) using a Sepharose-

coupledd mAb directed to Der p 1 (21). Elution was performed with 50%

ethyleneglycol/5mMol/LL lysine (pH 11). rDac g 5 and rproDer p 1 were purified from

yeastt culture supernatants by the same method as described for their natural

counterparts. .

Precedingg affinity-purification of rPr3 from dialyzed and concentrated yeast

culturee supernatant, Triton X-100 was added to a final concentration of 0.05% (v/v).

Affinity-purificationn was performed by an anti-Pr3 monoclonal antibody 12.8 (22)

coupledd to Sepharose. Bound protein was eluted with a 5mMol/L lysine buffer (pH

11.5)) containing 0.05% Tx-100.

Purityy of all three recombinant molecules was assessed by SDS-

PAGE/silverstainingg (Novex, San Diego, CA, USA). Protein concentrations were

determinedd using the BCA method as described by the manufacturer (Pierce,

Rockford,, IL, USA).

SD*S-PAGEI'silverstainingSD*S-PAGEI'silverstaining andglycan detection by Conj\ binding

Affinity-purifie dd allergens were separated by SDS-PAGE on 4-12% gradient

gelss (Novex) as described by the manufacturer, and silverstained according to the

Bexell Biotechnology silverstaining kit protocol (Bexel Biotechnology, Union City

141 1

SubstitutionSubstitution of Pichia-derived recombinant proteins with O-and K-linked mannose

Blvd,, Union City CA). Affinity-purified rPr3 was separated by SDS-PAGE on 8-18%

ExcelGell and silverstained according the manufacturer's instructions (Amersham

Pharmaciaa Biotech AB).

Purifiedd natural and recombinant proteins were electroblotted onto

nitrocellulosee membrane followed by an overnight incubation in Tris-buffered

saline/0.1%% Tween-20 (TBS-T). Subsequently, the blot was incubated with

Concanavalinn A (25 Hg/ml) (Sigma, St Louis, MO, USA) in TBS-T/1 mMol/L

MgCL/ll mMol/L CaCb for 90 min. After washing with the same buffer, the

membranee was incubated with 50 pg/ml horseradish peroxidase (HRP; Sigma) for 60' (23).. Glycoproteins were visualized with 1 tablet of DAB in 40 ml FLO (10 mg

Diaminobezidine-tetrahydrochloride)) (Kem-En-Tec, Copenhagen, Denmark). The

reactionn was started with 40 u.1 of 30% FLO? and ended by washing the blots in FLO.

SeparationSeparation of rDac g 5 by 2D gel electrophoresis

Fiftyy micrograms of rDac g 5 were dissolved in isoelectric focusing buffer (final

volume,, 400 ul) consisting of 5 Mol/L urea, 2 Mol/L thio-urea, 33 mMol/L 3-[(3-

cholamidopropyl)-dimethylammonio]-l-propanesulfonatee (CHAPS), 2 mMol/L

tributyll phosphine, 10 mMol/L dithiothreitol, 2% (v/v) carrier ampholytes (pH 3.5 to

10;; Sigma), and 0.4% (wt/vol) bromophenol blue. The first-dimension gel separation

wass carried out with Immobiline Dry Strips NL (pH 3 to 10; Amersham Pharmacia

Biotech).. The second-dimension separation was obtained by SDS-PAGE with a 12.5%

(wt/vol)) polyacrylamide resolving gel (width, 16 cm; length, 20 cm; thickness, 0.75

mm). .

TrypticTryptic digestion and isolation of glycopeptides from rDac g 5

Fiftyy u.g of rDac g 5 were run on 15% SDS-PAGE and subsequently stained

withh Coomassie Brilliant Blue. The protein band corresponding to the recombinant

allergenn was excised and cut into small fragments. Fragments were washed 3 times

withh 100 uL of a solution of 0.1 Mol/L NFÏ4HCO3 / CFLCN (1/1) for 15 minutes

eachh time and dried out. Protein was then digested with 7.5 u,g of trypsin in 200 uL of

0.055 mol/L NH4HCO3, at 37°C for 16 h. Two hundred uL of CH3CN were added

andd the supernatant was collected. Gel fragments were then washed with 50 uL of 0.1

Mol/ LL NH4HCO3, then with 50 \xL CH3CN again and finally with 50 uL 5% formic

acid.. Al l supernatants were pooled and lyophilized. Peptide separation by HPLC was

142 2

ChapterChapter 6

carriedd out on a Cis reverse-phase column (4.5x250 mm) with a linear gradient of

CH3CNN in TFA 0.1%. Fractions were collected, lyophilized and analyzed by MALDI -

TOF-MS. .

Matrix-AssistedMatrix-Assisted luiser Desorption loni^ation-Time of Flight-Mass Spectrometry

(MALDI-TOF-MS) (MALDI-TOF-MS)

MALD II TOF mass spectra were measured on a Micromass (Manchester, U-

K)) Tof. Spec. E. MALDI-TO F mass spectrometer equipped with a nitrogen laser

operatingg at 337 nm. Mass spectra were performed in positive ion mode using a

cyano-4-hydroxycinnamicc acic (10 mg/ml) as matrix. Each of the samples (2 jxL) was

mixedd with 2 jiL of the matrix solution in TFAiCFbCN (1.75: 0.75 ; v/v). Two iL of

thiss solution was applied on a stainless steel sample plate and allowed to dry under

vacuum.. Spectra were acquired in the reflectron mode. For collision induced

dissociationn (CID) and post source decay (PSD) experiments, selection of the

precursorr ion was carried out using a Bradbury-Nielsen ion gate with a m/Am=100

resolution.. CID was carried out by collision with helium at a pressure of 106 mbar in

thee collision cell. Extraction with a delay time of 600 ns was used. To record the full

PSDD spectra, reflector voltage was decreased into successive 25% steps leading to ten

spectrall segments using 30 laser shots per segment. PSD was calibrated using ACTH

(fragmentt 18-39).

FUSA.FUSA. for detection of specific IgG antibodies

Naturall and recombinant allergens were coated overnight at 4°C to microtiter

platess (Maxisorb, Nunc, Roskilde, Denmark) at a concentration of 0.25 (ig/100 ji l

coatingg buffer (0.1 mol/L NaHCCb / pH 8.5). Non-specific binding was decreased by

applicationn of high-performance ELISA (HPE) buffer (Sanquin, Amsterdam, The

Netherlands)) for dilution of the sera (1:30 to 1:800) and of the HRP-labeled

conjugatee goat anti-human IgG antibody (1:20.000) (Sigma). Color development was

initiatedd by adding 100 \i\ of 10 mg/ml TMB (3,3', 5,5'-tetramethylbenzidine) in

sodiumm acetate, pH 5.5 and 10 (j.1 3% H2O2. The reaction was stopped by adding 2 M

H2SO44 after which the absorbance was measured at 450 nm and 540 nm.

ELISA-inhibitionn experiments were performed by pre-incubation of serum with 2-0-

a-D-mannopyranosyl-D-mannopyranose,, 3-O-a-D-mannopyranosyl-D-

143 3

SubstitutionSubstitution of Pichia-derived recombinant proteins with O-and N-linked mannose

mannopyranosee or a-(1,3), a-(l,6)-Mannotriose-a-methyl-glycoside (Sigma) (0-16

mM),, prior to the addition to allergen-coated wells.

Forr detection of Pr3-speciflc IgG auto-antibodies, Pr3-specific mAb 12.8 (22>

wass coated to microliter plates and used to capture P. pastoris-derived rPr3 or natural

Pr33 from PMN lysates (—1 Wells were then incubated for 1 h at room temperature

withh human sera, diluted 1:20, 1:80, 1:320 and 1:1280 in a buffer containing 20

mMol/LL Tris-HCl (pH 6.8)/0.3 Mol/L NaCl/0.25% Tween-20/1.0% BSA/5.0%

normall goat serum. Bound IgG was detected by HRP-conjugated mAb against human

IgGG (Sanquin). Binding was expressed in arbitrary units, using a pool of Pr3-positive

ANCAA sera (set at 10,000 AU/ml) for construction of a standard curve. Sera were

consideredd to be positive for PR3 at titers >10 U/ml. ELISA inhibition was in this

casee performed with D-mannose or Saccharomyces terevisiae-derived mannan (Sigma) as

inhibitors. .

RASTRAST and RAST-inhibition

AA radioallergosorbent test (RAST) was performed as described previously (24'>.

Briefly,, both natural and recombinant allergens were coupled to CNBr-activated

Sepharosee 4B at 100 n.g/100 mg Sepharose. Human serum samples (50 (J.1) were

incubatedd with 0.5 mg Sepharose in a final volume of 300 (il PBS / 0.3% BSA / 0.1%

Tweenn (PBS-AT). After overnight incubation, unbound material was washed away and

500 ji l of 125I-labeled sheep antihuman IgE (SH25, kindly provided by S.O. Stapel,

Sanquin)) was added. After overnight incubation and washing, radioactivity was

measuredd in a gamma-counter. Results were expressed in international units/ml

(IU/ml)) IgE, which were calculated using a standard curve of a human/mouse

chimericc IgE antibody directed to Der p 2 and Sepharose-coupled rDer p 2 <25). A

resultt greater than 0.30 IU/ml was regarded as positive.

Forr RAST-inhibition (24) a pre-incubation of a human serum with 50 ul rDac

gg 5 (55 (ig/ml), nDac g 5 (7.5 \xg/ml) and Aspergillus fumigatus mould extract (9 mg/ml)

wass allowed, prior to addition of Sepharose-coupled rDac g 5. Samples were further

treatedd as described for the RAST.

144 4

ChapterChapter 6

Sera Sera

Seraa of healthy volunteers (n=39) were used to evaluate the specificity of

ELISAA for the measurement of IgG antibodies against yeast-derived glycoproteins

(rDacc g 5, rproDer p 1 and rPr3). In case of rPr3, sera of six patients with Wegener's

diseasee were used as positive control.

Too evaluate the specificity of rDac g 5 and rproDer p 1 RAST, sera of 9

patientss with allergic bronchopulmonary aspergillosis (ABPA) (kindly provided by S.O.

Stapel,, Sanquin) specific IgE were used.

Results s

ProductionProduction of purified natural and recombinant proteins

Thee major grass pollen group 5 allergen Dae g 5 (18), the pro-form of the

majorr house dust mite group 1 allergen Der p 1 (17), and the autoantigen Pr3 were

successfullyy expressed in the yeast Pichia pastoris and affinity-purified from culture

supernatants.. Natural homologues of Dae g 5 and Der p 1 were isolated from Dactylis

glomerataglomerata grass pollen extract and from Dermatophagoides pteronyssinus house dust mite

extract,, respectively. SDS-PAGE/silverstaining analysis revealed that rDac g 5 (figure

1,, panel A) and rproDer p 1 (figure 1, panel B) migrate at significantly higher apparent

M rr than their natural counterparts and are both detected with ConA. Natural Dae g 5

wass not detected by ConA, whereas nDer p 1 was only weakly stained.

Panell A Panell B

1 2 3 4 55 1 2 3 4 5

Figur ee la-b: SDS-

PAGE/silverstainingg and Con A

bindin g g

Panell A: lane 1-3 shows a SDS-

PAGE/silverstainingg of affinity-

purifiedd rDac 5.01, nDac g 5 and a

100 kDa protein ladder (Life

Technologies),, respectively. Showing

aa shift in apparent Mr for

recombinantt Dae g 5. Lane 4 and 5

representt the results obtained by

145 5

SubstitutionSubstitution qfPichia-derived recombinant proteins with O-and N-linked man nose

Conn A binding, showing positive staining of rDac g 5 and no staining of nDac g 5, respectively.

Panell B: lane 1-3 shows a SDS-PAGE/silverstaining of a precision protein ladder (Biorad), rpro-Der p 1

andd affinity-purified nDer p 1, respectively. Showing a shift in apparent molecular mass for recombinant

pro-Derr p 1, compared to the calculated Mr of 34 kDa. Lane 4 and 5 represent the results obtained by

Conn A binding, showing positive staining for both rpro-Der p 1 and nDer p 1.

MALDI-TOF-MS MALDI-TOF-MS

Gass chromatographic analysis of hydrolysed rDac g 5 demonstrated the

presencee of mannose exclusively (not shown). To further analyse the nature of glycan

moietiess on rDac g 5, the allergen was separated by 2D SDS-PAGE into three distinct

spots,, each of which were digested in-gel with trypsin. Resulting tryptic peptides were

analyzedd by MALDI-TO F MS. The tryptic peptides detected from all three spots were

identicall and covered most of the known sequence of Dae g 5 (fig. 2A). The N- and

C-terminall peptides were not found. In none of these three digests, glycopeptides

weree detected. Subsequently, rDac g 5 was separated by SDS-PAGE only, and the

resultingg protein band was digested in-gel with trypsin. Tryptic peptides were isolated

byy reverse-phase HPLC and individually analysed by MALDI-TO F MS. The same

trypticc peptides as found after 2D SDS-PAGE were present again. In addition, one

HPLCC peak was shown to be a mixture of three peptides with average Mr of 1955,

21177 and 2279 Da, respectively (fig.2B). The mass of these peptides could not be

matchedd to any part of the reported Dae g 5 sequence. The mass increments between

thee smallest and the larger two peptides were 162 Da and 324 Da, respectively. This

agreess exactly with the mass of one and two mannose residues. Finally, a peptide

correspondingg to the 36 N-terminal amino acids was found (3540 Da), mixed with a

seriess of four peptides with mass increments agreeing with substitutions of 4, 5, 6 or

77 mannose residues (fig. 2C).

Figur ee 2A: rDac g 5 amino acid sequence

ADAGYTPAAAATPATAGGKAMTEEQTLIEDVNAGFKAAVA AA ASS APPA

DKFKTFEATFTAACI<^VNIAAVATKVPLFVAKLDAAYA\'A.YKTATGPT P P

EAKYDAFVAALTEALRVIA GG ALEVHAVKPA A EL V PAAKIPAGELQIVD

KdDAAYIAATAANAAPANDKFTVFEGAFNKAIKESTGGAYESYKFIPT L L

EAAVKOAYAATVAAAPEVKYAVFEAALTKAITAMSEAQKVATPAAVA T T

GAATAAASAATGATAAAGGYK V V

146 6

V3 3

% %

5 5

pa a

mm -

08 8

CS S

S S

5: : :

2 2 5: : 5: :

' t t

ES S

!>5 5

u u

a a os s

+^ +^ en n

k k

00 0 " "

ChapterChapter 6

Figur ee 2a-c: MALDI-TO F mass spectra of glycopeptides isolated from rDac g 5 after trypti c

digestionn and separation by reverse-phase HPLC .

Figur ee 2a: Peptides detected by MALDI-MS : underlined N-and C-terminal peptides were not detected.

Sequencee in bold was observed in the 2-D gel experiments only. The rest of the sequence was detected in

bothh 1-D and 2-D gel experiments. Figure 2b: Glycopeptides (average Mr:2279, 2117 and 1955 Da)

differingg by one and two mannose residues. Figure 2c: Peptide 1 -36 Mr 3540 Da was found, mixed with

44 peptides of which the Mr is in agreement with substitutions of 4, 5, 6 or 7 Mannose residues.

InIn vitro diagnosis of Wegener's disease with natural and recombinant Pr3

Forr diagnosing Wegener's disease, IgG autoantibodies to Pr3 were measured

byy sandwich ELISA. For patients with established Wegener's disease, IgG titers

againstt natural Pr3 (nPr3) and rPr3 closely correlated (not shown). Healthy controls

weree negative in ELISA with nPr3 but frequentiy positive when rPr3 was used

(fig.3A).. This binding of IgG to rPr3 was inhibited up to 82% by yeast mannan (250

fj.g/ml)) and up to 99% with high concentrations of free mannose (50 mM) (fig. 3B).

Partiall inhibition of up to 43% was observed with a-(l,3), a-(l,6)-Mannotriose-a-

methyl-glycosidee (fig. 3C). Partial or no inhibition of IgG binding was observed for

patientss with Wegener's disease (<43% with mannan and <31% with mannose).

350 0

300 0

250 0

"gg 20 0

PP 15 0

100 0

50 0 IDD 1 « « bb c f W5

Samplee numbers \\'6 6

350 0

300 0

250 0

|| 200

PP 150 < <

100 0 50 0

0 0

cc f WS Samplee numbers

VV6 6

Figur ee 3a-c: ELIS A for detection of anti-Pr3 IgG auto-antibodies.

Figur ee 3a: Healthy controls (b, c and f) and patients with established Wegener's disease (W5 and W6)

weree tested in an ELISA with nPr3 and yeast-derived rPr3. The white bars represent the responses

towardss rPr3, the black bars the responses towards nPr3. Showing that rPr3 leads to false positive results.

Figur ee 3b: Healthy controls (b, c and f) and Wegener's disease patients (W5 and W6) were pre-incubated

withh 250 Hg/ml mannan (striped bars) and 50 mM mannose (grey bars), which shows that both mannan

149 9

SubstitutionSubstitution ofPichia-derived recombinant proteins with O-and N-linked mannose

andd mannose almost completely inhibit binding of IgG to rPr3 in sera of healthy controls, in contrast to

Wegener'ss sera, W5 and W6. The white bars represent the responses towards rPr3 and the black bars

towardss nPr3.

Inhibitionn in mM glycan

Figur ee 3c: In this graph, rPr3 was directly

coatedd to the microliter plate. Two healthy

controlss (pi 5 ( ° , A , D ) and p l7 ) were

pre-incubatedd with 2-O-a-D-mannopyranosyl-

D-mannopyranosee (A and A) , 3-O-a-D-

mannopyranosyl-D-mannopvranosee (a and )

andd a-(l,3), a-(l,6)-Mannotriose-a-mefhyl-

glycosidee (o and , respectively. False positive

bindingg towards rPr3 in both sera was partially

inhibitedd by Mannotriose.

Too further confirm that mannose-containing glycans on P. pastoris-derived

glycoproteinss can cause false-positive ELISA results, IgG antibodies were measured

againstt N-glycosylated rproDer p 1 and O-glycosylated rDac g 5 and against their

naturall homologues in sera of 39 healthy volunteers. IgG against the yeast-derived

recombinantt allergens was detected in most serum samples (fig.4).

• • • • •• • • ••• • • ••••• • • ••••• • • •• ••• • ••••• • • ••••• • • • •

•• •• • • • •• • •

•• •• • •• • •• • • • • •• • •• • • • •• •

• • • •

• •

• • • •

•• • •

•• ••• • •• ••• • •• ••• • •• • •• ••• • ••••• • • •• • •• • •

natt Dacg5 r-Dacg5 5

Figur ee 4: Detection

off IgG antibodies

againstt yeast-derived

glycanss in healthy

controll sera

Thee IgG binding of

399 healthy controls

towardss natural non-

glycosylatedd Dae g 5

andd its yeast-derived

O-linkedd glycosylated

counterpartt were

compared.. Showing a

significantt difference

150 0

ChapterChapter 6

inn response between both proteins. The same pattern is observed when nDer p 1 and rpro-Der p 1 were

compared. .

Thiss was not the case for the natural allergens. IgG titers against rPr3 and the two

unrelatedd allergens were significant correlated as analyzed by linear regression (fig.5).

1,6 6

ii '>4

^ PP i.2

°° a o,8 uu * 0,6

cc _£ 0,4

11 I 0,2 t z+St V j CC o k* * *

R2== 0.4054

1,22 1,4 1,6

E E 3 3 * *

l O O

o o IT) )

t t u u c c

1 1 - Q Q

< <

1.6 6

1.4 4

'c,'c, 1.2 88 1

oo O-8

frfr 0,6 JJ O.4

0,2 ZZ 0

>*>* X

00 0,2 0,4 0,6 (

0.661 1

00 0,2 0,4 0,6 0,8 1

Absorbancee 450/540nm Absorbance 450/540nm

N-linkedd rPr3 N-linked rPr3

Figur ee 5a-b: Comparison of IgG responses towards O-and N-linked glycans.

Figur ee 5a: Correlation between N-linked glycosylated rPr3 (x-axis) and O-linked glycosylated Dae g 5 (y-

axis).. Showing a clear correlation between both proteins, but with a preference for N-linked glycosylated

rPr3.. Figure 5b: Correlation between N-linked glycosylated rPr3 (x-axis) and N-linked glycosylated rpro-

Derr p 1 (y-axis). Showing a higher correlation between these two N-linked glycosylated proteins than

comparedd to O-linked Dae g 5. The highest responses are found with rpro-Der p 1, containing two

consensuss N-linked glycosylation sites.

InIn vitro diagnosis of allergy by KAST with natural and recombinant allergens

Too evaluate the potential interference of yeast glycans with the specificity of

RAST,, sera of ABPA patients with Aspergillus fumigatus-sped&c IgE antibodies (n=9)

weree tested in a RAST with nDac g 5 and rDac g 5 and/or nDer p 1 and rproDer p 1.

Twoo sera (22%) had significantly higher IgE titers to rDac g 5 than to nDac g 5: 2.1

IU/mll versus 0.6 IU/ml and 1.8 IU/ml and 0.5 IU/ml , respectively. These results

suggestt that IgE responses to yeast glycans can give rise to false-positive in vitro tests

inn patients with mould-and/or yeast-reactive IgE antibodies when yeast recombinant

proteinss are used.

151 1

SubstitutionSubstitution ofPichia-derived recombinant proteins with O-and N-linked mannose

IgEE binding to rDac g 5 was only partially inhibited by nDac g 5 (<35%) but

almostt completely by Aspergillus fumigatus extract (>88%) in one serum. This suggests

thatt mould glycans are cross-reactive with the glycans on the P. pastoris-derived grass

pollenn allergen (fig 6).

Figur ee 6: Specificity of in vitro

allergyy diagnosis by RAST

Thee interference of yeast glycans in

RASTT was demonstrated in 2/9

sera.. In this graph, it is shown that

inn one serum binding of IgE to

rDacc g 5 coupled to Sepharose

couldd be inhibited up to 88% by

AspergillusAspergillus extract (M). In both sera,

partiall inhibition (<35%) was

u u o o .s s p p d d .5 5 -? ?

11 no

80 0

611 1

4(1 1

2U U

0 0

pf l477 pf149

ABPAA patient numbers

obtainedd by nDac g 5 ) and complete inhibition (99%) was reached by rDac g 5 (D).

Discussion n

Glycosylationn of proteins expressed in the yeast P. pastoris has been studied in

somee detail. Both N- and O-linked glycans are dominated by mannose residues (3;7

i0;26;27)) \^e ha ve demonstrated that expression of the major grass pollen allergen Dae g

55 in P. pastoris results in substitution with mannose containing O-glycans not found in

thee natural allergen. The exact structure has not yet been elucidated, but is most likely

inn close agreement with earlier reports showing that P. pastoris substitutes serines and

threoniness with 1 to 4 mannose residues P;27-30)_ The N-terminal peptide of rDac g 5

wass shown to contain most likely 4 to 7 mannose residues. This suggests that two or

moree of the four threonines found in the first 36 amino acids of rDac g 5 are O-

glycosylated.. On the basis of earlier mass-spectrometric comparison of rDac g 5 and

nDacc g 5, the former was found to be 2500 Da larger than the natural allergen (18).

Thiss suggests that the recombinant allergen harbors more O-glycans than the ones

accountedd for in the N-terminal peptide (7 mannose residues: 1134 Da). Mannose

phosphatee diester groups have also been reported for P. pastoris-derived glycoproteins

152 2

ChapterChapter 6

t7-9).. Whether phosphorylation also plays a role for O-linked glycosylated rDac g 5 is

stilll under investigation.

Proteinss with N-glycosylation sites like the auto-antigen Pr3 and the major

housee dust mite allergen Der p 1 are both glycosylated by P. pastoris, most likely with

high-mannosee glycans. In this study, we did not analyze the structure of N-glycans on

bothh proteins, but clear recognition by ConA supports the presence of high-mannose

N-glycans.. When rPr3 was used in an ELISA for diagnosis of Wegener's autoimmune

disease,, around two-third of the healthy controls was falsely positive. This was not the

casee when neutrophil lysates were used as a source for nPr3. Several observations

supportt IgG binding to mannose-rich yeast glycans as the explanation for the low

specificityy of the rPr3-based ELISA. There was a strong correlation between IgG

responsess against rPr3 and two structurally completely unrelated allergen molecules,

i.e.. rDac g 5 and rproDer p 1. The only common structural feature between the three

moleculess is that they are all substituted with mannose containing glycans. The

correlationn found between IgG responses to O-glycans on rDac g 5 and N-glycans on

Pr33 suggests that IgG antibodies against yeast glycans are cross-reactive with the two

typess of glycan structures. The mannose-dominated nature of the IgG epitope

causingg false-positive ELISA results was further supported by the observation that

yeastt mannan, and even free mannose residues almost completely inhibit the binding

too Pr3.

Thee high prevalence of IgG antibodies directed to yeast glycans observed in

healthyy subjects (49% to 87%) is most likely explained by very regular environmental

andd dietary exposure to yeast and mould carbohydrate structures. In sera of a specific

patientt group 22% showed false positive IgE antibody responses to yeast glycans.

Severall other investigators have also reported a significant role for mannan-specific

IgEE antibodies in mould- and yeast-sensitized patients (31;32\ In our study, 2 sera out of

99 patients with ABPA demonstrated IgE antibodies against mould glycans that cross-

reactedd with the glycans on rDac g 5. Further serum screening will have to clarify

whetherr IgE antibodies against mannose-rich carbohydrate structures are relevant for

otherr mould or yeast-allergic patients, thereby decreasing specificity of RAST as a

diagnosticc tool for type I allergies.

Inn summary, this study has revealed that expression of proteins in Pichia

pastorispastoris results in their substitution with N-and O-glycans that are recognized by IgG

antibodiess in the majority of sera of healthy controls. This seriously decreases the

specificityy of diagnostic tests with such recombinant proteins. Glycosylation of

153 3

SubstitutionSubstitution of Pichia-derived recombinant proteins with O-and K-linked mannose

diagnosticc reagents produced in Pichia pastoris should therefore be carefully evaluated.

Iff prevention of unwanted glycosylation proves to be impossible, alternative

expressionn systems should be considered. Several new strains of E.co/i have become

availablee that better facilitate disulfide bridge formation and correct folding of

recombinantt proteins than the earlier strains. The choice between eukaryotic and

prokarvoticc expression systems depends on the characteristics of each individual

protein. .

Acknowledgements s

Thiss study was financially supported by Stallergènes S. A., Altadis, ANYA R

andd CNRS.

154 4

ChapterChapter 6

Referencee List

(1)) Ballou C. Structure and biosynthesis of the mannan component of the yeast celll envelope. Adv Microb Physiol 1976; 14(11):93-158.

(2)) Hayette MP, Strecker G, Faille C, Dive D, Camus D, Mackenzie DW et al. Presencee of human antibodies reacting with Candida albicans O-linked oligomannosidess revealed by using an enzyme-linked immunosorbent assay andd neoglycolipids. J Clin Microbiol 1992; 30(2):411-7.

(3)) Letourneur O, Gervasi G, Gaia S, Pages J, Watelet B, Jouvet M. Characterizationn of Toxoplasma gondii surface antigen 1 (SAG1) secreted fromm Pichia pastoris: evidence of hyper O-glycosylation. Biotechnol Appl Biochemm JID - 8609465 2001; 33(Pt l):35-45.

(4)) Podzorski RP, Gray GR, Nelson RD. Different effects of native Candida albicanss mannan and mannan-derived oligosaccharides on antigen-stimulatedd lymphoproliferation in vitro. J Immunol 1990; 144(2):707-16.

(5)) Kawai T, Suzuki Y, Eda S, Ohtani K, Kase T, Sakamoto T et al. Molecular andd biological characterization of rabbit mannan-binding protein. Glycobiologyy 1998; 8(3):237-44.

(6)) Herscovics A, Orlean P. Glycoprotein biosynthesis in yeast. FASEB J JID -88044844 1993; 7(6):540-50.

(7)) Hirose M, Kameyama S, Ohi H. Characterization of N-linked oligosaccharidess attached to recombinant human anfithrombin expressed in thee yeast Pichia pastoris. Yeast 2002; 19(14):1191-202.

(8)) Varki A, Kornfeld S. Structural studies of phosphorylated high mannose-typee oligosaccharides. J Biol Chem 1980; 255(22): 10847-58.

(9)) Ballou L, Hernandez LM, Alvarado E, Ballou CE. Revision of the oligosaccharidee structures of yeast carboxypeptidase Y Proc Natl Acad Sci U SAA 1990;87(9):3368-72.

(10)) Gemmili TR, Trimble RB. Overview of N- and O-linked oligosaccharide structuress found in various yeast species. Biochim Biophys Acta 1999; 1426(2):227-37. .

(11)) Ballou CE. A study of the immunochemistry of three veast mannans. J Biol Chemm 1970; 245(5):1197-203.

(12)) Spencer JF, Gorin PA. Mannose-containing polysaccharides of yeasts. Biotechnoll Bioeng 1973; 15(1):1-12.

155 5

SubstitutionSubstitution of Pkhia-derived recombinant proteins with O-and N-linked mannose

(13)) Specks U. What you should know about PR3-ANCA. Conformational requirementss of proteinase 3 (PR3) for enzymatic activity and recognition byy PR3-ANCA. Arthntis Res 2000; 2(4):263-7.

(14)) Chua KY, Stewart GA, Thomas WR, Simpson RJ, Dilworth RJ, Plozza TM ett al. Sequence analysis of cDNA coding for a major house dust mite allergen,, Der p 1. Homology with cysteine proteases. J Exp Med 1988; 167(l):175-82. .

(15)) Jacquet A, Haumont M, Massaer M, Daminet V, Garcia L, Mazzu P et al. Biochemicall and immunological characterization of a recombinant precursorr form of the house dust mite allergen Der p 1 produced by Drosophilaa cells. Clin Exp Allergy 2000; 30(5) :677-84.

(16)) Chapman MD, Platts-Mills TA. Purification and characterization of the majorr allergen from Dermatophagoides pteronyssinus-antigen PI. J Immunoll 1980; 125(2):587-92.

(17)) van Oort E, de Heer PG, van Leeuwen WA, Derksen NI, Muller M, Huveneerss S et al. Maturation of Pichia pastoris-derived recombinant pro-Derr p 1 induced by deglycosylation and by the natural cysteine protease Der pp 1 from house dust mite. Eur J Biochem 2002; 269(2):671-9.

(18)) van Oort E, de Heer PG, Lerouge P, Faye L, Aalberse RC, van Ree R. Immunochemicall characterization of two Pichia pastoris-derived recombinantt group 5 Dactylis glomerata isoallergens. Int Arch Allergy Immunoll JID - 9211652 2001; 126(3):196-205.

(19)) Qin S, Tang H, Zhao LS, He F, Lin Y, Liu L et al. Cloning of HBsAg-encodedd genes in different vectors and their expression in eukaryotic cells. Worldd J Gastroenterol 2003; 9(5): 1111-3.

(20)) van Ree R, Clemens JG, Aalbers M, Stapel SO, Aalberse RC. Characterizationn with monoclonal and polyclonal antibodies of a new major allergenn from grass pollen in the group I molecular weight range. J Allergy Clinn Immunol 1989; 83(1):144-51.

(21)) van der Zee JS, van Swieten P, Jansen HM, Aalberse RC. Skin tests and histaminee release with PI-depleted Dermatophagoides pteronyssinus body extractss and purified PI. J Allergy Clin Immunol 1988; 81(5 Pt l):884-96.

(22)) Goldschmeding R, van der Schoot CE, ten Bokkel HD, Hack CE, van den Endee ME, Kallenberg CG et al. Wegener's granulomatosis autoantibodies identifyy a novel diisopropylfluorophosphate-binding protein in the lvsosomess of normal human neutrophils. J Clin Invest 1989; 84(5):1577-87.

156 6

ChapterChapter 6

(23)) Faye L, Chrispeels MJ. Characterization of N-linked oligosaccharides by affinoblottingg with concanavalin A-peroxidase and treatment of the blots withh glycosidases. Anal Biochem 1985; 149(l):218-24.

(24)) Aalberse RC, Koshte V, Clemens JG. Immunoglobulin E antibodies that crossreactt with vegetable foods, pollen, and Hymenoptera venom. J Allergy Clinn Immunol 1981; 68(5):356-64.

(25)) Schuurman J, Perdok GJ, Lourens TE, Parren PW, Chapman MD, Aalberse RC.. Production of a mouse/human chimeric IgE monoclonal antibody to thee house dust mite allergen Der p 2 and its use for the absolute quantificationn of allergen-specific IgE. J Allergy Clin Immunol 1997; 99(4):545-50. .

(26)) Duman JG, Miele RG, Liang H, Grella DK, Sim KL, Castellino FJ et al. O-Mannosylationn of Pichia pastoris cellular and recombinant proteins. Biotechnoll Appl Biochem 1998; 28 ( Pt l):39-45.

(27)) Boraston AB, Sandercock LE, Warren RA, Kilburn DG. O-Glycosylation of aa Recombinant Carbohydrate-Binding Module Mutant Secreted by PICHIA PASTORIS.. J Mol Microbiol Biotechnol 2003; 5(l):29-36.

(28)) Mochizuki S, Hamato N, Hirose M, Miyano K, Ohtani W, Kameyama S et al.. Expression and characterization of recombinant human antithrombin II I inn Pichia pastoris. Protein Expr Purif 2001; 23(1 ):55-65.

(29)) Wartmann T, Stephan UW, Bube I, Boer E, Melzer M, Manteuffel R et al. Post-translationaii modifications of the AFET3 gene product: a component off the iron transport system in budding cells and mycelia of the yeast Arxulaa adeninivorans. Yeast 2002; 19(10):849-62.

(30)) Van den SP, Rudd PM, Proost P, Martens E, Paemen L, Kuster B et al. Oligosaccharidess of recombinant mouse gelatinase B variants. Biochim Biophyss Acta 1998; 1425(3):587-98.

(31)) Nermes M, Savolainen J, Kortekangas-Savolainen O. Nitrocellulose-RAST analysiss of allergenic cross-reactivity of Candida albicans and Saccharomycess cerevisiae mannans. Int Arch Allergy Immunol 1995; 106(2):: 118-23.

(32)) Kortekangas-Savolainen O, Kalimo K, Lammintausta K, Savolainen J. IgE-bindingg components of baker's yeast (Saccharomyces cerevisiae) recognized byy immunob lotting analysis. Simultaneous IgE binding to mannan and 46-48 kDD allergens of Saccharomyces cerevisiae and Candida albicans. Clin Exp Allergyy 1993; 23(3): 179-84. '

157 7

158 8