Distinct Spatiotemporal Dynamics of Peptidoglycan Synthesis between Mycobacterium smegmatis and Mycobacterium tuberculosis

ABSTRACT Peptidoglycan (PG), a polymer cross-linked by d-amino acid-containing peptides, is an essential component of the bacterial cell wall. We found that a fluorescent d-alanine analog (FDAA) incorporates chiefly at one of the two poles in Mycobacterium smegmatis but that polar dominance varies as a function of the cell cycle in Mycobacterium tuberculosis: immediately after cytokinesis, FDAAs are incorporated chiefly at one of the two poles, but just before cytokinesis, FDAAs are incorporated comparably at both. These observations suggest that mycobacterial PG-synthesizing enzymes are localized in functional compartments at the poles and septum and that the capacity for PG synthesis matures at the new pole in M. tuberculosis. Deeper knowledge of the biology of mycobacterial PG synthesis may help in discovering drugs that disable previously unappreciated steps in the process.

R od-shaped bacteria like Escherichia coli and Bacillus subtilis elongate by adding new peptidoglycan (PG) along the lateral body (1)(2)(3)(4). This so-called dispersed cell growth contrasts with zonal cell growth, which restricts the addition of cell building blocks to specific locations, including the cell poles. Polar growth prevails in Gramnegative Rhizobiales and in Gram-positive Actinomycetales, which include mycobacterial species (5)(6)(7)(8)(9)(10)(11). The selective advantage of polar growth over dispersed growth in certain species remains unknown, but it has been suggested that such growth may foster cell-to-cell heterogeneity and favor rejuvenation of progeny by selectively passing on intact proteins and cell wall components to newborn cells (8,(12)(13)(14).
PG is a rigid exoskeleton composed of strands of repeating units of the disaccharide N-acetyl-glucosamine (NAG)-N-acetyl-muramic acid (NAM) cross-linked by peptide chains (Fig. 1A). Penicillin and other ␤-lactams, such as carbapenems and cephalosporins, which commonly target the bacterial cell wall by halting peptidoglycan biosynthesis in replicating bacteria, are the antibiotics most widely used to treat Gram-positive and Gram-negative bacterial infections. ␤-Lactams often cause bacterial lysis (15,16). Disruption of the balance between PG synthesis and PG rupture by hydrolases generates a futile cycle that depletes cellular resources, contributing to cell death (17).
Despite the importance in medicine of antibiotics that target PG, the lack of adequate tools hindered deeper investigation and improved comprehension of PG dynamics, especially in live organisms, until the development of fluorescent D-alanine analogs (FDAA) (9,(24)(25)(26)(27)(28)(29)(30)(31). Two routes ensure the incorporation of FDAAs into bacterial PG: periplasmic editing of mature PG by L,D-transpeptidases or D,D-transpeptidases and cytosolic incorporation into PG precursors by intracellular ligases. Using chemical reporters of growth-either an FDAA or an amine-reactive dye-it was reported that Mycobacterium smegmatis elongates preferentially from the old pole (6-9), although this was disputed (32). However, the mode of elongation of wild-type (WT) M. tuberculosis remained largely unexplored.
Here, we used FDAA (Fig. 1B) to delineate fundamental differences in the modes of elongation of M. smegmatis and M. tuberculosis. A better definition of the repertoire of the synthetic machineries that incorporate PG into mycobacteria will help rationalize the modes of action of the growing number of ␤-lactam antimicrobials under consideration for use in TB therapy.  incorporation index was 0.09. In contrast to M. smegmatis, M. tuberculosis incorporated FDAAs heterogeneously: predominantly at one pole in some cells and at both poles in others ( Fig. 4B and C). Also in contrast to M. smegmatis, the incorporation index of HADA tended to be smaller in longer M. tuberculosis cells (Fig. 4C to E). This was indicative of variations in polar dominance as a function of cell cycle.
The quantification of FDAA incorporation into growing cells leads us to propose that M. smegmatis and M. tuberculosis have distinct modes of elongation (Fig. 4F). FDAA incorporation dominates at one pole in elongating M. smegmatis cells independently of their length. This is consistent with numerous reports that M. smegmatis elongates preferentially from the old pole (7,8,10,39). In contrast, FDAA incorporation into M. tuberculosis varies in polar dominance as a function of the cell's length. Considering that cell length increases from birth to separation into a progeny pair, we propose that FDAAs are incorporated chiefly at one pole immediately after cytokinesis but comparably at both poles just before cytokinesis (Fig. 4F). These differences likely reflect variations in PG synthetic pathways between the two species. Differences have been reported in the two species' penicillin-binding proteins (PBPs). M. smegmatis encodes three bifunctional PBPs, PonA1 to PonA3, whereas M. tuberculosis encodes only PonA1 and PonA2 (40). M. smegmatis requires PonA1 for in vitro growth, while the PonA1 homolog is dispensable in M. tuberculosis (41,42). However, such differences by themselves do not provide an explanation for our finding that in M. tuberculosis, the pole that incorporates FDAAs at a lower rate in a progeny cell at birth matures over the course of the cell cycle to become a pole from which cells elongate at a maximal rate under optimal conditions. The molecular mechanisms by which a cell pole matures remain poorly understood, although some studies have provided some routes of investigation. First, Meniche and colleagues documented that Wag31 is preferentially concentrated at the old pole in M. smegmatis when it is expressed at a native level (9). Thus, Wag31 may increasingly accumulate at the incipient pole to help recruit and anchor the elongation complex. Mycobacteria may deploy mechanisms to control the sequestration of Wag31 at the old pole or exclude Wag31 from the new pole. Second, the activity and/or the control of the location of the elongation machinery may be dictated by posttranslational modifications. In this vein, phosphorylation of the peptidoglycan synthase PonA1 governs the rate of polar elongation in M. smegmatis (43). Third, Rego and colleagues recently reported that LamA, a member of the mycobacterial division complex, inhibits growth at the incipient pole, thus contributing to asymmetric polar growth (44). Their work suggests that treatment of M. tuberculosis with rifampin in combination with an inhibitor of LamA might lead to faster mycobacterial killing. Finally, Eskandarian and colleagues showed that M. smegmatis cells present cell surface irregularities, so-called wave troughs (45). M. smegmatis progeny cells inherit a wave trough located near the old pole from their mother or grandmother For short-pulsed cells, the log 10 ii documents whether the incorporation of D-alanine analogs at the poles is symmetrical (log 10 ii of~0) or asymmetrical (log 10 ii of Ͼ0). (G) Log 10 ii values of all 340 cells in panel E relative to their lengths. Each dot represents a single cell profiled from 4 independent experiments. The yellow, blue, red, and green shaded areas represent the first, second, third, and fourth size quartiles, respectively. The red dots represent the median values for size and the log 10 ii values of all cells in each quartile. The blue line indicates the Spearman correlation. Cells bearing a septum were excluded from the analysis. (H) Cumulative distribution function of the incorporation index. For each quartile color coded in panel G, the curve reports the cumulative frequency cell counts as a function of the log 10 ii. This indicates the distribution of the log 10 ii values in each quartile and allows comparisons among them. The table includes the following: the Spearman correlation coefficient; its associated P value; the Kruskal-Wallis multiple-comparison test (a significant Kruskal-Wallis test indicates that a minimum of one quartile differs from the others); and the post hoc two-by-two comparison of each quartile with another using a Tukey-Kramer procedure. For the latter statistical test, the null hypothesis is that the data in two quartiles are from the same continuous distribution (value of 0); the alternative hypothesis is that the data in two quartiles are from different continuous distributions (value of 1). The result is 1 if the test rejects the null hypothesis at the 1% significance level and 0 otherwise. at birth. Asymmetric elongation results in the repositioning of the wave trough to the midcell. Once the wave trough is at midcell, the FtsZ ring forms; septum formation begins 30 min later, cytokinesis is completed after 20 min, and separation of the progeny pair occurs 40 min later. The authors proposed a model in which the division site is placed at a wave trough of the cell and the replicated chromosome negatively regulates the placement of the division site, leaving the midcell the preferred site for division by a mechanism akin to nucleoid occlusion (45). It is possible that the spatiotemporal regulation of the aforementioned mechanisms differs between the two species.
The combination of super-resolution imaging techniques and specific biological probes also has the potential to help increase the knowledge of the modes of action of antimycobacterial compounds by allowing the comparison of morphometric parameters from cells treated with antimycobacterial compounds with unknown modes of action to those of cells treated with compounds with known modes of action (46). Additionally, such studies on drug-treated cells or mutants may help us better understand the coordination of molecular events during cell elongation and cell division.
As an example, we treated M. smegmatis and M. tuberculosis with sublethal or bactericidal concentrations of meropenem for 4 and 24 h, respectively. For both species, FDAA incorporation decreased in proportion to the meropenem concentration and meropenem caused distinct morphometric abnormalities ( Fig. 5A and C to E). In meropenem-treated M. smegmatis cells, the FDAA signal was greater at the subpolar region than in the rest of the cell, except for the septum (Fig. 5A and C). One pole was affected more than the other, producing polar clubbing, and most cells accumulated FDAAs in the subpolar region of the pole that was clubbed ( Fig. 5A and C). Moreover, the proportion of M. smegmatis cells that bore a septum increased dramatically upon meropenem treatment (Fig. 5A and B). Most septa appeared as two foci rather than a segment that oriented perpendicularly to the longitudinal axis of a cell, suggesting that septation was incomplete (Fig. 5A). This was confirmed by 3-D-SIM (Fig. 5C). This has prompted us to speculate that two separate machineries initiate the formation of a septum and ensure its closure, as has been reported in Streptococcus pneumoniae strain D29 cells (47), and that meropenem inhibits the machinery responsible for septal closure. Visual inspection of meropenem-treated M. tuberculosis cells suggested that morphological alterations varied: one pole appeared chiefly altered in some cells, and the two poles were more similarly swollen in others. Moreover, FDAAs were well incorporated at the malformed pole(s) (Fig. 5D and E).
The spatiotemporal analysis of FDAA incorporation described here raises new questions about the enzymology, topology, and temporal regulation of PG synthesis in mycobacteria, particularly about the capacity for PG synthesis to mature in the new pole of mycobacterial cells after the separation of a progeny pair. Answers to these newly framed questions may lead to the identification of new drug targets and better understanding of the mechanisms of action of existing antimycobacterial drugs. porter construct mCherry was fused to the C terminus of Wag31. The nucleotide sequence TCGGGCTC GGGCTCGGGC encoding the protein linker SGSGSG was introduced between the nucleotide sequences encoding the protein of interest and the reporter. The plasmid pGMCZ-T0X-PTB38 wag31-mCherry allowed integration of Wag31-mCherry into the chromosome and constitutive expression under the control of the strong promoter TB38. Bacterial strains and culture conditions. M. smegmatis, M. smegmatis Wag31-mCherry, and M. tuberculosis were cultured at 37°C in Middlebrook 7H9 medium (BD Difco) supplemented with 0.2% glycerol, 0.5% bovine serum albumin fraction V, 0.05% Tween 80, 0.2% dextrose, 0.085% NaCl, and 0.02% tyloxapol. Zeocin was added at a final concentration of 25 g/ml when applicable. Cultures were grown in 50-ml Falcon tubes at a 1:100 liquid/air ratio.

Construction of M. smegmatis cells expressing
Synthesis of FDAAs and FLAA. All fluorescent D-alanine analogs used in this study were synthesized following the procedure described by Kuru and colleagues (48) and were purified using a Waters Acquity ultraperformance liquid chromatography (UPLC) system equipped with a 1.7-m Acquity UPLC BEH C 18 column (1.7-m particle size, 2.1 by 100 mm) using a 10-to-30% gradient of acetonitrile-water containing 0.01% trifluoroacetic acid (TFA) over a 5-min run, including a 3-min wash. The areas under the peaks were calculated using MassLynx software (Waters, Inc.). The degrees of purity of HADA, NADA, and HALA were Ͼ97%, Ͼ97%, and Ͼ98%, respectively. Heat inactivation. Cells were heat inactivated at 90°C for 15 min prior to a 4-h incubation in 7H9 medium supplemented with HALA or HADA. No CFU were recovered after heat treatment.
High-resolution microscopy. For imaging, bacterial suspensions were deposited on soft agar pads and visualized using a DeltaVision image restoration microscope (GE Healthcare) equipped with an Olympus IX-71 microscope with a 100ϫ/1.4 numeric aperture (NA) UPlanSApo objective and appropriate filter sets (for DAPI, excitation at 390/10 and emission at 435/48, and for fluorescein isothiocyanate [FITC], excitation at 475/28 and emission at 525/48), a pco.edge scientific complementary metal oxidesemiconductor (sCMOS) camera (PCO-Tech), and an Insight SSI 7-color solid-state illumination system.

3-D-SIM super-resolution microscopy.
For super-resolution imaging, bacterial suspensions were mixed with the same volume of mounting medium (ProLong gold antifade mountant; Thermo Fisher Scientific), and 10-l amounts were spread on microscope slides (75-by 25-mm Corning 2948) and covered with high-precision microscope cover glasses (1.5H; GmbH). Super-resolution microscopy data were acquired using a DeltaVision OMX V4/Blaze 3-D-SIM super-resolution microscope (GE Healthcare) housed in the Rockefeller University Bio-Imaging Resource Center (BIRC). This OMX system is fitted with a 100ϫ/1.40 UPLSAPO oil objective (Olympus), Evolve electron-multiplying charge-coupled device (EMCCD) cameras (Photometrics) used in EM gain mode at a gain of 170, 405-nm and 488-nm excitation lasers, and 436/31 and 528/48 emission filters. Optical sections were acquired at 125-nm intervals. The immersion oil refractive index was selected to optimize for the 488 channel and the ambient temperature. Structured illumination data sets were reconstructed using softWoRx software (GE Healthcare), employing optical transfer functions (OTFs) generated from point spread functions acquired from subresolution beads, channel-specific k0 (stripe rotation angle) values, and a Wiener filter of 0.002. The image registration parameters and OTFs were refined by the BIRC staff.
PG labeling using FDAAs. M. tuberculosis and M. smegmatis cultures were inoculated at an OD 600 of 0.1 to 0.3 into 7H9 medium supplemented with 1 mM HADA or 1 mM NADA. When sequential labeling was performed, bacterial suspensions perfused with the first FDAA were washed 3 times with 7H9 medium and further incubated in 7H9 medium supplemented with the second FDAA. For both single and sequential labeling, bacterial suspensions were washed 3 times with PBS-0.05% Tween 80 and fixed with 4% paraformaldehyde for 30 min and 4 h for M. smegmatis and M. tuberculosis, respectively, to ensure bacterial death for further imaging outside a contained environment. Of note, incubation with a first FDAA prior to the incubation with a second one helped us assess whether cells grew normally; most did in growth-permissive, rich medium. As such, the first label was not strictly required and did not serve the analysis detailed below. Image analysis and statistical analysis. We evaluated the incorporation of FDAAs into cells that did not bear a septum; cells that had a septum were excluded from the analysis. As such, this study reports the FDAA incorporation of cells spanning the cell cycle period that corresponds to 0.17 (0.5/3) cycles prior to cell separation to septum formation. Therefore, a minority of cells that have been included in the analysis may have just separated. Images were processed with the open-source program ImageJ (http://imagej.nih.gov/ij/) or Icy (http://icy.bioimageanalysis.org/) (49). Data were compiled using MatLab (MathWorks). Briefly, using ImageJ, we drew a segmented line that followed the longitudinal axis of a cell and measured the fluorescence intensities along it. Next, we used customized MatLab algorithms to automatically compute the logarithm of the ratio of the total fluorescence of the two halves of a cell. All cells were processed similarly. The statistical analysis was performed using MatLab.

ACKNOWLEDGMENTS
We thank Alison North (A.N.) and Kaye Thomas at the Bio-Imaging Center (Rockefeller University) for teaching and invaluable advice and help, Nicolas Chenouard and Lisa Roux for providing help with statistical analysis, and L. Botella, K. Burns-Huang, D. Schnappinger, K. Rhee, T. Lupoli, and S. Schrader (Weill Cornell Medicine) for careful editing of the manuscript. We thank George Sukenick and Joan Subrath from the NMR analytical Core Facility at MSKCC for help with NMR and UPLC. The Department of Microbiology and Immunology is supported by the William Randolph Hearst Foundation. The work at the Organic Synthesis Core Facility at MSKCC is partially supported by NCI core grant number P30 CA008748-48.
J.V. was supported by the Abby and Howard Milstein Program in Chemical Biology and Translational Medicine. The project was supported by award number S10RR031855 from the National Center for Research Resources. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health. No funders had any role in study design, data collection and interpretation, or the decision to submit the work for publication.