ABSTRACT
It was determined that the dTDP-rhamnose synthesis gene, rmlD, could be inactivated in Mycobacterium smegmatis only in the presence of a rescue plasmid carrying functional rmlD. Hence, dTDP-rhamnose biosynthesis is essential for the growth of mycobacteria and the targeting of dTDP-rhamnose synthesis for new tuberculosis drugs is supported.
The mycobacterial cell wall consists of a mycolic acid layer tethered to peptidoglycan via the polysaccharide arabinogalactan (3, 13, 14). Arbinogalactan is attached to peptidoglycan viaα-l-rhamnopyranosyl-(1→3)-α-d-N-acetyglucosaminosyl-1-phosphate (13). This structural arrangement strongly suggests that rhamnosyl residues are essential for the growth and viability of mycobacteria.
l-Rhamnosyl residues are not present in humans. They are synthesized in bacteria from dTDP-glucose via the enzymes RmlB, RmlC, and RmlD (5, 8, 15, 17). There is no salvage pathway for the formation of dTDP-l-rhamnose (dTDP-Rha) as there is with GDP-l-fucose (19, 20). The dTDP-Rha synthetic enzymes are soluble and have been studied by X-ray crystallography (1, 6, 7). Given these facts, the dTDP-Rha formation enzymes have significant potential as targets for new tuberculosis drugs. Therefore, it is important to directly demonstrate that an enzyme involved in the formation of these enzymes is essential.
The final enzyme in the series synthesizing dTDP-Rha is dTDP-6-deoxy-l-lyxo-4-hexulose reductase. Its gene, rmlD, is the first gene of an operon which also contains wbbL and manB (12). This arrangement occurs in all the mycobacterial genomes sequenced. The gene wbbL encodes the rhamnosyl transferase that inserts rhamnose into the cell wall and is also expected to be essential. Thus, a nonpolar rmlD knockout mutation was desired so that complementation would be straightforward.
The basic strategy used was to prepare a copy of rmlD interrupted with a kanamycin resistance cassette orientated in the same direction as rmlD with the hope that, after gene replacement, the downstream wbbL and manB genes would be transcribed from the kanamycin resistance promoter. The disrupted rmlD gene was then used to replace wild-type rmlD in the presence of an appropriate rescue plasmid.
Constructing the rmlD replacement plasmid (pFP201) and obtaining the first homologous recombination event.
A partial genomic DNA library of a mycobacterial lab strain (18) was constructed by isolation of SmaI fragments of approximately 3.5 kb and ligation into pCR-Blunt (Invitrogen, Carlsbad, Calif.). The rmlD gene was located by colony hybridization (9) using Mycobacterium tuberculosis rmlD (10) as a probe. Preliminary sequence data showed that the rmlD DNA sequence was almost identical to that of Mycobacterium smegmatis mc2155 (The Institute for Genomic Research website [http://www.tigr.org/ ]). Then a 1.6-kb DNA fragment (containing rmlD and 417 bp upstream and 328 bp downstream of it) was cut out by EcoRI and XhoI, filled in with the Klenow fragment, and then inserted into pCR-Blunt to generate pCR-rmlD. By using methods previously described (18), pCR-rmlD was used to construct pFP201 (Table 1), a plasmid with a temperature-sensitive (TS) origin of replication which carries rmlD::kan, sacB, and xylE. Although the orientation of kan in pFP201 was not controlled by the procedure used (18), it was shown in the case of pFP201 to be the same as that of rmlD by restriction enzyme digestion. Plasmid pFP201 was then electroporated into M. smegmatis mc2155, and a transformant was propagated in Luria-Bertani broth (LB)-kanamycin medium at 30°C, followed by plating onto LB-kanamycin plates at 42°C. Since the TS plasmid was able to replicate at 30°C but not at 42°C, the kanamycin-resistant colonies that appeared on the 42°C plates had necessarily integrated the kan gene into their chromosome. Analysis of colonies on these plates by Southern blotting revealed a colony arising from homologous recombination (Fig. 1) which was propagated for further experiments (M. smegmatis FP201) (Table 1).
Southern blot analysis of M. smegmatis. (A-1) Southern blot of DNA from the first-crossover strain, M. smegmatis FP201 (Table 1). DNA was cleaved with SmaI, and the probe was rmlD. (A-2) Origins of the fragments. (B-1) Southern blot of DNA from the second-crossover strain, M. smegmatis YM202 (Table 1). DNA was cleaved with SmaI, and the probe was rmlD. (B-2) Origins of the fragments. In the original blot, a fragment at 10 kb that originated from the rescue plasmid was faintly visible. Values to the sides of the blots are molecular sizes (in kilobases).
Key bacterial strains and plasmids
Construction of an rmlD rescue plasmid.
A rescue plasmid, pYM201 (Table 1), was constructed by digesting the M. tuberculosis bacterial artificial chromosome Rv3 clone (2) with NotI and Avr-II, yielding a 3,747-bp fragment containing the rmlD-wbbL-manB operon. The DNA ends were filled in (End Conversion Mix; Novagen, Madison, Wis.) and ligated to pSTBlue-1 (Novagen) to generate the pSTB1-M. tuberculosis rmlD operon. This plasmid was digested with EcoRI and XbaI to get a 1.5-kb fragment containing rmlD and approximately 550 bp of DNA upstream of the start codon. The fragment was filled in with the Klenow fragment and ligated to XbaI-cleaved and -blunted pCG76 (18) to generate pYM201 (Table 1). The plasmid had the same TS origin of replication used in pFP201.
Second-crossover attempts and events.
Single-colony isolates of M. smegmatis FP201 with and without rescue plasmid were grown in LB-kanamycin medium at 30°C and then plated onto LB-kanamycin-sucrose plates at 30°C. The resulting colonies were analyzed for their XylE phenotype (a yellow color develops in colonies expressing xylE when they are sprayed with catechol). Colonies that have undergone a second crossover should both be able to grow on sucrose and have lost the capacity for XylE enzyme production; colonies that can grow on sucrose but still express xylE are likely to arise from mutations in sacB rather than from the second-crossover event. Examination of the data revealed that only yellow colonies were obtained without the rescued plasmid but that 51% of the colonies were white (xylE absent) and 49% of the colonies were yellow when the rescue plasmid was present. Eighteen of these white colonies were analyzed by SmaI digestion and Southern blot analysis; all 18 showed bands at 2.3, 2.9, and 10.6 kb (the 10.6-kb band was from the rescue plasmid) as expected (Fig. 1) for the second single-crossover event. One colony, designated M. smegmatis YM202 (Table 1), was propagated for further experiments.
M. smegmatis YM202 will not grow at 42°C.
Curves indicating growth at 30 and 42°C were obtained (Fig. 2) for M. smegmatis YM202 containing the rescue plasmid pYM201 and, as a control, for wild-type M. smegmatis mc2155 containing pYM201. The results clearly showed that M. smegmatis YM202 was unable to grow at the temperature at which the rescue plasmid was lost, confirming that rmlD is essential for growth.
Growth curves of M. smegmatis strains at 30 and 42°C. The growth of the control strain, M. smegmatis mc2155, containing plasmid pYM201 at 30°C (▴) and at 42°C (▵), and the growth of the rmlD knockout strain, M. smegmatis YM202, containing plasmid pYM201 at 30°C (•) and at 42°C (○), are illustrated.
The wbbL gene is transcribed and translated in YM202.
To confirm transcription of wbbL, we prepared mRNA from M. smegmatis YM202 and showed that it hybridized with the wbbL DNA probe (data not shown). To confirm translation, we ran an enzyme assay for rhamnosyl transferase (16) and showed by thin-layer chromatography that the product of WbbL, α-l-Rha-(1→3)-α-d-GlcNAc-(1→P)-P-decaprenyl, was formed by membranes prepared from M. smegmatis YM202 (data not shown). Thus, the knockout is clearly a nonpolar event. In an earlier study (18), where the galactopyranose mutase gene (glf) was knocked out with the kanamycin cassette oriented in the direction opposite to the coding direction of glf, the mutation was polar since the gene downstream of glf, Rv3808c, was not transcribed. (In the original publication, the orientation of the kanamycin cassette was presented as being in the same direction as glf. However, the assignment was in error and an author correction to that effect has been published [see reference 18]).
Application of results to M. tuberculosis.
The experiments were done with M. smegmatis due to the fast-growth characteristics of this organism and the availability of a TS origin of replication for it. We have shown that the basic structures of the cell walls of all mycobacteria are indistinguishable from one another by 13C nuclear magnetic resonance and oligosaccharide profiling (4) and, in particular, that all mycobacteria have exactly the same linker structure with an identically positioned rhamnosyl residue (13). In addition, rmlA, rmlB, rmlC, and rmlD are found in the genomes of M. tuberculosis, Mycobacterium bovis, Mycobacterium avium, M. smegmatis, and Mycobacterium leprae (12). Thus, the identity of the rhamnosyl-containing linker, along with the genetics, strongly argues that dTDP-Rha formation enzymes are essential in all mycobacteria, including M. tuberculosis.
It is worthwhile to note that an understanding of the unique structural role of the l-rhamnosyl residue in the cell walls of M. tuberculosis cells (13) was required before RmlD emerged as an important drug target; such an insight is not available from the genome sequence alone and underscores the fact that fundamental organism-specific biochemistry is required to most effectively exploit genome data. The expression of rmlD (10) and development of microtiter-based assays to detect inhibitors of the Rml enzymes (12) have also been accomplished. With these assays, X-ray crystallographic studies (7), and the work reported herein, the framework to proceed in the development of drugs against dTDP-rhamnose synthesis is in place.
ACKNOWLEDGMENTS
This work was supported by funds provided through Public Health Service grant AI-33706, NIAID, NIH.
We thank Stewart Cole for providing the Institut Pasteur biological material, i.e., the M. tuberculosis bacterial artificial chromosome Rv3 clone (2).
FOOTNOTES
- Received 19 November 2001.
- Accepted 20 March 2002.
- Copyright © 2002 American Society for Microbiology