OUP user menu

Characterization of the alanine racemases from two Mycobacteria

Ulrich Strych , Rebecca L. Penland , Margarita Jimenez , Kurt L. Krause , Michael J. Benedik
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10547.x 93-98 First published online: 1 March 2001


d-Alanine is a necessary precursor in the biosynthesis of the bacterial peptidoglycan. The naturally occurring l-alanine isomer is racemized to its d-form through the action of a class of enzymes called alanine racemases. These enzymes are ubiquitous among prokaryotes, and with very few exceptions are absent in eukaryotes, making them a logical target for the development of novel antibiotics. The alanine racemase gene from both Mycobacterium tuberculosis and M. avium was amplified by PCR and cloned in Escherichia coli. Overexpression of the proteins in the E. coli BL21 system, both as native and as His-tagged recombinant products, has been achieved. The proteins have been purified to electrophoretic homogeneity and analyzed biochemically. A d-alanine requiring double knock-out mutant of E. coli (alr, dadX) was constructed and the cloned genes were able to complement its deficiencies.

  • Mycobacterium tuberculosis
  • Mycobacterium avium
  • Alanine racemase

1 Introduction

The Gram-positive bacterium Mycobacterium tuberculosis, the causative agent of tuberculosis, is a major human pathogen and thought to be responsible for 3 million deaths worldwide each year [1]. Mycobacterium avium constitutes part of the M. avium intracellular complex and, as such, is one of the most common infections in AIDS patients with advanced immunodeficiency [2]. The resurgence of these opportunistic organisms is primarily, but not exclusively, due to infections involving patients with AIDS [3]. The mortality and morbidity of many tuberculosis infections is further worsened by the increasing emergence of multi-drug resistance mycobacterial strains [4]. The treatment of multi-drug resistant tuberculosis is associated with higher treatment failure rates, higher mortality rates, and it also imposes a noticeable burden on the health care system [5]. The need for new antibiotics to fight this infection is unequivocal.

The uniquely complex mycobacterial cell envelope in particular is the focus of numerous ongoing drug discovery programs [6]. For instance, the pathways leading to the synthesis of many of the unique sugars and lipids in the mycobacterial cell wall are being investigated as potential targets for antimicrobial agents. Most prominently, mycolic acid biosynthesis and its inhibitors are currently studied in many groups [1,7,8].

Alanine racemase (EC is a proven antimicrobial target that deserves re-evaluation in the light of the advances in computer-based drug design. It catalyzes the reversible racemization between l- and d-alanine, an essential component of the peptidoglycan layer in both Gram-negative and Gram-positive bacteria. The trace amounts of d-alanine found in vertebrates can most likely be attributed to the breakdown of the food and intestinal flora [9].

This pyridoxal 5′-phosphate (PLP)-dependent enzyme is ubiquitous among prokaryotes, making it an attractive drug target. It is essential for bacterial growth, and, with only a few exceptions [7,8,10], is absent in eukaryotes. However, only one alanine racemase inhibitor, the structural d-alanine analogue, d-cycloserine, has been marketed clinically. Although it is an excellent inhibitor of Mycobacteria, as well as many other pathogenic bacterial species, serious side effects, especially CNS toxicity, have markedly limited its use [11,12].

Bacteria investigated to date have been found to possess either one or two alanine racemase genes. In Escherichia coli [13], Salmonella typhimurium [14], and Pseudomonas aeruginosa [15] two genes are present. The alr gene encodes the constitutively expressed biosynthetic enzyme, sufficient to provide enough d-alanine for cell wall biosynthesis [14]. The catabolic dadX gene encodes a second alanine racemase isozyme whose expression is subject to induction by l-alanine [16]. When l-alanine is abundant, the dadX gene product converts l-alanine to d-alanine, which through the subsequent action of d-amino acid oxidase and lactate dehydrogenase, is converted to pyruvate and is thus metabolized. All known alanine racemases have significant sequence similarity, with about 35–50% amino acid identity. In Mycobacteria including M. avium and M. tuberculosis only one alanine racemase gene is present.

In this paper we report the cloning, overexpression, purification and preliminary characterization of the alanine racemases from M. tuberculosis and M. avium. This represents a necessary step in the crystallization and subsequent structure-based approach to designing new antibacterial agents.

2 Materials and methods

2.1 Bacterial strains, plasmids and growth conditions

The plasmids and strains used in this study are described in Table 1. Cultures were routinely grown in LB medium or M9 minimal medium at 37°C. When l- or d-alanine was used as the carbon and energy source, it was supplemented at a final concentration of 50 mM. When used to support growth of d-alanine requiring strains in the presence of other substrates it was supplemented to 50 μM. When necessary, the growth medium was supplemented with 50 μg ml−1 ampicillin, 25 μg ml−1 kanamycin, 12 μg ml−1 tetracycline or 30 μg ml−1 chloramphenicol.

View this table:

Bacterial strains and plasmids

Strains and plasmidsGenotypeReference
E. coli
BL21(DE3), pLysSF, ompT, hsdSB (RB MB) gal dcm (DE3), CmRNovagen, Madison, WI, USA
MB1457DH5αrecA+laboratory stock
MB1547supE thi hsd Δ5 Δ(lac-proAB) endA F′[traD36 proAB lacIq lacZ ΔM15]laboratory stock
MB2034MC1000, dadXEC::frtthis work
MB2159MC1000, dadXEC::frt, alrEC::frtthis work
MC1000.1araD139 Δ(ara-leu)7679 galU galK Δ(lac)174 rpsL thi-1laboratory stock
pET26T7lac promoter controlled expression vector, KmRNovagen, Madison, WI, USA
pET28T7lac promoter controlled expression vector, allows His-tagging of the expressed protein, KmRNovagen, Madison, WI, USA
pMAK705Temperature-sensitive pSC101 vector, CmR[19]
pMAKblapMAK705, with bla gene integrated at unique ClaI site, CmR, ApRthis work
pMB1556pET17-NdeI-alrEC-EcoRI, ApRthis work
pMB1861pET17-NdeI-dadXEC-BamHI, ApRthis work
pMB1909pET26-XbaI-alrTB-short-EcoRIthis work
pMB2011pET26-XbaI-alrTB-long-EcoRIthis work
pMB2103pET28-XbaI-alrTB-short-EcoRIthis work
pMB2119pET28-XbaI-alrTB-long-EcoRIthis work
pMB2289pET28-NdeI-alrMA-BamHIthis work

2.2 DNA manipulations

DNA sequencing was performed on an ABI Prism 373 sequencer using the dye terminator cycle sequencing kit (Perkin-Elmer, Foster City, CA, USA). Primers complementary to the T7 promoter and terminator regions, as well as several gene specific primers (Table 2), were used to verify the sequence of the cloned fragments.

View this table:

Primers used in this study


2.3 PCR

PCR with PlatinumTaq DNA polymerase (Life Technologies) was performed in a GeneAmp 2400 thermal cycler (Perkin-Elmer). One μg of chromosomal DNA was used as template in a 50-μl touchdown PCR reaction [8]. DNA from M. tuberculosis was a gift from Dr. J. Musser, NIAID, NIH, Rocky Mountain Laboratories, Hamilton, MT, USA. Dr. J. Inamine, Colorado State University, Fort Collins, CO, USA, provided DNA from M. avium 2151. Oligonucleotides were obtained from Life Technologies and are listed in Table 2.

2.4 Overexpression and purification

The recombinant expression plasmids were transformed into E. coli BL21(DE3), pLysS. For the purification of the non-His-tagged alanine racemase, cells were grown, induced and harvested as described [15]. The subsequent purification was performed as described by Strych et al. [15] with the exception that the two Q-Sepharose purification steps were performed at pH 8.0 and pH 7.2, respectively. The alanine racemase containing fractions were identified by means of a spectrophotometric assay [17]. SDS–PAGE and isoelectric focusing confirmed the purity of alanine racemase in the peak fractions.

His-tagged alanine racemase was purified from cells harboring pET28-derived plasmids using Pharmacia's HisTrap™ columns, according to the manufacturer's instructions (‘optimized protocol’).

3 Results and discussion

3.1 Construction of an alr and dadX null mutant of E. coli MC1000

In order to test our cloned fragments for in vivo alanine racemase activity, we constructed a d-alanine auxotrophic double mutant of E. coli MC1000.1. The alrEC and dadXEC genes from E. coli MC1000.1 were amplified by colony PCR. Restriction sites added to the PCR primers (Table 2) allowed the cloning of the PCR products into the gene replacement vector pMAKbla. The dadXEC gene was digested with EcoRV, 666 bp downstream of the translational start codon and ligated with a 1-kb kanamycin resistance cassette flanked by frt recombination sites from pCP15 [18]. The gene replacement and the removal of the kanamycin cassette by FLP recombination were performed as described [18,19]. When the dadXEC allele was amplified from the resulting strain, MB2034, and analyzed by gel electrophoresis, it was found to be approximately 50–100 bp larger than the wild-type allele (Fig. 1). This was as predicted since the excision of the kan cassette leaves a 63-bp frt recombination site behind, interrupting the open reading frame. Also, in contrast to its parent strain, MB2034 failed to grow on minimal medium with l-alanine as the sole carbon and energy source, as would be expected for a dadX mutant. It did however grow in LB broth, demonstrating its ability to synthesize d-alanine using its alr gene.


Agarose gel electrophoresis comparing the PCR products after amplification of the alr and dadX alleles from E. coli wild-type MC1000.1 (WT) and the alanine racemase double mutant MB2034 (M). The mutant alleles run higher due to the insertion of the frt cassette into the wild-type genes.

In an analogous approach the alrEC gene was amplified and cloned into pMAKbla. Inserting the frt kanamycin resistance cassette into a SmaI site 633 bp downstream from the translational start codon interrupted this gene. Gene replacement and removal of the antibiotic resistance cassette were achieved as described above, but using dadX mutant strain MB2034 from above as host. The successful gene replacement was again verified by PCR. The resulting dadX and alr double mutant strain, MB2159, does not grow on LB medium; it requires the supplementation of 50 μM d-alanine. It thus is a d-alanine auxotroph, and as such is an excellent tester strain to assay for functional alanine racemases.

3.2 Amplification and cloning of the mycobacterial alanine racemase genes

The M. tuberculosis genome sequencing project entry for alr (GenBank accession no.: Z77165) suggests a GTG codon at the 5′ end of the alr open reading frame as the start codon for alanine racemase. However, when lined up with other alanine racemases (Fig. 2) an ATG codon 72 nucleotides (nt) further downstream appeared to us to be the more likely candidate.


Alignment of the N-terminal amino acid sequences of alanine racemases. PLP: pyridoxal phosphate attachment site (VxKA(D/N)(G/A)YGHG) (PROSITE accession no. PS00395). An arrow indicates our predicted N-terminal amino acid of the functional alanine racemase from M. tuberculosis. The sequences are taken from the following GenBank files: Mb_tub: M. tuberculosis (AF172731), Mb_lep: Mycobacterium leprae (U00020), Mb_avi: M. avium (AF214487), Mb_sme: Mycobacterium smegmatis (U70872), Ba_psy: Bacillus psychrosaccharolyticus (AB021683), Ba_sub: Bacillus subtilis (Z99106), Ba_ste: Bacillus stearothermophilus (MN19142), Li_mon: Listeria monocytogenes (AF038438), La_pla: Lactobacillus plantarum (Y08941), St_pne: Streptococcus pneumoniae (AF171873), Ec_alr: E. coli alr (AE000478), Sty_alr: S. typhimurium alr (M12847), Hm_inf: Haemophilus influenzae (U32831), Ec_dadX: E. coli-dadX (AE000217), Sty_dadX: S. typhimurium dadB (K02119), Kl_aer: Klebsiella aerogenes dadB (AF016253), Pa_alr: P. aeruginosa alr (AF165882), Pa_dadX: P. aeruginosa dadX (AF165881), St_aur: Staphylococcus aureus (Y16431), Tre_pal: Treponema pallidum (AE001242), He_pyl: Helicobacter pylori 26695 (AE000603), Ri_prow: Rickettsia prowazekii (AE235270).

Using two different oligonucleotides (MT-alr-short-F, MT-alr-long-F, Table 2), paired with a common reverse primer, MT-alr-R, both open reading frames were amplified separately. The longer amplification product comprises 1224 nt, and would encode a protein (ALRTB-long) of 408 amino acids with a molecular mass of 43 385 Da. The shorter open reading frame is 1152 nt, encoding a 384-amino acid protein (ALRTB-short) with a molecular mass of 40 721 Da. Restriction sites (XbaI and EcoRI) had been added to the primers, allowing the forced cloning of the PCR fragments into the T7lac promoter expression vectors pET26 and pET28 (leading to 6×His-fusion proteins). The resulting plasmids are listed in Table 1. DNA sequence analysis of the cloned PCR products confirmed the complete identity of both fragments with the sequence published in GenBank.

Comparing the translated E. coli alr and dadX sequences against the unfinished sequence of M. avium in GenBank identified the putative sequence of the alanine racemase gene from M. avium. Only one significant match was identified. It consists of an 1176-nt open reading frame, which would encode a 391-amino acid protein with a molecular mass of 41 243. It is 83%, 35% and 32% identical with the M. tuberculosis Alr, E. coli Alr and E. coli DadX proteins, respectively. The sequence of its PLP binding site is identical with that of M. tuberculosis (Fig. 2).

Primers containing NdeI and BamHI restriction sites were designed (Table 2), the gene was amplified and cloned into pET28, yielding pMB2289. DNA sequencing confirmed 100% identity with the sequence from the unfinished M. avium genome.

3.3 Complementation analysis

The in vivo activity of the gene products encoded by the cloned alanine racemase isoforms was investigated in a complementation experiment. The d-alanine auxotrophic strain MB2159 was transformed with the M. tuberculosis plasmids pMB1909, pMB2011, pMB2103, and pMB2119, and with the M. avium alr carrying pMB2289. In addition, plasmids encoding the cloned E. coli alanine racemases, pMB1556 and pMB1861, were used as positive controls. The original pET vectors, pET17, pET26 and pET28, served as negative controls. Cells were plated on LB medium with and without d-alanine supplementation, and scored for colony growth after 12 and 24 h at 37°C (Table 3). While the E. coli alanine racemase genes fully restored the wild-type phenotype, the alrTB-short and the alrAV alleles complemented the mutant slightly less efficiently, whereas the alrTB-long allele failed to complement.

View this table:

Complementation of a d-alanine auxotroph

Growth after 12 haGrowth after 24 h
MB2159 (alr, dadX)
pMB1556 (alrEC)++
pMB1861 (dadXEC)++
pMB1909 (alrTB-short)+
pMB2011 (alrTB-long)
pMB2103 (HIS6-alrTB-short)+
pMB2119 (HIS6-alrTB-long)
pMB2289 (HIS6-alrAV)+
  • aLB-agar plates without d-alanine were scored after incubation at 37°C, (−) indicates no visible growth and (+) indicates visible growth.

3.4 Overexpression, purification and biochemical characterization

The recombinant plasmids pMB1909 (alrTB-short), pMB2011 (alrTB-long), pMB2103 (alrTB-short-His), pMB2119 (alrTB-long-His) and pMB2289 (alrAV-long-His) were transformed into E. coli BL21(DE3), pLysS. The five alanine racemase variants were prepared as described under Section 2 (Fig. 3). The purified alanine racemases from E. coli pMB1909 and pMB2289 were characterized with respect to their reaction kinetics. The M. tuberculosis enzyme has a Km for d-alanine at 23°C of 1.1 mM and for l-alanine of 1.2 mM. The Vmax for the racemization (d- to l-alanine and l- to d-alanine) is 0.46 and 0.51 U mg−1, respectively, where one unit was defined as the amount of enzyme that catalyzed racemization of 1 μmol of substrate per minute. When these values were used, the calculated Keq for this reaction was 1.08, a number that is in qualitative agreement with the expected value for this chemically symmetric reaction [20]. Quantitatively and qualitatively similar results were obtained for the His-tagged alanine racemase expressed from pMB2103.


Purification of alanine racemase from M. avium (MA) and M. tuberculosis (TB) expressed in E. coli BL21(DE3), pLys. Lanes 1–5: uninduced controls (crude extracts). Lanes 7–11: crude extracts after induction with 1 mM IPTG and growth for 3 h. Lanes 13–17: purified alanine racemases. Lanes 1, 7, 13: ALRTB-short. Lanes 2, 8, 14: ALRTB-long. Lanes 3, 9, 15: ALRTB-short, His-tagged. Lanes 4, 10, 16: ALRTB-long, His-tagged. Lanes 5, 11, 17: ALRMA-short, His-tagged. Lanes 6, 12, 18: Molecular mass standard (Life Technologies).

The M. avium enzyme has Km values of 0.4 mM for d-alanine and 0.5 mM of l-alanine. The Vmax values for the racemization (d- to l-alanine and l- to d-alanine) are 1.4 and 1.2 U mg−1, respectively (Keq=0.93).

While the Km values for the mycobacterial enzymes are in the same range as those published for ALREC (Km (l-alanine)=0.97, Km (d-alanine)=0.42 mM), their maximum velocity is strongly reduced compared to the 28 (L→D) and 12 U mg−1 (D→L) determined for E. coli [13]. It is likely that this difference explains why the mycobacterial genes only allow slow growth of the E. coli auxotrophic mutant. Presumably the slower growth of Mycobacterium allow this reduced production of d-alanine to not be limiting.

When both enzymes were assayed in a time-dependent inactivation assay [21] both d-cycloserine and the racemic mixture of d- and l-cycloserine were effective inhibitors (data not shown).

ALRTB-long protein was also purified from cultures harboring pMB2011 and pMB2119, encoding the long form of the M. tuberculosis enzyme. Both proteins were enzymatically inactive, thus confirming the results of the complementation analysis.

In order to better understand the lack of activity of ALRTB-long, and consequently its failure to complement the d-alanine auxotroph, we performed a hydrophobicity analysis (Fig. 4). It appears that the N-terminal 24 amino acids form a pronounced hydrophobic region that may interfere with dimerization or ligand binding or folding. When we further compared the electrophysical properties of the two isoforms by isoelectric focusing, we did not observe any dramatic difference in their overall charge (Fig. 5). Both isoforms focus near their predicted theoretical isoelectric points (calculated at http://www.expasy.ch/) of 5.94 (ALTTB-long) and 6.08 (ALTTB-short). In future projects the comparison of the structure of both isoforms of the enzyme might contribute to a better understanding of the reaction mechanism.


Hydrophobicity analysis of the short and the long form of ALRTB according to Kyte–Doolittle [22]. A shaded box overlays the hydrophobic N-terminal region exclusively present in ALRTB-long. The hydropathy values were calculated for a window length of 17 and are indicated on the Y-axis.


Isoelectric focusing of ALRTB-short and ALRTB-long on a Pharmalyte 3-9 gel (Pharmacia). The separation was performed on Pharmacia's Phast Gel system. Lane 1: ALRTB-long. Lane 2: ALRTB-short. Lane 3: IEF standard (Bio-Rad).

Crystallization of the purified ALRTB from E. coli harboring pMB1909 was undertaken using the crystal screen reagent kit (Hampton Research, Riverside, CA, USA). Fourty-μg aliquots of the purified protein were mixed with each of the individual reagents and allowed to slowly evaporate at 4°C. Crystals of 0.01×0.02 mm were obtained in a solution of 2.2 M ammonium sulfate. These crystals are not yet suitable for diffraction analysis, but further experiments with purified protein from both mycobacterial species are currently underway.


This work was supported by NIH Grant AI46340, the Robert A. Welch Foundation, and Grants 3652-233 and 3652-272 from the Texas Advanced Research Program.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
View Abstract