Exposure to a Cutinase-like Serine Esterase Triggers Rapid Lysis of Multiple Mycobacterial Species*

Background: Mycobacteria possess a unique envelope with poorly understood architectural organization of constituent molecules. Results: Exposure to an esterase that targets trehalose mycolates triggers rapid lysis of mycobacteria. Conclusion: Trehalose mycolates are on the exposed surface of mycobacterial envelope with an indispensable structural role. Significance: The study identifies a novel enzyme-based lysis of mycobacteria, with a potential for new anti-tuberculosis applications. Mycobacteria are shaped by a thick envelope made of an array of uniquely structured lipids and polysaccharides. However, the spatial organization of these molecules remains unclear. Here, we show that exposure to an esterase from Mycobacterium smegmatis (Msmeg_1529), hydrolyzing the ester linkage of trehalose dimycolate in vitro, triggers rapid and efficient lysis of Mycobacterium tuberculosis, Mycobacterium bovis BCG, and Mycobacterium marinum. Exposure to the esterase immediately releases free mycolic acids, while concomitantly depleting trehalose mycolates. Moreover, lysis could be competitively inhibited by an excess of purified trehalose dimycolate and was abolished by a S124A mutation affecting the catalytic activity of the esterase. These findings are consistent with an indispensable structural role of trehalose mycolates in the architectural design of the exposed surface of the mycobacterial envelope. Importantly, we also demonstrate that the esterase-mediated rapid lysis of M. tuberculosis significantly improves its detection in paucibacillary samples.

Mycobacteria represent a group of ubiquitous and diverse bacterial species, many of which can establish chronic infections in humans (1). An estimated one-third of the world's population is infected by Mycobacterium tuberculosis, the etiologic agent of human tuberculosis (TB), 2 which causes about 1.7 million deaths worldwide every year (2). A distinct taxonomical classification for mycobacteria historically originated from their unique cellular morphology and acid-fastness, both directly attributed to their thick, hydrophobic, and atypically structured envelope (3)(4)(5)(6). The hydrophobicity of the envelope and consequently its low permeability to hydrophilic solutes contribute to a high level of intrinsic drug tolerance in mycobacteria (7). Not surprisingly, the envelope of mycobacteria has long remained a subject of intense investigation, yet its architectural organization remains to be fully understood.
The mycobacterial cell envelope is broadly stratified into a plasma membrane made of phospholipids, a core cell wall complex of covalently linked mycolyl-arabinogalactan-peptidoglycan (mAGP), and a membrane-like outer layer (6,8,9). The lipid bilayer in the membrane-like outer layer, also referred to as mycomembrane, is formed by mycolyl chains of mAGP and hydrocarbon chains of the outermost glycolipids, noncovalently associated with the cell wall core of the envelope (6,9). The outermost lipids, which vary across mycobacterial species, are extractable to a detectable level without breaching envelope integrity, although genetic dispensability of only a few of these lipids is determined (10).
The mycolyl esters of trehalose, trehalose monomycolates (TMM) and trehalose 6,6Ј-dimycolate (TDM) are among the most conserved and abundant noncovalently associated lipids of the mycobacterial envelope. The nascent mycolyl chains, synthesized in the cytoplasm by fatty-acid synthase systems I and II, are esterified to trehalose to produce TMM, which is subsequently transported across the membrane by MmpL3 (11,12). The abundant pool of TMM in the envelope is then utilized as a universal mycolyl donor by secreted mycolyltransferases, antigen 85 complex (Ag85A, Ag85B, and Ag85C), toward the constitutive synthesis of mAGP and TDM (11,(13)(14)(15). In addition, TMM is also used by the Ag85 complex to synthesize glucose monomycolate, which like other noncovalently linked mycolyl glycolipids, such as glycerol monomycolate and diarabino-dimycolyl glycerol, is produced under specific growth conditions depending on the type of carbon source in the medium (16 -18).
The essential role of TMM as a precursor of mAGP, the core cell wall component, makes it an indispensable lipid for mycobacteria, as recently validated by the viability loss due to specific inhibition of its MmpL3 transporter (12,19). Dispensability of TDM in the envelope, however, remains unclear, although it can be extracted from the bacilli in petroleum ether without any effect on their viability (20), but the extent of TDM depletion from the ether-treated bacteria remained unclear in these studies. Furthermore, TDM cannot be genetically depleted from mycobacteria due to the functional redundancies in the Ag85A, Ag85B, and Ag85C (13). A mutation in ag85A leads to partial depletion of TDM in Mycobacterium smegmatis, and inactivation or inhibition of Ag85C in M. tuberculosis leads to partial loss in mycolyl content of mAGP as well as the synthesis of TDM (14,21). Moreover, the partial loss in the mycolyl contents of cell wall components in the ag85A mutant could be complemented by expression of Ag85B and Ag85C (14). Although any of the three mycolyltransferases can be inactivated individually without affecting cellular viability, simultaneous inactivation of all three genes has not been possible. It is also noteworthy that their simultaneous intracellular depletion using antisense oligonucleotides severely retards the growth of M. tuberculosis (22).
TDM is studied in M. tuberculosis as one of the most potent immunomodulatory and granulomatogenic surface glycolipids (23,24), although it is highly abundant in at least nine other species of pathogenic and nonpathogenic mycobacteria analyzed so far (10,25,26). Its constitutive abundance in nonpathogenic mycobacterial species raises the possibility that TDM and other noncovalently linked mycolyl-glycolipids could have crucial structural contributions in the integrity of the mycobacterial envelope. Recently, we identified Msmeg_1529, an M. smegmatis enzyme of the serine esterase superfamily that can hydrolyze purified TDM from various mycobacterial species, including M. tuberculosis in vitro (27). This allowed us to question whether exogenous exposure to the purified recombinant esterase, henceforth called TDMH, may impact the integrity of the mycobacterial envelope. Here, we show that exposure to TDMH triggers an immediate release of free mycolic acids (FM) from noncovalently associated mycolyl-containing glycolipids, ultimately leading to rapid and extensive lysis of pathogenic species, such as M. tuberculosis, Mycobacterium bovis, and Mycobacterium marinum, as well as to a lesser extent of M. smegmatis and Mycobacterium avium. Although these findings highlight the structural contribution and importance of mycolyl glycolipids in the outer envelope of mycobacteria, they also open up new possibilities of improved detection and clearance of mycobacterial infections.

EXPERIMENTAL PROCEDURES
Strains and Media-Liquid cultures of M. tuberculosis (except mc 2 7000), M. bovis (BCG), and M. avium were grown at 37°C either in 7H90 ADC or Sauton's media with 0.05% Tween 80. For plate culture, 7H11 with OADC or Sauton's media agar was used. mc 2 7000 was grown similarly, except its media contained 100 g/ml pantothenic acid. For M. smegmatis, OADC was replaced with ADC in 7H9 broth or plate culture. M. marinum was grown at room temperature in 7H9OADC Tween 80 liquid medium or 7H11 OADC agar. Escherichia coli (DH5␣) was grown at 37°C in LB broth or LB agar.
Purification and in Vitro Lytic Activity of TDMH-The open reading frame of TDMH, Msmeg_1529, was cloned in pET21b expression system with His 6 tag, and the protein was purified on nickel-nitrilotriacetic acid affinity column as described previously (27). Functional activity of each batch of purified TDMH was determined through turbidometric measurement of bacterial lysis as described previously (28). For lytic activity, cells at OD of 0.5 (log phase) or OD of 3.0 (stationary phase) were harvested, washed, and resuspended in either phosphatebuffered saline with Tween 80 (PBST) or other culture media as required. Specified amounts of TDMH were mixed with the specified number of bacilli and incubated at 37°C. At regular time intervals, an aliquot was diluted and plated for enumeration. An equal volume of storage buffer was used as a negative control for each experiment. For inhibition of lytic activity of the enzyme, 100 g of either purified TDM (Sigma), purified apolar and polar lipids (27), or purified preparations of mAGP and PIM 2 (27) were dried in the reaction tube and suspended in 100 l of PBST by sonication at 55°C for 10 min. To the lipid suspension, cells (10 7 cfu/ml) and 0.8 M TDMH were added and incubated for 2 days at 37°C prior to enumerating the viable bacilli. For the plate assay, 500 l of 10 6 cfu/ml log phase cells was spread on a Sauton's agar plate, and 200 g of purified TDMH were spotted in the center. An equal volume of storage buffer was spotted as a negative control on a separate plate with bacilli.
ATP Release Assay-M. tuberculosis mc 2 7000 was grown to an OD of 0.5. Cells were harvested, washed with PBS, and diluted to 10 8 cfu/ml in PBST. 8 M TDMH was added to 1 ml of cell suspension and rotated at 37°C. At regular time intervals, 100 l samples were collected, and ATP content was determined by adding 100 l of ENLITEN luciferase/luciferin reagent (Promega), and the luminescence was measured using a Monolight 2010 luminometer.
Lipid Analysis-mc 2 7000 cultures (OD of 0.5) were labeled with [ 14 C]acetate for 6 h, washed, resuspended in 250 l of PBST at a density of 10 9 cfu/ml, and mixed with either 8 M TDMH or an equal volume of storage buffer at 37°C. At specific time intervals, apolar and polar lipids were extracted in petroleum ether as published previously (27). The extracted lipids equivalent to 5000 cpm from each sample were spotted on a one-dimensional TLC and developed in either CHCl 3 35,000 counts of each sample were analyzed as described previously (27). For the analysis of FM release in the presence of exogenous TDM and TMM, 100 g of TDM or 60 g of TMM were sonicated in 100 l of PBST prior to adding 50 g of TDMH (corresponding to 20-fold molar excess of lipid to the enzyme) to a final volume of 125 l. The lipid/enzyme mixture was incubated for 30 min, prior to adding it to a 125-l suspension of 14 C-labeled M. tuberculosis cells at 10 9 cfu/ml.
In Vitro Catalytic Activity of TDMH-[ 14 C]TDM or [ 14 C]TMM, equivalent to 100,000 cpm, was homogeneously suspended in the assay buffer as described earlier (27). The homogenate was then mixed with either 5 g of TDMH, 5 g of TDMH (S124A), or an equal volume of storage buffer and incubated at 37°C for 2 h. The lipids present in the reaction mixture were then sequentially extracted with an equal volume of petroleum ether followed by dichloromethane. The organic layer was dried and analyzed on TLC developed in CHCl 3 /MeOH (97:3, v/v) (27).
Nucleic Acid Detection and RT-PCR-For determining nucleic acids, 10 l of 10 7 cfu/ml M. tuberculosis (Erdman) was mixed with either 1 l of 13 g TDMH or 1 l of storage buffer in PBST supplemented with 26 mM L-asparagine (PBSTA) and incubated at 37°C. At various time intervals, the reactions were heat-inactivated at 80°C for 20 min and centrifuged at 14,000 rpm for 30 s, and 1 l of the reaction was placed in a nanophotometer to measure the NA contents. For RT-PCR, 10 l of 10 4 cfu/ml M. tuberculosis (Erdman) were mixed with either 13 g of TDMH or an equivalent volume of storage buffer in PBSTA, incubated for 30 min at 37°C, and then heat-inactivated at 80°C for 20 min. After centrifugation at 14,000 rpm for 30 s, 1 l of the supernatant from the treated mixture was directly added as a template to the molecular beacon-based RT-PCR described previously (30), with 500 nM each of the M. tuberculosis 16 S rRNA-specific forward and reverse primers, 200 nM molecular beacon, and 5 l of 2ϫ RT-PCR master mix (Applied Biosystems). The amplification conditions were as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 58°C for 60 s, and 72°C for 30 s in a StepOnePlus RT-PCR System (Applied Biosystems). The forward and reverse primer sequences were 5Ј-GAGATACTCGAGTGGCGAAC-3Ј and 5Ј-GGCCGGCTACCCGTCGTC-3Ј, respectively, and the molecular beacon consisted of a 5Ј-fluorescein-GCGCCCGC-GGCCTATCAGCTTGTTGGTGGCGC-dabcyl-3Ј. The statistical significance between treated and untreated samples was determined by a random intercept logistic regression model.
Electron Microscopy, Sample Preparation, and Analysismc 2 7000 cells (10 9 cfu/ml) were mixed with 8 M TDMH in PBST, harvested after a 12-h incubation, put on ice, and transferred to Leica EMPACT 2 (Leica Microsystems) membrane carriers (1.5 ϫ 0.2 mm) for high pressure freezing in EMPACT 2 (Leica), at an average pressure of 2000 bars (31). The frozen cells were then transferred to a freeze-substitution machine, Leica EM AFS (Leica), for a 5-day solvent substitution. Briefly, frozen samples were warmed up from Ϫ196 to Ϫ90°C over 3 days in a precooled (Ϫ90°C) 1% OsO 4 and 0.1% uranyl acetate mixture dissolved in acetone. Samples were then gradually warmed up to room temperature over 18 h and subsequently rinsed in acetone for further resin infiltration and embedding. Ultrathin sections (65 nm) were cut with a Reichart Ultracut and laid on 300 mesh carbon-coated EM grids. The thin sections were post-stained with 2% uranyl acetate in methanol for 10 min, followed by Reynold's lead citrate (32) for 7 min, and were examined with a Tecnai F20 electron microscope (FEI) equipped with a 4k ϫ 4k camera (Gatan).

Exposure to TDMH Leads to a Rapid Loss of M. tuberculosis
Viability-Approximately 10 6 cfu/ml M. tuberculosis H37Rv were incubated with increasing concentrations of the purified recombinant TDMH for 24 h in PBST. Fig. 1A shows that viability of the population was reduced by at least 100-fold after exposure to 0.8 M or higher concentrations of the enzyme (Fig.  1A). At a 10-fold excess concentration (8 M), the enzyme could reduce the viability of 10 8 cfu/ml of bacilli in less than 2 h, but higher density (10 9 cfu/ml) required longer periods of exposure (Fig. 1B).
We next evaluated TDMH activity in more complex chemical environments in which mycobacteria are routinely cultured or studied in the laboratory. TDMH activity varied widely from the highest in Sauton's media to the lowest in 7H9/OADC and DMEM/FCS (Fig. 1C), although no protein degradation was observed under these conditions (data not shown). Removal of the albumin-based supplement (OADC) restored TDMH activity in both the 7H9 and DMEM base but not to the levels observed in Sauton's broth medium (Fig. 1C). Because the activity in Sauton's media was even more pronounced than in PBS, we hypothesized that one or more components of Sauton's medium could potentially enhance the enzyme activity. Glycine has been previously found to sensitize mycobacteria against envelope-targeting agents like lysozyme (33). Although it is not clear how glycine could destabilize the cell envelope of mycobacteria, D-amino acids have been recently discovered to depolymerize amyloid-like protein fibers on the surface of Grampositive and Gram-negative bacterial envelopes (34). We therefore reasoned that a high concentration (26 mM) of L-asparagine (L-Asn) in Sauton's media could possibly sensitize mycobacteria to TDMH. Indeed, we found that an equivalent amount of L-Asn increases TDMH efficacy in PBST by 100-fold (Fig. 1D). To exclude the possible interference of Tween 80, known to influence the properties of mycobacterial envelope (35), we tested the activity of TDMH in PBS without Tween 80 and found no significant change in lytic activity of the enzyme (Fig. 1E). We next assayed the activity of the TDMH directly on a lawn of M. tuberculosis on a detergent-free Sauton's media agar plate. An unambiguous zone of clearance around the spotted area was observed after 3 weeks of incubation at 37°C (Fig.  1F). Overall, these results indicate that exposure to TDMH causes a very rapid and efficient loss of M. tuberculosis viability in diverse in vitro conditions. Viability Loss Is Due to Compromised Envelope Integrity and Cellular Lysis-The most likely explanation for loss of viability in TDMH-exposed bacteria is a breach of the envelope integrity and subsequent bacterial lysis, caused by the hydrolysis of target molecules. Lysis was evidenced by the clearance of a turbid suspension of M. tuberculosis after exposure to 8 M TDMH for 48 h (Fig. 2A). Lysis was further confirmed by the release of ATP from TDMH-exposed M. tuberculosis mc 2 7000 (Fig. 2B). The timing of ATP release in Fig. 2B was also consistent with the loss of cell viability at this density (Fig. 1B). We examined the integrity of TDMH-treated bacteria at an ultrastructural level by high pressure freezing and freeze-substitution electron microscopy. As expected, a distinct multiple layered architec-ture of the envelope, including an outer layer, was observed in the untreated bacteria (Fig. 2C, panels i-iii). In the TDMHexposed bacilli, the outer layer was visibly ablated at several places (Fig. 2C, panels iv-viii). We could also capture cells that were either in the process of losing the cytoplasmic content during lysis (Fig. 2C, panels v-vii) or had completely lost the cytosolic content and the outer cell envelope (Fig. 2C, panel  viii). The heterogeneous state of TDMH-treated bacilli and the partial loss of the envelope integrity can be attributed to the high density (10 9 cfu/ml) of bacilli exposed to subsaturating concentrations (8 M) of the enzyme as indicated by the slower lysis (Fig. 1B). These observations together are fully consistent with the idea that exposure to TDMH breaches the integrity of the cell envelope outer layer, thereby triggering bacterial lysis. Henceforth, activity of each TDMH preparation was determined by turbidometric measurement of cell lysis as described earlier for a bacteriophage lysin (28). TDMH-dependent lysis through exogenous disruption of the envelope implies that the anti-mycobacterial activity of the enzyme will be nonresponsive to the physiological state of the target bacilli. This was indeed supported by a comparable susceptibility profile of exponentially growing and stationary phase bacilli to TDMH exposure (Fig. 2D). We further confirmed the relationship between TDMH activity and bacterial lysis by engineering a point mutation in the catalytic site of TDMH and testing the lytic activity of the mutant enzyme. Cutinase-like serine esterases possess a highly conserved catalytic triad of serine (in a GXSXG motif), aspartate, and histidine (27). A catalytically inactive TDMH mutant was generated by replacing the catalytic Ser-121 by an Ala residue. As shown in Fig. 2E, TDMH(S124A) clearly failed to hydrolyze TDM (Fig. 2E). Importantly, exposure of M. tuberculosis to the mutant enzyme failed to lyse the culture (Fig. 2F). It can therefore be inferred that exposure to the TDM hydrolase activity caused envelope rupture and mycobacterial lysis.
Depletion of TMM and TDM during Early Phase of TDMH Exposure-Because TDMH hydrolyzes TDM in vitro to release FM (27), we next analyzed the levels of FM and TDM in enzyme-exposed M. tuberculosis cultures. To this aim, a high tuberculosis (mc 2 7000) suspension (10 8 cfu/ml) exposed to 8 M TDMH (also see Fig. 1B). As a negative control, ATP was measured in culture exposed to storage buffer. C, high pressure freezing and freeze-substitution electron microscopy of 10 9 cfu/ml M. tuberculosis (mc 2 7000) before (panels i-iii) and after a 12-h exposure to 8 M TDMH (panels iv-viii). Images in panels i-iii are independent fields showing low magnification of intact cells (panel i) and high magnification of their cross-sections before TDMH exposure. A close-up view of the envelope layers is shown in the inset of panel iii. The cell inner membrane and outer layer are marked as IM and OL. Images in panels iv-viii are independent fields showing various stages of lysing bacilli in a TDMH-exposed population. Bacteria shown are as follows: panel iv, with a lesion in the outer layer (marked with arrowhead); panel v, with a distorted cytoplasm, presumably moments before the cytoplasm ejection due to the membrane lesion; panels vi and vii, in the act of releasing cytoplasmic contents, and panel viii, completely devoid of cytoplasm. Arrows point to the lesion sites. Scale bars are 100 nm for panels ii-viii, and 1 m for panel i. D, exposure of log-phase (7-day old with OD 0.5) and stationary phase (35-day old with OD 3.0) cultures of H37Rv to 8 M TDMH. E, loss of in vitro hydrolytic activity against TDM in the catalytic serine mutant of TDMH (S124A). F, viability of 10 7 cfu/ml suspension of M. tuberculosis (H37Rv) exposed to 8 M TDMH (S124A). The error bars represent the standard errors of three independent experiments. bacterial density (10 9 cfu/ml) was used to obtain sufficient lipids in a small volume with an effective amount (8 M) of TDMH. Because lysis is slowed at this high cell density (Fig. 1B), the reactions were followed for an extended period of 6 days. The relative levels of FM in the TDMH/bacteria mixture progressively increased, although it remained almost unchanged in buffer-treated cells (Fig. 3, A and B). After a 48-h exposure, the FM levels were over twice the amounts present at the beginning of the exposure (Fig. 3E). Interestingly, although the level of FM started to accumulate very early, within the first hour of treatment, the depletion of TDM was essentially apparent only after a 48-h exposure, when its level reduced over 50% to the amount seen at the beginning (Fig. 3, A, B, and E). The delayed depletion of TDM suggests that FM is generated at the expense of another mycolyl ester, which is either directly targeted by the enzyme or indirectly utilized by the cells to replenish TDM. Besides TDM, tuberculosis treated with either buffer or TDMH for 2 and 6 h. D, radio-TLC of the lipids from TDMH and buffer-exposed cells developed in a polar solvent (CHCl 3 /MeOH/NH 4 OH, 80:20:2, v/v/v) to resolve TMM and TDM. Purified TMM and TDM were used as markers (M). E, relative change in the levels of TMM, TDM, and FM over a 48-h exposure in buffer and TDMH-exposed M. tuberculosis. The ratio between 0 (t 0 ) and a 48-h exposure was obtained from densitometric analysis of TLCs. The error bars denote standard error in the values from three independent experiments. F, relative change in the levels of TMM and TDM in normal and heat-killed M. tuberculosis exposed to TDMH over the first 6 h of exposure. The value at each time point, obtained from densitometric analysis of TLCs, represents the percentage of the lipid relative to t 0 (zero hour). The error bars denote standard error from two and three independent experiments for normal and heat-treated cells, respectively. G, TLC of apolar (developed in CHCl 3 /MeOH, 96:4, v/v) and polar glycolipids (developed in CHCl 3 /MeOH/H 2 O, 75:25:4, v/v/v) of heat-killed bacilli exposed to TDMH for 2 h. The major apolar and polar glycolipids, phenolic glycolipids (PGL), and PIMs are marked. For experiments in A-G, 10 9 cfu/ml M. tuberculosis were exposed to 8 M TDMH. mAGP and TMM are the two most dominant esters of mycolic acids in the mycobacterial cell wall. No significant change in the mycolyl content of mAGP was observed in TDMH-exposed cells during FM release (Fig. 3C), consistent with the lack of in vitro hydrolytic activity of the esterase against purified mAGP (27). We next examined whether FM was released at the expense of the TDM precursor TMM. Fig. 3D shows that the levels of TMM in TDMH-exposed cells indeed started to progressively decrease within 1 h of exposure, concomitant with the release of FM seen in Fig. 3B.
Because no loss of viability was observed during the first 2 h of exposure under these conditions (Fig. 1B), TMM depletion and FM release are very likely to be the pre-lysis events, suggesting that noncovalently cell wall-associated mycolyl esters are among the first targets of the enzymes. We also addressed whether any other lipids besides TMM are depleted during the early phase of TDMH exposure. Therefore, a comprehensive analysis was performed on both polar and apolar lipids from 10 9 cfu/ml of 14 C-labeled M. tuberculosis treated with either buffer or 8 M TDMH for 2 h. Besides TMM depletion, and concomitant release of FM, there was no significant depletion in any of the polar or apolar lipids in the TDMH-treated cells (Fig. 4A). However, we noted elevated levels of phosphatidylinositol and an unknown polar lipid in TDMH-exposed bacilli (Fig. 4A). Because there was no obvious depletion of any other lipids to account for the elevated levels of these polar lipids, it is likely that these are produced through anabolic pathways as an early cellular response to TDMH exposure. Taken together, these results suggest that TMM is the major lipid that is significantly depleted during the first 2 h of TDMH exposure. It is, however, noteworthy that in contrast to TDM, purified TMM could not be hydrolyzed by TDMH in vitro (Fig. 4B). This is rather surprising and difficult to reconcile with the in vivo effect of TDHM on cellular TMM. One possibility could be that TDM depletion during early exposure to the enzyme is countered by a response that induces its synthesis by utilizing TMM. In this scenario, TDM would be depleted earlier if heat-inactivated cells were exposed to the esterase. Indeed, we observed that the levels of TDM decreased to about 50% within the first hour when heat-inactivated cells were exposed to the enzyme (Fig.  3F). In addition, we also observed a slower and less pronounced depletion of TMM in the TDMH-exposed heat-killed bacilli (Fig. 3F). However, depletion of TMM remained significant in heat-killed cells, a decrease by about 40% within the first hour. No other glycolipids were altered in the enzyme-exposed heatkilled cells during the first 2 h of exposure (Fig. 3G).
To further substantiate the impact of the TDMH on cellular TDM, we next analyzed the effect of the exogenous addition of purified TDM on the enzyme-exposed M. tuberculosis. Addition of a 20-fold molar excess of purified TDM with respect to TDMH reduced the release of FM from the bacilli by about 30% within 2 h of exposure (Fig. 5A). However, a comparable molar excess of exogenous TMM had no effect on FM release (Fig.  5A), consistent with the inability of the enzyme to hydrolyze purified TMM in vitro (Fig. 4B). Similarly, a 200-fold molar excess of TDM with respect to TDMH could abrogate the lytic activity of the enzyme (Fig. 5B). The inhibition in the presence of equivalent weight of a TDM-containing apolar lipid fraction was partial (Fig. 5B). However, neither polar lipid nor purified mAGP and PIM, a major polar lipid of the cell envelope (Fig.  4A), had any effect on the lysis (Fig. 5B). The failure of mAGP to inhibit the lytic activity is consistent with the following: (a) the lack of any apparent loss of its mycolyl content in the exposed FIGURE 4. A, two-dimensional radio-TLC of total apolar and polar lipids equivalent to 35,000 cpm extracted from cells treated with either buffer or TDMH for 2 h. Apolar lipids were resolved in solvent (Sol.) systems A, C, and D, and polar lipids were resolved in solvent system E as described previously (39). cell (Fig. 3C), and (b) the inability of the enzyme to hydrolyze mAGP (27). Overall, inhibition of FM release and bacterial lysis by exogenous TDM and early depletion of TDM from the heatkilled bacilli during the esterase exposure support the idea that TDM is surface-exposed where it can directly interact with the exogenous enzyme. The requirement of a large molar excess of purified TDM for an apparent inhibition of TDMH activity on M. tuberculosis suggests that the conformation of the lipid in the envelope matrix could be more sensitive to the esterasemediated hydrolysis than in its aqueous form.
It is noteworthy that the extractability of TDM in petroleum ether without the loss of bacterial viability gives rise to a notion of its dispensability in mycobacteria (20,36). We therefore determined TDM levels in the following: (a) petroleum ether extracts; (b) the residual lipids obtained by re-extraction of petroleum ether treated cells in CHCl 3 /MeOH (2:1, v/v), and (c) total lipid extracts from cells in CHCl 3 /MeOH (2:1, v/v). To exclude the bias in lipid extraction, we measured the amount of total lipids in the three extracts. The amounts of lipids in conditions a and b were nearly equal to those obtained in condition c (Fig. 5C). Analysis of TDM on radio-TLCs in three independent extractions revealed that the ether-extracted fraction contained an average of about 5% of the TDM present in the total lipid mixture. Over 60% of the glycolipids were re-extracted along with the residual lipids of the ether-treated bacteria (Fig.  5D), indicating that only a limited amount of TDM is extractable in petroleum ether and that this level of depletion has no effect on mycobacterial viability.

TDMH Exposure Is Detrimental to Other Mycobacterial
Species-Because TDM and TMM are highly conserved glycoconjugates of mycobacteria (25,26), the question arises as to whether the exogenous TDMH exposure could be similarly detrimental to other mycobacterial species. We tested the activity of TDMH against four widely studied mycobacterial species, M. bovis (BCG), M. avium, M. marinum, and M. smegmatis, representatives of the M. tuberculosis complex as well as nontuberculous mycobacteria. Although loss of viability was observed in all four of them, M. smegmatis and M. avium were significantly more tolerant than the other species (Fig. 6A). Because mycolyl esters of trehalose are unique lipids of mycobacteria, TDMH exposure, as expected, had no effect on E. coli (Fig. 6A). The decreased sensitivity of M. smegmatis and M. avium suggests that their trehalose mycolates are either less exposed or recognized with a lower affinity by the enzyme. The tolerance observed in M. smegmatis could possibly be attributed to the fact that the enzyme is physiologically expressed by this species, and therefore, it may have evolved a mechanism to control the activity on its own envelope. In addition, the TDMH tolerance could also originate in the distinct species of mycolates conjugated to trehalose in M. smegmatis, for instance in the presence of shorter chain ␣Ј and absence of methoxy and keto species (11). Interestingly, M. marinum appears very sensitive to TDMH (Fig. 6A). The failure of TDMH(S124A) to lyse M. marinum cells (Fig. 6B) suggests that depletion of trehalose mycolates could be the likely trigger of lysis in this species as well. FIGURE 5. TDM is an early target of TDMH. A, quantitative analysis of relative change in FM levels in M. tuberculosis/TDMH mixture with respect to the buffer-exposed control, as described in A. Error bar indicates standard error of the densitometric data obtained from three independent experiments. B, lytic activity of 0.8 M TDMH against M. tuberculosis (H37Rv) in the presence of 100 g of either purified TDM, apolar lipids (apL), polar lipids (pL), PIMs, or the mAGP complex. Percentage of viability in each condition was calculated using buffer-treated cells as reference (100%). C, total radio lipids (counts/min) extracted from 100 mg of 14 C-labeled cells in either three sequential cycles of sonication in 2:1 CHCl 3 /MeOH (total) or petroleum ether (PE-extractable) followed by re-extraction through three cycles in 2:1 CHCl 3 /MeOH (residual). D, percentage of TDM in ether-extracted and residual lipids with respect to the total lipids. The values were determined by calculating the total number of TDM equivalent pixels in each of the samples obtained through densitometric analysis of radio-TLC. The error bars represent standard errors of three independent experiments.

TDMH Facilitates M. tuberculosis Detection at Low Density-
A rapid and efficient lysis of TDMH exposed M. tuberculosis, and a subsequent release of nucleic acids (NA) offers a unique opportunity to use this enzyme for achieving more sensitive detection in a nucleic acid amplification assay for TB diagnosis. NA content in the supernatant of TDMH-exposed bacilli indeed started to increase within 30 min of exposure, and it peaked at around 60 min (Fig. 7A). A prolonged exposure, however, decreased the NA content, presumably because of the activity of cellular nucleases released in the lysates. Interestingly, a cell suspension at this density treated with the storage buffer had low but detectable amounts of NA that remained unchanged during the exposure time, indicating the spontaneous lysis of bacteria at low frequency during heating (Fig. 7A). We next performed a molecular beacon-based 40-cycle real time-PCR (30) to detect NA corresponding to 16 S rRNA in samples with less than 10 bacilli. Out of 45 independent amplification reactions in paucibacillary samples, diluted in triplicate from 15 independent broth cultures, NA could be detected in only seven reactions without TDMH treatment (Fig. 7, B and  C). However, treatment of samples with TDMH prior to the amplification reaction facilitated the NA detection in 37 samples, a highly significant (p Ͻ 0.0001) improvement in the numbers of positive detection (Fig. 7C). Low frequency of NA detection in some of the untreated samples is consistent with the Poisson distribution pattern of dilution in which some dilutions have large enough bacilli to generate sufficient amounts of template due to autolysis during heating.

DISCUSSION
In this study, we report an unusual and unexpected consequence of mycobacterial lysis upon exposure to a cutinase-like serine esterase. The enzyme can hydrolyze purified TDM in vitro and sequentially deplete TMM and TDM from the envelope of the exposed cells. The early depletion of TMM and TDM indicates that these lipids are both exposed on the bacterial surface and therefore are among the first molecules encountered by the enzyme. Loss of trehalose mycolates could itself cause osmotic lysis and/or create openings for the enzyme to access the inner components and breach envelope integrity. Either of these situations therefore reveals a structural role of the glycolipids in maintaining the integrity of the mycobacterial envelope. Involvement of TMM is a more complex scenario to imagine because of the dynamic range of intermediate conformations it could assume during synthesis of terminal mycolyl esters, although a molecular subspecies of the glycolipid with a dedicated structural role cannot be ruled out. The role of TDM, however, is highly plausible because of its large abundance and its stable terminal structure as an end product. Given that the outermost layer of the envelope assumes a membrane-like configuration (9), TDM could likely be organized in a bilayer configuration constituting the outer leaflet of the mycomembrane, with trehalose as the polar head group at the environmental interface and mycolic acids forming the hydrophobic tail. The 60 -90-carbon mycolic acid could either be in an extended conformation to span the entire thickness (ϳ8 nm) of the mycomembrane or in a folded conformation to fit into the outer leaflet of the bilayer (8,9). The folded mycolic acid chain, presumably formed through interaction of the keto or methoxy groups with the lipid head group (8), is the likely conformation. This is extrapolated from the biophysical analysis of mycolic acid monolayers as well as from the visual evidence of hydrophobic interface in the cryo-electron microscopy of the outer membrane (9,37).
The likely integration of TDM in the outer membrane, its limited extractability from the intact envelope and disintegration of the envelope upon its depletion together raise a strong possibility that TDM, and therefore the outer membrane, could be indispensable for the overall integrity of mycobacteria. Moreover, this is consistent with the cryo-EM visualization of the dissolved inner membrane in the region where the outer membrane lipids are removed by detergent extraction (9,38). The indispensability of TDM is also suggested by growth inhibition of M. tuberculosis in the presence of a chemical inhibitor of Ag-85C that reduces the level of TDM without affecting the levels of mAGP and TMM (21).
Given the detrimental consequences of TDMH exposure, mycobacteria must have a mechanism to regulate its physiological activity. Regulation of TDMH is indeed manifested by a growth phase-dependent induction of its activity toward production of free mycolic acids during biofilm maturation of M. smegmatis (27,39). Furthermore, delayed depletion of TDM upon TDMH exposure also suggests a two-tier regulatory mechanism in which replenishment of the substrate could potentially salvage a dysfunctional regulation of tdmh expression. Although TDM hydrolase in other mycobacteria remains to be discovered, antibodies against some of the cutinase-like esterases, including the closest homologue of TDMH (Rv3452), are produced during active infection of M. tuberculosis in humans, indicating their expression during TB pathogenesis (40).
In summary, TDMH solves a long standing challenge of exogenously breaching the envelope of mycobacteria. The improved frequency of positive detection upon TDMH treatment of samples with small number of bacilli highlights the potential of the enzyme in elevating the sensitivity of TB diagnosis, particularly in paucibacillary infections.