synthetic and mechanistic studies with lipopeptide antibiotics

Synthetic and Mechanistic Studies with Lipopeptide Antibiotics

Laurens H. J. Kleijn

The cover image depicts Mount Ararat: “The new organism useful for the preparation of the A-21978C antibiotics [the blockbuster lipopepitide drug daptomycin] was isolated from a soil sample collected on Mount Ararat, Turkey.” (source: patent US4208403A)

Printed by GildeprintISBN: 978-94-6233-859-3

Printing of this thesis was supported by the Utrecht Institute of Pharmaceutical Sciences (UIPS), Utrecht, The Netherlands

Synthetic and mechanistic studies with lipopeptide antibiotics

Synthetische en mechanistische studies met antibiotische lipopeptiden

(met een samenvatting in het Nederlands)

Proefschrift

ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op dinsdag 9 januari 2018 des middags te 4.15 uur

door

Laurens Henri Johan Kleijn

geboren op 12 april 1989 te New York City, Verenigde Staten van Amerika

Promotor: Prof.dr R.J. Pieters

Copromotoren: Dr. N.I. Martin Dr. E.J. Breukink

And Everybody Knows that the Plague is ComingEverybody Knows that it’s Moving Fast

- Leonard Cohen / Sharon Robinson

Table of Contents

Chapter 1 9 The Cyclic Lipopeptide Antibiotics

Chapter 2 37Synthesis and Evaluation of Daptomycin Analogues

Chapter 3 55Synthesis and Mode of Action of Laspartomycin C

Chapter 4 77Target Recognition by the Calcium-Dependent Lipopeptide Antibiotic Laspartomycin C

Chapter 5 97Antibacterial Properties of a Lipid II Targeting Nisin-Derived Lipopeptide

Chapter 6 125SummarySamenvatting

Appendices 133List of Publications Curriculum VitaeAcknowledgements

9

Introduction

Chapter 1

The cyclic lipopeptide antibiotics

ABSTRACT: The cyclic lipopeptides comprise a number of clinically relevant classes of antibiotics that date back from the discovery of the polymyxins in 1947 to the recent introduction of the semisynthetic lipoglycopeptides. These natural products and derivatives thereof most often originate from soil-inhabiting and/or plant-derived organisms. The cyclic lipopeptides consist of peptide macrocycles that are acylated with a fatty acid lipid, and show great structural diversity owing to their nearly exclusive non-ribosomal synthesis production and/or post translational modification. This chapter presents a summary of the main classes of cyclic lipopeptide antibiotics with regard to their characteristic structural features, modes of action, clinical relevance, and the onset of bacterial resistance.

L. H. J. Kleijn, N. I. Martin, The Cyclic Lipopeptide Antibiotics 2017, In: Topics in Medicinal Chemistry. Springer, Berlin, Heidelberg. DOI: 10.1007/7355_2017_9

10

Introduction Chapter 1

1 Introduction

The cyclic lipopeptide antibiotics comprise a clinically relevant and structurally diverse group of compounds that operate via a number of distinct modes of action. While their chemical structures and properties may be considered unusual from the point of view of conventional drug development, three members of this compound class serve as cornerstones of the present medical system: the polymyxins, the calcium-dependent antibiotics, and the lipoglycopeptides. Each of these three is heavily relied upon in the treatment of serious, and often multi-drug resistant, systemic infections caused by both Gram-positive and Gram-negative pathogens. The cyclic lipopeptide antibiotics are generally poorly bioavailable via oral administration, restricting therapies to intravenous administration. The majority of targets exploited by these antibiotics are located on the outer surface of the bacterial cell, thus negating the need for them to cross bacterial membranes.

The cyclic lipopeptide antibiotics discussed in this chapter are natural products or semi-synthetic derivatives thereof isolated from both Gram-positive- and Gram-negative-producing organisms including Streptomyces, Bacillus and Actinoplanes. These lipopeptides are often tolerant to peptidases owing to their cyclic structures, and the presence of non-proteinogenic amino acids, including those with d-configurations. These unique features are made possible via the non-ribosomal synthesis pathways and post-translational modifications that contribute to the production of these structurally diverse compounds. The structural complexity of cyclic lipopeptide antibiotics means that their industrial production is rarely achieved by chemical synthesis. Rather, these compounds (which are often secondary metabolites) can be accessed via large-scale fermentation processes with producing strains selected to maximize production. Frequently, medicinal chemistry programs are implemented wherein the natural products are used as a starting material towards the development of semi-synthetic variants of which the lipoglycopeptides telavancin, oritavancin, and dalbavancin are examples that have made it to the clinic.

In general, lipopeptides, including those with antibacterial activity, are thought to be involved in several different microbial processes including bacterial swarming motility and the establishment of structured biofilms on solid surfaces.[1] These functions are proposed to be a consequence of the amphipathicity of such compounds, which can exhibit varying degrees of emulsifier and surfactant properties. Notable in this regard are the polymyxins which are known to have the capacity to lower surface and interfacial tension.[2] The amphipathic nature of cyclic lipopeptide antibiotics is

11


strongly influenced by the length of the lipid, and is often a key determinant of the therapeutic index relating antibiotic potency to toxicity. The lipid moiety of these compounds is generally positioned at the peptide N-terminus and can be saturated/unsaturated, linear/branched, and in some cases to bear additional functionalities such as hydroxyl groups that can participate in formation of the macrocycle.

In this chapter, the most prominent classes of cyclic lipopeptide antibiotics are summarized, focusing on their chemical structures, modes of action, clinical development, and the onset of bacterial resistance. Other recent reviews are available for further reading on these topics[3,4] as well as related compound classes and topics not covered here including: bacteriocins and antimicrobial peptides[5,6], lipopeptide biosynthesis[7], genome-mining[8,9] and other functions associated with lipopeptides.[10]

2 Colistin and the Polymyxins

2.1 Background The polymyxins were first identified in 1947 and represent the first clinically used class of lipopeptide antibiotics.[11] Originally isolated from fermentation of Paenibacillus polymyxa colistinus, the polymyxins are synthesized non-ribosomally resulting in the incorporation of non-proteinogenic amino acids like 2,4-diaminobutyric acid (DAB) and various d-amino acids. The class is characterized by a heptapeptide macrocyclic lactam core resulting from cyclization of the side chain of a DAB residue at position 4 with the C-terminus (Fig. 1). The exocyclic tripeptide is further modified by acylation of the N-terminus with one of a number of different branched, saturated lipids, most prominently 6-methyl-octanoic acid (Fig. 1). The preeminent members of the polymyxin class are the clinically used polymyxin B and polymyxin E (colistin) that vary only at position 6 (d-Phe vs d-Leu). The polymyxins are positively charged at physiological pH owing to the presence of five conserved DAB residues at positions 1, 3, 5, 8, and 9 with the other positions bearing a variety of other amino acids.[4]

2.2 Mode of actionThe highly cationic polymyxins elicit their bactericidal effect by binding the negatively charged lipid A portion of lipopolysaccharide (LPS) characteristic of the outer membrane of Gram-negative bacteria. The resulting displacement of LPS-bound divalent cations and subsequent outer membrane disruption leads to leakage of cellular content and cell death.[12] Additionally, the ability to bind and neutralize LPS give the polymyxins anti-endotoxic properties.[13,14]

12


2.3 Clinical useSince the 1950’s, the broad-spectrum activity of polymyxin B and colistin against Gram-negative aerobic pathogens was exploited in the development of a wide range of pharmaceutical preparations, with colistin use generally more widespread. Colistin itself (and as well pro-drug preparations) have found use as topical, ophthalmic, otic and injectable solutions for parenteral use as well as aerosol application for the treatment of lung infections (Fig. 2). Oral administration is also performed to achieve bowel decontamination as colistin is not absorbed by the gastrointestinal tract. Parenteral administration of colistin typically involves administration of the pro-drug “colistimethate sodium” (Fig. 2), a less potent and less toxic form of colistin generally used for intravenous and intramuscular applications as well as nebulization. Colistimethate sodium is commonly prepared by treating colistin with formaldehyde and sodium bisulfite. It undergoes hydrolysis in vivo [15]. After administration to rats, 61.1% ± 14.4% of the dose was recovered in urine during the first 24 hours with 32.6% ± 15.1% present as colistin.[16] After the 1970s, reports of nephrotoxicity and neurotoxicity resulted in a decline of parenteral colistimethate sodium treatment, limiting its use to the treatment of multidrug-resistant (MDR) infections in cystic fibrosis patients.[17] However, in the intervening years a significant body of data has amassed suggesting that the initially-reported toxicity, in particular with regard to nephrotoxicity, was overstated.[18] Taken together with the upswing

Figure 1. Structures of clinically-used polymyxin B and polymyxin E (colistin). All polymyxins contain a seven amino acid macrocycle, a number of conserved DAB residues and d-amino acids, as well as acylation of the N-terminus by a fatty acid tail.

13


of MDR Gram-negative infections in recent decades, the clinical application of intravenously administered colistimethate sodium has once again increased [19].

2.4 ResistanceIn recent years, an increase in polymyxin resistance in clinical isolates of Enterobacteriaceae and Acinetobacter baumannii has been reported [20,21]. Several mechanisms of resistance are known and generally involve modification of the molecular target lipid A with phosphoethanolamine or 4-amino-4-arabinose. These modifications introduce positive charges in lipid A, resulting in a reduced binding affinity for colistin and polymyxin B alike. The resistance mechanisms are mediated by chromosomal mutations that modulate two-component regulatory systems including pmrAB and phoPQ.[22,23] In 2016, the discovery of the MCR-1 resistance mechanism in Enterobacteriaceae demonstrated that modification of lipid A to introduce phosphoethanolamine can also be plasmid-mediated through the mcr-1 gene.[24] Worryingly, these developments reveal that the polymyxins may no longer be considered as reliable antibiotics of last resort for the treatment of infections due to MDR Gram-negative pathogens.

3 Daptomycin and the Calcium-Dependent Antibiotics

3.1 IntroductionThe family of calcium-dependent antibiotics (CDAs) consists of two main classes: 1) the lipodepsipeptides, epitomized by the clinically-used daptomycin and 2) the lipopeptides, which were the first members of the CDA family to be discovered dating to the 1953 discovery of amphomycin, a secondary metabolite of Streptomyces canus (Fig. 3).[25] Both the lipodepsipetides and the lipopeptides consist of a 10 amino acid macrocycle and an exocyclic region that

Figure 2. The colistin prodrug, colistimethate

14


is acylated at the N-terminus. Unlike most cationic antimicrobial peptides, the CDA bear an overall negative charge (–2 or –3) and contain a proposed calcium binding Asp-X-Asp-Gly motif. Interestingly, while the lipodepsipeptides and the lipopeptides both require the presence of free Ca2+ (1.25 mM) to effectively kill Gram-positive bacteria, they operate via distinct modes of action.

3.2 The CDA lipodepsipeptides

3.2.1 BackgroundDaptomycin was first isolated as minor fermentation product of the actinomycete Streptomyces roseosporus in 1987.[26] Daptomycin contains several d- and non-proteinogenic amino acids including ornithine (Orn), l-threo-3-methylglutamic acid, and l-kynurenine as well as a macrocyclic core due to cyclization of the C terminal kynurenine with the side chain of Thr4 to form an ester linkage (Fig. 3). The exocyclic N-terminal tripeptide is acylated with a 10-carbon aliphatic lipid tail, which is elegantly achieved on industrial scale by decanoic acid supplementation during the fermentation process. The four carboxylic acid side chains and the single amine side chain at the Orn position give daptomycin a net charge of –3 at physiological

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

O OHO

NH2OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

12

3

4

5

6 7 8

9

10

111213

D

D

DDaptomycin

O

HNO

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

NH2

O

OH

O

OH

O

OH

O

O

OHO

1

2

3

4 5 6

7

8

91011

Amphomycin A D

D

Figure 3. The clinically used CDA lipodepsipeptide daptomycin and non-clinical CDA lipopetide amphomycin.

15


pH. Daptomycin shares conserved structural motifs with its non-clinically used depsipeptide counterparts A54145 and CDA which also include the Asp-X-Asp-Gly motif at position 7–10 and the positioning of d- and achiral amino acids.[27]

3.2.2. Mode of actionDaptomycin, in complex with Ca2+, forms oligomers on the bacterial cytoplasmic membrane and interacts with the negatively charged phospholipid phosphatidyl glycerol (PG). Daptomycin’s action on the bacterial membrane results in membrane distortions and delocalization of essential membrane proteins involved in cell division accompanied by depolarization of the bacterial membrane.[28] In particular, daptomycin localization in fluid membrane domains displaces phospholipid synthase PlsX and the membrane protein MurG, which is responsible for biosynthesis of cell wall component lipid II.[29] Interestingly, daptomycin has also recently be found to act as a sensitizer to β-lactam antibiotics, most notably in the presence of β-lactamase inhibitors, leading to synergistic activity via an as of yet poorly understood mechanism.[30,31]

3.2.3 Clinical useDaptomycin’s spectrum of activity includes aerobic as well as anaerobic Gram-positive strains. It is clinically used to threat infections caused by a range of pathogens including staphylococci, enterococci, and streptococci. Daptomycin was first approved for clinical use in 2003 for the treatment of complicated skin and skin-structure infections followed in 2006 by FDA approval for treatment of S. aureus bacteremia and S. aureus right-sided endocarditis. Daptomycin is not suitable as a clinical treatment for infections due to S. pneumoniae given that daptomycin’s antibiotic action is inhibited in the presence of lung surfactant (presumably due to the high phosphatidylglycerol content of the lipids in the lung surfactant).[32] Driven primarily by its success in treating MRSA infections, daptomycin (marketed as Cubicin®) has achieved annual sales of more than 1 billion USD since 2013. [33] Administration of daptomycin is achieved via a once-daily 30 minute IV infusion. Its in vivo half-life is 8–9 hours. Excretion occurs primarily via the urine (78%) with half of the drug still intact. Daptomycin shows concentration-dependent antibiotic activity. The drug is generally well tolerated, which has prompted the exploration of higher dose regimens (8–12 mg kg-1) to increase the clinical success rate and to delay the onset of daptomycin resistance.[34,35]

3.2.4 ResistanceBacterial resistance to daptomycin is rare, but reports of emerging daptomycin non-susceptibility during therapy persist.[36] In nearly all of the case reports involving

16


daptomycin resistance, the patient had been treated with other antibiotics, in particular vancomycin, prior to daptomycin treatment. Daptomycin resistance mechanisms are both diverse and species-specific. A number of bacterial enzymes have been linked to daptomycin resistance including DltABCD (d-alanylation of wall teichoic acid), YycG (a histidine kinase), Cls2 (a cardiolipin synthase), and MprF (responsible for lysinylation of phosphatidylglycerol as well as its translocation to the outer cell membrane).[37,38] Modification of bacterial surface charge may result in decreased affinity of daptomycin for the bacterial membrane while enhanced bilayer rigidity, as a consequence of increased conversion of phosphatidylglycerol to cardiolipin, would prevent daptomycin’s entry to the inner membrane leaflet.[29] Interestingly, S. aureus is capable of releasing membrane phospholipids in response to daptomycin exposure to inhibit its antibiotic effect, a mechanism that can in turn be countered by co-administration of b-lactam antibiotics.[39]

3.3 The CDA lipopeptidesThe lipopeptide CDAs differ most significantly from the lipodepsipeptides in the composition of their peptide macrocycle. The lipopeptide CDA macrocycle is closed via an amide bond formed between the side chain of a l-2,3-diaminopropionic acid or l-2,3-diaminobutyric acid residue at position 2 and the peptide C-terminus (Fig 4.). In all known cases this amide linkage is located between the conformationally-restricted amino acids l-proline and d-pipecolic acid. In addition, the exocyclic region consists of only one amino acid that is N-terminally acylated with a branched, often unsaturated, fatty acid that is typically longer than that found in daptomycin and related lipodepsipeptides. The lipopeptide CDAs can be divided into the friulimicin/amphomycin and the laspartomycin sub-classes which differ primarily in their lipid motifs and the non-proteinogenic amino acids found at positions 2, 4, 9, and 10 (Fig. 4).[40,41]

The lipopeptide CDAs also show Ca2+-dependent antibiotic activity against Gram-positive bacteria, but operate via a mode of action that is distinct from that of daptomycin and the depsipeptides. Compelling biochemical evidence indicates that the lipopeptide CDAs inhibit peptidoglycan and wall teichoic acid biosynthesis by binding to and sequestering the bacterial cell wall precursor undecaprenyl phosphate (C55-P) (see chapter 3).[42,43] Furthermore, in the calcium-bound crystal structure of the amphomycin derivative tsushimycin, Ca2+ ions stabilize the active peptide conformation and in doing so are believed to facilitate binding to the negatively charged head group of C55-P (see chapter 4).[44] Interestingly (and in contrast to daptomycin) the antibiotic activity of the semi-synthetic amphomycin derivative MX-2401 is not affected by the presence of lung surfactant.[45] At present, no CDA lipopeptide or

17


derivative thereof has been approved for clinical use. Friulimicin B was evaluated in a 2007 phase I clinical trial but as of yet no further progress has been reported.[46]

4 Teicoplanin and the Lipoglycopeptides

4.1 BackgroundThe clinically used lipoglycopeptide antibiotics consist of both natural products and semi-synthetic compounds with strong antibacterial activity against Gram-positive pathogens. Their unique structural features are the result of post-translational modification of peptide precursors produced by non-ribosomal peptide synthases (NRPS) resulting in a characteristic glycosylated and acylated macrocyclic core. Teicoplanin (Fig. 5) was the first clinically-approved lipoglycopeptide, and since 1988 it has been marketed exclusively outside the USA. A fermentation product of Actinoplanes teichomyceticus, teicoplanin is biosynthesized from a heptapeptide precursor consisting completely of aromatic amino acids.[47]

In order to combat the rise of resistance to teicoplanin and the widely-used glycopeptide vancomycin, a new generation of lipoglycopeptides has been developed resulting in the clinical approval of the semi-synthetic glycopeptide derivatives

O

HNO

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

NH2

O

OH

O

OH

O

R3

O

O

R1

OHO

1

2

3

4 5 6

7

8

91011

Amphomycin A&

Friulimicin B

O

HN

O

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

OH

O

OH

O

OH

O

OH

O

O

1

2

3

4 5 6

7

8

91011

Laspartomycin C

R2

D

D

D

D

Figure 4. Representative members of the lipopeptide family of CDAs. Amphomycin and friulimicin (left) and laspartomycin (right). Amphomycin A: R1 = CH3, R2 = CH3, R3 = OH, friulimicin B: R1 = CH2(CH3)2, R2 = H, R3 = NH2.

18


telavancin, oritavancin, and dalbavancin (Fig. 5).[48,49] Telavancin is prepared from vancomycin through introduction of an N-decyl chain at the vancosamine moiety as well as addition of a phosphonomethylaminomethyl group at the C-terminal residue giving the peptide an overall charge of +2.[50] Oritavancin, also with +2 net charge, is obtained by incorporation of a 4-chlorobiphenyl moiety on chloroeremomycin, a fermentation product of Amycolatopsis orientalis that differs from vancomycin due to the presence of two 4-epi-vancosamine units.[51] Finally, dalbavancin with a +1 net charge results from inclusion of a 3,3-dimethylaminopropyl chain at the C-terminus of the natural glycopeptide A40926, a fermentation product of anactinomycete Nonomuraea species structurally similar to teicoplanin.[52]

OOOCl Cl

HN

ONH

O HN

O

OH

NH

O HN

O

NH2

OHOHHO

O

NH

HO

HN

O

HO

O

OOHO

HOOH

OHO

HN

NH

POH

OH

O

HN

Telavancin

OOOCl Cl

HN

ONH

O HN

O

OH

NH

O HN

O

NH2

OHOHHO

O

NH

O

HN

O

HO

O

OOHO

HOOH

O

HN

Cl

OHO

H2N

Oritavancin

OOO Cl

HN

ONH

O HN

ONH

O

NH2

OHOHO

O

NH

O

HN

O

HO

O

OHOHO

Teicoplanin

Cl

O

HOHO

HONH

O

NHAc

OH

O

OHOHOHOH

OO

HN

O

OHOHO

O

NH

HO

HN

O

HN

O

Cl

O

HOHO

HONH

O

O

OHOHOHOH

HO

HO

O

O

NH

O HN

ONH

O

NH

HO

HO

O

Cl

Dalbavancin

OH

12

34

56

7

123

45

6

7

12

34

56

7

12

34

56

7

O

D

DDD D

DDD

D

DDD

DDD

D

N

Figure 5. Clinically approved glycolipopeptide antibiotics teicoplanin, telavancin, oritavancin and dalbavancin.

19


4.2 Mode of actionLike the glycopeptides, the lipoglycopeptides interfere with bacterial cell wall biosynthesis pathway by inhibiting both the transglycosylation and the transpeptidation steps.[53] Specifically, these antibiotics block incorporation of lipid II as a substrate for the growing peptidoglycan by tightly binding to the d-Ala-d-Ala unit of lipid II through five well-defined hydrogen bonds.[54] Furthermore, the lipoglycopeptides display an enhanced lipid II binding interaction that is attributed to their higher membrane affinity, resulting from their lipid motifs, and/or the propensity to form homodimers. This increased affinity for lipid II allows the lipoglycopeptides, most notably oritavancin, to retain antibiotic activity, against vancomycin-resistant pathogens.[49] The antibiotic activity of the lipoglycopeptides towards vancomycin-resistant strains is further explained by secondary modes of action not relying on binding of d-Ala-d-Ala. Recent studies indicate that binding to the pentaglycine motif found in the lipid II of S. aureus as well disruption of bacterial membrane integrity, leading to membrane depolarization, are also features that contribute to the lipoglycopeptide antibiotic mechanism of action.[55,56]

4.3 Clinical useTeicoplanin’s clinical use includes the treatment of MRSA bacteraemia, where in Europe it is as commonly-used as is vancomycin. The two are equally efficacious, but with teicoplanin reportedly causing fewer adverse effects.[57] Among the new generation of lipoglycopeptide antibiotics, telavancin was the first to find its way to the clinic in 2009 and has since received clinical approval for several indications. Telavancin is approved for complicated skin and skin-structure infections (cSSSI) and hospital-acquired and ventilator-associated bacterial pneumonia (HABP/VABP). Both indications were expanded to include patients with concurrent S. aureus bacteremia in 2016. Oritavancin and dalbavancin were both approved for the treatment of acute bacterial skin and skin structure infections (ABSSSI) in 2014.

Both oritavancin and dalbavancin exhibit unusually long terminal half-lives of 245 hours and 346 hours respectively, which is attributed to high protein binding and intracellular accumulation allowing for a single dose treatment in the case of oritavancin (1200 mg) and a 2-dose regiment separated by 7 days (1000 mg and 500 mg) for dalbavancin.[58] Although telavancin also demonstrates high protein binding (93%), it has a higher degree of renal clearance and therefore requires a once daily dosing regiment (10 mg kg-1). While treatment with telavancin has been associated with nephrotoxicity, presumably due to the presence of solubility excipient hydroxypropylbetadex, adverse effects related to lipoglycopeptide therapy are generally mild and transient.[59]

20


4.4 ResistanceThe onset of bacterial resistance to lipopoglycopeptides used in the clinic has not yet been reported. However, in direct comparison with telavancin and dalbavancin, only oritavancin maintains strong antibiotic activity in vitro against VRE with phenotype VanA, which is responsible for the mutation of d-Ala-d-Ala to d-Ala-d-Lac in peptidoglycan precursors.[60] An increase in the prevalence of vancomycin-resistance with phenotype VanA among strains of S. aureus (VRSA), which also show reduced susceptibility to dalbavancin and telavancin, has the potential to limit lipoglycopeptide treatment options for multi-drug resistant S. aureus infections to oritavancin alone.[61]

5 Ramoplanin and the Lipoglycodepsipeptides

5.1 BackgroundAmong the lipoglycodepsipeptide group of antibiotics, the ramoplanin class comprising ramoplanin A1-A3 was first identified in 1984 from fermentation of Actinoplanes sp. ATCC 33076.[62] Prior to this, the structurally similar Enduracidins were isolated as fermentation products of Streptomyces fungicidicus B 5477 in 1968 (Fig. 6).[63] Both the ramoplanins and enduracins consist of a 16 amino acid macrocycle comprised of a number of non-proteinogenic amino acids, and share the same topology with identical backbone chirality. Both sets of peptide possess a net charge of +2 at physiological pH. The most significant structural difference between the two families is that the endurocidins are not glycosylated. In the ramoplanins the cycle is closed via an ester bond between the side chain of non-proteinogenic amino acid β-hydroxy-l-asparagine (2) and the C-terminal 3-chloro-4-hydroxyphenyl glycine (17), whereas the corresponding ester bond in the enduracidins is formed between l-threonine (2) and the 4-hydroxyphenyl glycine (17). The ramoplanins are further characterized by the unusual amino acids allo-Thr, d-allo-Thr, d-alanine, d-Orn, and both l- and d-4-hydroxyphenyl Gly, the latter of which is glycosylated at position 11 with the disaccharide mannosyl-(1,2)-α-d-mannose.[64] Both ramoplanin A2 and enduracidin A differ from their respective minor fermentation products in the length of the lipid. This lipid contains an E, Z-unsaturated motif in all cases and is typically branched. [65]

5.2 Mode of actionBoth ramoplanin and endurocidin are late-stage bacterial cell wall-synthesis inhibitors and operate by binding lipid II, most likely involving its pyrophosphate moiety, to prevent its incorporation in the peptidoglycan by bacterial transglycosylases.[66] Ramoplanin induces membrane depolarization at bactericidal

21


concentrations, equivalent to 3 times the minimum inhibitory concentration, suggesting that disturbance of bacterial membrane integrity is a secondary mode of action.[67] The similar spectra of antibiotic activity and the common antibacterial mechanism of ramoplanin and endurocidin (also shared with the ramoplanin aglycon) indicate that the dimannosyl moiety is not required for antibacterial activity.

5.3 Clinical developmentRamoplanin displays antibiotic activity against a range of Gram-positive pathogens with a spectrum of activity similar to vancomycin. Importantly, ramoplanin maintains activity against vancomycin-resistant enterococci (VRE) including teicoplanin-resistant strains. While ramoplanin is not suitable for parenteral administration due to toxicity issues relating to hemolysis, it can be safely administered orally as it is not absorbed systemically.[68] In this regard, ramoplanin has a significant record of clinical

HN

NH

HN

NH

HN

NH NH

O

NH2O

HNO

O

O

NH2

O

O

O

O

O

O

OH

H2NOH

OH

O

OH

OHNO

NHNH Cl

OH

O

NH

OHN

ONH

OHN

OH

ONH

OHO

O

H2N

HN

O

HO OHOHO

O

HOOH OH

OH

O

Ramoplanin A2

123

45

678

9

10

1112

1314

15 16

17

HN

NH

HN

NH

HN

NH NH

O

OHO

HNO

O

O

O

O

O

O

O

OH

H2NOH

OH

O

OH

OHNO

NHNH

OH

O

NH

OHN

ONH

OHN

OH

ONH

OHO

OH

HNO

Enduracidin A

123

45

678

9

10

1112

1314

15 16

17HN

HNHN

HN

OH2N

Cl ClHN

NH

NH

HO

HO

DDD

D

D D

DDD

D

D D

Figure 6. The main component of the ramoplanins and the enduracidins, ramoplanin A2 and endurocidin A.

22


evaluation culminating in FDA fast track designations and phase III clinical trials for both the treatment of both C. difficile-associated diarrhea (CDAD) and for prevention of nosocomical VRE infections by means of bowel decontamination.[69] However, clinical development was halted by its sponsor Oscient Pharmaceuticals, after which Nano Therapeutics Inc. obtained the development rights in 2009 with the intention to further evaluate oral ramoplanin for prevention of C. difficile infection relapse in a phase IIb clinical trial.[70] Of concern may be the propensity for growth promotion of indigenous and exogenous Gram-negative bacilli that is associated with ramoplanin therapy.[71] By comparison, enduracidin has been examined much less extensively despite positive early results showing safety at a twice daily dose of 100 mg in a trial with 20 patients aimed at treating MRSA infections conducted in 1973.[72] At present, enduracidin is reportedly used as a growth promoter in livestock, which may have implications for its future use in humans.[73]

5.4 ResistanceTo date, no indications of emergent ramoplanin resistance have been reported. Furthermore, the mechanism by which the producing strain of Actinoplanes protects itself from ramoplanin has yet to be elucidated.[74]

6 The Acyldepsipeptides

6.1 BackgroundResearchers at Eli Lily were the first to describe the family of acyldepsipeptides (ADEPs) as fermentation products of Streptomyces hawaiiensis NRRL 15010 displaying antimicrobial activity against Gram-positive bacteria.[75] In 2005 the chemical structure of the main component (designated ADEP 1) was reassigned and a novel antibiotic mode of action was established.[76] ADEP 1 is a hexapeptide consisting of a pentapeptide macrocycle that is closed via an ester bond between the side chain of l-Ser2 and the C-terminal 4-methyl-l-Pro6 (Fig. 7). ADEP 1 is further characterized by the presence of N-methyl-l-Ala4 and the unsaturated (E,E,E) C8 lipid present at the N-terminus of the exocyclic l-Phe1. The structurally similar enopeptin A was identified as a fermentation product of Streptomyces sp. RK-1051 in 1991, and consists of the same ADEP pentapeptide macrocycle but with a longer and amidated lipid.[77] Naturally occurring variants of both ADEP 1 and enopeptin A, namely A54556B and enopeptin B, have also been reported and contain l-Pro rather that l-4-methyl-Pro at position 6.[78]

23


6.2 Mode of actionOwing to their low number of backbone hydrogens, neutral charge, and low molecular weight compared to other cyclic lipopeptide antibiotics, the ADEPs are capable of acting on an intracellular bacterial target in Gram-positive bacteria. Activity against Gram-negative bacteria is limited to strains with defective efflux pumps, or the co-presence of a permealizing agent.[76] The ADEPs bind and activate caseinolytic protease (ClpP) and the resulting uncontrolled proteolysis in turn leads to inhibition of bacterial cell-wall synthesis and ultimately cell death.[76] Remarkably, ADEP binding to ClpP induces the formation of functional ClpP tetradecamers that are capable of independent proteolytic activity without interacting with the Clp-ATPases that it normally relies on. The specific ADEP binding site was determined based on the co-crystal structure obtained using a ClpP from B. subtilis ADEP binding stabilizes the multimeric ClpP complex and blocks access to the Ile-Gly-Phe loops that serve as binding sites for the ATP-ases. Additionally, ADEP binding induces opening of a gated pore allowing access of larger substrates to the proteolytic chamber.[79]

6.3 Preclinical developmentUpon preclinical evaluation ADEP 1 did not prove effective in mouse models of lethal bacterial infection, showing a high rate of clearance in addition to chemical instability and solubility problems.[78] Medicinal chemistry efforts to overcome these issues culminated in the generation of ADEP 4 and ADEP B315 that display significantly higher antibacterial potencies. The enhanced activity seen for these analogues is attributed to an increased conformational rigidity achieved by incorporation of a l-pipecolic acid residue at position 4 as well as the fluorine substituents on l-Phe1 (Fig. 8).[76,80] Both synthetic ADEPs demonstrate efficacy in various in vivo models including S. aureus, E. faecalis, and S. pneumoniae infections.[76,81] However, bacterial strains develop ADEP resistance in vitro with relative ease, with resistance frequencies comparable to those of rifampicin.[76] Interestingly, ADEP 4 administered in combination with rifampicin is capable of eradicating established biofilms in an S. aureus mouse infection model, demonstrating the potential for advancement of ADEPs in combination therapy.[82]

NHO

NO

O

NH

O

HN

O

O

N

ON

O

ADEP 1

12

34

5

6

NHO

NO

O

NH

O

HN

O

O

N

ON

O

NH

OHO

O

Enopeptin A

12

34

5

6

Figure 7. The natural products ADEP 1 and enopeptin A.

24


6.4 ResistanceAs a consequence of their distinct mode of action, no cross-resistance of clinical isolates against ADEPs and clinically used antibiotics has been reported to date. However, the bacterial ClpP is not essential for survival, which enables the onset of resistance against ADEPs through point mutations in ClpP that impair its proteolytic activity.[76] An alternative mechanism of resistance is the activation of the ABC transporter SclAB. There are indications for a target-substitution mechanism that is achieved by the expression of ClpP3, which is insensitive to ADEPs.[83]

7 Bacitracin

7.1 BackgroundBacitracin was discovered in 1945 and was creatively named after hospital patient “Traci I.” from which a producing strain of B. subtilis was isolated.[84] Although not consisting of a fatty acid-derived lipid, bacitracin fits within the realm of the lipopeptides in terms of its overall structure and antibacterial mode of action. The bacitracins consist of a seven-amino acid macrocycle closed via an amide bond between l-Lys6 and C-terminal d-Asn12 (Fig. 9). Bacitracin A, both the main and the most active component of the bacitracins, contains a thiazoline moiety formed via condensation of the side chain of Cys2 sidechain with the carbonyl group originating from the N-terminal l-Ile1. The peptide is further characterized by both l- and d-amino acids and an overall-neutral charge at physiological pH. Minor components of the bacitracin complex contain alternative aliphatic amino acids at positions 1, 5 and 8.[85]

7.2 Mode of actionBacitracin halts the bacterial cell wall synthesis as well as wall teichoic acid- and capsule synthesis in Gram-positive strains via a distinct mode of action involving the sequestration of undecaprenyl pyrophosphate (C55-PP). To elicit its antibacterial activity, bacitracin requires the presence of divalent cations, most notably Zn2+, to form a ternary complex that engulfs the pyrophosphate moiety

NHO

NO

O

NH

O

HN

O

O

N

ON

O NHO

NO

O

NH

O

HN

O

O

N

ON

O

F F

ADEP 4

F F

ADEP B315

12

34

5

6

12

34

5

6

Figure 8. Synthetic ADEP derivatives ADEP 4 and ADEP B315.

25


1

of C55-PP.[86] A crystal structure of bacitracin in complex with a truncated C55-PP variant (C10-P) shows the pyrophosphate moiety interacting with backbone residues as well as Na+ and Zn2+ ions, with the thiazoline specifically coordinating Zn2+.[87]

7.3 Clinical useIn addition to being active against cocci and bacilli strains as well as C. difficile, bacitracin inhibits the growth of the mold Neurospora crassa. Bacitracin is nephrotoxic, limiting its clinical use to topical treatments. However, it can be administered orally due to its limited uptake from the gastrointestinal tract. For topical application bacitracin is available as several ointments such as Polysporin®, a triple combination comprised of the zinc complex of bacitracin, polymyxin B sulfate, and neomycin sulfate. Although adverse events are rare, allergic reactions and anaphylactic shock can occur when bacitracin is applied to compromised skin barriers. [88]

7.4 ResistanceSeveral mechanisms that confer bacterial resistance to bacitracin are known. These mechanisms include activation of the plasmid-transferable bcrABC genes that encode an ABC-transporter. Alternatively, the bacterial pool of the bacitracin target C55-PP can be altered by up-regulation of the bacA gene resulting in the increased conversion of C55-PP to C55-P, or by overproduction of an undecaprenol (C55-OH) kinase resulting in the increased conversion of C55-OH to C55-P.[89] More recently, a bacitracin deactivating mechanism was described involving the aminohydrolase BahA that hydrolyses l-Asn12 to l-Asp12 to generate desamino bacitracin, which is devoid of antibiotic activity.[85,90

8 Lotilibcin

In the late 1990’s researchers at Wakamoto Pharmaceuticals described the isolation of a new class of lipodepsipeptides with potent anti-MRSA activity from a strain of

NH2

S

NO

NH

HN

NH

HN

NH

HN

O

O

O

O

O

OHOHN

NH2

NHOO

NHHN

O

N

HN

OHO

HN

O

O

O

H2N

12

34

56

7 8

9

10

1112Bacitracin A

DD

D

D

Figure 9. The main component of the bacitracins, bacitracin A.

26


the Lysobacter genus (strain number WAP-8294). Among the group of structurally similar “WAP-8296A” lipopeptides produced, the main component was identified and designated as WAP-8294A2, later named lotilibcin. Lotilibcin bears a net +1 charge at physiological pH and consists of a 12 amino acid macrocycle that is closed via an ester bond between the C-terminal N-methyl-l-Val12 and the hydroxyl moiety of the N-terminal (R)-3-hydroxy-7-methyloctanoic acid lipid (Fig. 10). The peptide cycle is further characterized by a number of unusual amino acids including threo-β-hydroxy-d-Asn2 and N-methyl-d-Phe5.

[91] Other members of the WAP-8294 family differ in the length of their lipids and degree of backbone N-methylation.[92] Lotilibcin is rapidly bactericidal against Gram-positive bacteria, and its in vitro antibiotic activity is enhanced in the presence of human serum. Although the mode of action of lotilibcin is not established, a key finding is that its antibiotic action is antagonized by addition of phosphatidylglycerol and cardiolipin but not by the vancomycin antagonist Ac-Lys(Ac)-d-Ala-d-Ala-OH, suggesting a mode of action involving the bacterial membrane.[92] Lotilibcin underwent a phase I clinical trial in 2011 sponsored by aRigen pharmaceuticals and subsequently by the Green Cross Cooperation, with the aim of development as a treatment for MRSA infection. However the current development status of lotilibcin is unclear.[70]

9 The Empedopeptins

Empedopeptin is an eight amino acid cyclic lipopeptide first isolated from the Gram-negative Empedobacter haloabium strain ATCC 31962 in 1984.[93] Bearing a net –1 charge, empedopeptin exhibits high aqueous solubility due to the numerous polar residues including the unusual amino acids threo-β-hydroxy-d-Asp5, trans-3-hydroxy-l-Pro7, and threo-β-hydroxy-l-Asp8 (Fig. 11). An ester bond closes the macrocycle between the C-terminal residue and the γ-hydroxylated C14 lipid functionality, for which the stereochemical orientation has yet to be determined. The amino acid stretch from Arg4 to threo-β-hydroxy-l-Asp8 is conserved in the more recently identified and structurally-related tripropeptins and plusbacins, produced by Lysobacter and Pseudomonas species respectively, with the exception of the l-/d- configuration of Ser6.

[94,95] Most prominent among the plusbacins and tripropeptins are plusbacin A3 and tripeptin C, both of which differ from empedopeptin at residues 1–3.

Empedopeptin and tripropeptin C display in vitro and in vivo activity against both aerobic and anaerobic Gram-positive pathogens.[93,94] Extensive mode of action studies performed with empedopeptin reveal a mechanism that targets the bacterial cell-wall synthesis pathway, likely similar to that of the structurally-related

27


plusbacins and tripeptins.[96] Empedopeptins form a complex with late stage cell wall precursors that contain a pyrophosphate unit, with the primary target being lipid II. Besides the pyrophosphate moiety, binding to lipid II involves interactions with the MurNAc moiety as well as parts of the pentapeptide and undecaprenyl chain. The antibiotic potency of the empedopeptins is enhanced 2–16 fold in the presence of Ca2+. Complex formation with bacterial targets occurs in a Ca2+-dependent manner that is not observed with other metal ions. The molecular interactions underlying the role of Ca2+ are not yet known, but it is likely that the positively-charged metal ion stabilizes the complex formed between the negatively-charged empedopeptin and the negatively-charged pyrophosphate moiety of lipid II. At present no members of this class of lipodepsipeptides or synthetic derivatives thereof have reached the stage of clinical development.

10 The Tridecaptins

The tridecaptins are a family of linear lipopeptides first discovered in 1978 as fermentation products of certain Bacillus polymyxa strains. These lipopeptides exhibit selective antibiotic activity against Gram-negative bacteria. Tridecaptin A1 is the best studied member of the tridecaptin family and consists of 13 l- and d-amino acids including both the l- and d- configuration of 2,4-diaminobutyric acid at positions 8 and 9 respectively, as well as d-allo-Ile12, and a N-terminus that is acylated with (3R,6S)-3-hydroxy-6-methyloctanoic acid (Fig. 12). Other members of the tridecaptin family maintain the same backbone stereochemistry while containing different aliphatic amino acids and variation of the N-terminal lipid.[4] The tridecaptins show in vitro and in vivo activity against many multi-drug resistant Gram-negative pathogen with exception of P. aeruginosa.[95,97]

HNHN

NH

HN

H2N

NHO

O

ONHO

O OH

NH2

O

H2N

N

O

OOO

NH

HO

O

HN

OHNH2

O

O

NH

HO

O

HNO

N

NHOO

12

3

4

5

6

78

9 10

11

12

Lotilibcin

D

DD

D

D

D

Figure 10. WAP-8294A2/Lotilibcin, the main component of the WAP-8294A compounds.

28


The antibacterial mode of action of the tridecaptins involves initial binding to lipid II at the inner membrane followed by disruption of the proton motive force leading to bacterial cell death. The form of lipid II used by Gram-negative bacteria typically differs from that of Gram-positive bacteria at amino acid 3 of the pentapeptide stem, wherein in Gram-negative bacteria a meso-diaminopimelic acid residue is found in place of lysine, giving the Gram-negative lipid II an additional negative charge at physiological pH. Recent findings have revealed that the tridecaptins selectively bind to Gram-negative lipid II. The complex formed between tridecaptin and lipid II involves the proximal isoprenyl units as well as the pentapeptide region, including a specific interaction between the γ-amino group of the l-2,4-diaminobutyric acid at position 8 of tridecaptin and the carboxylate of the meso-diaminopimelic acid of the Gram-negative lipid II. Unlike many of the other known lipid II-targeting lipopeptide antibiotics, interaction with the pyrophosphate moiety does not play a role in tridecaptin’s binding.[98]

NNH

OH

HN O

O

NO

O O

NH

HN

O

OOH

O

N

HN

NH

H2N

ONHOH

OO

HOHO

O123

4

5

67

8

Empedopeptin OH NNH

HN O

O

O O

NH

HN

O

OOH

O

N

HN

NH

H2N

ONHOH

OO

HOHO

O2

4

5

67

8

OH

HO

NHOH

O13

Plusbacin A3N

N

HN O

NH

HN

O

OOH

O

N

HN

NH

H2N

ONHOH

OO

HOHO

O2

4

5

67

8

OH

3

Tripropeptin C

O OHHN

O OO

1

OH

OH

OH

DD

D

D

D

D

D D

D

D

D

Figure 11. Lipodepsipeptide empedopeptin, plusbacin A3 and tripeptin C.

29


NH

HN N

H

HN

OH O

O

NH2

O

ONH

ONH

O

HNOH

NHO

NHNH2

O

NH2

NH

OHN

O

OHO

O

NHO

HN

O

NHO

HO

OH

12

34

5

6

7

89

1011

1213

Tripdecaptin A1

DD D

D

DD

Figure 12. The preeminent tridecapeptin, tridecapeptin A1.

11 Semisynthetic Lipopeptides derived from Lantibiotics and Teixobactin

In addition to the lipoglycopeptides oritavancin and telavancin, which were conceived from glycopeptide precursors (vide supra), other classes of semisynthetic lipopeptides have also been generated from natural product antibiotics. The lantibiotic deoxyactagardine B, a fermentation product of Actinoplanes liguriae NCIMB41362, was semi-synthetically modified by addition of 1,7-diaminoheptane at the C-terminal end to generate NVB302 and achieving a 4-fold increase of in vitro antibiotic potency against C. difficile compared with the parent compound (Fig. 13).[99] Mode of action studies with actagardine, a close homologue of deoxyatagardine B, point towards lipid II as the bacterial target with likely involvement of the pyrophosphate unit.[100] NVB302 demonstrated chemical stability in the GI tract in rats after oral administration and efficacy comparable to vancomycin in a C. difficile hamster survival model, culminating in advancement to the clinical development stage in 2011.[99] Despite the successful phase I study, the first such study to be conducted with a lantibiotic, the current status of NVB302 is unclear.[70]

Semi-synthetic analogues of the preeminent lantibiotic nisin also were reported recently wherein the proteolytically fragile C-terminal region of the peptide was replaced by a lipid. Nisin is a 34-amino acid peptide produced by Lactococcus lactis. It consists of 5 lanthionine rings (rings A–E) and exerts its bactericidal action against Gram-positive bacteria by binding to the pyrophosphate region of lipid II via the N-terminal nisin A/B motif comprised of residues 1–12.[101] Following lipid II binding, the C-terminal region of the nisin peptide inserts into the bacterial membrane leading to pore formation and rapid cell death.[101] Despite its potent

30


antibacterial activity, nisin is rapidly degraded in vivo by proteolysis, limiting its therapeutic application. Interesting, treatment of nisin with trypsin yields a fragment consisting of the N-terminal nisin A/B rings that, while devoid of antibiotic activity, is still capable of interacting with lipid II. The Martin lab recently found that modification of the nisin A/B fragment by the introduction of a variety of lipids at the C-terminal Lys12 restores in vitro antibiotic activity to levels on par with that of nisin. The mode of action of these nisin A/B lipopeptides is lipid II-mediated, but not identical to the action mechanism of parent compound nisin (see chapter 5).[102]

In addition to the lantibiotic-derived semisynthetic lipopeptides discussed above, a total synthesis strategy has also been applied in generating lipopeptide analogues of the recently discovered depsipeptide teixobactin.[103,104] Teixobactin also binds the lipid II pyrophosphate and demonstrates both potent in vitro activity and in vivo efficacy against a range of Gram-positive pathogens.[103] The teixobactin-inspired lipopeptide lipobactin 1 contains a linear C12 lipid in place of the N-terminal amino acid region comprising N-Me-d-Phe-Ile-Ser-d-Gln-d-allo-Ile and also includes an l-Arg at position 5 in place of the synthetically challenging l-allo-enduracididine found in the parent compound.[105] Lipobactin 1 was found to possess good in vitro antibiotic activity. However, neither mode of action nor in vivo studies have been reported.

11 Conclusions The onset of multi-drug resistant bacteria stresses the need for the development of antibiotics that operate via modes of action that are currently not employed in the clinical setting. Several classes of cyclic lipopeptide antibiotics have made their mark as effective clinical antibiotics used in the treatment of infections due to both Gram-positive and Gram-negative pathogens. In recently years the number of newly discovered classes of lipopeptide antibiotics with unique pharmacophores and modes of action has declined and those that are found are more often than not unsuitable for direct advancement to clinical development. However, semi-synthetic approached using natural products as starting point is a valuable strategy to extend the scope of cyclic lipopeptide antibiotics as exemplified by the clinical approval of the lipoglycopeptides telavancin, oritavancin and dalbavancin. In addition, advances in chemical peptide synthesis provide new tools for the optimization of pharmacophores as well as an alternative to fermentation for industrial production, further improving the future outlook for this important and diverse family of antibiotics.

31


References[1] A. F. Dexter, A. P. J. Middelberg, Ind. Eng. Chem. Res. 2008, 47, 6391–6398.[2] E. Rosenberg, E. Z. Ron, Appl Microbiol Biotechnol 1999, 52, 154–162.[3] T. Schneider, A. Müller, H. Miess, H. Gross, Int J Med Microbiol 2014, 304, 37–43.[4] S. A. Cochrane, J. C. Vederas, Med Res Rev 2016, 36, 4–31.[5] P. D. Cotter, R. P. Ross, C. Hill, Nature Rev Microbiol 2013, 11, 95–105.[6] J. L. Fox, Nature Biotechnol 2013, 31, 379–382.[7] G. Liu, K. F. Chater, G. Chandra, G. Niu, H. Tan, Microbiol. Mol. Biol. Rev. 2013, 77, 112– 143.[8] B. O. Bachmann, S. G. Van Lanen, R. H. Baltz, J Ind Microbiol Biotechnol 2014, 41, 175– 184.[9] X. Zhao, O. P. Kuipers, BMC Genomics 2016 17:1 2016, 17, 882.[10] J. M. Raaijmakers, I. de Bruijn, O. Nybroe, M. Ongena, FEMS Microbiol Rev 2010, 34, 1037–1062.[11] P. G. Stansly, R. G. Shepherd, H. J. White, Johns Hopkins Med J 1947, 81, 43–54.[12] M. J. Trimble, P. Mlynárčik, M. Kolář, R. E. W. Hancock, Cold Spring Harb Perspect Med 2016, 6, a025288.[13] B. A. Newton, Bacteriol Rev 1956, 20, 14–27.[14] T. Velkov, P. E. Thompson, R. L. Nation, J. Li, J Med Chem 2009, 53, 1898–1916.[15] E. G. Beveridge, A. J. Martin, Br J Pharmacol 1967, 29, 125–135.[16] J. Li, R. W. Milne, R. L. Nation, J. D. Turnidge, T. C. Smeaton, K. Coulthard, J. Antimicrob. Chemother. 2004, 53, 837–840.[17] S. P. Conway, M. N. Pond, A. Watson, C. Etherington, H. L. Robey, M. H. Goldman, Thorax 1997, 52, 987–993.[18] M. E. Falagas, M. Rizos, L. A. Bliziotis, K. Rellos, S. K. Kasiakou, A. Michalopoulos, BMC Infect Dis 2005, 5, 1.[19] M. E. Falagas, S. K. Kasiakou, Clin Infect Dis. 2005, 40, 1333–1341.[20] M. Monaco, T. Giani, M. Raffone, F. Arena, A. Garcia-Fernandez, S. Pollini, C. Network EuSCAPE-Italy, H. Grundmann, A. Pantosti, G. Rossolini, Eurosurveillance 2014, 19, 20939- 18.[21] Y. Cai, D. Chai, R. Wang, B. Liang, N. Bai, J. Antimicrob. Chemother. 2012, 67, 1607–1615.[22] E. A. Groisman, J. Kayser, F. C. Soncini, J. Bacteriol. 1997, 179, 7040–7045.[23] J. S. Gunn, K. B. Lim, J. Krueger, K. Kim, L. Guo, M. Hackett, S. I. Miller, Mol Microbiol 1998, 27, 1171–1182.[24] Y.-Y. Liu, Y. Wang, T. R. Walsh, L.-X. Yi, R. Zhang, J. Spencer, Y. Doi, G. Tian, B. Dong, X. Huang, et al., Lancet Infect Dis 2016, 16, 161–168.[25] B. Heinemann, M. Kaplan, R. Muir, I. Hooper, Antibiot Chemother (Northfield) 1953, 3, 1239–1242.[26] M. Debono, M. Barnhart, C. B. Carrell, J. A. Hoffmann, J. L. Occolowitz, B. J. Abbott, D. S. Fukuda, R. L. Hamill, K. Biemann, W. C. Herlihy, J. Antibiot. 1987, 40, 761–777.[27] L. Robbel, M. A. Marahiel, J. Biol. Chem. 2010, 285, 27501–27508.[28] J. Pogliano, N. Pogliano, J. A. Silverman, J. Bacteriol. 2012, 194, 4494–4504.[29] A. Müller, M. Wenzel, H. Strahl, F. Grein, T. N. V. Saaki, B. Kohl, T. Siersma, J. E. Bandow, H.-G. Sahl, T. Schneider, et al., Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 7077–7086.[30] A. Renzoni, W. L. Kelley, R. R. Rosato, M. P. Martinez, M. Roch, M. Fatouraei, D. P. Haeusser, W. Margolin, S. Fenn, R. D. Turner, et al., Antimicrob. Agents Chemother. 2017, 61, e01634–16.[31] K. E. R. Henson, J. Yim, J. R. Smith, G. Sakoulas, M. J. Rybak, Antimicrob. Agents Chemother. 2017, 61, e01564–16.[32] S. D. Taylor, M. Palmer, Bioorg Med Chem 2016, 24, 6253–6268.[33] M. May, Nature 2014, 509, S4–S5.[34] J. M. Munita, B. E. Murray, C. A. Arias, Int J Antimicrob Agents 2014, 44, 387–395.[35] C. Kosmidis, D. P. Levine, Expert Opin Pharmacother 2010, 11, 615–625.[36] S. Stefani, F. Campanile, M. Santagati, M. L. Mezzatesta, V. Cafiso, G. Pacini, Int J Antimicrob Agents 2015, 46, 278–289.

32


[37] A. S. Bayer, T. Schneider, H.-G. Sahl, Ann N Y Acad Sci 2013, 1277, 139–158.[38] W. R. Miller, A. S. Bayer, C. A. Arias, Cold Spring Harb Perspect Med 2016, 6, a026997.[39] V. Pader, S. Hakim, K. L. Painter, S. Wigneshweraraj, T. B. Clarke, A. M. Edwards, Nat Microbiol 2016, 2, DOI 10.1038/nmicrobiol.2016.194.[40] D. B. Borders, R. A. Leese, H. Jarolmen, N. D. Francis, A. A. Fantini, T. Falla, J. C. Fiddes, A. Aumelas, J. Nat. Prod. 2007, 70, 443–446.[41] C. Müller, S. Nolden, P. Gebhardt, E. Heinzelmann, C. Lange, O. Puk, K. Welzel, W. Wohlleben, D. Schwartz, Antimicrob. Agents Chemother. 2007, 51, 1028–1037.[42] T. Schneider, K. Gries, M. Josten, I. Wiedemann, S. Pelzer, H. Labischinski, H. G. Sahl, Antimicrob. Agents Chemother. 2009, 53, 1610–1618.[43] L. H. J. Kleijn, S. F. Oppedijk, P. ‘t Hart, R. M. van Harten, L. A. Martin-Visscher, J. Kemmink, E. Breukink, N. I. Martin, J Med Chem 2016, 59, 3569–3574.[44] G. Bunkoczi, L. Vertesy, G. M. Sheldrick, Acta Cryst D 2005, 61, 1160–1164.[45] E. Rubinchik, T. Schneider, M. Elliott, W. R. P. Scott, J. Pan, C. Anklin, H. Yang, D. Dugourd, A. Müller, K. Gries, et al., Antimicrob. Agents Chemother. 2011, 55, 2743–2754.[46] N. S. Ryder, Expert Opin Investig Drugs 2009, 19, 1–21.[47] D. Kahne, C. Leimkuhler, W. Lu, C. Walsh, Chem. Rev. 2005, 105, 425–448.[48] R. Leclercq, E. Derlot, J. Duval, P. Courvalin, N Engl J Med 2010, 319, 157–161.[49] F. Van Bambeke, Drugs 2015, 75, 2073–2095.[50] M. R. Leadbetter, S. M. Adams, B. Bazzini, P. R. Fatheree, D. E. Karr, K. M. Krause, B. M. T. Lam, M. S. Linsell, M. B. Nodwell, J. L. Pace, et al., J. Antibiot. 2004, 57, 326–336.[51] R. D. G. Cooper, N. J. Snyder, M. J. Zweifel, M. A. Staszak, S. C. Wilkie, T. I. Nicas, D. L. Muller, T. F. Butler, M. J. Rodriguez, B. E. Huff, et al., J. Antibiot. 1996, 49, 575–581.[52] A. Malabarba, R. Ciabatti, R. Scotti, B. P. Goldstein, P. Ferrari, M. Kurz, B. P. Andreini, M. Denaro, J. Antibiot. 1995, 48, 869–883.[53] D. Zeng, D. Debabov, T. L. Hartsell, R. J. Cano, S. Adams, J. A. Schuyler, R. McMillan, J. L. Pace, Cold Spring Harb Perspect Med 2016, 6, a026989.[54] J. Barna, D. H. Williams, Annu Rev Microbiol 1984, 38, 339–357.[55] S. J. Kim, L. Cegelski, D. Stueber, M. Singh, E. Dietrich, K. S. E. Tanaka, T. R. Parr Jr, A. R. Far, J. Schaefer, J Mol Biol 2008, 377, 281–293.[56] D. L. Higgins, R. Chang, D. V. Debabov, J. Leung, T. Wu, K. M. Krause, E. Sandvik, J. M. Hubbard, K. Kaniga, D. E. Schmidt Jr, et al., Antimicrob. Agents Chemother. 2005, 49, 1127–1134.[57] S. Svetitsky, L. Leibovici, M. Paul, Antimicrob. Agents Chemother. 2009, 53, 4069– 4079.[58] E. Wenzler, S. Liao, K. A. Rodvold, in Antibiotic Pharmacodynamics (Eds.: J.C. Rotschafer, D.R. Andes, K.A. Rodvold), Springer New York, New York, NY, 2016, pp. 285– 315.[59] R. Masterton, G. Cornaglia, P. Courvalin, H. M. Lode, J. Rello, A. Torres, Int J Antimicrob Agents 2015, 45, 213–220.[60] D. Sweeney, A. Stoneburner, D. L. Shinabarger, F. F. Arhin, A. Belley, G. Moeck, C. M. Pillar, J. Antimicrob. Chemother. 2016, DOI 10.1093/jac/dkw451.[61] J. A. Karlowsky, K. Nichol, G. G. Zhanel, Clin Infect Dis. 2015, 61, S58–S68.[62] S. Goto, S. Kuwahara, H. Zenyoji, N. Okubo, J. Antibiot. 1968, 21, 119–125.[63] B. Cavalleri, H. Pagani, G. Volpe, E. Selva, F. Parenti, J. Antibiot. 1984, 37, 309–317.[64] D. Shin, Y. Rew, D. L. Boger, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 11977–11979.[65] F. Castiglione, A. Marazzi, M. Meli, G. Colombo, Magn Reson Chem 2005, 43, 603–610.[66] X. Fang, K. Tiyanont, Y. Zhang, J. Wanner, D. Boger, S. Walker, Mol Biosyst 2006, 2, 69–76.[67] M. Cheng, J. X. Huang, S. Ramu, M. S. Butler, M. A. Cooper, Antimicrob. Agents Chemother. 2014, 58, 6819–6827.[68] D. K. Farver, D. D. Hedge, S. C. Lee, Ann Pharmacother 2005, 39, 863–868.[69] F. Van Bambeke, Curr Opin Investig Drugs 2006, 7, 740–749.[70] M. S. Butler, M. A. Blaskovich, M. A. Cooper, J. Antibiot. 2016, DOI 10.1038/ja.2016.72.[71] U. Stiefel, N. J. Pultz, M. S. Helfand, C. J. Donskey, Antimicrob. Agents Chemother. 2004, 48, 2144–2148.[72] M. Peromet, E. Schoutens, E. Yourassowsky, Chemotherapy 1973, 19, 53–61.

33


[73] D. G. McCafferty, P. Cudic, B. A. Frankel, S. Barkallah, R. G. Kruger, W. Li, Biopolymers 2002, 66, 261–284.[74] S. Walker, L. Chen, Y. Hu, Y. Rew, D. Shin, D. L. Boger, Chem. Rev. 2005, 105, 449–476.[75] K. H. Michel, R. E. Kastner, A54556 Antibiotics and Process for Production Thereof, 1985, US 4492650 A.[76] H. Brötz-Oesterhelt, D. Beyer, H.-P. Kroll, R. Endermann, C. Ladel, W. Schroeder, B. Hinzen, S. Raddatz, H. Paulsen, K. Henninger, et al., Nat Med 2005, 11, 1082–1087.[77] H. Koshino, H. Osada, T. Yano, J. Uzawa, K. Isono, Tetrahedron Lett 1991, 32, 7707–7710.[78] B. Hinzen, S. Raddatz, H. Paulsen, T. Lampe, A. Schumacher, D. Häbich, V. Hellwig, J. Benet Buchholz, R. Endermann, H. Labischinski, et al., ChemMedChem 2006, 1, 689–693.[79] B.-G. Lee, E. Y. Park, K.-E. Lee, H. Jeon, K. H. Sung, H. Paulsen, H. Rübsamen-Schaeff, H. Brötz-Oesterhelt, H. K. Song, Nat Struct Mol Biol 2010, 17, 471–478.[80] D. W. Carney, K. R. Schmitz, J. V. Truong, R. T. Sauer, J. K. Sello, J. Am. Chem. Soc. 2014, 136, 1922–1929.[81] M. Arvanitis, G. Li, D.-D. Li, D. Cotnoir, L. Ganley-Leal, D. W. Carney, J. K. Sello, E. Mylonakis, PLOS ONE 2016, 11, DOI 10.1371/journal.pone.0153912.[82] B. P. Conlon, E. S. Nakayasu, L. E. Fleck, M. D. LaFleur, V. M. Isabella, K. Coleman, S. N. Leonard, R. D. Smith, J. N. Adkins, K. Lewis, Nature 2013, 503, 365–370.[83] M. Gominet, N. Seghezzi, P. Mazodier, Microbiology 2011, 157, 2226–2234.[84] B. A. Johnson, H. Anker, F. L. Meleney, Science 1945, 102, 376–377.[85] L.-J. Ming, J. D. Epperson, J Inorg Biochem 2002, 91, 46–58.[86] D. R. Storm, J. L. Strominger, J. Biol. Chem. 1973, 248, 3940–3945.[87] N. J. Economou, S. Cocklin, P. J. Loll, Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 14207– 14212.[88] F. L. Lin, D. Woodmansee, R. Patterson, J Allergy Clin Immunol 1998, 101, 136–137.[89] A. Charlebois, L.-A. Jalbert, J. Harel, L. Masson, M. Archambault, PLOS ONE 2012, 7, DOI 10.1371/journal.pone.0044449.[90] A. C. Pawlowski, W. Wang, K. Koteva, H. A. Barton, A. G. McArthur, G. D. Wright, Nat Commun 2016, 7, DOI 10.1038/ncomms13803.[91] Azusa Kato, Seigo Nakaya, A. Yoshitami Ohashi, H. Hirata, K. F. and, K.-I. Harada, J. Am. Chem. Soc. 1997, 119, 6680–6681.[92] A. Kato, S. Nakaya, N. Kokubo, Y. Aiba, Y. Ohashi, H. Hirata, K. Fujii, K.-I. Harada, J. Antibiot. 1998, 51, 929–935.[93] M. Konishi, K. Sugawara, M. Hanada, K. Tomita, K. Tomatsu, T. Miyaki, H. Kawaguchi, R. E. Buck, C. More, V. Z. Rossomano, J. Antibiot. 1984, 37, 949–957.[94] H. Hashizume, R. Sawa, S. Harada, M. Igarashi, H. Adachi, Y. Nishimura, A. Nomoto, Antimicrob. Agents Chemother. 2011, 55, 3821–3828.[95] J. Shoji, H. Hinoo, T. Katayama, K. Matsumoto, T. Tanimoto, T. Hattori, I. Higashiyama, H. Miwa, K. Motokawa, T. Yoshida, J. Antibiot. 1992, 45, 817–823.[96] A. Müller, D. Münch, Y. Schmidt, K. Reder-Christ, G. Schiffer, G. Bendas, H. Gross, H.- G. Sahl, T. Schneider, H. Brötz-Oesterhelt, J. Biol. Chem. 2012, 287, 20270–20280.[97] S. A. Cochrane, X. Li, S. He, M. Yu, M. Wu, J. C. Vederas, J Med Chem 2015, 58, 9779– 9785.[98] S. A. Cochrane, B. Findlay, A. Bakhtiary, J. Z. Acedo, E. M. Rodriguez-Lopez, P. Mercier, J. C. Vederas, Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 11561–11566.[99] S. Boakes, M. J. Dawson, in Natural Products: Discourse, Diversity, and Design, John Wiley & Sons, Inc., Hoboken, NJ, USA, 2014, pp. 455–468.[100] H. Brötz, G. Bierbaum, K. Leopold, P. E. Reynolds, H.-G. Sahl, Antimicrob. Agents Chemother. 1998, 42, 154–160.[101] E. Breukink, I. Wiedemann, C. Van Kraaij, O. P. Kuipers, H. G. Sahl, B. de Kruijff, Science 1999, 286, 2361–2364.[102] T. Koopmans, T. M. Wood, P. ‘t Hart, L. H. J. Kleijn, A. P. A. Hendrickx, R. J. L. Willems, E. Breukink, N. I. Martin, J. Am. Chem. Soc. 2015, 137, 9382–9389.[103] L. L. Ling, T. Schneider, A. J. Peoples, A. L. Spoering, I. Engels, B. P. Conlon, A. Mueller, T. F. Schäberle, D. E. Hughes, S. Epstein, et al., Nature 2015, 517, 455–459.

34


[104] V. Ng, W. C. Chan, CHEM-EUR J 2016, 22, 12606–12616.[105] H. Yang, K. H. Chen, J. S. Nowick, ACS Chem. Biol. 2016, 11, 1823–1826.

35


36

Chapter 1

37

Synthesis and evaluation of daptomycin analogues

Chapter 2

Synthesis and evaluation of daptomycin analogues

ABSTRACT: The calcium-dependent lipopeptide antibiotics represent a promising new class of antimicrobials for use in combating drug-resistant bacteria. At present, daptomycin is the only such lipopeptide used clinically and displays potent antimicrobial activity against a number of pathogenic Gram-positive bacteria. Given the increasing need for new antibiotics, practical synthetic access to unnatural analogues of daptomycin and related antimicrobial lipopeptides is of value. We here report an efficient synthetic route combining solid- and solution-phase techniques that allows for the rapid preparation of daptomycin analogues. Using this approach, four such analogues, including two enantiomeric variants, were synthesized and their antimicrobial activities and hydrolytic stabilities evaluated.

P. ‘t Hart*, L. H. J. Kleijn*, G. de Bruin, S. F. Oppedijk, J. Kemmink, N. I. Martin, Org. Biomol. Chem. 2014, 12, 913–918. (* denotes equal first authorship)

L. H. J. Kleijn, F. M. Müskens, S. F. Oppedijk, G. de Bruin, N. I. Martin, Tetrahedron Lett 2012, 53, 6430–6432.

38

Synthesis and evaluation of daptomycin analogues Chapter 2

1 Introduction

The accelerated appearance of antibiotic resistance presents a serious and growing global health risk. Despite the increasing need for new antibacterial agents, only two mechanistically and structurally new antibiotics have reached the clinic in the past 40 years: linezolid and daptomycin.[1,2] While linezolid is of synthetic origin, daptomycin (Fig. 1) is a natural product isolated from fermentations of Streptomyces roseosporus. Daptomycin is rapidly bactericidal against Staphylococcus aureus, including methicillin-resistant S. aureus (MRSA), vancomycin-intermediate S. aureus (VISA) and vancomycin-resistant S. aureus (VRSA) strains.[3-5] Marketed under the trade name Cubicin, daptomycin is the first lipopeptide antibiotic of its kind to be approved for clinical use. Structurally unique, daptomycin is a cyclic depsipeptide composed of 13 amino acids (including non-proteinogenic and d-amino acids) and bears an N-terminal 10-carbon lipophilic tail.

While the precise mechanistic details of daptomycin’s antibacterial activity are unclear, it is known to disrupt aspects of bacterial cell membrane function.

In this regard, the current model for daptomycin’s mode of action involves interaction with the bacterial membrane leading to rapid depolarization and a loss of membrane potential resulting in bacterial cell death without rupturing the cell.[6,7] Daptomycin’s activity is calcium dependent with serum levels of free calcium (45–55 μg mL-1) sufficient to induce full antimicrobial activity.[8] Owing to the presence of four carboxylate side chains in the peptide, daptomycin is negatively charged at physiological pH and as such interacts with Ca2+ ions.[9] Upon binding to Ca2+ daptomycin is believed to oligomerize after which it is able to effectively disrupt the cell membranes of sensitive bacteria leading to defective cell division and cell wall synthesis.[10,11]

Given its role as the prototypical calcium-dependent lipopeptide antibiotic, daptomycin has garnered much attention from the synthetic community. Modified daptomycin analogues have been prepared via semi-synthesis[12-15] and the synthesis of daptomycin itself has been achieved using both chemo-enzymatic approaches or total synthesis. The Marahiel group demonstrated that linear precursor peptide thioesters could be converted to their corresponding daptomycin macrocycles by action of appropriate recombinant cyclase enzyme.[16] In addition, the three total syntheses of daptomycin have also been reported to date.[17-19] In all cases the synthetic routes described are labor intensive and make use of multi-step synthetic strategies. For these reasons we were drawn towards developing a convenient synthetic route for the preparation of daptomycin analogues so as to provide more

39


rapid access to structurally diverse derivatives of this important antibiotic.

In designing the daptomycin analogues to be prepared (Fig. 1) we opted to replace the l-threonine residue at position 4 with l-2,3-diaminopropionic acid. In doing so we were able to circumvent incorporation of the synthetically challenging

ester linkage between Thr4 and Kyn13 found in the natural daptomycin macrocycle by replacing it with the more accessible amide linkage.[17-19] In this regard, we also speculated that the corresponding macrocyclic amide analogue of daptomycin

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

ONH

O OHO

OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

D

D

DDap

amide linkage

Glu

12

3

4

5

6 7 8

9

10

111213

2

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

O OHO

OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

D

D

DKyn

NH2

3-MeGlu

1

12

3

4

5

6 7 8

9

10

111213

Daptomycin

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

ONH

O OH

OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

D

D

DDap

amide linkage

Glu

12

3

4

5

6 7 8

9

10

111213

3

NH

Trp

Thr

Kyn

NH2

Figure 1. Structure of daptomycin and synthetic analogues 2 and 3.

40


might show improved hydrolytic stability (vide infra). Aside from the amide for ester modification, we also incorporated l-glutamic acid at position 12 in place of the (2S,3R)-3-methyl glutamate normally found in daptomycin. While published preparations of (2S,3R)-3-methyl glutamate are available,[17-20] they are rather lengthy (>10 steps) and time consuming. In addition, Marahiel and coworkers previously investigated the same Glu for 3-MeGlu substitution in their chemo-enzymatic approach to daptomycin analogues and found it to have a relatively small effect on antimicrobial activity (MIC values increased by ca. 7-fold).[16] Finally, a second analogue, compound 3, was prepared wherein l-kynurenine at position 13 was replaced by the structurally similar l-tryptophan.

2 Results and discussion

2.1 Kynurenine synthesisThe incorporation of l-kynurenine in synthetic peptides can be achieved by on-resin treatment of a side chain protected l-tryptophan precursor with O3 at -78 °C.[21] While this method provides kynurenine in good yield, stringent ozonolysis conditions are required. Alternatively, a report by the group of Hoffman demonstrated that l-tryptophan can be converted to l-kynurenine under milder oxidation conditions via the β-3-oxindolylalanine intermediate. However, while leading to the formation of kynurenine, we found the yields achieved via this procedure to be somewhat variable. Furthermore, the reported use of reverse phase chromatography for the purification of both the polar intermediate and final product is not readily amenable to scale-up.

To address these issues we examined the use of Cbz-carbamate protected tryptophan 4 as a starting material to allow for a more convenient method of monitoring the subsequent conversions as well as product isolation (Scheme 1). To this end, Cbz-tryptophan was prepared via standard approaches after which it was treated with an oxidizing mixture of DMSO/HCl in AcOH to yield the expected diastereomeric mixture of Cbz-protected β-3-oxindolylalanines 5. A basic aqueous solution of 5 was then aerated, leading to formation of the protected kynurenine species 6 (conversion can be directly monitored by TLC). While the isolated yield of 7 was somewhat moderate (52%), its purification by conventional silica gel chromatography was straightforward and allowed for the reliable preparation of this material on large scale. Removal of the Cbz group by routine hydrogenation, followed by addition of an equimolar quantity of sulfuric acid and recrystallization from EtOH/H2O yielded pure kynurenine sulfate 8. In this manner, multi-gram

41


quantities of both L- and D-kynurenine were successfully prepared.[22] Fmoc- building block 9 was obtained without incident after treatment with Fmoc-Osu.

2.2 Peptide synthesisDaptomycin analogues 2 and 3 were generated using the combined solid- and solution-phase approach indicated (scheme 2). To avoid any racemization in the later cyclization step, the peptide synthesis began at resin-bound glycine 10, which in daptomycin corresponds to position 5. Employing an acid sensitive resin (2-chlorotrityl) and using standard SPPS techniques, the N-terminus of the peptide (intermediate 11) was first installed, including the C10 fatty acid tail. At this point, the diaminopropionic acid side chain was deprotected providing an attachment point for the kynurenine residue (position 13) after which the remainder of the peptide was assembled without incident to generate compound 12. Of particular note was the need to introduce DMB-protected glycine at position 10 to avoid aspartamide formation with the neighboring Asp residue. Upon completion, the intermediate protected peptide was cleaved from the resin using mild acid conditions and dissolved in CH2Cl2 at high dilution (0.5 mM) followed by treatment with BOP/DIPEA which led to clean formation of the desired macrocycle. Following deprotection and purification by RP-HPLC, daptomycin analogues 2 and 3 were obtained. 2.3 Biological activityMinimum inhibitory concentrations (MICs) activity of analogues 2 and 3 were determined and compared with authentic daptomycin (Table 1). Initially, the compounds were tested employing standard broth dilution assay employing S. aureus (ATCC 29213) as an indicator strain in the absence and presence of physiological Ca2+ levels (1.25 mM). Analogues 2 and 3 were both found to exhibit calcium-dependent antimicrobial activity albeit at a significantly reduced level

OH

O

HN

a)

NH2

OH

O

HN NHCbzOH

O

HN NHCbz

H

O

b)

OH

O

NHCbzONH2

d)

c)

OH

O

NH2ONH2 H2SO4

L-tryptophan4 5 6

7 8L-kynurenine

OH

O

NHFmocONH2

9

e)

Scheme 1. The conversion of l–tryptophan to l-kynurenine followed by generation of the Fmoc building block. Reagents and conditions: a) CbzCl, 1 M NaOH, H2O, 99%; b) DMSO, concentrated HCl, AcOH, 62%; c) air, 1 M NaOH, H2O, 52%; d) (i) H2, 10% Pd/C, 1,4-dioxane/aq. HCl; (ii) H2SO4, recrystallization from EtOH to H2O, 65%; e) Fmoc-Osu, 10% NHCO3/1,4-dioxane, 76%.

42


relative to that of daptomycin. After this first round of testing had been carried out, a report by the Taylor group demonstrated that the antibiotic potency of daptomycin analogs can improve significantly in the presence of higher Ca2+ concentrations.[19] They attribute this finding, which is only marginally seen in daptomycin itself, to diminished capacity of the daptomycin analogs to bind Ca2+. In line with these finding, compounds 2 and 3 showed significant improvement of antibiotic potencies upon retesting at in the presence of higher 5.0 mM and 10 mM Ca2+ concentrations.

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

ONH

O OHO

NH2OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

O

HN

O

NH O

HN

O

NH O

HN

O

HN

O

NH

O NHO

N

OHN

NH

O

O

HN

ONH

O OtBuO

NH2OtBu

O

OtBu

O

OtBuNHBoc

NBoc

O

NHTrt

O

OtBu

D

D

D

D

D

D

Dmb

OONH2

O

HN

O

NH O

HN

O

NH O

HN

NHAlloc

NBoc

O

NHTrt

O

OtBu

D

OO

FmocHN

OO a)

b), c)

d), e), f)

2

10

11

12

Scheme 2. Synthetic route developed for the preparation of daptomycin analogue 2. Reagents and conditions: (a) Fmoc-Xaa-OH (or decanoic acid), BOP, DIPEA, DMF; (b) Pd(PPh3)4, PhSiH3, CH2Cl2; (c) Fmoc-Xaa-OH, BOP, DIPEA, DMF; (d) HFIP–CH2Cl2; (e) BOP, DIPEA, CH2Cl2, 40 h; (f) TFA–TIS–H2O.

43


Nonetheless, compound 2 and 3 show significantly altered antimicrobial

activity in comparison with daptomycin. Marahiel and Taylor reported similarly attenuated activities for daptomycin analogues obtained via their respective chemo-enzymatic and synthetic approach.[18,19] The decrease of antibiotic activity seen in our compounds may be due to conformational changes/restrictions in the macrocycle that result from incorporation of the amide. The amide for ester substitution is also expected to impart greater hydrolytic stability. Thus, the serum stability of analogues 2 and 3 was also evaluated and compared with that of daptomycin. Each peptide was incubated with human plasma serum at 37 °C and sampled at specific time points. Under these conditions, daptomycin itself underwent significant degradation with an approximate 50% loss in the first 24 hours (Fig. 2). This degradation is presumably due to hydrolytic opening of the macrocyclic lactone as evidenced by the appearance of a new M + H2O species. By comparison, amide analogues 2 and 3 were much more stable under the same conditions with only minimal degradation detected over extended time periods of up to 48 hours.

compound no Ca2+ 1.25 mM Ca2+ 5.0 mM Ca2+ 10 mM Ca2+

1 (Daptomycin) >256 0.5 0.25 0.125

2 >256 160 16 4

3 >256 320 64 64

ent-2 >1280 >1280 ND ND

ent-3 >1280 >1280 >256 >256

Table 1. MIC determinations against S. aureus 29213 at various calcium concentrations. ND; not determined.

Figure 2. Serum stability of daptomycin (A) compared with analogues 2 (B) and 3 (C). Significant degradation of daptomycin is observed upon incubation at 37 °C while amide analogues 2 and 3 display increased stability.

44


2.4 Enantiomeric analoguesThe relative ease with which analogues 2 and 3 where assembled prompted us to also prepare the corresponding enantiomeric analogues ent-2 and ent-3. The preparation of biologically active peptides in both enantiomeric forms is an established approach used in establishing the role of chiral biomolecular targets.[23-25] If as some have proposed, daptomycin kills bacteria via membrane disruption without invoking a specific chiral bacterial target biomolecule, one would expect the “mirror-image” enantiomeric form of daptomycin to show an equal antibiotic activity. Conversely, should the enantiomeric form of daptomycin not show antibacterial activity, it can be taken as an indication that a chiral interaction with a specific biomolecular target is integral to daptomycin’s mode of action. To this end the enantiomeric analogues ent-2 and ent-3 were assembled using the appropriate stereochemically inverted amino acid building blocks. As expected, the enantiomeric analogues were shown to have identical retention times by analytical RP-HPLC (Fig. 3) and exhibited optical rotations of equal magnitude but opposite sign (Experimental table 2). In addition, the circular dichroism spectra and optical rotations obtained for the daptomycin analogues further supported their enantiomeric nature (Experimental fig. 4-5).

The antibacterial activities of ent-2 and ent-3 were next evaluated using the same assay as described above and revealed no detectable activity for both enantiomeric analogues under standard and higher Ca2+ concentrations (Table 1). These results indicate that a specific chiral interaction(s) is required for the activity of those analogues bearing the “native” daptomycin stereochemistry and may also support a similarly stereospecific mode of action for daptomycin itself. Plausible candidates for the requisite chiral target could include any number of membrane proteins or chiral phospholipids. In this regard, recent reports suggest that phosphatidyl glycerol may play a role. Specifically, in daptomycin-resistant strains of enterococci and S. aureus levels of phosphatidylglycerol are significantly reduced.

Figure 3. Overlays of analytical RP-HPLC traces obtained for enantiomeric daptomycin analogues (left) 2 and ent-2 and (right) 3 and ent-3.

45


3 Conclusions

In summary, we have developed a convenient approach employing both solid- and solution phase techniques for the preparation of analogues of the calcium-dependent lipopeptide family of antibiotics. While the antibacterial activity of the synthetic analogues was significantly reduced relative to that of daptomycin, their hydrolytic stability was greatly increased. The synthetic route here described should also be readily amenable to the production of new lipopeptide analogues with enhanced properties. Future work will be aimed at evaluating the effect of incorporating different amino acids in an attempt to increase antibacterial activity while maintaining the hydrolytic stability of these analogues. Support for this approach is evidenced by the observation that a tryptophan-for-kynurenine substitution as in compound 3 results to a more active analogue. In addition, our findings with the enantiomeric analogues implicate the involvement of a specific chiral target biomolecule(s) in bacterial strains that are sensitive to daptomycin.

46


Experimental Methods

Peptide synthesisLipopeptide analogue 2 was synthesized beginning with Fmoc-Gly loaded 2-chlorotrityl resin (410 mg, 0.25 mmol). Fmoc groups were removed with 20% piperidine in DMF (10 mL, 1 × 1 min., 1 × 30 min.). Coupling reactions were done in DMF (10 mL) with 4 equivalents of Fmoc-amino acid (or capric acid), 4 equivalents of BOP and 8 equivalents of DIPEA (1 hour). For removal of the Aloc protecting group, the resin was first washed with CH2Cl2 (2 × 10 mL) under argon after which PhSiH3 (0.74 mL, 6.0 mmol) in CH2Cl2 (4 mL) and Pd(PPh3)4 (74 mg, 0.06 mmol) in CH

2Cl

2 (12 mL) were added and the mixture swirled under argon for 1 hour. The reaction mixture was drained and the procedure was repeated. To remove residual palladium catalyst, the resin was then washed with CH2Cl2 (5 × 10 mL), a 0.5% solution of diethyldithiocarbamic acid trihydrate sodium salt in DMF (5 × 10 mL), and DMF (5 × 10 mL).

After Alloc removal, the remainder of the peptide was assembled on resin using standard SPPS approaches. The protected peptide was cleaved from the resin by addition of hexafluoroisopropanol (HFIP) in CH2Cl2 (1:4, 6 mL, 1 hour) and the filtrate collected. The resin was rinsed with additional HFIP–CH2Cl2 (2 × 3 mL) and the combined filtrates concentrated under vacuum. The crude linear peptide was dissolved in CH2Cl2 (500 mL, approximate peptide concentration of 0.5 mM) and BOP (0.22 g, 0.50 mmol, 2.0 equiv.) and DIPEA (0.17 mL, 1.0 mmol, 4.0 equiv.) were added. The cyclization reaction was monitored by TLC (5% MeOH in CH2Cl2) and was typically complete within 24 to 40 h. The reaction mixture was then concentrated, dissolved in EtOAc (250 mL) and washed with 1 M KHSO4 (2 × 200 mL) followed by saturated NaHCO3 (2 × 200 mL). The organic layer was dried over Na2SO4, filtered and concentrated under vacuum. The protected cyclic peptide was treated with a solution of TFA–TIS–H2O (95 : 2.5 : 2.5, 10 mL) for 1.5 hours. The mixture was then added to cold ether and the precipitated peptide collected by centrifugation and further washed with cold ether (2×).

Crude peptides were purified by RP-HPLC using a Maisch ReproSil-Pur C18-AQ column (250 × 22 mm, 10 μm) and employing a gradient of 30% to 70% Buffer B with a flow of 12 mL min-1 (Buffer A: 95% H2O, 5% MeCN, 0.1% TFA; Buffer B: 5% H2O, 95% MeCN, 0.1% TFA). Product containing fractions were pooled and lyophilized to yield between 5–10 mg of pure peptide (1.2–2.4% overall yield).

47


(S)-2-(((benzyloxy)carbonyl)amino)-3-(1H-indol-3-yl)propanoic acid (5)l-Tryptophan (10.24 g, 50 mmol) was dissolved in 1M NaOH (50 mL) and stirred at 0 °C. Benzyl chloroformate (7.15 mL, 50.1 mmol) and 1M NaOH (50 mL) where then simultaneously added in drop-wise fashion. The mixture was stirred for 1 hour at

room temperature. The solution was acidified with 6M HCl to pH 1 after which the product was extracted with EtOAc (3x200 mL). The organic layers were combined, dried by Na2SO4 and evaporated. Compound 5 was obtained a yellow solid and used directly in the following step (16.72 g, 99%). Rf 0.43 (95:4:1 CH2Cl2/MeOH/AcOH); Mp: 132-134 oC; 1H NMR (DMSO-d6): δ 7.57-7.53 (m, 2H), 7.35-7.22 (m, 6H), 7.16-7.15 (m, 1H), 7.09-7.04 (m, 1H) 7.00-6.95 (m, 1H), 5.02- 4.92 (m, 2H), 4.28-4.21 (m, 1H), 3.22-3.16 (m, 1H), 3. 04-2.96 (m, 1H); 13C NMR (DMSO-d6): δ 174.2, 156.5, 137.4, 136.6, 128.8, 128.2, 128.0, 127.6, 124.2, 121.4, 118.8, 118.6, 111.9, 110.5, 65.8, 55.5, 27.4; LRMS (ESI): calc. [M+H]+

339.13, found 339.35.

(2S)-2-(((benzyloxy)carbonyl)amino)-3-(2-oxoindolin-3-yl)propanoic acid (6)Cbz-l-Trp (5) (33.20 g, 98 mmol) was dissolved in AcOH (100 mL) and a mixture of DMSO (18 mL) and concentrated HCl (50 mL) was added. The resulting blue solution was stirred for 70 minutes. The mixture was then diluted with water (500 mL) and

extracted with EtOAc (3x400 mL). The combined organic layers were washed with brine (800 mL), dried with MgSO4, and evaporated. Compound 6 was purified by column chromatography (50:1:0 to 25:1:0.1, DCM/MeOH/AcOH), yielding the product as a white solid (23.14 g, 67%). The product is a diastereomeric mixture of (S)-2-(((benzyloxy)carbonyl)amino)-3-((S)-2-oxoindolin-3- yl)propanoic acid and (S)-2-(((benzyloxy)carbonyl)amino)-3-((R)-2-oxoindolin-3-yl)propanoic acid. Rf 0.14 (95:4:1 CH2Cl2/MeOH/AcOH); Mp: 160-164 oC; 1H NMR (DMSO-d6): δ 7.88-7.85 (m, 1H), 7.76-7.73 (m, 1H), 7.34-7.07 (m, 14H), 6.95-6.91 (m, 2H), 6.85-6.79 (m, 2H), 5.08-5.00 (m, 4H), 4.55-4.53 (m, 1H), 4.39 (m, 1H), 3.39 (m, 2H) 2.25-1.89 (m, 4H); 13C NMR (DMSO-d6): δ 179.2, 178.9, 174.2, 173.8, 156.8, 156.6, 143.1, 142.9, 137.4, 129.8, 129.3, 128.8, 128.3, 128.3, 128.2, 128.1, 124.8, 124.4, 121.8, 121.7, 109.9, 109.7, 66.0, 65.9, 52.0, 51.8, 42.7, 42.0, 32.8; LRMS (ESI): calc. [M+H]+

355.13, found 355.35.

(S)-4-(2-aminophenyl)-2-(((benzyloxy)carbonyl)amino)-4-oxobutanoic acidThe Cbz-protected oxindolylalanine (6) (10.29 g, 29 mmol) was dissolved in 1M NaOH (500 mL) and air was bubbled through the mixture while stirring for 4 hours. The solution was then acidified with 6M HCl on ice to pH 1 followed by extraction with

OH

O

HN NHCbz5

OH

O

HN NHCbz

H

O6

OH

O

NHCbzONH2

7

48


EtOAc (3x 250 mL). The organic layers were combined, dried with Na2SO4, and concentrated under vacuum. The product was purified by column chromatography (100:4:0.4, DCM/MeOH/AcOH), yielding a light green solid (5.21 g, 52%). Rf 0.35 (95:4:1 CH2Cl2/MeOH/AcOH); Mp: 154-162 oC; 1H NMR (DMSO-d6): δ 7.71-7.69 (m, 1H), 7.50-7.47 (m, 1H), 7.33-7.20 (m, 7H), 6.77-6.74 (m, 1H), 6.55-6.50 (m, 1H), 5.02 (s, 2H), 4.58-4.57 (m, 1H) , 3.452-3.284 (m, 2H); 13C NMR (DMSO-d6): δ 198.4, 173.9, 156.3, 151.7, 137.4, 134.8, 131.6, 128.8, 128.2, 128.1, 117.4, 116.7, 114.9, 65.9, 50.3; LRMS (ESI): calc. [M+H]+

343.13, found 343.40.

(S)-2-amino-4-(2-aminophenyl)-4-oxobutanoic acid sulfate, L-Kynurenine sulfateCbz-protected l-kynurenine (7) (3.56 g, 10.0 mmol) was dissolved/suspended in a mixture of dioxane (60 mL) and 0.5 M HCl (240 mL). 10% Pd/C (300 mg) was added and the mixture was stirred under H2 (balloon pressure) overnight. The

catalyst was then removed by filtration over celite and the filtrate concentrated under vacuum. A slight excess of 1M H2SO4 (12.3 mL) was added and the solution was concentrated under vacuum once more. The solid that remained was then recrystallized from 66% EtOH (250 mL) overnight at 4 oC. The precipitated product was collected by filtration and dried under vacuum to yield pure l-kynurenine sulfate as a light yellow solid (2.44 g, 65%). Mp: 185 oC (decomp.); 1H NMR (D2O): δ 8.03-8.01 (m, 1H), 7.65-7.60 (m, 1H), 7.51-7.46 (m, 1H), 7.36-7.34 (m, 1H), 4.41-4.38 (m, 1H), 3.82-3.80 (m, 1H); 13C NMR (DMSO-d6): δ 199.2, 171.0, 162.3, 135.8, 132.0, 130.8, 129.3, 126.8, 125.0, 48.4, 38.9; LRMS (ESI): calc. [M+H]+

209.09, found 209.25; [α]D20 + 9.1 (c 1.0, H2O) [literature[26], [α]D

20: + 9.6 (c 1.0, H2O)].

(R)-2-amino-4-(2-aminophenyl)-4-oxobutanoic acid sulfate, d-Kynurenine sulfate d-Kynurenine sulfate was prepared in analogous fashion but beginning from d-tryptophan. The analytical data obtained for d-kynurenine sulfate (and all intermediate compounds generated in its preparation) were identical to those obtained

in the synthesis of l-kynurenine sulfate with the exception of the optical rotation. As expected, the optical rotation for the d-kynurenine sulfate prepared is of equal magnitude but opposite sign to that measured for the l-enantiomer: [α]D

25 - 9.2 (c 1.0, H2O) [literature[26], [α]D

20 - 9.5 (c 1.0, H2O)].

OH

O

NH2ONH2 H2SO4

8

OH

O

ONH2 H2SO4

D-kynurenineNH2

49


S)-2-((((9H-fluoren-9-yl)methoxy)carbonyl)amino)-4-(2-aminophenyl)-4-oxobutanoic acid

FmocOsu (2.30g, 6.80 mmol) in dioxane (10 mL) was added slowly to a cooled solution of L-kynurenine sulfate salt (1.99 g, 6.48 mmol) in 10% NaHCO3 (12 mL) and dioxane (10 mL) and the mixture was stirred overnight at room temperature. The

mixture was concentrated, diluted with water (100 mL) and washed with Et2O (3x80 mL). The aqueous phase was acidified (pH=3.5) with citric acid and extracted with EtOAc (3x100 mL). The combined organic fractions were washed with 1M KHSO4 (3x80 mL) and brine (3x80 mL), dried with Na2SO4 and concentrated to yield Fmoc-l-kynurenine as an orange solid (2.12 g, 4.93 mmol, 76%). Rf: 0.46 (CH2Cl2/MeOH, 9/1). ESI-MS: [M+H+]: calc: 431.16, found: 431.25; 1H NMR (300 MHz, CDCl3): δ 7.74 (d, 7,5 Hz, 2H), 7.66 (d, 8.4 Hz, 1H), 7.58 (d, 7.5 Hz, 2H), 7.39-7.29 (m, 5H), 6.64 (m, 2H), 5.93 (d, 7.2, 1H), 4.79 (m, 1H), 4.47-4.32 (m, 2H), 4.24-4.19 (m, 1H), 3.80 (dd, 16.2 Hz, 1H), 3.52 (dd, 15.3 Hz, 1H). 13C NMR (300 MHz, CDCl3): δ 199.2, 176.3, 156.4, 150.6, 143.9, 143.7, 141.3, 135.1, 131.1, 135.1, 131.1, 127.7, 127.1, 125.2, 120.0, 117.5, 117.0, 116.1, 67.3, 67.0, 50.1, 47.1, 41.3. Optical rotation (1.00 g 100mL-1, CHCl3): Fmoc-(l)-Kyn: [α]D

25= +97.8±0.7 °; Fmoc-(d)-Kyn [α]D

25= -84.0±0.2 °.

Minimum inhibitory concentration assays The daptomycin analogues were tested against S. aureus (ATCC 29213) as an indicator strain. Two-fold serial dilutions of each compound were made in microtiter plates using Mueller–Hinton broth. Each well was inoculated at 1 × 106 CFU and incubated at 37 °C for 16 h followed by visual determination of MIC values. In these assays calcium- and magnesium-free Mueller–Hinton broth (Fluka) was supplemented with CaCl2 (final Ca2+ concentration 1.25 mM, 5.0 mM or 10.0 mM) and with MgCl2 (final Mg2+ concentration 10 mg L-1) or alternatively with MgCl2 alone to assess the effect of calcium on antibacterial activity.

Serum stability assays Daptomycin and analogues 2 and 3 were added to human plasma serum to a final concentration of 150 μg mL-1, followed by incubation at 37 °C. At 0, 24 and 48 hours a 100 μL aliquot was taken and added to a 200 μL volume of methanol to precipitate plasma proteins (methanol also contained 0.075 mg mL-1 ethylparaben as an internal standard). The mixture was vortexed for 5 seconds and allowed to stand at room temperature for 15 minutes. After centrifugation at 13 000 rpm for 5 minutes, the supernatant was removed and used for HPLC analysis. Analysis of the samples was performed by analytical RP-HPLC with a Maisch ReproSil-Pur 120

OH

O

NHFmocONH2

7

50


C18-AQ column (250 × 4.6 mm, 5 μm) and a linear gradient of 0–100% buffer B over 48 minutes at a flow rate of 1 mL min-1 (Buffer A: 95% H2O, 5% MeCN, 0.1% TFA; Buffer B: 5% H2O, 95% MeCN, 0.1% TFA). Both the internal standard and peptides were detected at 220 nm allowing for the amount of intact peptide to be calculated based upon the ratio of peak areas corresponding to the internal standard and the intact peptides. P. ‘t Hart performed this experiment.

Circular dichroism spectroscopy CD spectra for 2, ent-2, 3 and ent-3 were recorded at a peptide concentration of 60 μM in 20 mM HEPES buffer (pH 7.4) (Experimental Fig. 4). Samples were measured at room temperature over the 210–250 nm range at a scan rate of 20 nm min-1 and a bandwidth of 1 nm. The spectra were converted to molar ellipticities in units of deg × cm2 dmol-1.

51


Analytical data of synthesized peptides

Figure 4. CD spectra. Left: daptomycin analogue 2 (solid green) and ent-2 (dashed green). Right: analogue 3 (solid red) and ent-3 (dashed red).

Table 2. Analytical data for compounds 1 (daptomycin), 2, ent-2, 3, and ent-3.

Table 3. NMR chemical shift assignment for compound 2 (DMSOd6). * ambiguous assignment.

compound Rt (min) [α]D25 [M + H]+ calc. [M + H]+ found

1 (daptomycin) 30.2 N.D. 1620.7182 1620.7167

2 29.3 −7.98 (0.28, H2O) 1591.7029 1591.6997

3 29.3 +8.70 (0.28, H2O) 1591.7029 1591.6995

ent-2 29.6 −9.45 (0.65, H2O) 1587.7080 1587.7069

ent-3 29.6 +9.24 (0.49, H2O) 1587.7080 1587.7063

Residue Hα(Cα) Hβ (Cβ) Sidechain

Tail CH2 1.07-1.19 (28.5), CH2 1.20 (31.0), CH3 0.84 (13.7)

Trp-1 4.43 (53.8) 2.90/3.05 (26.9) γ2 7.15 (123.4), δ4 7.56 (118.2), δ5 6.95 (117.9),

δ6 7.03 (120.5), δ7 7.30, (111.0) HN 10.77

Asn-2 4.59 (49.5) 2.69/2.53 (35.8)*

Asp-3 4.53 (49.5)* 2.69/2.53 (35.8)*

Dap-4 4.24-4.27 (51.9)* NA

Gly-5 3.72/3.85 (42.0)

Orn-6 4.20 (52.3) NA γ 1.58 (22.9)

Asp-7 4.53 (49.5)* 2.69/2.53 (35.8)*

Ala-8 4.20 (48.2) 1.22 (17.2)

Asp-9 4.53 (49.5)* 2.69/2.53 (35.8)* ---

Gly-10 3.72/3.85 (42.0) --- ---

Ser-11 4.24-4.27 (51.9)* 3.60 (61.5) ---

Glu-12 4.45 (49.7) NA γ2.26 (429.9)

Kyn-13 4.24-4.27 (51.9)* 2.90/3.05 (26.9) γ3 6.74 (116.6), γ4 7.22 (133.9), γ5 6.53 (114.2),

γ6 7.72 (131.0)

52


Table 4. NMR chemical shift assignment for compound 3 (DMSOd6). * ambiguous assignment.

Residue Hα(Cα) Hβ (Cβ) Sidechain

Tail CH2 1.07-1.19 (28.5), CH2 1.19 (31.0), CH3 0.84 (13.7)

Trp-1 4.47 (53.8) 3.09/2.92 (26.9) δ1 7.16 (123.4), ε3 7.58 (118.1), ζ3 6.96 (117.9), η2 7.04

(120.5), ζ2 7.31 (110.9), HN 10.78

Asn-2 4.60 (49.4) 2.71/2.52 (35.9)* ---

Asp-3 4.53 (49.4)* 2.71/2.52 (35.9)* ---

Dap-4 4.16 (51.9) 3.64/3.03 (39.8) ---

Gly-5 3.87/3.67 (42.2) --- ---

Orn-6 4.33 (51.8) 2.75 (38.2) γ 1.56 (23.1), δ 2.75 (38.3)

Asp-7 4.53 (49.4)* 2.71/2.52 (35.9)* ---Ala-8 4.22 (48.4) 1.18 (17.3) ---

Asp-9 4.53 (49.4)* 2.71/2.52 (35.9)* ---

Gly-10 3.87/3.67 (42.2) --- ---

Ser-11 4.40 (54.9) 3.60 (61.5) ---

Glu-12 4.35 (51.7) 1.90/1.71 (26.9) γ 2.21 (29.7)

Trp-13 4.35 (54.0) 3.09/2.92 (26.9) δ1 7.07 (123.4), ε3 7.47 (118.0), ζ3 6.96 (117.9), η2 7.04

(120.5), ζ2 7.31 (110.9) HN 10.71

53


1

References[1] T. Roemer, C. Boone, Nat. Chem. Biol. 2013, 9, 222–231.[2] G. D. Wright, Nature Rev Microbiol 2007, 5, 175–186.[3] D. Patel, M. Husain, C. Vidaillac, M. E. Steed, M. J. Rybak, S. M. Seo, G. W. Kaatz, Int J Antimicrob Agents 2011, 38, 442–446.[4] H. S. Sader, R. N. Jones, Diagn. Microbiol. Infect. Dis. 2009, 65, 158–162.[5] H. S. Sader, T. R. Fritsche, R. N. Jones, Diagn. Microbiol. Infect. Dis. 2009, 13, 291–295.[6] J. Pogliano, N. Pogliano, J. A. Silverman, J. Bacteriol. 2012, 194, 4494–4504.[7] A. Müller, M. Wenzel, H. Strahl, F. Grein, T. N. V. Saaki, B. Kohl, T. Siersma, J. E. Bandow, H.-G. Sahl, T. Schneider, et al., Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 7077– 7086.[8] G. M. Eliopoulos, C. Thauvin, B. Gerson, R. C. Moellering Jr, Antimicrob. Agents Chemother. 1985, 27, 357–362.[9] R. H. Baltz, V. Miao, S. K. Wrigley, Nat. Prod. Rep. 2005, 22, 717–741.[10] R. H. Baltz, Current Opinion in Chemical Biology 2009, 13, 144–151.[11] L. Robbel, M. A. Marahiel, J. Biol. Chem. 2010, 285, 27501–27508.[12] J. Siedlecki, J. Hill, I. Parr, X. Yu, M. Morytko, Y. Zhang, J. Silverman, N. Controneo, V. Laganas, T. Li, et al., Bioorg. Med. Chem. Lett. 2003, 13, 4245–4249.[13] J. Hill, J. Siedlecki, I. Parr, M. Morytko, X. Yu, Y. Zhang, J. Silverman, N. Controneo, V. Laganas, T. Li, et al., Bioorg. Med. Chem. Lett. 2003, 13, 4187–4191.[14] Y. He, J. Li, N. Yin, P. S. Herradura, L. Martel, Y. Zhang, A. L. Pearson, V. Kulkarni, C. Mascio, K. Howland, et al., Bioorg. Med. Chem. Lett. 2012, 22, 6248–6251.[15] S. Yoganathan, N. Yin, Y. He, M. F. Mesleh, Y. G. Gu, S. J. Miller, Org. Biomol. Chem. 2013, 11, 4680–4685.[16] Jan Grünewald, Stephan A Sieber, Christoph Mahlert, A. Uwe Linne, M. A. Marahiel, J. Am. Chem. Soc. 2004, 126, 17025–17031.[17] D. C. Alexander, P. Brian, G. M.-F. Coeffet-Le, X. He, V. Kulkarni, C. Leitheiser, I. B. Parr, D. Ritz, R. H. Baltz, C. P. Inc, et al., 2007.[18] H. Y. Lam, Y. Zhang, H. Liu, J. Xu, C. T. T. Wong, C. Xu, X. Li, J. Am. Chem. Soc. 2013, 135, 6272–6279.[19] C. R. Lohani, R. Taylor, M. Palmer, S. D. Taylor, Org. Lett. 2015, 17, 748–751.[20] Claire Milne, Amanda Powell, John Jim, Majid Al Nakeeb, A. Colin P Smith, Jason Micklefield, J. Am. Chem. Soc. 2006, 128, 11250–11259.[21] C. T. T. Wong, H. Y. Lam, X. Li, Org. Biomol. Chem. 2013, 11, 7616–7620.[22] L. H. J. Kleijn, F. M. Müskens, S. F. Oppedijk, G. de Bruin, N. I. Martin, Tetrahedron Lett 2012, 53, 6430–6432.[23] D. Wade, A. Boman, B. Wåhlin, C. M. Drain, D. Andreu, H. G. Boman, R. B. Merrifield, Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 4761–4765.[24] L. Sando, S. Troeira Henriques, F. Foley, S. M. Simonsen, N. L. Daly, K. N. Hall, K. R. Gustafson, M. I. Aguilar, D. J. Craik, ChemBioChem 2011, 12, 2456–2462.[25] F. Dettner, A. Hänchen, D. Schols, L. Toti, A. Nußer, R. D. Süssmuth, Angew. Chem. Int. Ed. 2009, 48, 1856–1861.[26] J. L. Warnell, C. P. Berg, J. Am. Chem. Soc. 1954, 76, 1708–1709.

54

Chapter 2

55

Synthesis and Mode of Action of Laspartomycin C

Chapter 3

Synthesis and mode of action of laspartomycin C

ABSTRACT: Laspartomycin C is a lipopeptide antibiotic with activity against a range of Gram-positive bacteria including drug-resistant pathogens. We report the first total synthesis of laspartomycin C as well as a series of structural variants. Laspartomycin C was found to specifically bind undecaprenyl phosphate (C55-P) and inhibit formation of the bacterial cell wall precursor lipid II. While several clinically used antibiotics target the lipid II pathway, there are no approved drugs that act on its C55-P precursor.

L. H. J. Kleijn, S. F. Oppedijk, P. 't Hart, R. M. van Harten, L. A. Martin-Visscher, J. Kemmink, E. Breukink, N. I. Martin, J. Med. Chem. 2016, 59, 3569–3574.

56

Synthesis and Mode of Action of Laspartomycin C Chapter 3

1 Introduction

Laspartomycin C is a cyclic lipopeptide that belongs to the family of calcium-dependent antibiotics (CDAs).[1] Originally isolated from Streptomyces viridochromogenes, laspartomycin C has been found to possess antibacterial activity against a variety of Gram-positive pathogens including methicillin- resistant Staphylococcus aureus (MRSA), vancomycin-intermediate S. aureus (VISA), vancomycin-resistant S. aureus (VRSA), and vancomycin-resistant enterococci (VRE).[2,3] The structure of laspartomycin C was elucidated in 2003 showing it to be a member of the CDA family (Fig. 1).[1,4] In the same year daptomycin (2) became the first CDA to receive FDA approval for clinical use, and it is now a widely used antibiotic of last resort.[5] In addition, surotomycin, a semisynthetic analog of daptomycin, is currently undergoing phase 3 clinical trials for the treatment of Clostridium difficile infections.[6] In light of emerging bacterial resistance to conventional antibiotics, interest in the CDAs has grown steadily over the past 2 decades. This increased interest is driven by findings that indicate that the various CDAs operate via modes of action unlike those of conventional antibiotics.[7,8]

Laspartomycin C, also known as glycinocin A, consists of a 10 amino acid cyclic core and an N-terminal exocyclic region.[4] The macrocycle contains a number of nonproteinogenic amino acids, including l-2,3-diaminopropionic acid (l-2,3-Dap) as well as d-amino acids such as d-pipecolic acid (d-Pip) and d-allo-threonine. Also present in the macrocycle is the Asp-X-Asp-Gly motif, which is implicated in Ca2+ binding and conserved among all known CDAs. Laspartomycin C bears an unsaturated and branched C15 fatty acid tail linked to the exocyclic N-terminal aspartic acid residue. By comparison, daptomycin contains a larger exocyclic tripeptide unit terminated with a straight chain, fully saturated C10 lipid. In addition, while the 10 amino acid macrocycle in daptomycin is formed via an ester linkage between its C-terminal residue (l-kynurenine) and a threonine side chain, the laspartomycin C macrocycle is closed via an amide linkage between its C-terminal proline and the side chain of the l-2,3-Dap residue at position 2.

Interestingly, while all CDAs require the presence of Ca2+ to achieve their antimicrobial activity, they do not all act via the same antibacterial mechanism.[7] Daptomycin is the most potent of the known CDAs and acts on the bacterial membrane where the proposed formation of daptomycin oligomers is believed to induce membrane perturbation.[8-11] Alternatively, the CDAs amphomycin and friulimicin B were recently shown to exert their bactericidal effect by inhibiting bacterial cell wall synthesis through complex formation with the essential cell wall

57


precursor undecaprenyl phosphate (C55-P).[12,13] While laspartomycin C shares structural similarities with amphomycin and friulimicin B (all three contain amide-linked macrocycles flanked by d-Pip and Pro residues at positions 3 and 11), the mechanistic details of its antibiotic mode of action have not been reported. Furthermore, while multiple syntheses of daptomycin have been reported,[14-16] to date no other CDAs have been prepared in synthetic fashion. In this study we report the first total synthesis of laspartomycin C and we here describe the application of a range of biochemical and biophysical approaches in characterizing the antibacterial mechanism of laspartomycin C.

In addition, a series of lipopeptides that comprise elements of both daptomycin and laspartomycin C were synthesize and evaluated with regard to their antibiotic activities (Fig. 2). The conserved motif of conformationally restricting amino acids Pro and d-Pip in laspartomycin C, amphomycin and friulimicin likely plays an important role in establishing the biologically relevant conformation of the peptide. We therefore set out to investigate whether incorporating these amino acids would improve the antibacterial activity of our previously prepared daptomycin “amide

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

O OHO

NH2OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

12

3

4

5

6 7 8

9

10

111213

D

D

D

O

HN

O

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

OH

O

OH

O

OH

O

OH

O

O

1

2

3

4 5 6

7

8

91011

D

D

1

2Daptomycin

Laspartomycin C

Figure 1. Structures of laspartomycin C (1) and daptomycin (2) indicating N-terminal lipids (red) and conserved Asp-X-Asp-Gly motif (blue). The peptide macrocycles are formed biosynthetically via cyclization of the C-terminal residue with the side chain of l-2,3-Dap2 in laspartomycin C or Thr4 in daptomycin (linkages shown in green).

58


analogues” 3 and 4 (chapter 2).[17] As illustrated in Figure 2, analogs 5 and 6 contain a Gly to d-Pip mutation in comparison with 3 and 4. Analogue 7 bears a Pro in place of the Kyn normally present in daptomycin, and in analogue 8 both Pro and d-Pip residues are introduced as in laspartomycin C. Analogue 12 represents a hybrid structure wherein the laspartomycin C macrocyle is augmented with the daptomycin exocyclic tripeptide unit and C10 lipid tail.


2.1 Synthesis of daptomycin/laspartomycin C analoguesIn order to arrive at a synthesis method that is applicable to the preparation of laspartomycin C and all intended analogues, we opted for the convenient end-to-side chain cyclization strategy (scheme 1). In the case of analogue 8, which served as example to explore this synthesis strategy, the C-terminal Pro13 was immobilized on 2-chlorotrityl resin followed by standard automated Fmoc solid-phase peptide synthesis under HBTU/HOBt/DiPEA coupling conditions. However, LC-MS analysis

O

HN

O

NH O

HN

O

NH O

O

HN

O

NH

O NHO

N

OHNNH

O

OtBuO

OtBu

O

OtBu

O

OtBuNHBoc

NBoc

O

NHTrt

O

OtBu

O

N

NNH

O

OO

NFmoc

OO

NHAlloca)

b), c)O

HN

O

NH O

HN

O

NH O

O

HN

O

NH

O NHO

N

OHNNH

O

OtBuO

OtBu

O

OtBu

O

OtBuNHBoc

NBoc

O

NHTrt

O

OtBu

O

N

NNH

O

OHO

NH2d), e)

8

Dmb

Dmb12

34

5

6 7 8

9

10

1112

13

12

34

5

6 7 8

9

10

1112

13

Scheme 1. The initial end-to-side chain cyclization synthesis strategy (a) Fmoc SPPS; (b) Pd[(C6H5)3P]4, C6H5SiH3, CH2Cl2, 1 h; (c) (CF3)2CHOH, CH2Cl2, 1 h; (d) BOP, DIPEA, CH2Cl2, 16 h; (e) TFA, TIS, H2O, 1 h.

59


of the protected linear peptide showed predominantly the presence of a 9 amino acid intermediate with Fmoc-deprotected d-Pip5 at the N-terminus demonstrating that achieving acylation of the sterically hindered secondary amine present in d-Pip5 as part of this particular amino acid sequence would require alternative conditions.

Accordingly, the more stringent HATU/HOAt/NMM mixture was applied to couple Fmoc-Dap(Alloc)-OH to d-Pip allowing for assembly of the full linear peptide. The Alloc group was removed using palladium tetrakis followed by a mild HFIP/CH2Cl2 cleavage from the resin leaving the protecting groups intact. Finally, the peptide was cyclized in solution and deprotected using TFA/TIS/H2O. Unfortunately, LC-MS analysis of crude 8 showed two compounds with the desired m/z in 70:30 ratio. To rule out that conformational isomers had been generated, both compounds were purified using reverse phase preparative HPLC. Next, a 1:1 mixture of the purified compounds was subjected to LC-MS analysis at both 25 °C and 60 °C. At both temperatures, the 1:1 ratio was maintained demonstrating that no interconversion of confomers took place. The observed 70:30 diasteriomeric ratio observed in crude peptide 8 could therefore be attributed to racemization of Fmoc-Dap(Alloc)-OH under the employed HATU/HOAt/NMM conditions.

In parallel, an alternative strategy for introducing the Dap-d-Pip sequence into the peptide was explored. Generation of the Fmoc-Dap(Alloc)-d-Pip-OH dipeptide using solution phase chemistry, followed by coupling of the dipeptide to resin bound Lys6, would circumvent the need to generate the challenging Dap-d-Pip amide bond on-resin. The synthesis of the Fmoc-protected dipeptide was surprisingly challenging and as such three different strategies were explored (scheme 2).

Synthesis route 1 involved treatment of Fmoc-Dap(Alloc)-OH with DCC and NHS to generate the Fmoc-Dap(Alloc)-OSu intermediate. The OSu-ester was subsequently reacted with H-d-Pip-OH under either basic aqueous (sat. NaHCO3/acetone) or basic organic (Et3N, CH2Cl2/MeOH) conditions, leading only to the recovery of starting materials. Route 2 involved generating the methyl ester of d-Pip using SOCl2 and MeOH. Fmoc-Dap(Alloc)-OH was then successfully coupled to H-d-Pip-OMe to generate the protected dipeptide in 72% yield after silica gel column chromatography. Finally, Fmoc-Dap(Alloc)-d-Pip-OMe was treated with LiOH, which unfortunately led to the removal of the Fmoc-group rather than achieving the intended selective hydrolysis of the methyl ester. A final effort (route 3) involved employment of HClO3 and sat. NaHCO3 to generate H-d-Pip-OtBu. Again, the corresponding Fmoc-Dap(Alloc)-d-Pip-OtBu dipeptide could then be generated without incident using BOP and DiPEA coupling and after silica gel

60


column chromatography the compound was isolated in 95% yield. Subsequent treatment with TFA resulted both in removal of the tert-butyl protective group (ca. 65%) and in formation of an unidentified side product (ca. 35%). However, after isolation the Fmoc-Dap(Alloc)-d-Pip-OH dipeptide showed a significant degree of racemization of the d-Pip α-carbon. The racemization as well as the formation of the side product had likely occurred via an oxazolonium ion intermediate that is known to form in certain pipecolic acid peptide sequences under strong acidic conditions.[18] This finding prompted us to abandon the “dipeptide” strategy in favor of an alternative SPPS strategy.

We next opted to initiate the peptide synthesis from the alternative starting point of Gly10 in an effort to easy the coupling between Dap4 and d-Pip5 (scheme 3). This strategy holds the inherent advantage that the C-terminal Gly10 cannot undergo racemization during the cyclization reaction and all intended analogues as well as laspartomycin C contain a glycine at that position in the macrocycle. The synthesis began with immobilization of Fmoc(Dmb)-Gly-OH on the 2-chlorotrityl resin (the Dmb protecting group was employed to prevent possible aspartamide formation). Interestingly, treatment of the resin-bound Fmoc(Dmb)-Gly10 with piperidine:DMF did not result in complete Fmoc deprotection nor did treatment with a more potent mixture of DBU:piperidine:DMF. However, treatment with ethanolamine:DMF did lead to full Fmoc removal indicating that deprotection of resin-bound Fmoc(Dmb)-Gly requires a less sterically hindered base. After coupling of aspartic acid, the Fmoc

HN

O OH

HN

O OtBu

AllocHN

FmocHNO

f) g) N

O OtBu

AllocHN

FmocHNO

OHa)

AllocHN

FmocHNO

OSub)

AllocHN

FmocHNO

N

O OH

HN

O OH

HCl.HN

O OMe

AllocHN

FmocHNO

c) d)N

O OMe

AllocHN

FmocHNO

N

O OH

e)

AllocHN

FmocHNO

N

O OH

h)

Route 1

Route 2

Route 3

H

Scheme 2. Explored Fmoc-Dap(Alloc)-d-Pip-OH synthesis strategies. Reagents and conditions. (a) DCC, NHS, THF; (b) d-Pip, sat. NaHCO3/acetone or d-Pip, Et3N, CH2Cl2/MeOH; (c) SOCl2, MeOH, 96%; (d) Fmoc-Dap(Alloc)-OH, BOP, DiPEA, CH2Cl2, 72%; (e) LiOH, THF/H2O; (f) HClO3,

tBuAc, NaHCO3, NaHCO3 (sat.), 56%; (g) Fmoc-Dap(Alloc)-OH, BOP, DiPEA, CH2Cl2/DMF (99/1), 95%; (h) TFA/CH2Cl2.

61


loading of the resin was determined spectrophotochemically (0.52 mmol g-1). Fmoc SPPS was then employed to arrive at d-Pip5, which could now conveniently be coupled to Fmoc-Dap(Alloc)-OH without incident under standard BOP/DiPEA or HBTU/HOBt/DiPEA coupling condition. It is likely that the position of d-Pip5 relative to the resin, with the Dap4-d-Pip5 bond forming reaction now representing the 6th amino acid coupling, reduced steric hindrance by or aggregation of the growing peptide to render the N-terminal d-Pip5 more reactive towards electrophiles. Using

O

HN

O

NH O

HN

O

NH O

O

HN

O

NH

O NHO

N

OHNNH

O

OtBuO

OtBu

O

OtBu

O

OtBuNHBoc

NBoc

O

NHTrt

O

OtBu

O

N

NNH

O

ONH Dmb1

23

4

5

6 7 8

9

10

111213

a)

O

HN

O

NH O

HN

O

NH O

O

HN

O

NH

O NHO

N

NH2

NH

O

OtBuO

OtBu

O

OtBu

O

OtBuNHBoc

NBoc

O

NHTrt

O

OtBu

O

N

NNH

O

ONH Dmb1

23

4

5

6 7 8

9

10

111213

O

O

O

HN

O

NH O

HN

O

NH O

O

HN

O

NH

O NHO

N

O

OtBu

O

OtBuNHBoc

NBoc

O

NHTrt

O

OtBuN

NH

O

NHAllocDmb1

23

4

5

6 7 8

9

10

O

OFmocN

DmbO

O

b), c)

d), e) f)8

Scheme 3. The synthesis strategy (a) Fmoc SPPS; (b) Pd[(C6H5)3P]4, C6H5SiH3, CH2Cl2, 1 h; (c) Fmoc SPPS; (d) (CF3)2CHOH, CH2Cl2, 1 h; (e) BOP, DIPEA, CH2Cl2, 16 h; (f) TFA, TIS, H2O, 1 h.

62


standard Fmoc SPPS, the remainder of the linear peptide stretch was obtained with N-terminal acylation achieved using decanoic acid. Removal of the side chain Alloc protecting group was followed in turn by addition of the remaining three amino acids. Cleavage from the resin using mild acidic conditions yielded the protected peptide, which was directly subjected to solution phase cyclization. Complete conversion from the linear to the cyclic peptide was achieved within 12 h as evidenced by HPLC analysis followed by global deprotection and HPLC purification. Using this methodology, lipopeptides 5-9 were synthesized with no detectable racemization in an average yield of 7.1%.

2.2 Synthesis of laspartomycin CThe same synthesis strategy could now be applied to the synthesis of natural product laspartomycin C. While analogues 5-9 bear a N-terminus acylated with commercially available decanoic acid, laspartomycin C contains (E)-13-methyltetradec-2-enoic acid 16, which was obtained in 6 steps from commercially available undec-10-enoate 10 (scheme 4). Conversion of methyl ester 10 to alcohol 11 was achieved using a Grignard reaction followed by treatment with BF3.Et2O to obtain intermediate 12.[19] Oxidation of the double bond with NaBH4 and H2O2 yielded alcohol 13.[19] Treatment with PCC provided aldehyde 7 which was used in the subsequent Horner–Wadsworth–Emmons reaction yielding exclusively E-alkene 15.[20] Finally, saponification gave (E)-13-methyltetradec-2-enoic acid 16 in 53% yield after silica gel chromatography suitable for SPPS.[21]

The SPPS part of the laspartomycin C synthesis occurred without incident yielding 10.3 mg (6.6%) of pure material (scheme 4). In order to compare our synthetic material with natural laspatomycin C, we obtained its reported producing strain S. viridochromogenes ssp. komabensis (ATCC-29814). The producing strain was cultured over the course of six days applying various growth media as described in published procedures dating back to the original work of Naganawa et al. from 1968.[1,2,22] However, while the S. viridochromogenes species generally showed good growth, no production of laspartomycin C could be detected using either LC-MS or bio-assay. Alternative strategies to stimulate the production of the secondary metabolite including addition of N-acetyl glucosamine or exogenous bacteria to the media were unsuccessful. Consequently, we preformed a full NMR analysis of our synthetic laspartomycin C in order to demonstrate its equality to the fermentation product.

NMR analysis including sequential assignment of the amino acid residues showed that the synthetic peptide was identical to natural laspartomycin C, confirming the

63


previously assigned chemical structure (Experimental table 4).[1,4] Characteristic 13C chemical shifts of Pro at positions β (29.1 ppm) and γ (24.2 ppm) confirm a trans orientation as was also reported for natural laspartomycin C.[23,24] Similarly, 1H and 13C chemical shifts at d-Pip positions α (1H 4.80 ppm, 13C 55.8 ppm) and ε (1H 2.88/4.36 ppm, 13C 39.5 ppm), along with strong NOESY correlations between the α proton of d-pip and Dap (1H 4.66 ppm), demonstrate the cis conformation of the d-Pip residue.[23]

2.3 Biological activity The antibacterial activities of laspartomycin C, daptomycin, and the chimeric analogues were determined against Staphylococcus aureus 29213 and Staphylococcus simulans 22 (Table 1). The MICs were measured at various calcium

O

Oa) b)

7 7

HO

7 7

c)OH

7O

e)

7O

O

7OH

Of)

O

HN

O

NH O

NNH

O

HN

O

N

O NHO

NNHAlloc

O

OtBu

O

OtBu

O

OtBu

O

9Dmb

Dmb

O

O

FmocN

DmbO

Og)

O

HN

O

NH O

NNH

O

HN

O

N

O NHO

N

NH2

NH

O

ON

NH

OH

O

OtBu

O

OtBu

O

OtBu

O

O

9Dmb

Dmb

O

Oh)i)

1

d)

10 11 12 13

14 15 16

Scheme 4. Total synthesis of laspartomycin C. Top: synthesis of the C15-lipid (a) (i) MeMgBr, THF, -84 °C; (ii) 3M HCl, 97%; (b) BF3.Et2O, Et3HSi, CH2Cl2, 91%; (c) (i) I2, NaBH4, THF; (ii) H2O2, 3M NaOH, 39%; (d) PCC, CH2Cl2, 85%; (e) triethylphosphonoacetate, NaH, THF, 76%; (f) 1M NaOH, tBuOH, 60 °C, 53%. Bottom: peptide synthesis strategy (g) Fmoc SPPS; (h) (i) Pd[(C6H5)3P]4, C6H5SiH3, CH2Cl2, 1 h; (ii) Fmoc SPPS; (i) (i) (CF3)2CHOH, CH2Cl2, 1 h; (ii) BOP, DIPEA, CH2Cl2, 16 h; (iii) TFA, TIS, H2O, 1 h (Fmoc-d-allo-Thr was employed without side chain protection and incorporated without incident).

64


concentrations to investigate the influence this has on antibiotic potency. The antibacterial activity of laspartomycin C and daptomycin increases significantly at higher calcium concentrations, an effect that was also reported by Taylor and co-workers who observed that elevated calcium concentrations were required to achieve full potency with their daptomycin analogues. The influence of introducing conformationally restricted amino acids in the macrocycle was investigated with analogues 5-9. In general, incorporation of d-Pip and Pro residues was detrimental, resulting in large increases in MIC or complete loss of antibacterial activity. Of particular note is the observation that compound 9, a daptomycin−laspartomycin C hybrid, is completely inactive. The decreased activity of these analogues does not, however, appear to be due to an inability to bind to calcium. The circular dichroism spectra obtained (experimental Fig. 7) indicate that all analogues undergo conformational changes upon mixing with calcium, as was previously observed for daptomycin and the amphomycin-derived MX-2401. Particularly intriguing are d-Pip and Pro bearing analogues 8 and 9, which, although devoid of activity, exhibit calcium-induced conformational changes very similar to that of laspartomycin C. These findings prompted us to further investigate the antibacterial mechanism of laspartomycin C.

2.4 Mode of action studiesWe began by examining the effect of laspartomycin C on bacterial cell wall biosynthesis by specifically looking for accumulation of the cytoplasmic lipid II precursor UDP-MurNAc-pentapeptide in response to administration of the antibiotic.[13] As is clearly seen in figure 3, when S. aureus cells are treated with laspartomycin C there is a significant accumulation of UDP-MurNAc-pentapeptide, an effect not seen with daptomycin. These findings indicate that the target of laspartomycin C lies downstream of UDP-MurNAc-pentapeptide, implicating

Table 1. Minimum inhibitory concentrations (μg mL−1) measured against indicator strains S. simulans 22 and S. aureus 29213 in the absence and presence of various CaCl2 concentrations (1.25-10.0 mM).

S. aureus 29213 S. simulans 22

compound no Ca2+ 1.25 mM 5.0 mM 10 mM no Ca2+ 1.25 mM 5.0 mM 10 mM

2 (daptomycin) >256 0.5 0.25 0.125 32-64 0.063 0.031 0.016-0.031

1 (laspartomcyin C) >256 8 4 2 >128 8 4 ≤1

3 >256 256 16 4 >128 32 4 1

4 >256 >256 64 64 >128 64 8 4

5 >256 >256 128 32 >128 128 16 8

6 >256 256 32 8 >128 64 4 4

7 >256 >256 >256 >256 >128 >128 >128 >128

8 >256 >256 >256 >256 >128 >128 >128 >128

9 >256 >256 >256 >256 >128 >128 >128 >128

65


one of the subsequent membrane-associated steps involved in bacterial cell wall biosynthesis.

We then investigated whether the antimicrobial activities of laspartomycin C, daptomycin, and the active synthetic analogues were antagonized by various bacterial cell wall precursors including lipid I, lipid II, UDP-MurNAc-pentapeptide, UDP-GlcNAc, undecaprenyl pyrophosphate (C55-PP), and undecaprenyl phosphate (C55-P). The antibiotics including vancomycin as a positive control were pre-incubated with the various bacterial cell wall precursors and administered to S. simulans 22, which was used as an indicator strain (Table 2). As expected, the activity of vancomycin was fully antagonized by lipid I, lipid II, and UDP-MurNAc-pentapeptide, each of which contains the d-Ala-d-Ala motif recognized by vancomycin. In contrast, daptomycin and analogues 5−9 were not antagonized by any of the bacterial cell wall precursors. For laspartomycin C, however, both C55-P and C55-PP gave an indication of antagonism. Upon repeating the assay with the water-soluble C15-P and C15-PP, it became clear that only the monophosphate species is capable of antagonizing the activity of laspartomycin C.

Figure 3. HPLC traces for the UDP-MurNAc-pentapeptide accumulation assay. Treatment of S. aureus 29213 with laspartomycin C results in accumulation of UDP-MurNAc-pentapeptide, an effect not observed with daptomycin

Table 2. Inhibition of antibiotic activity by selected antagonists.

Compound Lipid I Lipid II MurNAca GlcNAcb C55-P C55-PP C15-P C15-PP

1 (lasp C) - - - - ± ± + -

5-9 - - - - - - - -

2 (daptomycin) - - - - - - - -

Vancomycin + + + - - - - -

(-) Antibiotic activity unaffected; (+) Antibiotic activity antagonized; (±) Antagonization unclear. aUDP-MurNAc-pp, bUDP-GlcNAc. Assays performed by S. Oppedijk.

66


Next, a TLC-based binding assay was employed to assess the binding of laspartomycin C to C55-P. Mixing of C55-P with laspartomycin C in Ca2+-containing buffer led to formation of a surprisingly stable complex that could be clearly visualized by TLC (Fig. 4). By comparison, there was no indication of C55-P binding by daptomycin. When the same TLC experiment was performed with lipid II, no complex formation was observed further indicating that laspartomycin C selectively targets C55-P. We next investigated whether laspartomycin C is capable of inhibiting lipid II synthesis by means of an in vitro assay. To do so, C55-P was pretreated with laspartomycin C at a variety of concentrations, followed by addition of UDP-GlcNAc, UDP-MurNAc-pentapeptide, and M. flavus membrane vesicles known to contain the known to contain the lipid II-producing enzymes MraY and MurG. Under these conditions laspartomycin C blocked lipid II synthesis in a dose-dependent manner while incubation with daptomycin had no effect lipid II synthesis (Experimental fig. 8). These findings support a mode of action for laspartomycin C wherein the sequestration of C55-P leads to blocked formation of lipid II and prevention of bacterial cell wall biosynthesis.

To obtain a more quantitative understanding of the binding of C55-P by laspartomycin C, we turned to isothermal titration calorimetry (ITC). ITC was used to study the interaction of laspartomycin C with large unilamellar vesicles (LUVs) comprising 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1.5 mol % C55-P, a system mimicking a membrane environment. Titration of the C55-P containing LUVs into a solution of laspartomycin C results in an isotherm that appears to be the combination of two distinct binding events (Fig. 5). The initial stage of the isotherm (left side) shows an endothermic event resulting from the interaction of laspartomycin C with the DOPC membrane vesicles. Similarly, titration of “empty” DOPC vesicles, not containing C55-P, into a laspartomycin C solution resulted in an isotherm displaying an identical interaction (Experimental fig. 9) However, as the quantity of C55-P injected increases, a second binding event is

Figure 4. TLC analysis of C55-P incubation with 0.5−2 eq. of laspartomycin C and daptomycin. Laspartomycin C forms a stable complex with C55-P (red box) contrary to daptomycin. The brightness and contrast of the figure have been adjusted to enhance visibility of the laspartomycin C reference band (the original unadjusted figure is included in the publication). S. Oppedijk performed the experiment.

Dap 0.5 eq. 1.0 eq. 2.0 eq. Lasp 0.5 eq. 1.0 eq. 2.0 eq. C55-P

67


apparent (Fig. 5, enlarged). This is ascribed to the binding of C55-P by laspartomycin C and indicates a remarkably high affinity interaction with a dissociation constant in the low nanomolar range (Kd = 7.3 ± 3.8 nM). The ITC approach employed further reveals that enthalpic (ΔH = −9.8 ± 0.6 kJ mol-1) and entropic (ΔS = 122.9 ± 4.7 J mol-1 K-1) factors contribute to the tight binding of C55-P by laspartomycin C.

3 Conclusion

In summary, we here describe the total synthesis of laspartomycin C by means of a flexible route that also allows for the preparation of structural analogues. While the synthesis of daptomycin, the preeminent depsipeptide CDA, has been previously described[14-16], our synthesis of laspartomycin C represents the first of its kind among the macrolactam subfamily of CDAs. Hybrid structures combining aspects of laspartomycin C and daptomycin were also prepared and evaluated for antibacterial activity. In all cases these variants were less active than either parent compound, suggesting a significant difference in the modes of action of laspartomycin C and daptomycin. Following up on these findings we established that unlike daptomycin, laspartomycin C exerts its antibiotic effect by tightly complexing C55-P. In further assessing this interaction, we also report the first ITC-based characterization of a C55-P-targeting CDA and determined the thermodynamic parameters governing the binding of laspartomycin C to C55-P. Of particular note is the low nanomolar Kd value associated with laspartomycin C’s binding to its structurally simple phospholipid target.

At present, daptomycin is the only clinically approved CDA and our findings show it to be a generally more potent antibiotic than laspartomycin C and the structural analogue here investigated. That said, laspartomycin C’s ability to kill a range of

0.0 0.5 1.0

-10

-5

0

5

10

molar ratio

kJ m

ol-1

of i

njec

tant

0.50 0.75 1.00 1.25

-10

-5

0

molar ratio

kJ m

ol-1

of i

njec

tant

Figure 5. Isothermal titration calorimetry. DOPC vesicles containing 1.5 mol % C55-P were titrated into a solution of laspartomycin C in HEPES buffer containing CaCl2. Fitting of the second binding curve (red) provides a Kd of 7.3 ± 3.8 nM. Each point is the average of three independent experiments.

68


Gram-positive pathogens[2,3] via a mode of action different from that of daptomycin indicates that it may have potential for development. At present, no clinically used antibiotic acts via a C55-P targeting mode of action. In this regard our synthetic route provides the means for future structure−activity relationship studies with this interesting class of CDAs. While the therapeutic potential of C55-P targeting peptides requires further validation, such compounds may be of value in addressing the growing threat posed by antibiotic-resistant bacteria.

69



Preparation of Laspartomycin C and Analogues. 2-Chlorotrityl resin (5.0 g, 1.60 mmol g-1) was loaded with Fmoc(Dmb)-Gly-OH. Resin loading was determined after coupling of the second amino acid because complete Fmoc deprotection of resin bound Fmoc(Dmb)-Gly required nonstandard conditions: Fmoc(Dmb)-Gly 2-chlorotrityl resin (6.0 g) was thus treated with ethanolamine:DMF while shaking vigorously (1:4 v:v, 1 × 30 min, 1 × 90 min) followed by washing with DMF. Overnight coupling of Fmoc-Asp(tBu)-OH (3.7 g, 9.0 mmol), BOP (4.0 g, 9.0 mmol), and DiPEA (3.1 mL, 18.0 mmol) in DMF followed by end-capping with Ac2O:DiPEA:DMF (0.5:0.5:9 v:v:v, 20 mL) yielded Fmoc-Asp(tBu)-(Dmb)-Gly 2-chlorotrityl resin (0.52 mmol g-1 as determined spectrophotometrically).

Linear precursor peptides encompassing Gly8 to Asp1 were assembled via standard Fmoc solid-phase peptide synthesis (SPPS) via manual synthesis (resin bound AA:Fmoc-AA:BOP:DiPEA, 1:4:4:8 molar equiv.) or automated synthesis (resin bound AA:Fmoc-AA:HBTU:HOBt:DiPEA, 1:4:3.75:3.75:8 molar equiv.) typically on 0.25 mmol scale. NMP or DMF was used as solvent, and Fmoc deprotections were carried out with piperidine:DMF or piperidine:NMP (1:4 v:v). Amino acid side chains were protected as follows: Boc for Orn and Trp, Trt for D-Asn, Alloc for DAP, tBu for Asp, Glu, and d-Ser, Dmb for Gly in Asp-Gly sequences. Kyn and d-allo-Thr were introduced without side chain protection. Following coupling and Fmoc deprotection of Asp1, N-terminal acylation was achieved by coupling (E)-13-methyltetradec-2-enoic acid using the same coupling conditions used for the SPPS.

The resin-bound, Alloc protected intermediate was next washed with CH2Cl2 and treated with Pd(PPh3)4 (74 mg, 0.06 mmol) and PhSiH3 (0.74 mL, 6.0 mmol) in CH2Cl2 (ca. 10 mL) under argon for 1 h. The resin was subsequently washed with CH2Cl2 (5 × 10 mL), followed by a solution of diethyldithiocarbamic acid trihydrate sodium salt (5 mg mL-1 in DMF, 5×10 mL), and DMF (5×10 mL). The remaining three amino acids where added via standard Fmoc SPPS with removal of the final Fmoc protecting group to yield the complete linear resin-bound peptide with a free N-terminal amine. The resin was treated with (CF3)2CHOH:CH2Cl2 (1:4, 10 mL) for 1 h and rinsed with additional (CF3)2CHOH:CH2Cl2 and CH2Cl2. The combined washings were then evaporated to yield the linear protected peptide with free C- and N-termini. The residue was dissolved in CH2Cl2 (250 mL) and treated with BOP (0.22 g, 0.5 mmol) and DiPEA (0.17 mL, 1.0 mmol), and the solution was stirred overnight after which TLC indicated complete cyclization. The reaction mixture was

70


concentrated and directly treated with TFA:TIS:H2O (95:2.5:2.5, 10 mL) for 60−90 min. The reaction mixture was added to Et2O:hexanes (1:1), and the resulting precipitate washed once more with

Et2O:hexanes (1:1). The crude cyclic peptide was lyophilized from tBuOH:H2O (1:1) and purified with reverse phase HPLC by applying a gradient of 25−65% buffer B (buffer A, H2O:MeCN:TFA, 95:5:0.1 v:v:v; buffer B, H2O:MeCN:TFA, 5:95:0.1 v:v:v) over 1 h with a flow rate of 12 mL min-1 on a C18 Maisch 250 mm x 22 mm column. Pure fractions were pooled and lyophilized to yield the desired cyclic lipopeptide products in >95% purity (based on analytical HPLC analysis) as white powders, typically in 10−20 mg quantity (4.2−9.3% yield based on resin loading).

(E)-13-methyltetradec-2-enoic acid. Steps a-c and steps d-e in scheme 4 were carried out according to literature procedures.[19,20] Step f was carried out in accordance with the published synthesis of the same compound from a

similar ester precursor.[21] 1H NMR (400 MHz, CDCl3): δ 12.03 (br, 1H), 7.07 (td, J1 = 15.6 Hz, J2 = 7.2 Hz, 1H), 5.81 (d, J = 15.6 Hz, 1H), 2.21 (q, J = 7.2 Hz, 2H), 1.56-1.42 (m, 3H), 1.25 (m, 12H), 1.13 (m, 2H), 0.85 (d, J = 6.4 Hz, 6H). 13C NMR (CDCl3): δ 172.5, 152.4, 120.9, 39.2, 32.4, 30.1, 29.8, 29.7, 29.5, 29.3, 28.1, 28.0, 27.6, 22.8; HR-MS [M-H+]: Calc. 239.2017, found 239.2023.

MIC determinations Minimum inhibitory concentrations (MICs) were determined by broth microdilution according to CLSI guidelines. Blood agar plates were inoculated with glycerol stocks of S. aureus 29213 and S. simulans 22 followed by incubation for 16 hours at 37 °C and 30 °C respectively. Cation adjusted Mueller-Hinton broth (MHB) containing 10 mg L-1 Mg2+ was inoculated with individual colonies of S. aureus and S. simulans, and incubated for 16 hours at 220 RPM. The peptides were dissolved in MHB (10 mg L-1 Mg2+) and serially diluted on polypropylene microtiter plates with a volume of 50 μL per well. Inoculated MHB (2x105 CFU mL-1) containing 10 mg L-1 Mg2+ and varying concentrations of Ca2+ was added to reach a total volume of 100 μL per well. The microtiter plates were sealed with an adhesive membrane and after 16 hours of incubation at 37 °C or 30 °C and 220 RPM the wells were visually inspected for bacterial growth. All reported MIC values result from two or more measurements.

7OH

O

16

71


Circular Dichroism CD spectra were recorded for 60 μM peptide solutions in 20 mM HEPES (pH=7.4) buffer with and without 5.0 mM CaCl2 (Experimental fig. 7). Data collection was limited to the 210-260 nm range because extensive scattering occurs below 210 nm under these conditions. The experiments were performed using a 1.0 mm cuvet, a bandwith of 1 nm and a scan speed of 20 nm min-1. The average of 10 scans was baseline corrected by subtracting the average elipticity over 255-260 nm and units were converted to mean residue molar elipticities (deg cm2 dmol-1).

Figure 7. Circular dichroism: The peptides in the absence (red) and presence (dashed-blue) of 5.0 mM CaCl2.

UDP-MurNAc-pentapeptide accumulation assayS. aureus 29213 was grown until OD600 = 0.5 in TSB supplemented with CaCl2 (5.0 mM). Chloramphenicol (130 μg mL-1) was added and after incubation for 15 minutes at 37 °C, the culture was divided in 5 mL aliquots. Antibiotics were added at 10xMIC (laspartomycin C, daptomycin) and one aliquot remained untreated. After 60 minutes, cells were separated from the medium and extracted with boiling d-H2O (1 mL) for 15 minutes. The suspensions were spun down and the supernatant was lyophilized. The resulting material was analyzed by HPLC applying a gradient from 100% eluent A (50 mM NaHCO3:5 mM Et3N, pH = 8.3) to 75% eluent A over 15 minutes using a C18 column (eluent B: MeOH). Formation of UDP-MurNAc-pentapeptide was confirmed by comparison with authentic material by HPLC, and LC-MS analysis applying the same gradient with an adjusted eluent A (50 mM NH4HCO3:5 mM Et3N, pH = 8.3).

-20

-10

0

10

20

30

40

50

210 220 230 240 250 260[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Laspartomycin C (1)

-40

-30

-20

-10

0

10

20

30

40

50

210 220 230 240 250 260

[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Daptomycin (2)

-50

-40

-30

-20

-10

0

10

20

30

40

210 220 230 240 250 260

[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Compound 3

-30

-10

10

30

50

70

90

210 220 230 240 250 260[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Compound 5

-5

0

5

10

15

20

25

30

210 220 230 240 250 260

[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Compound 6

-5

0

5

10

15

20

210 220 230 240 250 260

[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Compound 7

-10

-5

0

5

10

15

20

25

30

35

40

210 220 230 240 250 260

[]

10-3

(deg

cm

2 dm

ol-1

)

Wavelength (nm)

Compound 9

-10

-5

0

5

10

15

20

25

30

35

40

210 220 230 240 250 260

[]

10-3

(deg

cm

2 dm

ol-1

)Wavelength (nm)

Compound 9

72


Antagonization assay Appropriate amounts of antagonists lipid I, lipid II, UDP-MurNAc-pentapeptide, UDP-GlcNAc, C55-P, C55-PP, C15-P and C15-PP in CHCl3:MeOH (1:1) or MeOH were evaporated and redissolved in MHB or MQ-H2O. Antagonists were added to the peptides in MHB to achieve a 5-fold excess of antagonist relative to the concentration of peptide antibiotic. These solutions were then added to bacterial cultures of S. simulans 22 in MHB (107 CFU mL-1) to reach a final concentration of peptide antibiotic corresponding to 8 times the MIC. Each experiment was executed in triplicate and after incubation for 16 hours at 30 °C and 220 RPM bacterial growth was inspected visually. S. Oppedijk performed this experiment.

C55-P and Lipid II binding experiment C55-P (5 nmol) or lipid II (5 nmol) were incubated for 2 hours with laspartomycin C or daptomycin (0.5-2 equivalents) in 75 μL 100 mM Tris-HCl (pH=7.5) containing 0.1% TX100, MgCl2 (13 mM) and CaCl2 (13 mM). Treatment with PyOAc buffer (pH=4.2) was followed by BuOH extraction. The BuOH phase was evaporated, the residue was dissolved in CHCl3:MeOH (1:1) and subsequently analyzed by TLC with iodine visualization (CHCl3:MeOH:H2O:NH4OH, 88:48:10:1). S. Oppedijk performed this experiment.

Lipid II synthesis inhibition assay C55-P (5 nmol) was mixed with 0.5, 1 or 2 equivalents of peptide in 100 mM Tris-HCl (pH=7.5) containing 0.1% TX100, MgCl2 (13.3 mM) and CaCl2 (13.3 mM). An excess of UDP-GlcNAc and UDP-MurNAc-pentapeptide was added followed by M. flavus membrane vesicles in Tris-HCl to reach a total volume of 75 μL. Quenching with 6 M PyOAc buffer (pH=4.2) after 2 hours was followed by extraction with BuOH. The BuOH phase was evaporated, the residue dissolved in

Figure 8. TLC analysis of the lipid II synthesis assay. The figure shows the result of supplementing the lipid II synthesis reaction with nisin, daptomycin and laspartomycin C in 0.5-2 eq. relative to C55-P. Two reference bands are included: pure lipid II (lipid II) and the lipid II synthesis reaction mixture in the absence of antibiotics (“Control”). The presence of laspartomycin C blocks the formation of lipid II in a dose dependent manner (see red box) while daptomycin has no effect on lipid II production. The lantibiotic nisin is known to inhibit lipid II formation and was therefore used as positive control.

73


CHCl3:MeOH (1:1) and subsequently analyzed by TLC with iodine visualization (CHCl3:MeOH:H2O:NH4OH, 88:48:10:1) (Experimental fig. 8). S. Oppedijk performed this experiment.

ITC experiments DOPC (10 mM) and DOPC:C55-P (10 mM:0.15 mM) vesicles were obtained after appropriate volumes of lipid stock solutions in CHCl3:MeOH (1:1) were evaporated under vacuum and suspended in 20 mM HEPES buffer containing 5.0 mM CaCl2. The vesicles were extruded using filters with a 0.2 μm pore size and degassed by stirring for 20 minutes under reduced pressure. A laspartomycin C solution (25 μM) in the same buffer was degassed using ultrasound. Vesicles were titrated into the laspatomycin C solution in triplicate at 25 °C with a reference power of 2.0 μcal sec-1 over 26 titrations of 1.5 μL with exception of the first titration (0.5 μL) (Experimental fig. 9). The observed titration peaks were integrated using NITPIC software. Fitting of the C55-P:laspartomycin C binding curves using SEDPHAT software excluding the first 13 data points provided the thermodynamic parameters. The reported parameters are averages of three independent experiments and errors were estimated by Monte Carlo simulation using standard deviations of the individual experiments. The final 4 data points of the isotherms were zeroed to 0 kJ mol-1 and GraphPad Prism was used to generate figure 5.

Figure 9. Representative ITC thermograms resulting from titration of DOPC vesicles in HEPES buffer into a solution of laspartomycin C in HEPES buffer (left) and titration of C55-P containing DOPC vesicles into a laspartomycin C solution (right).

0.0 0.5 1.0

-8.4

0.0

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0 10 20 30 40 50 60 70

Time (min)

��

Molar Ratio

��

��

��

0 20 40 60 80

0.00

0.05

0.10

0.15

0.20

-0.05

0.00

0.05

0.10

0.15

0 10 20 30 40 50 60 70

��

��

��

Δ � �� Δ� ��

��

��

��

��

74



Table 3. Analytical data for compounds 1 (laspartomycin C) and analogues 5-9.

compound yield: mg (µmol, %) Rt (min) [M + H]+ calc. [M + H]+ found

1 (laspartomycin C) 10.3 (8.3, 6.6) 38.2 1247,6518 1247,6507

5 10.5 (6.4, 7.7) 29.1 1641,7544 1641,7580

6 19.1 (11.6, 9.3) 28.6 1645,7493 1645,7442

7 15.7 (10.1, 8.1) 28.3 1552,7278 1552,7290

8 9.7 (6.5, 4.2) 27.9 1498,6809 1498,6820

9 11.6 (7.8, 6.3) 31.6 1479,7114 1479,7117

Table 4. NMR chemical shift assignment compound 1 (laspartomycin C) (5 mM in DMSOd6).

Residue NH Hα(Cα) Sidechain

Tail - 5.93 (123.7) CβH (6.63, 142.9), CγH2 (2.12, 31.0), CδH2 (1.38, 27.5),

CεH2 (1.26, 28.5), CζH2-CιH2, (1.22-1.28, 28.2-29.3),

CκH2 (1.23, 26.5), CλH2 (1.12, 38.2), CμH (1.38, 27.5),

2CνH3 (0.84, 22.2)

Asp-1 8.13 4.61 (49.1) CβH2 (2.50/2.63, 35.9)

Dap-2 8.24 4.66 (48.2) CβH2 (3.10/3.57, 39.5), NγH (7.50)

d-Pip-3 - 4.80 (55.8) CβH2 (1.54/2.18, 26.4), CγH2 (1.40/1.51, 20.1), CδH2

(1.22/1.51, 24.2), CεH2 (2.88/4.36, 39.5)

Gly-4 8.08 3.65/4.00 (41.5) -

Asp-5 8.25 4.59 (49.2) CβH2 (2.53/2.74, 35.8)

Gly-6 8.13 3.76 (41.8) -

Asp-7 8.33 4.50 (49.7) CβH2 (2.56/2.71, 35.6)

Gly-8 7.88 3.68/3.74 (41.7) -

d-allo-Thr-9 7.88 4.29 (58.1) CβH (3.82, 66.4), CγH3 (1.03, 19.3)

Ile-10 7.73 4.31 (54.0) CβH (1.73, 35.7), CγH2 (1.07/1.50, 24.1), CγH3

(0.87, 14.5), CδH3 (0.78, 10.4)

Pro-11 - 4.19 (59.3) CβH2 (1.72/2.02, 29.1), CγH2 (1.81/1.92, 24.2), CδH2

(3.52/3.77, 46.9)

75


References[1] D. B. Borders, R. A. Leese, H. Jarolmen, N. D. Francis, A. A. Fantini, T. Falla, J. C. Fiddes, A. Aumelas, J. Nat. Prod. 2007, 70, 443–446.[2] H. Naganawa, M. Hamada, K. Maeda, Y. Okami, T. Takeuchi, H. Umezawa, J. Antibiot. 1968, 21, 55–62.[3] W. V. Curran, R. A. Leese, H. Jarolmen, D. B. Borders, D. Dugourd, Y. Chen, D. R. Cameron, J. Nat. Prod. 2007, 70, 447–450.[4] F. Kong, G. T. Carter, J. Antibiot. 2003, 56, 557–564.[5] H. S. Sader, R. K. Flamm, R. N. Jones, Diagn. Microbiol. Infect. Dis. 2013, 75, 417–422.[6] N. Yin, J. Li, Y. He, P. Herradura, A. Pearson, M. F. Mesleh, C. T. Mascio, K. Howland, J. Steenbergen, G. M. Thorne, et al., J. Med. Chem. 2015, 58, 5137–5142.[7] M. Strieker, M. A. Marahiel, ChemBioChem 2009, 10, 607–616.[8] L. Robbel, M. A. Marahiel, J. Biol. Chem. 2010, 285, 27501–27508.[9] D. Jung, A. Rozek, M. Okon, R. E. W. Hancock, Chem. Biol. 2004, 11, 949–957.[10] J. Pogliano, N. Pogliano, J. A. Silverman, J. Bacteriol. 2012, 194, 4494–4504.[11] A. Müller, M. Wenzel, H. Strahl, F. Grein, T. N. V. Saaki, B. Kohl, T. Siersma, J. E. Bandow, H.-G. Sahl, T. Schneider, et al., Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 7077–7086.[12] E. Rubinchik, T. Schneider, M. Elliott, W. R. P. Scott, J. Pan, C. Anklin, H. Yang, D. Dugourd, A. Müller, K. Gries, et al., Antimicrob. Agents Chemother. 2011, 55, 2743–2754.[13] T. Schneider, K. Gries, M. Josten, I. Wiedemann, S. Pelzer, H. Labischinski, H. G. Sahl, Antimicrob. Agents Chemother. 2009, 53, 1610–1618.[14] D. C. Alexander, P. Brian, G. M.-F. Coeffet-Le, X. He, V. Kulkarni, C. Leitheiser, I. B. Parr, D. Ritz, R. H. Baltz, C. P. Inc, et al., 2007.[15] C. R. Lohani, R. Taylor, M. Palmer, S. D. Taylor, Org. Lett. 2015, 17, 748–751.[16] H. Y. Lam, Y. Zhang, H. Liu, J. Xu, C. T. T. Wong, C. Xu, X. Li, J. Am. Chem. Soc. 2013, 135, 6272–6279.[17] P. 't Hart, L. H. J. Kleijn, G. de Bruin, S. F. Oppedijk, J. Kemmink, N. I. Martin, Org. Biomol. Chem. 2014, 12, 913–918.[18] C. Rubini, A. Osler, A. Calderan, A. Guiotto, P. Ruzza, J. Pept. Sci 2008, 14, 989–997.[19] M. B. Richardson, S. J. Williams, Beilstein J. Org. Chem. 2013, 9, 1807–1812.[20] T. Shioiri, N. Irako, Tetrahedron 2000, 56, 9129–9142.[21] A. G. Myers, D. Y. Gin, D. H. Rogers, J. Am. Chem. Soc 1994, 116, 4697–4718.[22] Y. Wang, Y. Chen, Q. Shen, X. Yin, Gene 2011, 483, 11–21.[23] F. Kong, K. Janota, J. S. Ashcroft, G. T. Carter, Records of Natural Products 2010.[24] H. Kessler, Angew. Chem. Int. Ed. 1982, 21, 512–523.

76

Chapter 3

77

Target recognition by the calcium-dependent lipopeptide antibiotic laspartomycin C

Chapter 4

Target recognition by the calcium-dependent lipopeptide antibiotic laspartomycin C

ABSTRACT: The calcium-dependent antibiotics (CDAs) are an important emerging class of antibacterial drugs. We here report the crystal structure of a CDA, laspartomycin C, in complex with calcium and the ligand geranyl-phosphate, at a resolution of 1.28 Å. The structure is the first to be reported for an antibiotic that binds the essential bacterial phospholipid undecaprenyl phosphate and provides insight for the design of antibiotics capable of exploiting this unique bacterial target.

L. H. J. Kleijn*, H. C. Vlieg*, T. M. Wood, J. Sastre Toraño, B. J. C. Janssen†, N. I. Martin†, (2017) Angew. Chem. doi: 10.1002/ange.201709240 * equal contribution, † shared corresponding authorship

78

Target recognition by the calcium-dependent lipopeptide antibiotic laspartomycin CChapter 4

1 Introduction

Key to addressing the growing threat of drug-resistant bacteria is the identification and characterization of antibiotics that operate by unexploited mechanisms.[1,2] Notable in this regard are the calcium-dependent antibiotics (CDAs), which have gained clinical prominence due to their activity against multi-drug resistant pathogens.[3-5] While a variety of antibacterial mechanisms have been ascribed to the various CDAs, an atomic-level understanding of the recognition of their bacterial target(s) remains elusive. At present, daptomycin is the only clinically used CDA and its mode of action is the topic of ongoing investigation.[6] By comparison, the structurally similar CDAs laspartomycin C, friulimycin B, tsushimycin, and amphomycin have more clearly understood antibacterial mechanisms.[7-9] It was recently reported that laspartomycin C forms a high-affinity complex (Kd = 7.3 ± 3.8 nM) with the bacterial cell wall precursor undecaprenyl phosphate (C55-P) and in doing so inhibits peptidoglycan biosynthesis, ultimately leading to cell death (Chapter 3).[9] Notably, the sequestration of C55-P is a mechanism not exploited by any current clinically used antibiotic making it an attractive target for further study. To date, the only structural insights available for the C55-P binding CDAs are provided by the unliganded structure of tsushimycin.[10] To achieve a deeper understanding of the high-affinity C55-P binding and specificity, we here report the crystal structure of laspartomycin C in complex with calcium and geranyl-phosphate (C10-P, a soluble C55-P analogue) at a resolution of 1.28 Å. The target specificity of the laspartomycin C was also studied by examining its interaction with other common phospholipids. Furthermore, the stereochemical implications of laspartomcyin C target binding revealed by the crystal structure were probed by synthesizing and testing the enantiomeric form of the antibiotic. Not only is the structure here reported the first for a CDA in complex with both calcium and its biomolecular target, it is also the first for any antibiotic that targets C55-P.


2.1 Target-bound laspartomycin C crystal structureWhile laspartomycin C can be isolated from fermentation of S. viridochromogenes it was found to be more convenient to prepare the compound by synthetic means as previously described (Chapter 3)[9] Crystals of the laspartomycin C/Ca2+/geranyl phosphate (C10-P) complex were grown and diffracted to 1.28 Å resolution (Experimental table 3). The refined structure reveals a stoichiometry of the complex of a single laspartomycin C molecule bound to one geranyl-phosphate ligand along

79


with two calcium ions playing key roles in mediating recognition (Fig. 1).

Laspartomycin C adopts a C-shaped amphipathic fold with polar residues (Asp5, Asp7, d-allo-Thr9) aligning the top plain and aliphatic residues (d-Pip3 and Pro11) forming the hydrophobic bottom plane from which the fatty acid side chain protrudes downwards (Fig. 1, bottom). The cavity created within the laspartomycin C macrocycle envelops the phosphate head group of the C10-P, which is held in place by hydrogen bonding to the peptide backbone and chelation of both calcium ions (Fig. 1). The C10-P isoprenyl tail appears to be stabilized by hydrophobic interactions with the laspartomycin C fatty acid side chain as it exits the binding pocket. The complex is well resolved except for the outermost parts of the C10-P isoprenyl tail and the laspartomycin C fatty acid side chain. In the crystal lattice, the C10-P isoprenyl units and fatty acid side chains of multiple complexes cluster together to form continuous hydrophobic sheets. This may also hint at the biologically relevant orientation that laspartomycin C adopts on the bacterial cell surface in its sequestration of C55-P at the membrane interface.

Figure 1. Laspartomycin C forms a 1:1:2 complex with C10-P and Ca2+. (Top, left) The structure of laspartomycin C and C10-P. (Top, right) Structure of the ternary complex with laspartomycin C (green stick representation), two bound Ca2+ ions (orange spheres) and bound water molecules (red spheres) and the C10-P ligand (dark green ball and stick representation). (Bottom) Surface distribution of polar regions of the complex. The “top” and the sides of the complex are hydrophilic whereas the bottom face is much more hydrophobic. The C10-P ligand is bound within the cavity of the laspartomycin C macrocycle (green).

90° 180°

Side TopBottom

Ternary complex

HNO

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

OH

O

OH

O

OH

O

OH

O

O2

3

4 5 6

7

8

91011

D

D

1

O9

OPOH

O OH

Laspartomycin C C10-P

Thr9

Pro11

Dap2

Asp7

Gly6

Gly8

Ca

Ca

C10-P

Fatty acid

Ile10

Asp5

Asp1

Gly4

D-Pip3

80


Asp1

C10-P

H2O

Fatty acid

H2O

Asp7

Ca

C10-PDap2

Ile10

Asp5 Gly8

Gly6Ca

2.3

2.9

3.0

3.2

2.2O

OH2

OH2

O

OO

O

O

O

O

O

OO

O

PO O

O

NH

HN

NH

CaCa

2.2 Molecular recognitionThe structure of the complex provides an explanation for the high-affinity binding of C55-P by laspartomycin C (Fig. 2). Target recognition takes place via direct interactions of the phospholipid head group with the laspartomycin C backbone and both calcium ions. The backbone and side chain amides of Dap2 as well as the backbone amide of Gly8 form hydrogen bonds with three of the four phosphate oxygens (Experimental table 3). Of key importance are the two calcium ions that interact with the phosphate moiety as part of an octahedral coordination symmetry. The central Ca2+ is coordinated by 5 interactions with laspartomycin C involving four backbone carbonyls (Dap2, Gly6, Gly8, Ile10) and one aspartic acid side chain (Asp5) with all distances (2.3 Å) in agreement with expected Ca-O coordination distances (Fig. 2, Experimental table 4).[11] The peripheral Ca2+ is bound via interactions provided by two aspartic acid side chains (Asp1, Asp7) and the N-terminal fatty acid carbonyl group along with two H2O molecules that complete the octahedral coordination (Fig. 2).

Figure 2. Target recognition by laspartomycin C. (Top) The C10-P phosphate headgroup interacts via the coordinated Ca2+ ions and three laspartomycin C backbone amides. (Bottom, left) The central Ca2+ ion is coordinated by backbone carbonyl and side chain interactions as well as the C10-P head group. (Bottom, right) The peripheral Ca2+ ion is coordinated by two side chain interactions, the fatty acid carbonyl, and the C10-P head group with octahedral coordination completed by two water molecules.

81


O

HN

O

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

N

NH

OH

O

OH

O

OH

O

OH

O

O

O

HN

O

NH O

NNH

O

HN

O

N

O NHO

N

NH2

NH

O

O

N

NH

OtBu

O

OtBu

O

OtBu

O

OtBu

O

O

O

ODmb

Dmb

9

FmocN

DmbO

Oa)

3

O

HN

O

NH O

NNH

O

HN

O

N

O NHO

NNHAlloc

O

OtBu

O

OtBu

O

OtBu

O

O

ODmb

Dmb

9

b)

c) 9

DD1

2

3

4 5 6

7

8

91011D

D

D

D

ent-Laspartomycin C

Scheme 1. Synthesis of ent-laspartomycin C (a) Fmoc SPPS; (b) (i) Pd[(C6H5)3P]4, C6H5SiH3, CH2Cl2, 1 h; (ii) Fmoc SPPS; (c) (i) (CF3)2CHOH, CH2Cl2, 1 h; (ii) BOP, DIPEA, CH2Cl2, 16 h; (iii) TFA, TIS, H2O, 1 h.

The binding of C10-P by laspartomycin C as revealed by the crystal structure is not intrinsically dependent on a chiral interaction. To confirm this, we synthesized the enantiomeric form of laspartomycin C (Scheme 1). As expected, circular dichroism (CD) analysis of laspartomycin C and its enantiomer yielded identical ellipticities of opposite sign with clear effects observed in the absence and presence of Ca2+ (Fig. 3). Additionally, the CD spectra of both peptides indicate a second significant conformational change upon addition of 1.0 equivalent of C10-P. Addition of excess C10-P did not result in further elliptic changes, in line with a 1:1 peptide:C10-P binding stoichiometry (Experimental table 4). Finally, antibiotic testing of ent-laspartomycin C in parallel with laspartomycin C showed identical antibiotic minimum inhibitory concentrations (MICs) for both compounds against a panel of five Gram-positive pathogens (Experimental table 1).

82


The specificity of laspartomycin C binding to phosphate monoesters was evaluated using an antagonization assay wherein the antibiotic peptide was pre-mixed with various phospholipids and the effect on activity assessed (Fig. 4, experimental table 2). Mixing laspartomycin C with inorganic phosphate (HPO4

2-) at concentrations corresponding to normal serum levels resulted in no observable antagonization. As expected, treatment of laspartomycin C with 1.0 equivalent of C10-P led to loss of antibiotic activity while the activity of daptomycin was not inhibited by C10-P addition even at higher concentrations. We also assessed the antagonization potential of a C6 truncated variant of the aliphatic mammalian lipid phosphatidic acid (C6-PA). These studies revealed that C6-PA is also capable of blocking laspartomycin C’s antibiotic action with 1.0 molar equivalent eliciting complete antagonization. This effect appears to be specific for phosphate monoesters.

When the common lipid phosphodiester phosphatidyl glycerol (PG) was added to laspartomycin C in equimolar amount no antagonization was observed. In fact

Figure 4. Antagonization of antibiotic activity of laspartomycin C and daptomycin by inorganic phosphate and selected phosphoesters. Activity of the antibiotic, administered at 8xMIC, was antagonized (+) or unaffected (-) by the presence of ≤2.0 molar eq. antagonist. Antagonization by Na2HPO4 was assessed employing a fixed 1.25 mM concentration corresponding to 24 and 500 molar eq. for laspartomycin C and daptomycin, respectively.

O O

O

O

O

POH

O

OH

C6-PA

O O

O

O

O

PO

O

OH OHOH

13

6

6

PG

compound Na2H2PO4 C10-P C6-PA PG

laspartomcyin C - + + -daptomycin - - - +

Figure 3. CD spectra of laspartomycin C and its enantiomer show inverted ellipticities. laspartomycin C and ent-laspartomycin C without (left) and with (middle) 5.0 mM CaCl2. Right: laspartomycin C and ent-laspartomycin C in the presence of 5.0 mM CaCl2 and 1.0 eq. C10-P.

83


only after adding a large excess of PG (8.0 molar equivalents) was the activity of laspartomycin C diminished. By comparison, the activity of daptomycin is more readily antagonized by PG with 2.0 molar equivalents leading to complete loss of antibiotic action. Interestingly, PG is found in high concentrations in mammalian lung surfactant and is the likely cause of daptomycin’s ineffectiveness in treating lung infections.[12]

The laspartomycin C crystal structure provides key insights into the structural feature that differentiate the antibiotic mechanisms of the lipopeptide and lipodepsipeptide CDA subclasses. In the case of laspartomycin C the side chain of Dap2 closes the peptide macrocycle via an amide bond with the C-terminal proline. This newly formed amide linkage plays a role in target recognition as it contributes a hydrogen bonding interaction with the phosphate group in the binding pocket (Fig. 2). Conversely, for lipodepsipeptide CDAs like daptomycin an ester linkage is found at this position, resulting from cyclization of the C-terminus with a threonine side chain. The ester linkage found in daptomycin is unable to serve as a H-bond donor and instead a repulsive electrostatic oxygen-oxygen interaction is expected to arise when encountering a phosphate monoester.

2.3 The complex as dimer The laspartomycin C/Ca2+/C10-P ternary complex exists as a dimer in the asymmetric unit related by a two-fold rotation and is stabilized by direct and indirect interactions between the two ternary units (Fig. 5, experimental table 5). We also found evidence for the formation of the dimeric species in solution based on mass spectrometry studies (Fig. 6). The two ternary units in the dimer are nearly identical to each other with only a substantial difference in the side chain rotamer for d-allo-Thr9. As illustrated in Fig. 5, there are a number of interactions unique to the dimer. A hydrogen bond is observed between the d-allo-Thr9 backbone amide of one laspartomycin C molecule and the Asp7 side chain carboxylate of the other. Additional indirect hydrogen bonding interactions are mediated by the d-allo-Thr9 backbone carbonyl with the water molecules coordinated by the peripheral Ca2+ of the other ternary unit. The fatty acid side chain of one laspartomycin C molecule has hydrophobic interactions with the Pro11 side chain of the other molecule in the dimer. In addition, the absence of a sidechain in Gly8 prevents steric hindrance from occurring at the dimer interface.

In the dimer complex the two C10-P phosphate head groups are fully coordinated and completely sequestered from the solvent. The two C10-P head groups also directly interact through a hydrogen bond (Fig. 5). In addition, the isoprenyl

84


tails of the two C10-P ligands have hydrophobic interactions with each other and with the Pro11 side chains in the apposing laspartomycin C monomers. The more distal part of the isoprenyl tails are not stabilized in the complex and disordered in the crystals. The full coordination of the C10-P headgroup and the interactions of laspartomycin C with the headgroup-proximal part of the isoprenyl tails in the dimer complex is consistent with the high-affinity laspartomycin C – undecaprenyl phosphate interaction and explains the specificity of laspartomycin C for phosphate monoesters with both long and short hydrocarbon tails.

Figure 6. CE-MS detection (m/z) of laspartomycin C complexes as monomer (*) and dimer (**). (Left) laspartomycin C in the absence of Ca2+ ionizes as monomer (M-H+ calc. 1245,6372). (Middle) Laspartomycin C in the presence of Ca2+ ionizes as both monomer (M-H+ calc. 1283,5842) and dimer (M-H+ calc. 1284.0859) of a 1:1 complex with Ca2+. (Right) Upon mixing with C10-P and Ca2+, the 1:1:2 Laspartomycin C/C10-P/Ca2+ complex forms (the second Ca2+ is apparently bound more tightly by the presence of the C10-P ligand) and again ionizes as both the monomer (M-H+ calc. 1555,6332) and dimer (M-H+ calc. 1556.1349).

1283

.5

1284

.0

1284

.5

1285

.0

1285

.5

1284 1286

1555

.629

7

1556

.133

5

1556

.637

3

1557

.137

415

57.6

349

5 1556 15571285

1245

.632

51246.6357

1247

.637

812

48.6

413

1245 12491247 1283

Laspartomycin C Laspartomycin C + Ca2+ Laspartomycin C + Ca2+ + C10-P

****

**

*

Figure 5. The laspartomycin C/Ca2+/C10-P complex is a dimer in the crystal. Two views of the dimer interface, with cross-dimer interactions shown as dotted lines. The two laspartomycin C chains of the two ternary complexes that make op the dimer are colored green and cyan. For clarity, in the side view only the interactors closet to the viewer are labeled.

Thr9

Thr9

Asp7

Top

Asp7

Ca CaC10-P C10-P

Ca

Ca

Side

C10-P C10-P

Thr9

CaCa

Asp7

90°

85


Figure 7. (Top) Proposed orientation of the laspartomycin C dimer in the membrane (grey gradient). (Bottom) Table of structural features that are important for dimer formation and Ca2+ binding in the laspartomycin C dimer are conserved (green) among the calcium dependent lipopeptides and lipodepsipeptides. Residue d-Ser1 (red) in the lipodepsipeptide CDA1b, in place of an Asp or Asn residue, represents the single exception.

The calcium dependent lipopeptides

The calcium dependent lipodepsipeptides

Restrictions based on the Laspartomycin C dimer

Residue Laspartomycin C Friulimicin B/Amphomycin/Tsushimycin

Daptomycin A54145A CDA1b

-3 lipid lipid -

-2 Trp Trp -

-1 lipid lipid D-Asn D-Glu lipid First carbon should be C=O

1 Asp Asn/Asp Asp hAsn D-Ser Asp or Asn

2 Dap Dab Thr Thr Thr L-aa closing the cycle

3 D-Pip D-Pip Gly NMe-Gly D-Trp D-aa or Gly

4 Gly MeAsp Orn Ala Asp L-aa; sidechain allowed

5 Asp Asp Asp Asp Asp Asp or Asn

6 Gly Gly D-Ala D-Lys D-Hpg D-aa or Gly

7 Asp Asp Asp MeOAsp Asp Asp

8 Gly Gly Gly Gly Gly Gly, sidechain not allowed

9 D-allo-Thr D-Dab D-Ser D-Asn D-pAsn D-aa or Gly, polar functionality

10 Ile Val MeGlu Glu Glu L-aa or Gly

11 Pro Pro Kyn Ile Trp Should be hydrophobic

A straightforward model of how laspartomycin C is positioned on the bacterial membrane follows from the C10-P bound dimer structure. The two laspartomycin C fatty acid side chains and the two C10-P isoprenyl tails are all orientated perpendicular to a hydrophobic plane formed by the bottom sides of two laspartomycin C molecules in the dimer (Fig. 7) This plane is likely oriented parallel to the cell surface. The C10-P head groups are sequestered above the hydrophobic

86


Laspartomycin CFriulimicin B R=NH2 Amphomycin R=OHTsushimycin R=OH

Daptomycin

A54145A CDA1b

lipid

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

O OHO

NH2OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

D

D

D

HNO

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

NH2

O

OH

O

OH

O

R

O

O

OHO

1

2

3

4 5 6

7

8

91011

D

D

lipid O

1

2

3

4 5 6

7

8

91011

-1-2

lipid O

HN

O NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

OH

O

OH

O

OH

O

OHO

O

1

2

3

4 5 6

7

8

91011

D

D

O

HN

O NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

OHO

O

OH

O

OH

OH D

D

1

2

3

4 5 6

7

8

91011

lipid O

HN

O NHO

HNO

NH O

N

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

O OH

O

OH

O

OH

NH

O

D D

D

1

2

3

4 5 6

7

8

91011-1

-2

OH

O

HONH2

O

NH2

O

O NH2 O

NH

O

NH2O

P OHO

OH

OHHO O

OHN D

Figure 8. The calcium dependent lipopeptides and lipodesipeptides included in figure 7.

bottom plane within the core of the laspartomycin C dimer. This suggests that when bound to C55-P, laspartomycin C is slightly embedded in the bacterial membrane and that the hydrophobic side chains of d-Pip3 and Pro11 contribute to interactions with the hydrophobic part of the lipid bilayer. Structural features that are important for dimer formation and Ca2+ binding are predominantly conserved within the CDA family (Fig. 7-8). Whether this indicates that the structures of all CDA’s are similar and whether they all form dimers awaits further experimental verification.

87


3 Conclusion

The laspartomycin C structure here reported is the first of a CDA bound to its bacterial target and provides a clear explanation for the two conserved calcium-binding sites common to all CDAs. Furthermore, it is also is the first structure to be reported for an antibiotic that binds the essential bacterial phospholipid C55-P. At present, no clinically-used antibiotics operate via C55-P binding. Our findings provide a structural blueprint for the design of new antibiotics capable of exploiting this unique bacterial target.

88



Ent-laspartomycin C was synthesized and purified using previously described methodology[9] in 14.0 mg (11.2 µmol) quantity followed by characterization with analytical HPLC, HR-MS, 2D NMR (TOCSY, HSQC, NOESY) and CD (experimental fig. 9, 11-12).

MIC assaysMinimum inhibitory concentrations were determined in accordance with CLSI guidelines.[13] Antibiotic stocks in DMSO were diluted 50x in cation-adjusted Mueller Hinton broth (CAMHB; 10 mg L-1 Mg2+, 50 mg L-1 Ca2+, 0.002 v/v % TWEEN 80) and serially diluted in polypropylene 96-well plates to reach a volume of 50 µL per well. Bacterial cells were cultured in TSB until the exponential growth phase (OD600 = 0.5) before dilution in CAMHB and addition to the wells (50 µL) to reach a final CFU concentration of 5x105 mL-1. After overnight incubation (35 °C, 250 RPM) the plates were inspected visually for growth. MIC determinations against enterococci were carried out in cation-adjusted LB medium (10 mg L-1 Mg2+, 50 mg L-1 Ca2+, TWEEN 80 0.002 v/v %) to ensure reliable bacterial growth (Experimental table 1).

Table 1. MICs of laspartomycin C, ent-laspartomycin C and daptomycin against a panel of Gram-positive pathogens: MRSA USA300, S. aureus 29213, B. subtilis 168, E. faecalis E4125, E. faecium E155.

compound MRSA S. aureus B. subtilis E. faecalis E. faecium

laspartomcyin C 8 8 8 32 16

ent-laspartomcyin C 8 8 8 32 16

daptomycin 0.5 0.5 1 4-8 4

Circular dichroismCD spectra were recorded for 60 μM laspartomycin C and ent-laspartomycin C solutions in 20 mM HEPES (pH = 7.4) buffer in the presence and absence of 5.0 mM CaCl2 and 0-4.0 eq. C10-P. Data collection was limited to the 210-260 nm range because extensive scattering occurs below 210 nm under these conditions. The experiments were performed using a 1.0 mm cuvet, a bandwith of 1 nm and a scan speed of 20 nm min-1. The average of 10 scans was baseline corrected by subtracting the average elipticity over 255-260 nm and units were converted to mean residue molar elipticities (deg cm2 dmol-1) (Experimental Fig. 9).

89


Figure 9. CD spectra. Left: laspartomycin C in the presence of 5.0 mM CaCl2 and 0-4.0 eq. C10-P. Middle: ent-laspartomycin C in the presence of 5.0 mM CaCl2 and 0-4.0 eq. C10-P. Right: control measurement with laspartomycin C and ent-laspartomycin C in the absence of CaCl2 mixed with 1.0 eq. C10-P.

Antagonization assaysAntagonists were dissolved in MHB before serial dilution on a polypropylene 96-well plate to reach a volume of 50 µL per well. Unadjusted MHB was used to prevent the formation of C10-P:Ca2+ species that show poor aqueous solubility in the absence of laspartomycin C. Antibiotics in CAMHB were added (25 µL) to reach a final concentration of 8xMIC with antagonists present in up to 32 molar equivalents relative to the antibiotics. Na2HPO4 was applied at a fixed concentration of 1.25 mM corresponding to 24 and 500 molar eq. for laspartomycin C and daptomycin respectively. The antibiotic stocks contained extra cations and TWEEN 80 to achieve CAMHB medium composition (vide supra). After mixing for a minimum of 30 minutes, MRSA USA300 in CAMHB (25 µL) was added to reach a final concentration of 5x105 CFU mL-1. The plates were incubated overnight (35 °C, 250 RPM) and inspected visually for growth. A volume of 10 µL was transferred from each well onto blood agar plates and incubated overnight at 37 °C to confirm antagonization of antibiotic activity.

Table 2. Antagonisation of laspartomycin C and daptomycin by increasing quantities of phosphoesters. (+) antibiotic activity andtagonised; (-) antibiotic activity unaffected; (N.D). not determined.

Laspartomycin C 0 eq. 0.5 eq. 1.0 eq. 2.0 eq. 4.0 eq. 8.0 eq. 16.0 eq. 32.0 eq.

C10-P - - + / - + + + N.D. N.D.

C6-PA - - + + + + N.D. N.D.

PG - - - - - + + +

Daptomycin 0 eq. 0.5 eq. 1.0 eq. 2.0 eq. 4.0 eq. 8.0 eq. 16.0 eq. 32.0 eq.

C10-P - - - - - - N.D. N.D.

C6-PA - - - - - - N.D. N.D.

90


CE-TOF-MS analysis Laspartomycin C samples 1-3 (vide infra) were prepared in water and analyzed with capillary electrophoresis - time-of-flight mass spectrometry (CE-TOF-MS). CE was performed on a Beckman PA800 (Sciex, Framingham, MA, USA) with a background electrolyte (BGE) consisting of 20 mM formic acid and 22 mM ammonium hydroxide (NH4OH; pH = 8.3) for analysis of sample 1, and with the same BGE with an additional 1.25 mM CaCl2 for the analysis of sample 2 and 3. Samples were injected hydrodynamically into a 90 cm bare fused-silica capillary with 50 µm internal diameter by applying a pressure of 1 psi for 12 s. A 30-kV separation voltage with an additional pressure of 2 psi was used for separation, resulting in a constant 16 µA current for sample A and 18 µA for sample B and C. The temperature of the capillary was maintained at 25 °C during analysis. Detection was performed on a Bruker TOF-MS (Bruker Daltonics, Bremen, Germany) equipped with a co-axial sheath liquid sprayer (Agilent Technologies, Waldbronn, Germany). The sheath liquid consisted of MeOH:H2O 1:1 (v/v) containing 10 mM NH4OH and was used at a flow of 5 µL min-1. The MS settings were optimized for laspartomycin C signal intensity. The capillary voltage of the MS was 3.5 kV, the nebulizer gas pressure 5.8 psi, the dry gas flow 4.0 L min-1, the dry temperature 180 °C and the scan range m/z 50-3000 at a spectral rate of 0.5 Hz. Sample 1: laspartomycin C (100 µM); sample 2: laspartomycin C (100 µM) + CaCl2 (1.25 mM); 3: Laspartomycin C (100 µM) + CaCl2 (1.25 mM) + C10-P (100 µM). J. Sastre Toraño conducted the CE-TOF-MS measurement.

Crystallization and data collection C10-P was solubilized in PEG-200 and mixed with a solution of laspartomycin C in HEPES buffer (5 mM, pH = 7.5, 10 mM CaCl2) to achieve a 1:2 peptide:ligand molar ratio (7.2 mM:14.4 mM) with a final PEG-200 concentration of 10% (v/v). This solution was contact dispensed in 100 nL drops by a Gryphon LCP robot (Art Robbins Instruments) and 200 nL of reservoir solution, supplemented with 10% v/v PEG-200, was added on top. Crystals were grown at 20 °C under sitting drop vapor diffusion conditions against a reservoir solution containing either 20 mM CaCl2, 0.10 M NaAc pH = 4.6 and 40% v/v MPD or 20 mM CaCl2 and 37.5% v/v MPD. Datasets were collected at the ERSF beamline ID23-2 (wavelength 0.873 Å, temperature 100 K) to 1.60 Å and 1.28 Å resolution respectively. Crystals for anomalous data collection were grown against a reservoir solution of 0.20 M CdCl2 and 40% v/v MPD and a dataset was collected in-house using a rotating anode X-ray source (Bruker, wavelength 1.541 Å). H. C. Vlieg collected the data.

91


Table 3. Crystallographic data collection and refinement statistics.

Native Anomalous (Cd2+) High resolution

Data collectionSpace group P 43 3 2 P 43 3 2 P 43 3 2

Cell dimensions

a, b, c (Å) 56.88, 56.88, 56.88 56.79, 56.79, 56.79 56.93, 56.93, 56.93

α, β, γ (°) 90, 90, 90 90, 90, 90 90, 90, 90

Resolution (Å) 32.84-1.60 (1.64-1.60) 32.79-1.76 (1.82-1.76) 28.47-1.28 (1.30-1.28)

Rmerge 0.172 (1.691) 0.106 (0.497) 0.067 (0.699)

I / σI 15.6 (3.3) 4.91 (0.91) 15.0 (2.9)

Completeness (%) 100 (100) 96.97 (71.60) 100 (100)

Redundancy 22.3 (22.5) 87.89 (5.81) 9.4 (9.6)

Refinement

Resolution (Å) 1.28

No. reflections 8650 (438)

Rwork / Rfree 0.1393 / 0.1641

No. atoms 213

Protein 144

Ligand/ion 52

Water 17

B-factors (Å2) 23.1

Protein 18.3

Ligand/ion 33.9

Water 30.7

R.m.s. deviations

Bond lengths (Å) 0.038

Bond angles (°) 3.274

Structure solution and refinementThe data for the native Ca2+ datasets were integrated in iMOSFLM followed by scaling and merging using AIMLESS.[14,15] The anomalous data for the Cd2+ derivative was processed using EVAL.[16] The laspartomycin C ternary complex structure was solved by single isomorphous replacement with anomalous scattering using the Cd2+ derivative and the native Ca2+ dataset in SHELX c/d/e and autobuilding in ARP/wARP.[17,18] The structure was further refined to high resolution against the 1.28 Å dataset in REFMAC and manual building in Coot.[19,20] Custom restraints for the non-standard amino acids and C10-P were generated by adapting existing restraints from the CCP4 monomer library. Restraints for the laspartomycin C fatty acid tail were generated using Ghemical and eLBOW.[21,22] Peptide bonds that were not automatically recognized by the refinement software were defined as trans-peptide bonds. Figures were prepared using PyMOL Molecular Graphics System (DeLano Scientific LLC) and the structure has been deposited to the Protein Data Bank with accession code 5O0Z (Experimental table 3-5, fig. 10). H. C. Vlieg and B. J. C. Janssen solved the crystal structure.

92


Table 5. Dimer interface coordination distances.*

Atom monomer 'A' Atom monomer 'B' Distance, Å

Phosphate headgroup Phosphate headgroup 2.5

Thr9 carbonyl Coordinated water** 2.8

Thr9 amide Asp7 side chain 2.9

Figure 10. The laspartomycin C dimer superposed onto dimers of tsushimycin. The structure laspatomycin C in the ternary laspartomycin C/Ca2+/C10-P dimer (green and cyan) matches the structure of to the calcium-bound unliganded tsushimycin dimer solved previously (grey).[10] When comparing the laspartomycin dimer with the six tsushimycin dimer structures the maximum root mean square deviation (rmsd) of all 22 Cα atoms is 0.57 Å, this is less than the maximum rmsd of the tsushimycin dimers compared with each other.

* The interactions at the dimer interface are symmetrical; each unique interaction is listed only once. ** The Thr9 carbonyl interacts with two coordinated waters, both are 2.8 Å.

Ligand Distance

Central calcium coordination (Å)LaspC: Dap2 backbone carbonyl 2.3

LaspC: Asp5 side chain oxygen 2.3

LaspC: Gly6 backbone carbonyl 2.3

LaspC: Gly8 backbone carbonyl 2.3

LaspC: Ile10 backbone carbonyl 2.3

C10-P: head group oxygen 2.3

Peripheral calcium coordination

LaspC: lipid amide bond carbonyl 2.3



C10-P: Head group oxygen 2.2

Water 2.4

Water 2.4

Ligand Distance

Head group oxygens (Å)LaspC: Gly8 backbone amine 3.4

LaspC: Gly8 backbone amine 3.2

Central calcium 2.3

Peripheral calcium 2.2

Connecting oxygen

LaspC: Dap2 backbone amine 3.3

LaspC: Dap2 amine side chain 3.0

Table 4. Ca2+ (left) and C10-P (right) coordination distances.

93



-0.2

0.3

0.8

1.3

1.8

0 10 20 30 40 50 60

Abs

Time (min)

-0.20

0.20.40.60.8

11.2

0 10 20 30 40 50 60

Abs

Time (min)

laspartomycin C

Ent-laspartomycin C

1247.6507 m/z

1247.6529 m/z

Figure 11. Analytical HPLC and HR-MS for laspartomycin (Top) and ent-laspartomycin C (Bottom). HRESI-MS [M+H]+ calc: 1247.6518 m/z.

Figure 12. The NOESY spectrum (5 mM, DMSOd6, 500 MHz) for ent-laspartomycin C (right) matches the NOESY spectrum of laspartomycin C (left).

laspartomycin C Ent-laspartomycin C

8.4

8.4

8.2

8.2

8.0

8.0

7.8

7.8

7.6

7.6

7.4

7.4

ω2 - 1H (ppm)

8 8

6 6

4 4

2 2

ω1

- 1 H (

ppm

)

Spectrum: LK50-NOESY500-01032016User: laurenskleijn Date: Wed Mar 2 10:57:17 2016Positive contours: low 5.00e+05 levels 1 factor 1.40

8.4

8.4

8.2

8.2

8.0

8.0

7.8

7.8

7.6

7.6

7.4

7.4

ω2 - 1H (ppm)

8 8

6 6

4 4

2 2

ω1

- 1 H (

ppm

)

Spectrum: LK14c-NOESY500-07102014User: laurenskleijn Date: Wed Mar 2 11:06:01 2016Positive contours: low 5.30e+05 levels 1 factor 1.40

94


References[1] M. F. Chellat, L. Raguž, R. Riedl, Angew. Chem. Int. Ed. 2016, 55, 6600–6626.[2] R. Tommasi, D. G. Brown, G. K. Walkup, J. I. Manchester, A. A. Miller, Nat. Rev. Drug Discov.

2015, 14, 529–542.[3] T. Schneider, A. Müller, H. Miess, H. Gross, Int J Med Microbiol 2014, 304, 37–43.[4] M. Strieker, M. A. Marahiel, ChemBioChem 2009, 10, 607–616.[5] L. H. J. Kleijn, N. I. Martin, 2017, 47, 1–27.[6] A. Müller, M. Wenzel, H. Strahl, F. Grein, T. N. V. Saaki, B. Kohl, T. Siersma, J. E. Bandow,

H.-G. Sahl, T. Schneider, et al., Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 7077–7086.[7] T. Schneider, K. Gries, M. Josten, I. Wiedemann, S. Pelzer, H. Labischinski, H. G. Sahl,

Antimicrob. Agents Chemother. 2009, 53, 1610–1618.[8] E. Rubinchik, T. Schneider, M. Elliott, W. R. P. Scott, J. Pan, C. Anklin, H. Yang, D. Dugourd,

A. Müller, K. Gries, et al., Antimicrob. Agents Chemother. 2011, 55, 2743–2754.[9] L. H. J. Kleijn, S. F. Oppedijk, P. 't Hart, R. M. van Harten, L. A. Martin-Visscher, J. Kemmink,

E. Breukink, N. I. Martin, J. Med. Chem. 2016, 59, 3569–3574.[10] G. Bunkoczi, L. Vertesy, G. M. Sheldrick, Acta Cryst D 2005, 61, 1160–1164.[11] M. M. Harding, IUCr, Acta Crystallogr. Sect. D 2006, 62, 678–682.[12] J. A. Silverman, L. I. Mortin, A. D. G. VanPraagh, T. Li, J. Alder, J Infect Dis 2005, 191, 2149–

2152.[13] Standard, Approved, and Ninth Edition. "CLSI Document M07-A9.", Wayne, PA: Clinical

and Laboratory Standards Institute, 2012.[14] T. G. G. Battye, L. Kontogiannis, O. Johnson, H. R. Powell, A. G. W. Leslie, IUCr, Acta

Crystallogr. Sect. D 2011, 67, 271–281.[15] P. R. Evans, G. N. Murshudov, IUCr, Acta Crystallogr. Sect. D 2013, 69, 1204–1214.[16] A. M. M. Schreurs, X. Xian, L. M. J. Kroon-Batenburg, IUCr, J Appl Crystallogr 2010, 43,

70–82.[17] G. M. Sheldrick, IUCr, Acta Crystallogr. Sect. D 2010, 66, 479–485.[18] G. Langer, S. X. Cohen, V. S. Lamzin, A. Perrakis, Nature Protocols 2008, 3, 1171–1179.[19] G. N. Murshudov, P. Skubák, A. A. Lebedev, N. S. Pannu, R. A. Steiner, R. A. Nicholls, M. D.

Winn, F. Long, A. A. Vagin, IUCr, Acta Crystallogr. Sect. D 2011, 67, 355–367.[20] P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, IUCr, Acta Crystallogr. Sect. D 2010, 66,

486–501.[21] T. Hassinen, M. Peräkylä, Journal of Computational Chemistry 2001, 22, 1229–1242.[22] N. W. Moriarty, R. W. Grosse-Kunstleve, P. D. Adams, IUCr, Acta Crystallogr. Sect. D 2009,

65, 1074–1080.

95


96

Chapter 4

97

Antibacterial Properties of a Lipid II Targeting Nisin-Derived Lipopeptide

Chapter 5

Antibacterial properties of a lipid II targeting nisin-derived lipopeptide

ABSTRACT: Nisin-derived lipopeptides consisting of the nisin A/B ring system and a lipid motif at the C-terminus are rapidly bactericidal against VRE faecium. Here, we describe the synthesis of a nisin-derived lipopeptide and characterize its effect on the bacterial cell envelope of VRE. In addition, we provide a detailed investigation of its lipid II mediated mode of action and assess the potential for the development of bacterial resistance.

Manuscript in preparation: L. H. J. Kleijn, D. Morales Angeles, M. C. Viveen, R. W. L. Willems, D.-J. Scheffers, N. I. Martin, Working title: Antibacterial Properties of a Lipid II Targeting Nisin-Derived Lipopeptide

98

Antibacterial Properties of a Lipid II Targeting Nisin-Derived Lipopeptide Chapter 5

1 Introduction

Natural products have shaped the field of antibiotic pharmaceuticals more than in any other therapeutic area. With the exception of a few fully synthetic classes of antibiotics including the oxazolidinones and quinolones, the majority of clinically used antibiotics are either natural products or derivatives thereof.[1] However, there is a also a large number of well-studied natural product antibiotics that have not made it to the clinic. Among them is the lantibiotic nisin (Fig. 1).[2] A fermentation product of Lactococcus lactus, nisin was first reported in 1928.[3] Nisin consists of 34 amino acids and belongs to the so called lanthipeptide class known for their characteristic lanthionine rings which result from extensive post-translational modifications (Fig. 1).[4,5] Owing to its potent antibiotic activity against Gram-positive pathogens including enterococcal and stalphylococcal species as well as Clostridium difficile there has been much interest in nisin.[6,7] Of particular note is nisin’s ability to act as a rapidly bactericidal agent. Furthermore, little bacterial resistance to nisin has been reported despite its widespread and decades-long use as food preservative.[8] Nisin’s antibacterial mode of action is distinct from the action mechanisms of current clinical antibiotics and is mediated by cell-wall precursor lipid II.[9] Lipid II is unique to bacteria and is the final precursor to the bacterial cell wall. It consists of the carbohydrate moieties N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc), a pentapeptide (l-Ala-γ-d-Glu-l-Lys-d-Ala-d-Ala) and an isoprenyl lipid carrier which is responsible for anchoring in the bacterial inner membrane (Fig. 2A).

Nisin’s antibiotic action is driven by a unique molecular recognition event involving binding of lipid II at the pyrophosphate region by the N-terminal nisin A/B ring system.[10,11] Specifically, 5 hydrogen bonds form between the nisin backbone hydrogens (Dhb2, Ala3, Ile4, Dha5, Abu8) and two pyrophosphate oxygens.

Nisin1

Ile

Dhb AlaIleDha

LeuAla Abu

Pro GlyAla Lys Abu

GlyAla

LeuMet

Gly

Ala Asn Met Lys Abu

Ala Abu

Ala His

Ala Ser

Ile

HisVal

Dha

Lys

NH2

COOH

S

S

S S

S

AB

CD

E

Figure 1. Schematic representation of the lantibiotic nisin with rings A-E. Dhb = dehydrobutyrine, Dha = dehydroalanine, Abu = aminobutyric acid.

99


Additional interaction between the nisin A/B ring system and the MurNAc and isoprenyl portions of lipid II strengthen the interaction. Following lipid II binding, the C/D/E ring system of nisin inserts in the bacterial membrane to form large membrane destabilizing nisin:lipid II aggregates resulting in cell death.[10,12]

However, in spite of its selective and unique mode of action, nisin is not suited for development as clinical antibiotic for systemic use. As a large peptide with a number of positively charged amino acids nisin is particularly vulnerable to proteolytic degradation.[13]

1.1 The nisin-derived lipopeptidesTo utilize the antibiotic properties of nisin while addressing its drawbacks, a novel class of nisin-derived lipopeptide antibiotics was conceived in the Martin group.[14]

Central to these nisin-derived lipopeptides is the nisin A/B fragment (nisinAB), which as product of proteolitic degradation of nisin is resistant to further proteolysis. NisinAB itself is devoid of antibiotic activity, but a single-step chemical modification whereby a lipid is introduced at the C-terminus restores the antibiotic activity (Fig. 2B). A previous report detailed the first characterization of the nisinAB lipopeptides including their spectrum of antibiotic activity.[14]

OOAcHN

O

O NHOPO

O

O

OHO

AcHN

HO

PO

O

NH

O

HN

O

OH

OH

O

O

OHN

H3N

O

HN

O

O

O

7

3

NisinAB

Lipid II

HN

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NH2

Lipid

NisinAB lipopeptides

A

B

A

B

Figure 2. (A) Lipid II with its pyrophosphate region (highlighted in yellow) that is recognized by nisinAB. (B) General structure of the nisinAB lipopeptides.

100


Here, a new nisin-derived lipopeptide is evaluated with regard to its bacterial killing properties, mode of action, and propensity to induce the development of bacterial resistance.


2.1 SynthesisThe lipopeptide detailed here consists of nisinAB functionalized with the aromatic chloro-bis-phenyl (CBP) functionality at the C-terminus (scheme 1). The function of the lipid is to increase affinity of the peptide for bacterial membranes (i.e. act as a membrane anchor) and to increase binding to lipid II for which aromatic functionalities are particularly suited.[14,15]

To selectively isolate nisinAB (2), the inherent instability of nisin towards proteolytic enzymes is exploited by means of treatment with trypsin (Scheme 1). After HPLC purification, full NMR assignment (1H, 13C) of the peptide backbone and amino acids confirmed the chemical structure of nisinAB (2). Characteristic downfield chemical shifts result from the C-C double bond containing amino acids Dhb2 (CβH 6.31, 126.8) and Dha5 (CβH2 5.51/6.00, 104.1). Next, 2 was reacted with Boc2O and DiPEA in dry MeOH to obtain intermediate 3 with the Boc protecting groups in place on both the N-terminus and the Lys12 side chain. Lysine protection prevents the formation of a cyclic side product that can result from intramolecular amide bond formation between the lysine side chain amine and the activated C-terminus during the subsequent coupling reaction.[14] Installment of the CBP lipid amine is performed using BOP and DiPEA followed by routine acid treatment (TFA/H2O/TiS) to remove the Boc protecting groups. After HPLC purification, the structure of nisinAB-CBP (4) was confirmed by full NMR characterization and HR-MS analysis.

2.2 Characterization of bactericidal activity against VREBacteremia caused by Enterococci primarily result from E. faecalis and E. faecium species. The former category is more prominent, but E. faecium bacteremia is more difficult to treat resulting in 20-46% mortality when the species is vancomycin-resistant (VRE).[16] One of the factors that contribute to poor clinical outcome is the lack of bactericidal treatment options.[17] Here, we assess the ability of nisinAB-CBP (4) to act as a bactericidal agent against vancomycin-resistant E. faecium in a time-kill assay. Clinically used antibiotic daptomycin, which is considered a good treatment option for bacteremia caused by VRE due to its bactericidal properties, is

101


included in the study.[18]

Bacterial cultures of E. faecium (VRE strain E155) were challenged with nisinAB-CBP (4) at concentrations corresponding to 2-fold, 5-fold, and 10-fold the minimum inhibitory concentration (MIC) and compared with daptomycin and

Cl

HN

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NH2

NisinAB-CBP4

b)OH

O

HN OBocHN

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NHBoc

3

a)

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NH2

OH

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

HN

O

O NH

NH2

O

NH

OHNO

NH

OHN

O

NH

S

O

HNO

S

NH

O

HN

O

NH2

O NH

S

O

H2N

HN

O NH

S

O

HN

HNO

O

NH

S

HN

O

NHN

NH

O

OHN

HO

OHNNH

OHN

O

N NH

O

NHH2N

OOH

Nisin

NisinAB

1

2

a)

Scheme 1. Synthesis of nisinAB-CBP (4) from nisin. Reagents and conditions: a) trypsin, tris buffer, 30°C, 48 hrs, 39%; b) Boc2O, DIPEA, MeOH, 57% c) (i) Chloro-bis-phenylamine.HCl, BOP, DiPEA, DMF, (ii) TFA/H2O/TiS, 27%.

102


nisin, both administered at a single concentration (10xMIC) (Fig. 3A). After addition of the antibiotics, the amount of viable bacterial cells was determined at selected time point over a 20-hour timeframe. After just 1 hour, exposure to nisin and the higher nisinAB-CBP (4) concentrations (5xMIC, 10xMIC) had resulted in a 4-log10 reduction of viable cells corresponding to 99.99% killing.[6] The same reduction in cell viability was observed for nisinAB-CBP (4) at low concentration (2xMIC) after 20 hours, but not for daptomycin (10xMIC) matching previous studies.[19-21]

The bactericidal action of nisin and nisinAB-CBP (4) is even more pronounced when the time-kill assay in performed at high bacterial density (Fig. 3B). Nisin and nisinAB-CBP (4) (even at the lower multiples of the MIC) achieve a strong and lasting reduction in the amount of viable cells while treatment with daptomyicin results in an initial 2-log10 reduction followed by restored bacterial growth.

2.3 Lysis of VREIn order to establish whether nisinAB-CBP (4) acts as a lytic antibiotic agent, as is known to be the case for nisin [22], a simple lysis assay was performed (Fig. 4). A high inoculum of VRE E155 in the log phase was treated with various concentrations of nisinAB-CBP (4) with nisin and daptomycin also tested as control compounds.[23] For daptomycin, which acts as a bactericidal agent without induction of lysis,[24] only a moderate decrease in turbidity was observed (ΔOD600 = -30±6 %) after incubation for 18 hours. NisinAB-CBP (4) demonstrated stronger lytic activity at a concentration of 2xMIC (ΔOD600 = -46±3%), and high levels of lysis at 5xMIC (ΔOD600 = -81±1 %) and 10xMIC (ΔOD600 = -78±1 %) corresponding to the degree of lysis seen after treatment with nisin (ΔOD600 = -71±5 %).[25] Upon transfer of the treated bacterial cultures to glass tubes lysis is observed as transparent medium.

Figure 3. Time-kill curves against VRE E155 for nisinAB-CBP (4), nisin and daptomycin at (A) low and (B) high cell density. (A) curves for nisinAB-CBP (4) (5xMIC, 10xMIC) and nisin overlap.

0 2 4 6 8 10 12 14 16 18 202

3

4

5

6

7

8

9

10

Time (hours)

Col

ony

coun

t (Lo

g 10 C

FU m

L-1)

A

0 2 4 6 8 10 12 14 16 18 202

3

4

5

6

7

8

9

10

11

Time (hours)

Col

ony

coun

t (Lo

g 10 C

FU m

L-1)

NisinAB-CBP (4) (5xMIC)NisinAB-CBP (4) (10xMIC)Nisin (10xMIC)Daptomycin (10xMIC)Growth control

NisinAB-CBP (4) (2xMIC)

B

103


Next, transmission electron microscopy (TEM) was used to visualize the effects of the antibiotics on the E. faecium cells (Fig. 5). TEM allows for analysis of the bacterial cells under realistic conditions as chemical fixation and sample dehydration as part of the sample preparation procedure is not required, which prevents the introduction of artificial morphological changes. Untreated cells maintained normal cell morphology over the course of the experiment, but mesosome-like structures (intracytoplasmic membrane inclusions) were observed after 1 hour indicating rearrangements in the cytoplasmic membrane. However, treatment with nisin at a concentration of above (2xMIC) and below MIC (0.5xMIC) immediately results in bacterial capsule detachment. This capsule shredding phenomenon, well-characterized in the Gram-positive S. pneumonia, has been described as a bacterial defense mechanism in response to antimicrobial peptide (e.g. polymyxin B) exposure.[26] Capsule shredding, mediated by autolysin LytA, facilitates invasion of epithelial cells and the bloodstream by S. pneumonia.[26]

After 30 minutes of supra-MIC treatment with nisin, membrane defects appear and after 1 hour significant structural damage is seen in most, but interestingly not all cells. At the sub-MIC concentration of nisin, the same effects with delayed onset are seen on a smaller fraction of the cells.

TEM analysis of NisinAB-CBP (4) reveals similarities and differences with regard to the effect on E. faecium cells compared with nisin (Fig. 6). No bacterial capsule shredding is observed, but mesosome-like stuctures form immediately at both sub- and supra-MIC levels of antibiotic. Like nisin, NisinAB-CBP (4) at supra-MIC concentration has a profound effect on cell morphology as evidenced by the observed disfigurement at time point 2 hours. This effect on the cells takes longer to manifest, but contrary to nisin, all cells in the culture are affected (no healthy cells are observed after 2 hours). At sub-MIC concentration, the manifestation of

Figure 4. Assessment of lysis of VRE E155 by nisinAB-CBP (4), nisin and daptomycin. Left: OD600 measurement as a function of lytic activity. Right: Treatment with nisinAB-CBP (4) and nisin results in clearing of the medium in contrast to daptomycin (all at 10xMIC).

0 180.0

0.2

0.4

0.6

0.8

1.0

1.2

Time (hours)

OD

600

Growth controlNisinAB-CBP (4) (2xMIC)NisinAB-CBP (4) (5xMIC)NisinAB-CBP (4) (10xMIC)Nisin (10xMIC)Daptomycin (10xMIC)

(+) NisinAB-CBP (4)

Daptomycin Nisin

104


Figure 5. TEM analysis of VRE E155 cells untreated and treated with nisin. The untreated cells remain healthy and start to develop mesosome-like structures after 60 minutes (black arrows). Exposure to 2xMIC nisin leads to immediate capsule detachment, membrane defects after 30 minutes (white arrows) and significant structural damage after 1 hour to most of the cells. At 0.5xMIC, the effects take longer to manifest. M.C Viveen performed the TEM experiment.

Untreated Nisin (0.5xMIC) Nisin (2xMIC)

500nm

500nm 500nm 200nm

500nm 500nm

500nm 200nm 500nm

500nm 500nm 200nm

T = 0 min

T = 30 min

T = 60 min

T = 120 min

105


NisinAB-CBP 4 (0.5xMIC) NisinAB-CBP 4 (2xMIC)

500nm

500nm

500nm

500nm

500nm

500nm

500nm

500nmT = 0 min

T = 30 min

T = 60 min

T = 120 min

Figure 6. TEM analysis of VRE E155 cells treated with nisinAB-CBP (4) (0.5xMIC and 2xMIC). Mesosome-like stuctures form immediately at both sub- and supra-MIC levels and after 2 hours significant structural damage is observed. M.C Viveen performed the TEM experiment.

106


mesosome-like structures and cell shape alterations are less pronounced and affect only a minority of the bacteria in the cell culture.

2.4 Mode of actionTo rationalize the rapid bactericidal activity and pronounced effects on cell morphology of nisinAB-CBP (4), we performed a series of mode of action studies. Previous studies showed that nisin-derived lipopeptides act via a lipid II mediated mechanism, but contrary to nisin, the nisin-derived lipopeptides did not permeabilized large unilamellar vesicles (LUVs) containing 0.2% lipid II. This pointed towards an occlusion type of mode of action that depletes the bacterial pool of lipid II to inhibit cell-wall synthesis.

Following the TEM analysis that revealed the profound effects of nisinAB-CBP (4) on the morphology of live bacterial cells, we set out to quantify the ability of nisinAB-CBP (4) to depolarize the plasma membrane of live bacterial cells of Bacillus subtilis 168. Membrane depolarization was assessed using the dye DiSC3 following incubation of the cells with either nisin or nisinAB-CBP (4) (Fig. 7A). The observed strong ability of nisin to depolarize the B. subtilis membrane (EC50 = 0.19 µg mL-1) was in line with published reports.[27] Interestingly, nisinAB-CBP (4) also induced membrane depolarization at a higher, but biologically relevant concentration (EC50 = 2.80 µg mL-1).

Depolarization of bacterial membranes by nisin correlates strongly with the formation of membrane disrupting lipid II clusters and is accompanied by the formation of pores. We therefor assessed the ability of nisinAB-CBP (4) to form pores using a methodology based on the Live/Dead bacterial viability assay involving DNA staining with dyes SYTO9 and propidium iodide (Fig 7B). Upon pore formation, the influx of propidium iodide quenches fluorescence of the membrane permeable DNA-stain SYTO9. In line the depolarization assay, nisinAB-CBP (4) demonstrated the ability to form pores at a concentration higher than seen for nisin. Initially, this finding appears contradictory in the context of the reported inability of nisin-derived lipopeptides to permeabilize vesicles spiked with lipid II in a carboxyfluorescein efflux assay.[14] However, this same apparent discrepancy was reported for several nisin variants that bear amino acid mutations.[27] The phenomenon is attributed to the absence of the relevant biological framework in unilamellar vesicles like the presence of membrane proteins, the lack of a membrane potential or differences in lipid membrane components.[27]

107


Figure 7. (A) Depolarization of the B. subtilis 168 membrane caused by nisin and nisinAB-CBP (4) as determined using the dye DiSC3(5). Each data point in the curve (one of two data sets) represents two measurements. (B) Pore formation in B. subtilis 168 resulting from nisin and nisinAB-CBP (4) exposure as determined using the Live/Dead bacterial viability assay (Life Technologies). Each data point in the curve (one of two data sets) represents two measurements. (C) Fluorescence microscopy of B. subtilis 168 after incubation with nisin or nisinAB-CBP (4) followed by staining of lipid II with fluorescent vancomycin. Lipid II clusters are observed as dots. The amount of cells with clusters is indicated as percentage of the total cells counted: [4 µg mL-1]: nisin n = 347, nisinAB-CBP (4) n = 302; [30 µg mL-1]: nisin n = 271, nisinAB-CBP (4). Images were inverted for clarity (D) Summary of B. subtilis 168 data. *No accurate 95% CI could be determined due to steepness of the slope. D. Morales Angeles performed the experiments detailed here.

CompoundMIC

[µg mL-1]ΔΨ dissipation EC50 (95% CI) [µg mL-1]

Pore formation EC50

(95% CI) [µg mL-1]% cells with lipid II

clusters at 30 µg mL-1

Nisin 4 0.19 (0.16-0.23) 0.57 (0.53-0.61) 100

NisinAB-CBP (4) 4 2.80 (2.59-3.02) 2.61 (2.26-X*) 27.5

Nisin NisinAB-CBP (4) Untreated

3

0 µg

mL-

1

4 µg

mL-

1

38.9% 0.7%

100% 27.5%

1 0 - 3 1 0 - 2 1 0 - 1 1 0 0 1 0 1 1 0 2

0

5 0

1 0 0

C o n c e n tra tio n

Re

sp

on

se

(A

. U

.)

N is in

N is in A B -C B P

1 0 - 2 1 0 - 1 1 0 0 1 0 1 1 0 2

0

5 0

1 0 0


Re

spo

nse

(A

. U

.)

N is in

N is in A B -C B P

Nisin

NisinAB-CBP (4)

1 0 - 3 1 0 - 2 1 0 - 1 1 0 0 1 0 1 1 0 2

0

5 0

1 0 0


Re

sp

on

se

(A

. U

.)

N is in

N is in A B -C B P

Concentration (µg mL-1)

Res

pons

e (A

U)

Nisin

NisinAB-CBP (4)

Concentration (µg mL-1)

Res

pons

e (A

U)

1 0 - 3 1 0 - 2 1 0 - 1 1 0 0 1 0 1 1 0 2

0

5 0

1 0 0


Re

sp

on

se

(A

. U

.)

N is in

N is in A B -C B P

A B

C

D

108


Finally, the propensity of nisinAB-CBP (4) to form clusters with lipid II was assessed using fluorescently labeled vancomycin to act as lipid II stain (Fig 7C). Clustering entails the formation of non-physiological antibiotic/lipid II aggregates throughout the bacterial membrane that contain up to 106 lipid II molecules.[12] The formation of (in particular large) lipid II aggregates correlates strongly with membrane disruption and bacterial cells death, which leaves their biological relevance with regard to the antibiotic mode of action beyond question.[12]

Fluorescence microcopy analysis of untreated B. subtilis cells shows localization of lipid II in the septum and along on the cell edges as is in line with normal cell envelope function. Treatment with nisin results in concentration dependent clustering of lipid II as evidenced by the formation of spots and loss of defined fluorescent cell edges. At an antibiotic concentration of 30 µg mL-1, 100% of all observed cells exposed to nisin for 10 minutes show lipid II clusters. At the same antibiotic concentration, NisinAB-CBP (4) also leads to lipid II clustering in B. subtilis affecting 27.5% of all counted cells after 10 minutes of incubation.

Combining the observed effects on B. subtilis, it is evident that the antibiotic action of nisinAB-CBP (4) is lipid II mediated in a manner reminiscent of parent compound nisin. However, nisin is a superior agent with regard to the ability to form lipid II aggregates and to depolarize bacterial membranes. In fact, the primary action of nisin on lipid II resulting in the formation of clusters may be detrimental to the degree that cell wall synthesis inhibition by lipid II, through means of lipid II sequestration, may not contribute to the action mechanism.[12,27,28] In that regard, the equal potency of nisin and NisinAB-CBP (4) against B. subtilis 168 warrant the question whether NisinAB-CBP (4) does inhibit bacterial cell wall synthesis through lipid II occlusion.

To assess this, we performed a UDP-MurNAc-pentapeptide accumulation assay. When live bacterial cells are exposed to an antibiotic that interferes with the membrane-bound stages of bacterial cell wall synthesis, the ‘upstream’ precursor UDP-MurNAc-pentapeptide accumulates as a result.[23] UDP-MurNAc-pentapeptide is the last soluble precursor to the cell wall, which conveniently allows for monitoring of the accumulation effect by analytical HPLC analysis of the bacterial cell content.

The assay was performed using at a lower than customary antibiotic concentrations of 1xMIC in order to minimize lytic activity of nisin and NisinAB-CBP (4). As lysis is a concentration dependent phenomenon, and significant display of lysis would

109


obscure the interpretation of the assay results, the optimum conditions for detection of cell wall synthesis are at lower antibiotic concentration. The vancomycin-susceptible E. faecium E980 strain was used in the study to allow for the use of vancomycin as positive control antibiotic. As is seen in figure 8, NisinAB-CBP (4) has a clear inhibitory effect on the synthesis of the cell wall when administered to cells at a concentration of 1xMIC. This accumulation effect is also seen at concentrations of 0.5xMIC and 2xMIC (data not included here). Exposure to nisin however does not lead to detectable accumulation at 1xMIC, nor at 0.5xMIC and 2xMIC.

The UDP-MurNAc-pentapeptide accumulation assay confirms that NisinAB-CBP (4) inhibits bacterial cell wall synthesis at concentrations that are biologically relevant. Taken together with the finding that NisinAB-CBP (4) forms lipid II clusters that are associated with detrimental membrane damage, this signifies that NisinAB-CBP (4) operates via multiple modes of action.

2.5 Molecular target interactionThe “occlusion” and “clustering” antibiotic modes of action described above imply that a strong binding event takes place between NisinAB-CBP (4) and lipid II. Considering the molecular recognition of the nisin A/B system for specifically the pyrophosphate unit of lipid II the question is warranted whether an interaction with undacaprenyl pyrophosphate (C55-PP) takes place as well. C55-PP is the membrane bound residue that remains part of the lipid carrier cycle after incorporation of the GlcNAc-MurNAc-pentapeptide portion of lipid II into the peptidoglycan.

We assessed whether lipid II, C55-PP and C10-PP (a truncated variant with increased aqueous solubility) have the ability to antagonize the antibiotic action of nisin and NisinAB-CBP (4). Antibiotics were added to a culture of E. faecium

NisinAB-CBP (4)NisinVancomycinUntreated

Figure 8. Accumulation of UDP-MurNAc-pentapeptide in E. faecium E980 detected by analytical HPLC. Accumulation is observed after treatment with NisinAB-CBP 4 (1xMIC) and positive control vancomycin (10xMIC), but not with nisin (1xMIC).

110


E980 at a concentration of 8xMIC with antagonists present at 5 molar eq. As expected, the bactericidal activity of control antibiotic daptomycin was unaffected by all antagonists while vancomycin’s activity was inhibited only by the presence of lipid II (table 1). Lipid II also completely inhibited the activity of both nisin and nisinAB-CBP (4). Interestingly, while leaving the activity of nisin unaffected, antagonists C55-PP and C10-PP did have a moderate effect on the antibiotic activity of nisinAB-CBP (4). The presence of both C55-PP and C10-PP resulted in the prevention of further bacterial growth, but the bactericidal action that is normally seen in nisinAB-CBP (4) was inhibited as indicated by the presence of viable cells at the end of the antagonization assay (Experimental fig. 10).

To provide additional support for the proposed pyrophosphate-directed binding interactions, we evaluated whether nisinAB-CBP (4) and nisin are capable of forming a soluble complex with lipid II and C10-PP that is sufficiently stable to allow for mass spectrometry detection (Table. 2). Both antibiotics were mixed with either lipid II or C10-PP in H2O / MeCN. With exception of the combination nisin / lipid II, which forms a complex that is insufficiently soluble in aqueous systems, the samples remained clear. A complex between nisinAB-CBP (4) and lipid II in 1:1 stoichiometry was clearly detected confirming the molecular interaction.

Table 1. Antagonisation of antibiotic activity of nisinAB-CBP (4), nisin, daptomycin and vancomycin by selected (truncated) cell wall precursors with E. faecium E980. (+) antibiotic activity antagonized; (+/-) antibiotic activity affected (inhibition of bactericidal activity); (-) antibiotic activity unaffected .

Compound No antagonist Lipid II C55-PP C10-PP

NisinAB-CBP (4) - + +/- +/-Nisin - + - -

Daptomycin - - - -

Vancomycin - + - -

Table 2. MS detection of antibiotic/ligand complexes. ND no data, insufficient solubility.

SampleStochiometry of

detected complex Formula Calculated Detected

NisinAB-CBP (4) / lipid II 1:1 C158H247ClN22O38P2S22- 1610.8344 1610.8319

Nisin / lipid II ND ND ND ND

NisinAB-CBP (4) / C10-PP 1:1 C74H112ClN14O19P2S2- 1661.6839 1661.6820

1:1 C74H111ClN14O19P2S22- 830.3383 830.3384

1:2 C84H132ClN14O26P4S2- 1975.7529 1975.7492

Nisin / C10-PP 1:1 C153H252N42O44P2S72+ 1833.8141 1833.8144

1:1 C153H253N42O44P2S73+ 1222.8785 1222.8799

1:2 C163H272N42O51P4S72+ 1990.8483 1990.8287

1:2 C163H273N42O51P4S73+ 1327.5680 1327.5736

111


Moving to C10-PP as ligand, a 1:1 complex with both nisinAB-CBP (4) and with nisin was detected. In addition, the MS analysis revealed species corresponding to antibiotic / C10-PP complexes in 1:2 stoichiometry albeit at 10-fold lower ion counts than the corresponding 1:1 species indicating lower prevalence and/or stability.

2.6 Bacterial resistanceTo assess the propensity of for resistance development a 30-day serial passage assay experiment was carried out. VRE E155 was passaged once daily in the presence of a concentration gradient of nisinAB-CBP (4), nisin, and daptomycin. The E. faecium strain developed a tolerance for daptomycin in both independent experiments resulting in a ≥512-fold and 16-fold reduction in accordance with previous observations with published data.[29] Gratifyingly, for both nisinAB-CBP (4) and nisin no resistant mutants could be detected.

3 Conclusion

Nisin-derived lipopeptide nisinAB-CBP (4) has a rapidly bactericidal and lytic action against VRE faecium. It operates via both bacterial cell wall synthesis inhibition and via lipid II-mediated membrane disruption. A 30-day serial passage assay did not result in the detection of a resistant VRE faecium mutant.

Statement of clarity: The intellectual property described in this chapter is covered by a patent granted to Utrecht University (WO2016116379A1). This patent has been licensed to Karveel Pharmaceuticals (a company co-founded by L. H. J. Kleijn and N. I. Martin).

Figure 9. Serial passage assay over 30 days carried out in independent duplicates: VRE E155 was challenged with nisinAB-CBP (4), nisin, and daptomycin. Only exposure to daptomycin resulted in the development of bacterial resistance.

0 5 10 15 20 25 300.5

1248

163264

128256512

Time (days)

Fold

incr

ease

in M

IC

0 5 10 15 20 25 300.5

1248

163264

128256512

Time (days)

Fold

incr

ease

in M

IC

NisinAB-CBP (4)NisinDaptomycin

112



NisinAB (2) Nisin (600 mg, 0.18 mmol) was dissolved in 250 ml Tris buffer (25 mmol NaOAc, 5 mmol Tris acetate, 5 mmol CaCl2, pH 7.0) and the solution was cooled on ice for 15 min. Trypsin (50 mg) was added and the mixture was stirred at RT for 15 min. The mixture was then heated to

30 °C for 16 hours. Another 50 mg of trypsin was added and after an additional 24 hours the reaction mixture was acidified with HCI (1 N) to a pH of 4 and solvents were removed in vacuo. NisinAB (2) was purified from the mixture using preparative HPLC. Product fractions were lyophilized to obtain a white powder (80 mg, 39%). Preparative HPLC: Maisch Reprospher 100 C8-Aqua, 250 mm x 20 mm, 10 μm. 0-60% gradient buffer B over 70 minutes, flow 12 min-1. Buffer A: H2O/MeCN/TFA (95/5/0.1), Buffer B: H2O/MeCN/TFA (5/95/0.1). Product eluted at 54% buffer B. Full NMR assignment included (vide infra).

Boc-NisinAB(Boc) (3) Boc2O (50 mg, 229 µmol) and DIPEA (51 µL, 293 µmol) were added to a solution of 2 (100 mg, 86.9 µmol) in dry MeOH (30 ml) and the mixture was stirred for 4.5 hours. The reaction mixture was concentrated, dissolved in H2O/MeCN/TFA (70/30/0.1) and

purified by preparative HPLC using a C18 column to yield 68.9 mg (51.0 µmol) of white powder (57% yield). HPLC gradient: 35-75% buffer B over 53 minutes with a flow rate of 12 mL min-1. Buffer A: H2O/MeCN/TFA (95/5/0.1), Buffer B: H2O/MeCN/TFA (5/95/0.1). ESI-MS: C61H100N13O17S2

+ [M+H]+: calc. 1350.6796, found 1350.6818.

NisinAB-CBP (4) Chloro-bis-phenylamine hydrochloride (2.5 mg, 9.8 μmol), BOP (4.3 mg, 9.8 μmol) and DiPEA (5.1 μL, 29.3 μmol) were added to a solution of 3 (11.0 mg, 8.14 μmol) in dry CH2Cl2 (5 ml). A few drops of DMF were added to ensure solubility. The mixture was stirred for 45 min, concentrated and the residue was treated with TFA/TiS/H2O (95/2.5/2.5) for 1 hour and precipitated in MTBE/hexanes

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NH2

OH

NisinAB2

OH

O

HN OBocHN

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NHBoc

3

113


(1:1), centrifuged (5 min, 4500 rpm). The pellet was dissolved in H2O/t-BuOH (1:1) and lyophilized. The lyophilized powder purified via preparative HPLC using a gradient of 35-75% buffer B over 53 minutes at a

flowrate of 12 mL min-1. (Buffer A: H2O/MeCN/TFA (95/5/0.1), Buffer B: H2O/MeCN/TFA (5/95/0.1). Yield: 3 mg (2.2 µmol, 27%). ESI-MS: C64H94ClN14O12S2

+ [M+H]+ calc: 1349.6300; found: 1349.6408. Full NMR assignment included (vide infra).

MIC assaysMinimum inhibitory concentrations were determined in accordance with CLSI guidelines.[30] MIC determinations were performed in cation-adjusted TSB (CAMHB; 10 mg L-1 Mg2+, 50 mg L-1 Ca2+) for testing of enterococcus and in LB medium for testing against B. subtilis. Bacterial cells were cultured until the exponential growth phase (OD600 = 0.5) before mixing with antibiotic dilution series in a polypropylene 96-well plate to reach OD600 = 0.0025. After sealing with an adhesive membrane, plates were incubated overnight (37 °C, 250 RPM) and inspected for bacterial growth. Bactericidal activity was assessed by plating well content (10 µL) onto blood agar plates followed by overnight incubation at 37 °C.

Time dependent killingA freshly inoculated culture of VRE E155 in cation-adjusted TSB (50 mg L-1 CaCl2, 10 mg L-1 MgCl2) was grown to the log phase and applied directly (to OD600 = 0.34) or after dilution (OD600 = 0.0025) in cation-adjusted TSB (low inoculum) to Nisin-CBP (2x, 5x and 10xMIC), nisin (10xMIC) and daptomycin (10xMIC). The cultures (2 mL each) were incubated at 37 °C (200 RPM) and at time point 0, 1, 2, 4 and 20 hours, a sample of 100 µL was transferred to Eppendorf tubes and spun down (10,000 RPM, 1 minute). After removal of the supernatant, the pellets were suspended in 0.9% NaCl2 to prevent antibiotic carryover followed by 10-fold serial dilution and plating on blood agar plates. After overnight incubation, colonies were counted. The experiment was carried out with biological duplicate and the reported CFU counts represent the mean of the replicate experiments.

Cl

HN

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NH2

NisinAB-CBP4

114


Lysis assaysLysis was assessed by administration of Nisin-CBP (4) (2x, 5x and 10xMIC), nisin (10xMIC), and daptomycin (10xMIC) to log-phase VRE E155 cultures (2 mL, OD600 = 0.80) followed by incubation at 37 °C (150 RPM) for 18 hours in culture tubes. OD600 values were determined and the cultures were transferred to glass tubes and photographed. The experiment was carried out with biological duplicate and the reported OD600 values with standard deviation represent the mean of the replicate experiments.

TEM analysisAn overnight culture of E. faecium (VRE E155) in TSB was used to prepare a fresh log-phase culture in cation-adjusted TSB (50 mg L-1 CaCl2, 10 mg L-1 MgCl2). The log-phase culture was diluted (OD600=0.05), transferred to a 96-wells plate and either left untreated or threated with nisin and nisinAB-CBP (4) at both 0.5xMIC and 2xMIC. The cultures were incubated at 37 °C on a plate shaker (200 RPM) and directly (unfixed) used for negative staining TEM at selected time points. Fifty mesh copper TEM girds with Formvar support-film were freshly carbon-coated prior to use. The girds were placed on a drop of bacterial suspension for 30 seconds, rinsed once on a drop of milli-Q H2O and contrasted with 2% methylamine tungstate. The samples were immediately examined after staining using a Tecnai 12 (Fei) at 80kV. M.C Viveen performed the experiment and R. Scriwanek optimized the TEM images with regard to contrast.

Membrane depolarizationMembrane potential dissipation was assessed using the membrane potential sensitive dye DiSC3(5).[27] In this assay, dissipation of the ΔΨ leads to redistribution of DiSC3(5) in the bacterial membrane counteracting the initial quenching caused by the presence of the potassium ionophore valinomycin. This results in a ΔΨ dissipation dependent increase in fluorescence.[31,32]

B. subtilis 168 was grown overnight in LB at 37 °C, diluted (1:100) and grown until the log phase. Measurements were performed in PIPES/NaOH buffer (50 mM, pH 7) in 96 well plates. DiSC3(5) (6 µM) and valinomycin (0.2 µM) from stock solutions in EtOH were added followed by the addition of B. subtilis (OD600 = 0.3). Finally, nisin and nisinAB-CBP (4) were added before brief mixing. Fluorescence was measured directly with a Synergy MX fluorescence plate reader at 630 nm excitation and 670 nm emission wavelengths. Data was normalized and fitted with dose-response relationship using GraphPad Prism (version 7.03). D. Morales Angeles performed the experiment.

115


Pore formationPore formation was measured using methodology adapted from the Live/Dead bacterial viability assay (Life Technologies).[27] Upon antibiotic induced membrane permeabilization, propidium iodide enters the cell and quenches fluorescence of the cell-penetrating DNA-staining dye SYTO9.

B. subtilis 168 was grown overnight in LB at 37 °C, diluted (1:100) and grown until the log phase. Measurements were performed in PIPES/NaOH buffer (50 mM, pH 7) in 96 well plates. To a two-fold dilution series of nisin and nisinAB-CBP (4) was added SYTO9 (4.2 µM), propidium iodide (25 µM), and finally B. subtilis (OD600 = 0.3). Data was obtained with a Synergy MX fluorescence plate reader at 485 nm excitation and 530 nm emission wavelengths. Data was normalized and fitted with dose-response relationship using GraphPad Prism (version 7.03). D. Morales Angeles performed the experiment.

Lipid II clusteringB. subtilis 168 was grown overnight in LB at 37 °C, diluted (1:100) and grown until the log phase. Cells were spun down, resuspend in PBS, and incubated with nisin and nisinAB-CBP (4) (4 and 30 μg ml-1) for 10 minutes. Lipid II was stained by applying a 1:1 mixture of vancomycin and BODIPY-vancomycin at a concentration of 1 µg/ml. Pictures were obtained using a Nikon Ti-E inverted microscope equipped with a CFI Plan Apochromatic DM 100× oil objective. Digital images were recorded using a Hamamatsu Orca Flash 4.0 camera, analysed using ImageJ.[27] D. Morales Angeles performed the experiment.

UDP-MurNAc-pentapeptide accumulation assay[33]

E. faecium E980 was grown until OD600 = 0.5 in TSB before addition of chloramphenicol (130 µg mL-1). After incubation for 15 minutes at 37°C, the culture was divided in 5 mL aliquots. Antibiotics were added at 1xMIC (nisin, nisinAB-CBP (4)) or 10xMIC (vancomycin) and one aliquot remained untreated. After incubation for 30 minutes at 37°C, cells were separated from the medium and extracted with boiling d-H2O (1 mL) for 15 minutes. The suspensions were spun down and the supernatant was lyophilized. The resulting material was analyzed by HPLC applying a gradient from 100% eluent A (50 mM NH4HCO3:5 mM Et3N, pH = 8.3) to 75% eluent A over 15 minutes using a C18 column (eluent B: MeOH).

116


Antagonization assaysAntagonists (5 eq. relative to the antibiotics) were dissolved in cation-adjusted TSB (50 mg L-1 CaCl2, 10 mg L-1 MgCl2) supplemented with 0.002% TWEEN 80 and mixed with antibiotics on a polypropylene 96-well plate. After 15 minutes, E. faecium E980 in the exponential growth phase was added to reach a CFU count of 5x105 mL-1 and an antibiotic concentration of 8xMIC. The plates were incubated overnight (37 °C, 250 RPM) and inspected for bacterial growth. A volume of 10 µL was transferred from each well onto blood agar plates and incubated overnight at 37 °C to further asses antagonization of antibiotic activity (Experimental fig. 10).

Figure 10. After overnight incubation, the content of the wells (10 µL) was transferred to blood agar to further assess antagonization for nisinAB-CBP (4) (left) and nisin (right). Left: The presence of viable cells (C55-PP, C10-PP) indicates partial antagonization of the antibiotic activity of nisinAB-CBP (4) by the pyrophosphate species.

Lipid II

C55-PP

C10-PP

No antagonist

Nisin

Lipid II

C55-PP

C10-PP

No antagonist

NisinAB-CBP (4)

MS analysisNisinAB-CBP (4) and nisin (both 100 µM) were mixed with lipid II or C10-PP (both 200 µM) in H2O/MeCN (1:1, 100 µL). The samples were subjected to both positive and negative ionization MS analysis using a 6560 Ion Mobility Q-TOF LC/MC (Agilent Technologies) Serial passage assays A culture of VRE E155 at OD600 = 0.5 was diluted 1:100 and applied to antibiotic dilution series of NisinAB–CBP (4), nisin, and daptomycin in cation-adjusted TSB medium (50 mg L-1 CaCl2, 10 mg L-1 MgCl2). After overnight incubation at 37°C, the bacterial cultures corresponding to 0.25xMIC were diluted 1:100 into the newly prepared antibiotic dilution series. The experiment was performed with biological replicates and for either replicate the MIC was determined in duplicated. After 30 passages, cross-resistance for all 8 passaged strains was assessed without incubation

117


in drug-free medium. Antibiotic challenge to daptomycin led in one of the replicates to a ≥512 fold increase of MIC corresponding to the highest concentration tested.

118



Table 3. NMR chemical shift assignment compound nisinAB (2) (5.0 mM in DMSOd6).

Residue NH Hα(Cα) Additional (1H, 13C)

Ile1 7.64/8.08 3.79 (56.6) CβH (1.89, 35.9), CγH2 (1.18, 23.5), CγH3 (0.98, 14.2), CδH3 (0.89, 10.8)

Dhb2 8.16 - CβH (6.31, 126.8), CγH3 (1.71, 12.8)

D-Alas3 8.17 4.39 (54.1) CβH2 (2.94/3.03, 34.0)

Ile4 7.94 4.10 (58.8) CβH (2.02, 33.3), CγH2 (1.08/1.37, 24.1), CγH3 (0.88, 15.7),

CδH3 (0.78, 10.3)

Dha5 8.84 - CβH2 (5.51/6.00, 104.1)

Leu6 8.78 4.23 (51.1) CβH2 (1.63-1.70, 37.6), CγH2 (1.55, 23.8), Cδ1H3 (0.85, 21.5),

Cδ2H3 (0.89, 22.6)

Alas7 7.95 4.64 (54.0) CβH2 (2.79/3.07, 34.4)

D-Abus8 8.27 4.97 (56.9) CβH (3.45, 48.2), CγH3 (1.18, 21.8)

Pro9 - 4.18 (63.0) CβH2 (1.64/2.24, 28.2), CγH2 (1.75/1.88, 25.4), CδH2 (3.26/3.40, 47.7)

Gly10 8.19 3.41/4.16 (41.7) -

Alas11 7.59 3.84 (53.4) CβH2 (2.97/3.39, 36.4)

Lys12 8.21 4.18 (51.1) CβH2 (1.55/1.74, 30.2), CγH2 (1.34, 21.8), CδH2 (1.52, 26.1),

CεH2 (2.76, 38.5), NH2 (7.65)

Figure 11. Assigned HSQC spectrum for nisinAB (2) (5.0 mM in DMSOd6).

7 6 5 4 3 2 1

F2 ppm

120

110

100

90

80

70

60

50

40

30

20

10

0

F1

ppm

Dhb2β

Dha5β Dha5β

*

D-Abu8αAs

7α D-As3α

L6αK12αG10α

P9α

I4α

As11α

I11α

D-Abu8α

P9δ

As11β

D-As3β

As7β

K12ε

P9β

L6β

I4β

I1β

Dha5γ

P9γ

K12δK12δ

I4δI1δ

I1γ I4γ

L6δ1

L6δ2

K12γ D-Abu8γ

L6γ

I4γI1γ

130

119


Residue NH Hα(Cα) Additional (1H, 13C)

Ile1 7.66/8.11 3.80 (56.4) CβH (1.89, 35.8), CγH2 (1.18, 23.5), CγH3 (0.97, 14.2), CδH3 (0.88, 10.9)

Dhb2 8.16 - CβH (6.29, 126.1), CγH3 (1.71, 12.8)

D-Alas3 8.17 4.38 (53.9) CβH2 (2.92/3.02, 33.9)

Ile4 7.92 4.09 (57.6) CβH (2.02, 33.3), CγH2 (1.07/1.37, 24.1), CγH3 (0.88, 15.7),

CδH3 (0.78, 10.4)

Dha5 8.86 - CβH2 (5.50/6.00, 103.6)

Leu6 8.77 4.22 (51.4) CβH2 (1.63-1.69, 37.5), CγH2 (1.55, 23.8), Cδ1H3 (0.84, 21.5),

Cδ2H3 (0.88, 22.6)

Alas7 7.95 4.63 (53.8) CβH2 (2.77/3.06, 34.3)

D-Abus8 8.26 4.94 (56.7) CβH (3.41, 48.0), CγH3 (1.14, 21.7)

Pro9 - 4.19 (62.8) CβH2 (1.64/2.24, 28.2), CγH2 (1.76/1.88, 25.4), CδH2 (3.26/3.48, 47.6)

Gly10 8.16 3.34/4.20 (41.6) - Alas

11 7.62 3.85 (53.4) CβH2 (2.95/3.42, 36.3)

Lys12 8.13 4.26 (51.9) CβH2 (1.53/1.72, 31.1), CγH2 (1.31, 21.8), CδH2 (1.51, 26.2),

CεH2 (2.76, 38.5), NH2 (7.66)

Lipid 8.46 4.31 (41.4) (CγH)2 (7.33, 127.4), (CδH)2 (7.63, 126.2), (CηH)2 (7.51, 128.6),

(CθH)2 (7.68, 128.0)

Table 4. NMR chemical shift assignment compound nisinAB-CBP (4) (5.7 mM in DMSOd6).

8

8

6

6

4

4

2

2

ω2 - 1H (ppm)

150 150

100 100

50 50

0 0

ω1

- 13C

(pp

m)

Spectrum: SNP011028-HSQCC13-09062017User: laurenskleijn Date: Tue Jun 27 16:47:49 2017Positive contours: low 1.00e+05 levels 14 factor 1.40

Dha5γ I4δI1δI1γ

I4γL6δ1

L6δ2

K12γD-Abu8γ

I1γ

P9β

L6β

I4β

I1β

P9γ

K12δK12δ

L6γ

I4γAs11β

D-As3β As

7β

K12εD-Abu8α

As7α

D-As3α L6αK12α G10α

P9α

I4α

As11α

I11α D-Abu8α

P9δ

* **

*

*

Lipid Hα *

*

*

*

Dhb2β

Dha5β Dha5β

Lipid Hγ

Lipid HηLipid Hθ

Lipid Hδ

Figure 12. Assigned HSQC spectrum for nisinAB-CBP (4) (5.7 mM in DMSOd6).

120


Figure 13. Assigned COSY spectrum for nisinAB-CBP (4) (5.7 mM in DMSOd6).

9.0

9.0

8.5

8.5

8.0

8.0

7.5

7.5

ω2 - 1H (ppm)

8 8

6 6

4 4

2 2ω

1 - 1 H

(pp

m)

Spectrum: SNP011028-DQFCOSY-09062017User: laurenskleijn Date: Tue Jun 27 11:04:32 2017Positive contours: low 2.00e+06 levels 14 factor 1.40

Leu6 NH

Gly10 NH

D-Ala3 NH

Ala11 NH

D-Abu8 NH

Lipid NH

Ile1 NHLys12 NH2Lys12 NH

Ala7 NHIle4 NH

Leu6 Hα

Lipid Hα

D-Abu8 Hα

Gly10 Hα

D-Ala3 Hα

Ile1 HαLys12 Hα

Ala7 Hα

Ile4 Hα

Ala11 Hα

Lys12 Hε

Lipid-Hη

Lipid-HδLipid-Hθ

Lipid-Hγ

121


References[1] G. D. Wright, Nat. Prod. Rep. 2017, 34, 694–701.[2] E. Gross, J. L. Morell, J. Am. Chem. Soc. 1971, 93, 4634–4635.[3] E. O. W. L A Rogers, J. Bacteriol. 1928, 16, 211.[4] P. G. Arnison, M. J. Bibb, G. Bierbaum, A. A. Bowers, T. S. Bugni, G. Bulaj, J. A.

Camarero, D. J. Campopiano, G. L. Challis, J. Clardy, et al., Nat. Prod. Rep. 2013, 30, 108–160.

[5] P. J. Knerr, W. A. van der Donk, Annu. Rev. Biochem. 2012, 81, 479–505.[6] W. Brumfitt, M. R. J. Salton, J. M. T. Hamilton-Miller, J. Antimicrob. Chemother.

2002, 50, 731–734.[7] A. Bartoloni, A. Mantella, B. P. Goldstein, R. Dei, M. Benedetti, S. Sbaragli, F.

Paradisi, J Chemother 2004, 16, 119–121.[8] N. E. Kramer, E. J. Smid, J. Kok, B. de Kruijff, O. P. Kuipers, E. Breukink, FEMS

Microbiol. Lett. 2004, 239, 157–161.[9] E. Breukink, I. Wiedemann, C. Van Kraaij, O. P. Kuipers, H. G. Sahl, B. de Kruijff,

Science 1999, 286, 2361–2364.[10] S.-T. D. Hsu, E. Breukink, E. Tischenko, M. A. G. Lutters, B. de Kruijff, R. Kaptein,

A. M. J. J. Bonvin, N. A. J. van Nuland, Nat Struct Mol Biol 2004, 11, 963–967.[11] E. Breukink, B. de Kruijff, Nat. Rev. Drug Discov. 2006, 5, 321–332.[12] K. M. Scherer, J.-H. Spille, H.-G. Sahl, F. Grein, U. Kubitscheck, Biophys. J. 2015,

108, 1114–1124.[13] A. J. Van Heel, M. Montalban-Lopez, O. P. Kuipers, Expert Opin Drug Metab

Toxicol 2011, 7, 675–680.[14] T. Koopmans, T. M. Wood, P. 't Hart, L. H. J. Kleijn, A. P. A. Hendrickx, R. J. L.

Willems, E. Breukink, N. I. Martin, J. Am. Chem. Soc. 2015, 137, 9382–9389.[15] R. D. G. Cooper, N. J. Snyder, M. J. Zweifel, M. A. Staszak, S. C. Wilkie, T. I. Nicas,

D. L. Muller, T. F. Butler, M. J. Rodriguez, B. E. Huff, et al., J. Antibiot. 1996, 49, 575–581.

[16] E. P. Balli, C. A. Venetis, S. Miyakis, Antimicrob. Agents Chemother. 2014, 58, 734–739.

[17] D. Soule, M. M. Climo, Curr Treat Options Infect Dis 2016, 1–14.[18] P. A. Moise, G. Sakoulas, J. A. McKinnell, K. C. Lamp, D. D. DePestel, M. J. Yoon,

K. Reyes, M. J. Zervos, Clin Ther 2015, 37, 1443–1453.e2.[19] H. Hanberger, L. E. Nilsson, R. Maller, B. Isaksson, Antimicrob. Agents

Chemother. 1991, 35, 1710–1716.[20] G. Sakoulas, A. S. Bayer, J. Pogliano, B. T. Tsuji, S. J. Yang, N. N. Mishra, V. Nizet,

M. R. Yeaman, P. A. Moise, Antimicrob. Agents Chemother. 2012, 56, 838–844.[21] G. A. McKay, S. Beaulieu, F. F. Arhin, A. Belley, I. Sarmiento, T. Parr, G. Moeck, J.

Antimicrob. Chemother. 2009, 63, 1191–1199.[22] G. Bierbaum, H.-G. Sahl, Arch. Microbiol 1985, 141, 249–254.[23] L. L. Ling, T. Schneider, A. J. Peoples, A. L. Spoering, I. Engels, B. P. Conlon, A.

Mueller, T. F. Schäberle, D. E. Hughes, S. Epstein, et al., Nature 2015, 517, 455–459.

[24] N. Cotroneo, R. Harris, N. Perlmutter, T. Beveridge, J. A. Silverman, Antimicrob. Agents Chemother. 2008, 52, 2223–2225.

[25] E. Severina, A. Severin, A. Tomasz, J. Antimicrob. Chemother. 1998, 41, 341–347.[26] C. C. Kietzman, G. Gao, B. Mann, L. Myers, E. I. Tuomanen, Nat Commun 2016,

7, 10859.[27] M. B. Tol, D. Morales Angeles, D.-J. Scheffers, Antimicrob. Agents Chemother.

2015, 59, 3683–3686.

122


[28] H. E. Hasper, N. E. Kramer, J. L. Smith, J. D. Hillman, C. Zachariah, O. P. Kuipers, B. de Kruijff, E. Breukink, Science 2006, 313, 1636–1637.

[29] K. L. Palmer, A. Daniel, C. Hardy, J. Silverman, M. S. Gilmore, Antimicrob. Agents Chemother. 2011, 55, 3345–3356.

[30] Standard, Approved, and Ninth Edition. "CLSI Document M07-A9.", Wayne, PA: Clinical and Laboratory Standards Institute, 2012.

[31] C. A. Schneider, W. S. Rasband, K. W. Eliceiri, Nat. Methods 2012, 9, 671–675.[32] R. Rink, J. Wierenga, A. Kuipers, L. D. Kluskens, A. J. M. Driessen, O. P. Kuipers,

G. N. Moll, Appl. Environ. Microbiol. 2007, 73, 5809–5816.[33] L. H. J. Kleijn, S. F. Oppedijk, P. 't Hart, R. M. van Harten, L. A. Martin-Visscher,

J. Kemmink, E. Breukink, N. I. Martin, J. Med. Chem. 2016, 59, 3569–3574.

123


124

Chapter 5

125

Summary / Samenvatting

Chapter 6

Summary Samenvatting

126

Summary / Samenvatting Chapter 6

Summary

The cyclic lipopeptides comprise a number of clinically relevant classes of antibiotics that date back from the discovery of the polymyxins in 1947 to the recent introduction of the lipoglycopeptides. These natural products and natural product derivatives generally originate from soil-inhabiting producing organisms. They consist of peptide macrocycles that are modified with a lipid, and show great structural diversity owing to their non-standard biosynthetic pathways involving non-ribosomal synthesis and/or post-translational modifications. In recent years the number of newly discovered classes of lipopeptide antibiotics with unique modes of action has declined and those that are found are generally not suitable for direct advancement to clinical development. However, natural products can be used as starting point to extend the scope of cyclic lipopeptide antibiotics as exemplified by the clinical approval of the semi-synthetic lipoglycopeptides telavancin, oritavancin and dalbavancin. Chapter 1 presents a summary of the main classes of cyclic lipopeptide antibiotics with regard to their characteristic structural features, modes of action, clinical relevance and the onset of bacterial resistance.

One of the clinically used cyclic lipopeptide antibiotics is daptomycin, which belongs to the class of calcium dependent antibiotics (CDAs) (Fig. 1). Daptomycin is a blockbuster drug with potent activity against Gram-positive pathogens most notably against methicillin-resistant S. aureus. Chapter 2 reports a synthetic route for the preparation of analogs of daptomycin and the CDAs. Several variants of daptomycin and their mirror-images were generated and evaluated with regard to their antimicrobial activities. In order to obtain these analogs, we developed a synthetic route for the preparation of the unusual amino acid kynurenine that is present in daptomycin (Fig. 1, residue 13). The synthesis route developed allows for

O

HN

O

NH O

HN

O

NH O

HN

NH

O

HN

O

NH

O NHO

HN

OHNNH

O

O

HN

OO

O OHO

NH2OH

O

OH

O

OHNH2

NH

O

NH2

O

OHO

12

3

4

5

6 7 8

9

10

111213

D

D

DDaptomycin

Figure 1. The structure of clinically used antibiotic daptomycin.

127


the gram-scale preparation of both l- and d-kynurenine from the corresponding l- or d-tryptophan.

Apart from the “lipodepsipepides” subclass of CDAs to which daptomycin belongs, there exists the “lipopeptides” subclass that is currently not represented in the clinic. Laspartomycin C (Fig. 2) is member of the lipopeptide CDA family. In chapter 3, we described the first total synthesis of laspartomycin C and detail its previously unknown mode of action. Laspartomycin C binds undecaprenyl phosphate (C55-P) and inhibits formation of the bacterial cell wall precursor lipid II. In further assessing this interaction, we report the first ITC-based characterization of a C55-P-targeting CDA and determined the low nanomolar Kd value associated with laspartomycin C’s binding to its phospholipid target.

To understand the molecular recognition that is at play in laspartomycin C’s interaction with C55-P we obtained a crystal structure of the antibiotic bound to both calcium and its biomolecular target (chapter 4). The complex revealed by the high-resolution crystal structure consists of laspartomycin C bound to two calcium ions and geranyl phosphate (C10-P), a truncated variant of bacterial target C55-P (Fig. 3). The crystal structure reveals the role of the positively charged calcium ions in establishing a stable complex between the negatively charged laspartomycin C peptide and the negatively charged phosphate head group present in C55-P. Oxygen atoms in the phosphate group interact with both the peptide backbone and the calcium ions. The crystal structure is the first that details the interaction between a CDA and its bacterial target. The antibiotic/target interaction that we observed in the crystal structure implies that the mirror-image of laspartomycin C would interact with C55-P in identical manner resulting in equal antibiotic potency. To test this we synthesized the enantiomer of laspartomycin C. The mirror-image variant of laspartomycin C was indeed equally active and demonstrated the same interaction

O

HN

O

NH O

NNH

O

HN

O

NH

O NHO

HN

OHNNH

O

ON

NH

OH

O

OH

O

OH

O

OH

O

O

1

2

3

4 5 6

7

8

91011

D

D

Laspartomycin C

OP

O OH

OH

7

3

undecaprenyl phosphate (C55-P)

Figure 2. The structure of calcium dependent antibiotic laspartomycin C and its bacterial target C55-P.

128


with calcium and C10-P based on circular dichroism analysis. In addition, we found that laspartomycin C is selective for phosphate monoester binding and as such does not interact with inorganic phosphate, which is present in the blood of mammals. Laspartomycin C is however, capable of binding phosphate monoesters other than C55-P (e.g. phosphatidic acid) that are predominantly present in intracellular spaces in mammals, and therefor outside of reach of the lipopeptide that is not suspected of possessing membrane permeating properties.

Another natural product antibiotic that is currently not clinically used is the well-studied lantibiotic nisin. A fermentation product of Lactococcus lactis, nisin was first reported on in 1928. Of particular note is nisin’s ability to act as a rapidly bactericidal agent via a unique lipid II mediated mode of action. Despite nisin’s widespread use as food preservative for decades few reports of bacterial resistance are known. In spite of these promising features, nisin suffers from stability issues most prominently due to degradation by proteases that are present in living systems. To utilize the antibiotic properties of nisin while addressing its drawbacks, a novel class of nisin-derived lipopeptide antibiotics was conceived in the Martin group. Central to the nisin-derived lipopeptides is the nisin A/B fragment (nisinAB), which can be produced by enzymatic degradation of full-length nisin. NisinAB can be synthetically modified with a lipid to generate nisinAB lipopeptides with antibiotic activity on par with that of nisin (Fig. 4). In chapter 5, we describe the synthesis and mode of action of a new nisinAB lipopeptide. Like nisin itself, the nisinAB lipopeptides operate via a similar, but not identical lipid II mediated action mechanism that involves binding to the pyrophosphate region (Fig. 4). Upon

A

B

C10-PLipid tail

Asp1 Asp7

Gly6

Gly8Asp5

Ile10

Dab2

Figure 3. A high-resolution (1.28 Å) crystal structure was obtained for laspartomycin C in complex with two calcium ions and its phospholipid target C10-P (the soluble variant of C55-P). Only the interactions involving Ca2+ are shown.

129


binding lipid II, the nisinAB lipopeptide kills bacteria via a multifaceted mode of action involving both the formation of membrane-damaging lipid II aggregates and cell wall synthesis inhibition. Furthermore, serial passage assays with a vancomycin-resistant enterococcal stain revealed that resistance to the nisinAB lipopeptide does not readily appear even after prolonged (30 days) exposure to the antibiotic.

In summary, the work in this thesis focuses on the characterization of lipopeptide antibiotics that operate via underexplored and underexploited modes of action. Following chemical synthesis and microbiological evaluation of lipopeptide antibiotics, biophysical and biochemical tools were employed for mode of action studies ranging from studies aimed at unraveling the molecular recognition mechanism to the effect on whole bacterial cells.

Samenvatting

Sinds de ontdekking van de polymyxins in 1947 hebben meerdere klassen cyclische lipopeptiden toepassing gevonden als klinisch antibioticum. Deze lipopeptiden zijn natuurlijke stoffen of afgeleiden daarvan die doorgaans door in de grond levende bacteriën worden aangemaakt. De productie van deze lipopeptiden gaat gepaard met bacteriële synthese methodiek die verder rijkt dan ribosomale biosynthese met een overvloed aan structurele diversiteit tot gevolg. Het aantal nieuw ontdekte natuurlijke stoffen met antibiotische werking neemt echter af. Ook zijn natuurlijke stoffen vaak ongeschikt voor het gebruik als klinisch medicijn vanwege toxiciteit, labiliteit of onvoldoende activiteit. Evenwel kunnen deze natuurlijke stoffen chemisch aangepast worden met optimalisatie van de vereiste medicinale

OOAcHN

O

HN OOPO

O

OH

O

HONHAc

HO

PO

O

OHHO

OH

NisinAB

Lipid II

HN

O

HN OH2N

HN

OHN

O

NH O

HN

HN O

O

HNS

S

ON

O

NHHN

O

O

HNO

NH2

Lipid

NisinAB lipopeptides

A

B7

3

L-Ala

D-GluL-Lys

D-AlaD-Ala

Figure 4. General structure of nisinAB lipopeptides that act on lipid II (left). The structure of bacterial lipid II with the pyrophosphate region, which is bound by the nisin A/B system, highlighted in yellow (right).

130


eigenschapen als doel. De recentelijk in gebruik genomen medicijnen telavancine, oritavancine en dalbavancine dienen hiertoe als voorbeeld. Hoofdstuk 1 bevat een overzicht van de karakteristieken en werkingsmechanismen van de voornaamste klassen lipopeptiden antibiotica, alsmede een beschrijving van het klinische gebruik en de ontwikkeling van bacteriële resistentie.

Het lipopeptide daptomycine is een veelgebruikt klinisch antibioticum dat met name wordt toegepast voor de behandeling van infecties veroorzaak door al dan niet methicilline-resistente Staphylococcus aureus (Fig. 1). Hoofdstuk 2 behandelt de synthese van nieuwe daptomycine varianten en het testen van de antibacteriële activiteit. Voor het verkrijgen van het niet-proteinogene aminozuur kynurenine dat in daptomycine aanwezig is (Fig. 1, aminozuur 13) werd een synthese route voor optisch zuiver kynurenine ontwikkeld, uitgaande van l- of d tryptofaan als startmateriaal.

Daptomycine behoort tot de klasse van antibiotica waarvan de werking calcium-afhankelijk is en tot de subklasse van lipodepsipeptiden. Tot die klasse behoort ook een subklasse bestaande uit natuurlijke stoffen zonder klinische toepassing, de subklasse van lipopeptiden. Tot deze subklasse behoort laspartomycine C (Fig. 2). Hoofdstuk 3 behandelt het bepalen van het tot nu toe onbekende werkingsmechanisme van laspartomycine C. Ook bevat het de synthese van laspartomycine C, tevens de eerste synthese van een stof uit deze klasse. Laspartomycine C verstoort de celwand synthese door het essentiële membraancomponent undecaprenylfosfaat (C55-P) aan zich te binden. Met isotherme titratie calorimetrie kon worden bepaald dat de binding tussen laspartomycine C en C55-P een hoge affiniteit heeft (een laag-nanomolaire dissociatieconstante).

Om tot een begrip te komen van het moleculaire herkenningsmechanisme tussen laspartomycine C en C55-P werd een kristalstructuur verkregen (hoofdstuk 4). De hoge-resolutie kristal structuur bestaat naast laspartomycine C uit twee calcium ionen en geranylfosfaat (C10-P), een oplosbare variant van het bacteriële target C55-P (Fig. 3). Het moleculaire complex legt de rol van de positief geladen calcium ionen bloot bij het stabiliseren van de interactie tussen de beide negatief geladen bindingspartners C10-P en laspartomycine C. Zuurstof atomen van de fosfaat groep gaan zowel met de peptide backbone als met de calcium ionen interacties aan. De verkregen kristalstructuur is de eerste die de interactie tussen een antibioticum met een calcium afhankelijk werkingsmechanisme en het bacteriële target weergeeft. De interactie tussen laspartomycin C en de bacteriële bindingspartner is gespeend van een chirale component wat impliceert dat de enantiomere vorm van

131


laspartomycine C tot een identieke interactie in staat is resulterende in een identieke antibacteriële activiteit. Dit kon worden bevestigd na het synthetiseren van de spiegelbeeld isomeer van laspartomycince C middels het testen van de antimicrobiële activiteit en het bestuderen van de interactie met C10-P met circulair dichroïsme. Ook werd de selectiviteit van laspartomycine C voor het binden van fosfaat mono-esters vastgesteld op basis van de kristalstructuur wat in overeenkomst is met het experimentele resultaat dat er geen interactie plaatsvind met het anorganisch fosfaat dat in de bloedbaan van zoogdieren aanwezig is. Wel is laspartomycine C in staat andere in zoogdieren aanwezige fosfaat mono-esters te binden zoals het intracellulair aanwezige fosfatidezuur, ware het niet dat aan laspartomycine C geen membraan doorkruisende eigenschappen worden toegedicht.

Een andere natuurlijke stof met antibiotische eigenschappen zonder klinische toepassing is nisine, dat voor het eerst in 1928 werd beschreven als fermentaat van de melkzuurbacterie Lactococcus lactis. Nisine heeft sterke bacteriedodende eigenschappen en ondanks het wijdverspreide gebruik van nisine als conserveermiddel is er geen noemenswaardige bacteriële resistentie opgetreden. Desondanks is nisine ongeschikt voor ontwikkeling als medicijn voor systemische toediening met name vanwege gevoeligheid voor afbraak door in het lichaam aanwezige proteasen. Om de ongunstige eigenschappen van nisine te adresseren met behoud van de sterke antibacteriële activiteit werden in het Martin laboratorium de klasse van nisin-afgeleide lipopeptiden gegenereerd. Deze lipopeptiden bestaan uit met lipiden gemodificeerde varianten van het A/B ringsyteem (nisineAB) dat uit nisine geïsoleerd kan worden (Fig. 4). Hoofdstuk 5 bevat een beschrijving van de synthese en een uitgebreide studie naar het werkingsmechanisme van een nisinAB lipopeptide. Het nisineAB lipopeptide werkt in op membraancomponent ‘lipid II’ (Fig. 4) op vergelijkbare, maar niet identieke wijze als nisine: het werkingsmechanisme is tweeledig en omvat zowel het vormen van membraan verzwakkende lipid II-aggregaten als het verstoren van de celwand synthese. Ook werd getracht mutanten van vancomycineresistente enterokok te genereren met resistentie tegen het nisinAB lipopeptide met als veelbelovend resultaat dat dit niet gelukt is.

Samenvattend, het werk in dit proefschrift omvat onderzoek naar antibiotische lipopeptiden met klinisch-onbenutte werkingsmechanismen. Volgend op chemische synthese en microbiologische analyse werden biofysische en biochemische technieken aangewend om tot mechanistische inzichten te komen variërend van moleculaire herkennings mechanismen tot aan de werking op bacteriële cellen.

132

Chapter 6

133

Appendices

Appendices

List of Publications Curriculum Vitae

Acknowledgements

134

List of Publications

From this thesis

Kleijn*, L. H. J., Vlieg*, H. C., Wood, T. M, Sastre Toraño, J., Janssen†, B. J. C., Martin†, N. I. (2017) A High-Resolution Crystal Structure that Reveals Molecular Details of Target Recognition by the Calcium-Dependent Lipopeptide Antibiotic Laspartomycin C. Angew. Chem. doi: 10.1002/ange.201709240

Kleijn, L. H. J., and Martin, N. I. (2017) The Cyclic Lipopeptide Antibiotics. In: Topics in Medicinal Chemistry. Springer, Berlin, Heidelberg.

Kleijn, L. H. J., Oppedijk, S. F., ‘t Hart, P., van Harten, R. M., Martin-Visscher, L. A., Kemmink, J., Breukink, E., and Martin, N. I. (2016) Total Synthesis of Laspartomycin C and Characterization of Its Antibacterial Mechanism of Action. J. Med. Chem. 59, 3569–3574.

Koopmans, T., Wood, T. M., ‘t Hart, P., Kleijn, L. H. J., Hendrickx, A. P. A., Willems, R. J. L., Breukink, E., and Martin, N. I. (2015) Semisynthetic Lipopeptides Derived from Nisin Display Antibacterial Activity and Lipid II Binding on Par with That of the Parent Compound. J. Am. Chem. Soc. 137, 9382–9389.

‘t Hart*, P., Kleijn*, L. H. J., de Bruin, G., Oppedijk, S. F., Kemmink, J., and Martin, N. I. (2014) A Combined Solid- and Solution-Phase Approach Provides Convenient Access to Analogues of the Calcium-Dependent Lipopeptide antibiotics. Org. Biomol. Chem. 12, 913–918.

Kleijn, L. H. J., Müskens, F. M., Oppedijk, S. F., de Bruin, G., and Martin, N. I. (2012) A Concise Preparation of the Non-Proteinogenic Amino Acid l-Kynurenine. Tetrahedron Lett. 53, 6430–6432.

Other

Rossin, R., van Duijnhoven, S. M. J., Hoeve, ten, W., Janssen, H. M., Kleijn, L. H. J., Hoeben, F. J. M., Versteegen, R. M., and Robillard, M. S. (2016) Triggered Drug Release from an Antibody–Drug Conjugate Using Fast “Click-to-Release” Chemistry in Mice. Bioconjugate Chem. 27, 1697–1706.

135

Manuscripts in preparation

Kleijn, L. H. J., Morales Angeles, D., Viveen, M. C., Willems, R. W. L., Scheffers, D.-J., Martin, N. I. Working title: Antibacterial Properties of a Lipid II Targeting Nisin-Derived Lipopeptide

Bolt*, H. L., Kleijn*, L. H. J., Martin†, N. I. Cobb†, S. L. Synthesis of Antibacterial Nisin-Peptoid Hybrids Using Click Methodology. Manuscript submitted.

* Denotes shared first authorship, † denotes equal corresponding authorship

Patent

Kleijn, L. H. J., Koopmans, T., Martin, N. I, Wood, T. M. (2016) “Nisin-based compounds and use thereof in the treatment of bacterial infections”, WO2016116379A1.

136

Curriculum Vitae

Laurens Henri Johan Kleijn, borne on April 12th 1989 in New York City, attended Barlaeus grammar school in Amsterdam after which he studied chemistry at Utrecht University. Having developed and interest in organic and medicinal chemistry, he continued his studies in Utrecht to obtain a cum laude masters degree in Drug Innovation. He was awarded an NWO graduate grant PhD scholarship to fund his PhD research focused on studying lipopeptide antibiotics. He co-founded Karveel Pharmaceuticals, which aims to develop first-in-class antibiotics, and acts as CSO. He is married and has a son.

synthetic and mechanistic studies with lipopeptide antibiotics

Documents