Mycothiol Biosynthesis and Metabolism

Mycothiol (MSH; 1-d-myo-inosityl-2-(N-acetyl-l-cysteinyl)amido-2-deoxy-α-d-glucopyranoside (AcCys-GlcN-Ins)) is a novel thiol produced at millimolar levels by mycobacteria and other actinomycetes that do not make glutathione. We developed methods to determine the major components of MSH (AcCys, Cys-GlcN, AcCys-GlcN, Cys-GlcN-Ins, GlcN-Ins) in cell extracts.Mycobacterium smegmatis was shown to produce measurable levels (nmol/g of residual dry weight) of AcCys (∼30), Cys-GlcN-Ins (∼8), and GlcN-Ins (∼100) but not Cys-GlcN (<3) or AcCys-GlcN (<80) during exponential growth in Middlebrook 7H9 medium. The level of GlcN-Ins declined 10-fold in stationary phase and ∼5-fold in 7H9 medium lacking glucose. Incubation in 10 mm AcCys produced 50- and 1000-fold increases in cellular Cys and AcCys levels, respectively, a 10-fold decrease in GlcN-Ins and a transient 3-fold increase in Cys-GlcN-Ins. These results exclude Cys-GlcN and AcCys-GlcN as intermediates in MSH biosynthesis and implicate GlcN-Ins and Cys-GlcN-Ins as key intermediates. Assay of GlcN-Ins/ATP-dependent ligase activity with Cys and AcCys as substrates revealed that Cys was at least an order of magnitude better substrate. Based on the cellular measurements, MSH biosynthesis involves assembly of GlcN-Ins, ligation with Cys to produce Cys-GlcN-Ins, and acetylation of the latter to produce MSH.

Preparation of 1-D-myo-Inosityl-2-amino-2-deoxy-␣-D-glucopyranoside (GlcN-Ins)-GlcN-Ins was purified from M. echinospora (NRRL B-12180) by a modification of the method of Maehr et al. (9). M. echinospora was cultured on 0.5% yeast extract, 3% Todd Hewitt media, and 0.25% sucrose to late log phase and collected by centrifugation. Bulk cell pellets were frozen at Ϫ70°C until the material was processed. A 235-g pellet of M. echinospora was suspended in 1 liter of 50% acetonitrile-water (60°C) containing 20 mM H 2 SO 4 (pH 5 cell suspension). The suspension was adjusted to pH 2.5 with concentrated H 2 SO 4 and then disrupted using a Bransonic sonicator at ϳ70% maximum power for 15 min without cooling. The cell extract was cooled on ice, and the cell debris was removed by centrifugation at 6000 ϫ g for 15 min at 4°C. The supernatant was reduced to 250 ml using a rotary evaporator and clarified by centrifugation as above. The supernatant contained 240 mol of thiol by assay with 5,5Ј-dithiobis(2-nitrobenzoic acid) (10).
Next, MSH was recovered from the extract. DTT (250 mol) was added, and the extract was adjusted to pH 7.9 with concentrated NH 4 OH. After clarification by centrifugation, the supernatant was passed over a 2-thiopyridine-activated thiolpropyl-agarose column (3), the column was eluted with DTT, and the mycothiol was purified by preparative HPLC after derivatization with 5,5Ј-dithiobis(2-nitrobenzoic acid).
The unbound extract (310 ml) from the thiol affinity column (pH 7.6) was applied directly to a 2.5 ϫ 18-cm Amberlite IRC50 (Mallinckrodt) column that had been prewashed with 3 M NH 4 OH followed by H 2 O. The effluent pH was 10.3 when the sample was applied. The column was washed with water and eluted with 350 ml of 1 N NH 4 OH followed by 550 ml of 3 M NH 4 OH. The fractions from both effluents were found to contain GlcN-Ins as assayed by TLC with ninhydrin detection as previously reported (9). The fractions containing GlcN-Ins were combined and lyophilized. The residue was redissolved in water (pH ϳ 9) and applied to a 1 ϫ 12-cm AG1-X8 (Bio-Rad, 200 -400 mesh) anion exchange column in the hydroxyl form. The column was eluted in water, and the GlcN-Ins-containing fractions were again pooled. The fractions containing GlcN-Ins (R f 0.33) were also contaminated by higher R f ninhydrin-positive materials.
The GlcN-Ins-containing fractions were again combined, dried by lyophilization, and dissolved in 0.1% trifluoroacetic acid, water. This sample was applied to a 20-ml Sep Pak C18 cartridge (Waters) equilibrated in the same solvent. The GlcN-Ins was eluted with 0 -5% methanol gradient in aqueous 0.1% trifluoroacetic acid. The GlcN-Ins-containing fractions gave a single ninhydrin spot on TLC, free of the high R f contaminant. However, amino acid analysis following acid hydrolysis (3) revealed that GlcN was a minor component and that another amine eluting after arginine was a major component. To remove this contaminant, the dried GlcN-Ins-containing fractions were applied to a 1 ϫ 12-cm Biorex 70 (sodium form, 100 -200 mesh, Bio-Rad) column. The GlcN-Ins was eluted in water, and the GlcN-Ins-containing fractions were pooled. Amino acid analysis showed that the basic contaminant had been completely removed. The product gave a small peak at mass 341 (molecular ion) and a major peak at mass 364 (molecular ion plus sodium) on electrospray mass spectroscopy. The 1 H NMR was consistent with the expected structure, and all peaks could be assigned by analogy with the spectrum for MSH (Table I). A stock solution of this material, which contained 0.5 mg of GlcN-Ins by NMR analysis relative to a known concentration of acetone as internal standard, was used for all experiments with GlcN-Ins.
Determination of GlcN and GlcN-Ins-AccQ-Fluor was dissolved in acetonitrile to 10 mM as recommended by the manufacturer. Standard solutions of D-glucosamine⅐HCl and purified GlcN-Ins were prepared at 3.1 to 200 M in water. For the derivatization of GlcN and GlcN-Ins, 6 l of the standard amine was diluted to 30 l in 200 mM HEPES (pH 8.0), 15 l of acetonitrile; 15 l of 10 mM AccQ-Fluor were added, and the mixture vortexed. After 1 min at room temperature, samples were heated for 10 min at 60°C. The samples were diluted 4-fold with water and stored at Ϫ70°C. To determine GlcN and GlcN-Ins levels in cells, a sample of cell suspension was chilled on ice and pelleted at 4°C using a SS34 or GSA rotor in a Sorvall RC-5 centrifuge. The cell pellet was lysed by the addition of 450 l of warm acetonitrile and heating at 60°C for 2 min. Next, 490 l of water, 10 l of 1 M HEPES (pH 8), and 50 l of 100 mM NEM in acetonitrile were added, the mixture was heated at 60°C for 10 min, and the container was immediately chilled in ice. The sample was pelleted by brief centrifugation in a microfuge at 4°C, and the supernatant was removed to an Eppendorf microcentrifuge tube on ice. A 15-l aliquot was immediately taken from the supernatant for derivatization with AccQ-Fluor by the addition of 12.5 l of 1 M HEPES (pH 8.0), 42.5 l of water, 23.8 l of acetonitrile, and 31.2 l of 10 mM AccQ-Fluor, and the mixture was vortexed immediately. A second 30-l aliquot of cell extract was taken at the same time and prepared in the same fashion except that the volumes of water and acetonitrile added were 35 and 16.8 l, respectively. After 1 min at room temperature, the samples were heated for 10 min at 60°C, diluted 4-fold with water, and stored at Ϫ70°C. The cell pellet was dried in a tared tube to determine the residual dry weight (RDW) of the extract. The AccQ-Fluor-derivatized samples were analyzed by HPLC utilizing a Waters 600E solvent delivery system equipped with a Waters WISP Model 710B autoinjector, Laboratory Data Control Fluorometer III, and a Nelson Model 444 data collection system. Separation was obtained on a Beckman Ultrasphere IP (250 ϫ 4.6 mm) analytical column equipped with a Brownley HPLC guard column containing an OD-GU 5-m C-18 cartridge using the following linear gradients: 0 min, 100% A (0.1% trifluoroacetic acid in water); 10 min, 100% A; 50 min, 60% B (0.1% trifluoroacetic acid in methanol); 53 min, 100% B; 57 min, 100% B; 60 min, 100% A; 70 min, reinjection. The flow rate was 1 ml min Ϫ1 , and the effluent was monitored by fluorescence with a 254-nm excitation filter and a 370 -700-nm emission filter. This protocol is designated HPLC method 5. AccQ-Fluor-derivatized amines had the following retention times: GlcN, 12 min; GlcN-Ins, 24 min.
mBBr Derivative of Cys-GlcN-Ins (CySmB-GlcN-Ins)-The bimane derivative of mycothiol (MSmB) was prepared in pure form as described previously (3). Approximately 10 mg of MSmB was hydrolyzed in 2 ml of 6 N HCl for 2 h at 60°C, and the hydrolysate was purified by HPLC on a preparative C-18 Vydac column (22 ϫ 250 mm). The sample was eluted in 0.1% trifluoroacetic acid in water (mobile phase A) with a 1%/min linear gradient from 0 to 40% B (0.1% trifluoroacetic acid in methanol) at a flow rate of 5 ml min Ϫ1 . The effluent was monitored by fluorescence detection with excitation at 370 nm and emission at 418 -700 nm. CySmB-GlcN-Ins was collected at 32 min and repurified twice to minimize CySmB-GlcN contamination. The identity of this peak was confirmed with electrospray mass spectrometry, which yielded a major peak at 657 daltons (molecular ion ϩ sodium). The final yield of 0.63 mg of mB-Cys-GlcN-Ins was estimated assuming ⑀ ϭ 5300 at 390 nm, the value reported for the mBBr derivative of glutathione (11).
Determination of Thiols-Labeling of cell extracts with mBBr was carried out as described previously (5). HPLC analysis of Cys, MSH, AcCys-GlcN, ACys, and H 2 S was performed using method 1A, a modification of the previously described method 1 (12)  the HPLC conditions were identical to those of method 5 above with the following modifications: solvent B was 7.5% methanol in acetonitrile, and fluorescence detection was accomplished with a 370-nm excitation filter and a 418 -700 emission filter (designated method 5A). Assay of ATP-dependent Ligase Activity with GlcN-Ins Plus Cys/ AcCys-A minor modification of the protocol described by Bornemann et al. (6) was used. M. smegmatis MC 2 155 was grown to mid-log phase in Middlebrook 7H9 medium with 0.4% glucose and 0.05% Tween. Cells were pelleted and washed with 50 mM sodium phosphate (pH 7.5) containing 1 mM DTT and resuspended in the same buffer at a concentration of 0.25 g of wet weight/ml. Cells were lysed by sonication on ice and pelleted by ultracentrifugation at 100,000 ϫ g for 30 min in a Beckman tabletop ultracentrifuge. Assay of the supernatant was conducted in a final volume of 600 l containing 60 l of supernatant, 50 M GlcN-Ins, 100 M Cys or AcCys, 100 M sodium acetate, 1 mM ATP, 1 mM MgCl 2 , 50 mM sodium phosphate (pH 7.5), 1 mM DTT, and 35 M each of the protease inhibitors phenylmethanesulfonyl fluoride, N-␣-ptosyl-L-phenylalanylchloromethyl ketone, and N-␣-p-tosyl-L-lysinechloromethyl ketone. The mixture was incubated at 30°C, and 100 l samples were removed at 0 and 60 min for thiol analysis. One sample was mixed with 4 l of 100 mM mBBr and allowed to react for 5 min at room temperature before acidification with 0.5 l of 5 M methanesulfo-nic acid to quench the reaction. A second control sample was reacted 5 min with 5 mM NEM before treatment with mBBr as above.

RESULTS
Analysis of GlcN and GlcN-Ins-Several standard methods of fluorescent amine labeling in conjunction with HPLC were tested initially with GlcN as potential ways to determine GlcN and GlcN-Ins. Reaction with dansyl chloride, fluorescamine, 12-(N-methyl-N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)) chloride, and CBQCA (Molecular Probes) provided markedly poorer sensitivity than labeling with AccQ-Fluor, a reagent recently developed for analysis of amino acids (13). For labeling of amino sugars, a buffer of slightly lower pH than employed with amino acids was chosen so that the amino sugars, whose ammonium forms have lower pK a values than those of most amino acids, would have a competitive advantage during labeling of cell extracts. Fig. 1 (top panel ) shows the chromatogram obtained with a standard GlcN sample, and Fig. 2 presents a representative calibration curve.
To test the method with GlcN-Ins, it was necessary to obtain an authentic sample. GlcN-Ins was originally isolated from M. echinospora by Maehr et al. (9), and an initial description of its synthesis was recently reported by Bornemann et al. (6). Isolation appeared to be the simpler route and had the advantage of permitting simultaneous isolation of MSH in significant amounts with a minor modification of the protocol. However, at the end of the published isolation protocol our GlcN-Ins sample, although apparently pure by TLC, proved to be substantially contaminated by another amine. Pure GlcN-Ins was obtained by adding an additional purification step; the overall yield was poor but adequate for the present purposes.
The chromatogram obtained from a standard solution of GlcN-Ins after labeling with AccQ-Fluor is shown in Fig. 1  (middle panel), and a representative standard curve is included in Fig. 2. Fig. 1 also presents the chromatogram obtained from the labeling of a sample of a 50% acetonitrile extract of the M. echinospora used for the purification of GlcN-Ins. The peak for GlcN-Ins corresponds to a content of ϳ0.8 mol/g of RDW. From this value it can be calculated that the recovery in our purification of GlcN-Ins was only 8%. No doubt this can be improved with suitable modifications of the purification protocol and the use of AccQ-Fluor labeling with HPLC analysis to more closely monitor the purification process.
Analysis of the Thiol Components of Mycothiol-Cellular thiol levels were analyzed by HPLC after fluorescent labeling of the thiol moiety with mBBr (14). Quantitative values for the cellular thiol levels were obtained after correction for any nonthiol fluorescent background identified in control samples in which thiols were blocked by reaction with NEM before treatment with mBBr. Standards were prepared by labeling of the commercial (Cys, AcCys) or synthetic (Cys-GlcN, AcCys-GlcN) thiols or of isolated MSH (3). A standard for labeled Cys-GlcN-Ins was obtained by partial hydrolysis of MSmB, purification of CySmB-GlcN-Ins by preparative HPLC, and quantitation of the standard based upon the absorbance of the bimane label. All of these bimane derivatives could not be separated from each other, from reagent-derived components, and from fluorescent cellular materials using a single HPLC protocol, and many different HPLC separations were tested before two methods were found that provided analyses for the thiols of interest without major coeluting peaks in the NEM control sample. Fig.  3 illustrates the analysis used to obtain values for Cys-GlcN-Ins and Cys-GlcN. In cell extracts, no peaks were observed for Cys-GlcN, and reported values represent the limits of detection (Fig. 3B). Peaks for Cys-GlcN-Ins were observed with cell extracts as illustrated in Fig. 3B. An unidentified thiol peak of comparable size to the Cys-GlcN-Ins peak and eluting at 26.7 min is also apparent in Fig. 3B. Fig. 4 illustrates the analysis for Cys, AcCys, AcCys-GlcN, MSH, and H 2 S; the Cys-GlcN-Ins and Cys-GlcN derivatives were only slightly retained and could not be separated under these conditions. The peaks for Cys and MSH have small coeluting peaks present in the NEM control (Fig. 4, lower panel), which amounted to 10 and 2%, respectively, of the Cys and MSH peaks measured in the sample (Fig.  4, middle panel). In calculating the Cys and MSH content, small corrections for these control values were incorporated.  a Numbers designated with Ͻ represent detection limits where no discernable peak was present. Values designated with Յ represent measurement of a discernable peak at the retention time for the indicated component but for which independent verification of the structure was not available; the value represents an upper limit for the content. Note that the ␣ and ␤ epimers of bimane derivatives of Cys-GlcN and AcCys-GlcN interconvert to give an equilibrium mixture producing two partially resolved peaks with similar retention times (Figs. 3A and 4).
Analysis of Selected Bacteria-Results for analysis of the full range of potential MSH intermediates in several bacteria harvested in log phase growth are given in Table II. Analysis of extracts from M. smegmatis MC 2 155 revealed the presence of a measurable GlcN-Ins content. To check whether the values are sensitive to the sample size analyzed, possibly as the result of depletion of the labeling agent by the plethora of amines present in cells, we analyzed 30 l of supernatant from extracts (1 ml volume) of M. smegmatis MC 2 155 with increasing sample size from 3.9 to 61 mg RDW of cell pellet residue. For samples sizes up to 30 mg of RDW, there is an apparent slight decrease in measured GlcN-Ins content with increasing sample size (0.13%/mg RDW), but the change is within the uncertainties in the measurements. However, above 30 mg RDW, a marked decrease in measured value was apparent. To prevent underestimates of GlcN-Ins content, all quantitative determinations were made with samples corresponding to Յ30 mg RDW.
M. smegmatis MC 2 6, the parent strain of M. smegmatis MC 2 155, also produced GlcN-Ins in significant amount. Cys was also found, but neither strain appeared to produce comparable amounts of other potential intermediates of MSH metabolism, including AcCys, AcCys-GlcN, Cys-GlcN, or Cys-GlcN-Ins. It had been previously shown that S. aureus and E. coli did not produce MSH (5). These were reexamined using the present methods to ascertain whether they might produce one or more of the intermediates involved in mycothiol biosynthesis and especially GlcN-Ins. The results (Table II) indicate that neither of these bacteria produce any component of MSH at measureable levels, other than Cys and GlcN.
Variation in Metabolite Levels During Growth-Levels of potential MSH intermediates were determined for M. smegmatis MC 2 155 as a function of growth phase (Fig. 5). The MSH level remained remarkably constant through exponential growth and into stationary phase, as did the Cys level at about 1% the MSH level. However, in stationary phase, there did appear to be a significant drop in the GlcN and GlcN-Ins levels and an increase in the H 2 S level.
Effect of Nutrient Supplementation and Deprivation Upon Intermediate Levels-To test whether an increase in cellular Ins content would influence the levels of MSH or potential MSH intermediates, we examined M. smegmatis MC 2 155 incubated with Ins. We initially tested Ins uptake by adding 37 M [ 14 C]Ins to a culture at A 600 ϭ 1.0 and determining the loss of radiolabel from the medium and its appearance in cells. After 2.5 h, 77% of the counts had been lost from the medium to the cells, showing that M. smegmatis efficiently imports Ins. Next we harvested cells after growth to mid-log phase, resuspended them at A 600 ϭ 1.05 in 7H9 media with glucose and 1 mM Ins at 37°C, and took samples for analysis at 0, 0.5, and 2.5 h (Fig.  6A). The GlcN-Ins content was initially more than 3-fold greater than normal but fell to about half normal at 0.5 h before returning to normal at 2.5 h. Smaller reciprocal changes occurred in the Cys levels, whereas the MSH and H 2 S levels did not change to a significant extent. The A 600 value increased 40% over the 2.5 h incubation, a change comparable with that found for exponential cultures without added Ins.
To test whether changes in cellular Cys levels would influence MSH or MSH intermediates, we incubated cells in medium containing 10 mM AcCys. Because AcCys is readily taken up by passive diffusion, we expected this to produce markedly elevated levels of cellular AcCys and, through deacylation, Cys. The results (Fig. 6B) showed this to be the case; the cellular AcCys level was elevated 1000-fold over the normal level, and the Cys level was increased 50-fold. This was accompanied by a 5-to 10-fold decrease in GlcN-Ins content, whereas the H 2 S level roughly doubled. The Cys-GlcN-Ins content was elevated 20-fold over normal immediately after the start of incubation but then fell a factor of 70 over the next 2.5 h. Least affected was the MSH content, which was elevated ϳ40% after 2.5 h. The A 600 value changed little (5%) over the 2.5-h incubation period, indicating that AcCys significantly inhibits growth under these conditions. Incubation of cells in the absence of glucose (Fig. 6C) resulted in a decrease in GlcN-Ins level to values 13-30% of normal during the initial 2.5-h incubation, but continuation of the incubation resulted in a return to normal values by the end of 8 h (data not shown). The MSH level increased to 40% over normal by 2.5 h and remained at that value to 8 h. The value of A 600 declined slightly (10%) during the 2.5-h incubation period, showing that glucose deprivation arrested cell growth.
Assay of Enzyme Activity for ATP-dependent Ligation of Cys and AcCys to GlcN-Ins- Bornemann et al. (6) demonstrated the presence of enzyme activity in extracts of M. smegmatis capable of converting Cys plus GlcN-Ins to Cys-GlcN-Ins in the presence of ATP and the accompanying production of MSH, which was enhanced by acetate or acetyl-CoA. They were unable to assess AcCys as substrate in this reaction, because it was rapidly converted to Cys. We conducted analogous assays with an undialyzed supernatant fraction from extraction of M. We note first that whereas AcCys and GlcN-Ins were found at measureable levels, Cys-GlcN was never detected in cell extracts, even when cellular Cys was present at high levels in cells incubated in AcCys. This indicates that step 3 of Scheme I is not involved in the biosynthesis of MSH and that steps 7, 8, and 11 are therefore also not involved. Because AcCys-GlcN was not detected in cells during normal growth, AcCys-GlcN also appeared not to be involved in the biosynthesis path leading to MSH, but this conclusion is less certain because of the limits of the analytical method to measure low levels of AcCys-GlcN. However, the failure to observe AcCys-GlcN during incubation of cells in AcCys, where levels of AcCys are markedly elevated, argues against step 4, and therefore step 9, as components of the pathway leading to MSH. This leaves steps 5 and 6 as the remaining possible second steps in MSH biosynthesis. If step 6 is involved in MSH production, then elevation of the cellular Cys level by incubation of cells in AcCys should produce an elevation in the level of Cys-GlcN-Ins, and this is observed (Fig. 6B). This is consistent with biosynthesis of MSH occurring via steps 2, 6, and 10 but does not exclude step 5 as a concurrent pathway. The finding that cell extracts incubated with GlcN-Ins and Cys produce Cys-GlcN-Ins at a rate more than 10-fold faster than the rate that extracts incubated with GlcN-Ins and AcCys produce MSH demonstrates that the route via steps 6 and 10 is the favored path leading to MSH.
The present conclusions on the biosynthetic pathway from measurements of cellular metabolite levels are in accord with those of Bornemann et al. (6) based on in vitro measurements with cell extracts incubated with synthetic GlcN-Ins. Although their extracts converted AcCys to Cys at a rate too great for them to distinguish which of these substrates was favored for reaction with GlcN-Ins, we did not experience that difficulty, and our results clearly indicated that Cys was the preferred substrate.
An estimate of the cellular concentrations of GlcN-Ins and Cys-GlcN-Ins is useful for comparison to the K m values for the enzymes using these substrates. Assuming a ratio of cell water to cell RDW of 3, the cellular concentrations of GlcN-Ins and Cys-GlcN-Ins during exponential growth (Table II) are calculated as 33 Ϯ 14 and 2.6 Ϯ 1.6 M, respectively. The value for GlcN-Ins is about one-fourth the value of K m estimated for GlcN-Ins in its conversion to Cys-GlcN-Ins with saturating Cys and ATP, using a partially purified extract from M. smegmatis (6). This suggests that changes in the cellular GlcN-Ins level will produce a corresponding change in the rate of conversion of GlcN-Ins to Cys-GlcN-Ins. The observed high ratio of GlcN-Ins to Cys-GlcN-Ins indicates that the effective first order rate constant of conversion of the latter to MSH is more than an order faster than the corresponding rate of formation from GlcN-Ins during exponential growth.
One postulated function of MSH is to serve as a slowly autoxidizing cellular reserve of Cys (3). If this is correct then there must be a pathway by which Cys is regenerated from MSH, and this would be expected to be different from the reverse of the biosynthesis route. We know that M. smegmatis is capable of converting intracellular AcCys to Cys (Fig. 6B), and we can estimate from the data in Table II that the AcCys level in exponential cells is ϳ4 M. The source of the AcCys is unclear but may involve step 5 of Scheme I as a degradative pathway of MSH. The glucose deprivation experiment (Fig. 6C) was conducted in part to see whether loss of a primary energy source needed to synthesize MSH might result in changes in intermediate levels, indicative of the MSH-degradative pathway. However, under these conditions the MSH level did not decline but actually increased by 40%, and the AcCys level exhibited little change. Thus, no clear insight into the degradative pathway was obtained from this experiment.
The present results demonstrate that the MSH level is very tightly regulated in M. smegmatis. Only minor changes in concentration were observed in the transition from exponential to stationary phase (Fig. 5) during glucose deprivation (Fig. 6C) and during inositol or AcCys/Cys supplementation (Fig. 6, A  and B). The steady-state level of MSH depends upon the net balance in its rate of production and its rate of degradation. These can be influenced at a variety of levels, from feedback inhibition/activation of key enzymes in the pathway to regulation of gene expression, and elaboration of the mechanisms involved will be of interest as more is learned about the enzymes of MSH metabolism.
In conclusion, the present studies have provided some important tools for the study of MSH metabolism. Their application to M. smegmatis has provided evidence that GlcN-Ins and Cys-GlcN-Ins are intermediates in the biosynthetic pathway and suggested that AcCys and GlcN-Ins may be involved in the degradative pathway. The HPLC analysis methods presented here complement our immunoassay methods for MSH (8), and together these methods provide important tools for identifying and analyzing mycobacterial mutants blocked in MSH biosynthesis, a topic of subsequent papers in this series.