Synthetic Peptidoglycan Motifs for Germination of Bacterial Spores

Certain Gram-positive bacteria—as exemplified by Bacillus anthracis, the causative agent of anthrax—can produce dormant and environmentally resistant spores under conditions of nutritional limitation. These spores exit from dormancy via the process of germination that is triggered by exposure to specific molecules.^[1] While the precise chemical nature of these molecules, known as germinants, varies according to the organism, they are typically nutrients.^[2] Recently, we reported that supernatants from cultures of growing bacteria and constituents of the cell wall could serve as germinants of dormant B. subtilis and B. anthracis spores.^[3] Since fragments of the cell wall are released in the course of bacterial growth, the presence of these molecules in the milieu as germination signals might be physiologically relevant.

The major constituent of the cell wall is the bacterial peptidoglycan. The peptidoglycan backbone is comprised of repeating units of N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM). A distinctive pentapeptide is appended to the NAM unit, which contains unusual amino acids such as D-Glu, D-Ala and meso-diaminopimelate (DAP). We describe herein the study of three Lys- (compounds 1a, 2a and 3a) and three DAP-containing peptidoglycan fragments (compounds 1b, 2b and 3b) as spore germinants (Scheme 1). This study utilizes synthetic compounds with defined structures, with variations in the number of rings (1 to 4), different muramic acid moieties (muramic acid vs anhydromuramic acid), and different amino acids (Lys vs meso-DAP) at the third position of the stem pentapeptide, which is important for the recognition events. We document for the first time that the required minimal structural motif for germination of spores is the presence of one NAG-NAM(peptide) unit. Compounds 2a/b and 3a/b that possess this minimal structural motif are potent spore germinants at the low nanomolar level.

Scheme 1.

Structures of synthetic peptidoglycans with L-Lys and meso-DAP-containing pentapeptides used for spore germination study

The DAP-containing compounds 1b, 2b, and 3b were prepared in multistep syntheses. For the synthesis of the pentapeptide, a preparation of meso-DAP was required. Whereas syntheses of DAP have been reported, only four publications have addressed the suitable functionalization of DAP for incorporation into peptidoglycan variants.^[4–7] Schmidt and coworker^[7] used the Wittig-Horner reaction of L-glutamate aldehyde and a phosphoryl glycine derivative to produce the desired C7 carboxylic acid derivative after C-C double bond reduction and chromatographic separation of diastereomers. Boons and coworkers^{[5, 6]} synthesized meso-DAP using cross metathesis between allyl glycine and vinyl glycine derivatives, followed by reduction of the double bond of the resulting compound. Fukase et al^[4] applied the Kocienski-modified Julia olefination, utilizing an aldehyde and a sulfone, both of which were derived from D-serine. The existing methods for DAP preparations proved not to be suitable in our hands for the large-scale reactions and orthogonal protection scheme that were needed. meso-Diaminopimelic acid (meso-DAP) contains two sterogenic centers with configurations of 2S and 6R. Although it is a meso compound, 2S- and 6R-carbons should be differentiated to be incorporated into the pentapeptide backbone. That is, the 2S carbon serves as the main chain C_α and will be connected to the D-Ala-D-Ala backbone and the 6R carbon will be a part of the side chain functionality. Hence, for the synthesis of meso-DAP, we needed to construct two sterogenic centers and the synthetic strategy should allow for orthogonal protection of the two carboxylates and the two amino groups. The convergent syntheses utilized the suitably protected DAP-based peptide and the corresponding saccharide units. We incorporated orthogonal protection scheme of meso-diaminopimelic acid to suit the need of our global deprotection strategy at the end of the synthesis into the method of Hernández and Martín,^[8] for construction of the template leading to the DAP structure. This approach uses L-aspartic acid in two Wittig reactions, followed by Sharpless epoxidation of allyl alcohol using Ti(O-iPr)₄ and ethyl L-tartarate and its selective epoxide ring opening with azide. This 12-step route for the preparation of known compound 4^[8] was amenable to multi-gram scale. The transformations of 4 to the desired pentapeptide 11 are depicted in Scheme 2. We converted compound 4 to the benzyl monoester 5, by periodic acid oxidation of the diol in the presence of a catalytic amount of ruthenium trichloride, followed by treatment with benzyl bromide and potassium bicarbonate. However, selective hydrolysis of the methyl ester in 5 prior to attachment of D-Ala-D-Ala proved challenging. Many conditions,^[9] including potassium trimethylsilanolate, barium hydroxide, LiBr in the presence of triethylamine, resulted in the undesirable hydrolysis of the benzyl ester and the formation of the diacid. Directed hydrolysis of the methyl ester was tried, which involves deprotection of the Boc group under acidic condition, conversion of the free amine to the trifluoroacetamide ((CF₃CO)₂O, THF), intramolecular base-catalyzed ring closure (NaH, THF) to the corresponding unstable oxazolidinone, and hydrolytic aqueous work-up to result in the corresponding carboxylic acid. Although this clever approach for selective intramolecular hydrolysis of the methyl ester has proven to be successful for a similar compound in the literature,^[10] it did not work for compound 5 in our hands. Instead of looking for different ester group for C-1, we decided to attach D-Ala-D-Ala to C-1 prior to chemistry at C-7 (oxidation and ester formation). After hydrolysis of the methyl ester in 4 using barium hydroxide, the diol was converted to the acetonide 7. The carboxylate was subsequently activated with N-hydroxysuccinimide and the fragment was coupled to D-Ala-D-Ala-OBn to give 8. The treatment of 8 with aqueous acetic acid gave the desired diol without affecting the Boc group. The resulting diol was subsequently transformed to the benzyl ester 9 via a two-step treatment (oxidative cleavage of the resultant diol, and benzyl protection of the carboxylate). The Boc group in compound 9 was subsequently removed by treatment with trifluoroacetic acid (TFA) and the resultant deprotected amine was coupled with Boc-L-Ala-γ-D-Glu(ONHS)-OBn to give the pentapeptide 10. The Boc group in 10 was removed by TFA treatment, before the coupling reaction with the suitable saccharide derivatives (vide infra).

Scheme 2.

Synthesis of meso-DAP-containing pentapeptide 11

Since the stereo center at C-6 in the meso-DAP derivatives was newly formed, its absolute stereochemistry needed to be established. The diol in 4 was oxidized to the carboxylate. Several ester and amine protective groups for C-7 carboxylate and N1, respectively, were tried in order to attempt crystallization and ultimately structure determination by X-ray. The effort was successful with compound 12, whose X-ray structure confirmed that nitrogens at C-2 and C-6 exist as syn to each other and their respective configurations were 2S and 6R, as desired.^[11]

Completion of the syntheses of compounds 1, 2 and 3 start with the orthogonally protected derivatives 13, 14 and 15, which were individually synthesized from D-glucosamine in 6, 21 and 31 steps, respectively, by methodology developed in our lab.^{[12, 13]} In each case, the lactyl moiety was activated by the formation of succinimide esters, which underwent reaction with 11 to give the corresponding peptidoglycan derivatives (compounds 16, 17, and 18, Scheme 3). The protective groups in sugar and peptide moieties were designed for global removal in one pot by treatment with acetic acid, followed by hydrogenolysis. Compounds 1b, 2b and 3b were prepared for the first time for this study in convergent syntheses involving 29, 44 and 63 synthetic steps, respectively. When L-Lys was used in place of meso-DAP in the protected pentapeptide during the coupling reaction with derivatives 13, 14 and 15, compounds 1a, 2a and 3a were produced.^{[12, 13]}

Scheme 3.

Synthesis of meso-DAP-containing fragments of the peptidoglycan 1b, 2b, and 3b

Germination of B. subtilis spores by the synthetic samples was assessed. This response requires PrkC, a well-conserved bacterial Ser/Thr membrane kinase, which contains an extracellular domain capable of binding peptidoglycan.^[3] B. subtilis peptidoglycan is DAP-based and B. subtilis spores responded only to meso-DAP-containing peptidoglycan fragments. However, B. subtilis spores lacking PrkC but expressing the PrkC homolog from S. aureus (PrkC_Sa), responded to both L-Lys- and meso-DAP-containing peptidoglycan fragments. We therefore addressed whether the identity of the residue at the third position (Lys vs meso-DAP) was responsible for this specificity by assaying the ability of synthetic muropeptides of both types to germinate spores expressing only one of the kinase homologs. The B. Subtilis system recognized only DAP-containing variants. In contrast the system with the S. aureus protein recognizes both DAP- and Lys-containing samples. The synthetic meso-DAP-containing peptidoglycan analogs 2b and 3b germinated spores expressing either native PrkC or S. aureus PrkC (Figure 1a and 1b) at an EC₅₀ of approximately 10 nM. In contrast, L-Lys-containing synthetic molecules (2a and 3a) only stimulated germination of spores expressing the PrkC_Sa homolog (Figure 1b). Thus, both classes of synthetic compounds are biologically active for the respective target proteins.

Figure 1.

Dose response curve for spore germination using the synthetic samples. Wild-type B. subtilis spores expressing the native PrkC (a) and B. subtilis spores expressing the S. aureus PrkC protein (b) were incubated with indicated concentrations of synthetic samples and the percentage germination was determined. The dashed gray line was used for EC₅₀ determination. 19a, R = Lys-pentapeptide = L-Ala-γ-D-Glu-L-Lys-D-Ala-D-Ala, 19b, R = DAP-pentapeptide = L-Ala-γ-D-Glu-m-DAP-D-Ala-D-Ala

Compounds 1a and 1b lacking the NAG residue failed to stimulate germination entirely, indicating that the presence of the NAG-NAM disaccharide is a strict minimal structural motif for germination

Hence, the third amino acid (Lys or meso-DAP) appears to be crucial for the recognition event by the given target protein and that the binding requires the presence of one NAG-NAM unit. Interestingly, the critical role of this third amino acid residue is also observed with the eukaryotic proteins that recognize peptidoglycan.^[14] For example, Nod1 is an intracellular pattern recognition molecule activated during bacterial infection that is stimulated by meso-DAP-containing peptidoglycan, but not Lys-containing peptidoglycan.^[15–17] A mechanistic explanation of such discrimination is found in the binding of two Peptidoglycan Recognition proteins (PGRPs) from Drosophila to a meso-DAP-containing muropeptide, where the side-chain carboxylate of meso-DAP forms a bidentate salt bridge with a conserved arginine residue in the PGRP.^{[18, 19]} Although there is no sequence similarity between the PGRPs and the extracellular domain of the bacterial kinase that binds peptidogylcan, future structural studies with different kinases and their respective ligands should resolve whether the kinases use a similar strategy for ligand discrimination. As much as 50% of the bacterial cell wall is turned over in the course of normal growth through the action of a muralytic family of lytic transglycosylases. These enzymes facilitate formation of a transient oxocarbenium species at the C1 of the NAM unit, which entraps the C6 hydroxyl group, resulting in what is referred to as NAG-1,6-anhydromuramyl-peptide (19a and 19b).^[20–23] This process releases two sugars at a time from the polymeric cell wall. Both 19a and 19b have been prepared in our laboratories,^[20] hence we explored if they could serve as spore germinants. The ability of 19a and 19b to stimulate germination was significantly less than that of 2 and 3 (Figure 1). This result indicates that the products of lytic transglycosylases are not involved in stimulating the kinase activity and further suggest that peptidoglycan fragments produced by other muralytic enzymes are responsible for the germination event. Secreted murein hydrolases analogous to the muralytic Resuscitation Promoting Factors, which play an important role in M. tuberculosis pathogenesis, stimulate activity of these kinases,^[24] although their reaction products have not been characterized. The maximal observed germination with these synthetic compounds (~45%) is similar to that reported previously for purified samples from digested bacterial cell wall,^[3] indicating that only a fraction of the spores might be able to respond. Nevertheless, it is clear that synthetic compounds 2 and 3 are potent stimulants of spore germination. These peptidoglycan derivatives reveal that two or more sugar moieties are required in the germinant and that the single-sugar variants and the anhydromuramyl derivatives, which are products of turnover of peptidoglycan, fail to stimulate germination. These reagent will be useful mechanistic tools for further exploration of the details of the complex processes involved in germination of bacterial spores. In addition, these molecules will be useful in elucidating the mechanisms underlying the essential function of PrkC homologs in pathogenesis and antibiotic resistance of a number of other important non-spore-forming pathogens such as S. aureus^{[25, 26]} and E. faecalis.^[27]

Experimental Section

Germination Assay

Spore germination assays were performed with compounds 1a/b-3a/b and 19a/b, at concentrations ranging from 50 μM to 1 fM.^[3] Briefly, we incubated 10⁶ B. subtilis wild-type spores or B. subtilis spores expressing PrkC_Sa with a synthetic compound in 50 μL of germination buffer (10 mM Tris-HCl, pH 8.0, 1 mM glucose) for 60 min at 37 °C. Since spores that have initiated germination, but not dormant spores, are temperature sensitive, the response to the compounds was assessed by subjecting the reactions to wet heat (80 °C, 20 min), followed by incubation at 37 °C overnight on LB agar plates to determine survival. The percentage germination was calculated by calculating the ratio of the CFUs obtained following incubation with germinant to that obtained following incubation with buffer alone. EC₅₀ values were determined for active compounds using the dose-response curves.

Syntheses of compounds 1b, 2b, and 3b

Detailed chemical syntheses of compounds 1b, 2b, and 3b are given in Supporting Information.