OUP user menu

Killing activity and rescue function of genome-wide toxin–antitoxin loci of Mycobacterium tuberculosis

Amita Gupta
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01400.x 45-53 First published online: 1 January 2009


Toxin–antitoxin (TA) loci are typically two-component systems that encode a stable toxin, which binds an essential host target leading to cell growth arrest and/or cell death, and an unstable antitoxin, which prevents the cytotoxic activity of the toxin. The ubiquitous presence of these loci in bacterial genomes, along with their demonstrated toxicity not only in the native but also in heterologous systems, has provided the possibility of their use in wide-spectrum antibacterials. Mycobacterium tuberculosis contains nearly 40 TA loci, most of which are yet to be characterized. Here we report the heterologous toxicity of these TA loci in Escherichia coli and show that only a few of the M. tuberculosis-encoded toxins can inhibit E. coli growth and have a killing effect. This killing effect can be suppressed by coexpression of the cognate antitoxin. This work has identified functional TA pairs for sequences that are presently unannotated in the mycobacterial genome. These toxins need to be further tested for their activity in the native host and other organism backgrounds and growth environments for utilization of their antibacterial potential.

  • toxin–antitoxin loci
  • Mycobacterium tuberculosis
  • cell death
  • persistence


Toxin–antitoxin (TA) modules are parasitic devices, which were initially identified as mediators of plasmid stability in host cells (Gerdes et al., 1986). TA modules typically encode two gene products: a stable toxin that binds an essential host target leading to cell growth arrest and/or cell death, and an unstable antitoxin that prevents the activity of the toxin and is itself susceptible to degradation by an intracellular host protease (Van Melderen et al., 1994; Lehnherr & Yarmolinsky, 1995; Engelberg-Kulka & Glaser, 1999). Recently, chromosomal genes similar to the plasmid-borne addiction modules have been identified in numerous bacterial species (Pandey & Gerdes, 2005).

The TA loci have been classified into seven groups namely, ccd, parDE, mazEF, relBE, vapBC, phd/doc and higBA (Gerdes et al., 2005). The toxin components of different groups have different modes of action. For example, the RelE, HigB and MazF toxins and recently the VapC toxin have been shown to act as ribonucleases and to inhibit translation (Pedersen et al., 2003; Zhang et al., 2003; Christensen-Dalsgaard & Gerdes, 2006). The phd/doc system is also proposed to act at the level of translation (Hazan et al., 2001). The CcdB toxin of the F-episome is a gyrase poison, which inhibits DNA replication (Critchlow et al., 1997; Dao-Thi et al., 2005). Gyrase has also been shown to be the target for the ParE toxin of the broad host-range RK-2 plasmid parDE system (Jiang et al., 2002).

The chromosomal TA loci have been described as mediators of programmed cell death (PCD) in bacteria (Jensen & Gerdes, 1995; Yarmolinsky, 1995). Cellular stresses such as starvation, DNA damage, antibiotics, oxidative stress and high temperature have all been shown to lead to antitoxin insufficiency and toxin activation, resulting in cell death (Christensen et al., 2001, 2004; Hazan et al., 2001, 2004; Sat et al., 2001, 2003; Kolodkin-Gal & Engelberg-Kulka, 2006). However, other studies have suggested that, rather than mediating PCD, TA systems induce a reversible bacteriostatic state in the host that facilitates adaptation to diverse stress conditions encountered by the host (Pedersen et al., 2002). Recent reports have also described TA modules as mediators of persistence and the multidrug tolerance phenotype in bacteria (Keren et al., 2004). This new role of TA loci in bacterial physiology could be of immense importance in the life cycle of several pathogens that exhibit a dormant phenotype in the host.

Mycobacterium tuberculosis, one of the most devastating human pathogens, which survives under various stressful conditions in human tissues, exhibits both long-term dormancy, generally termed as latent infection, and persistence of infectious bacteria in host tissue. Thirty-eight putative TA loci (Table 1) have been identified in M. tuberculosis through homology searches (Pandey & Gerdes, 2005). These loci are conserved in other pathogenic strains of the Mycobacterium family including Mycobacterium bovis and Mycobacterium avium. In contrast, Mycobacterium smegmatis, a fast-growing nonpathogenic relative of M. tuberculosis, has only two TA loci, while Mycobacterium leprae, an obligate intracellular pathogen, does not have a single intact TA locus.

View this table:
Table 1

Chromosomal TA loci in Mycobacterium tuberculosis H37Rv

Sl. no.TA locusToxinProteinLengthAntitoxinProteinLength
1relBE1RelE1Rv 1246c97RelB1Rv 1247c89
2relBE2RelE2Rv 286687RelB2Rv 286593
3relBE3RelE3Rv 335885RelB3Rv 335791
4higBA1HigB1Rv 1956149HigA1Unannotated117
6mazEF2MazF2Rv 0659c102MazE2Rv 0660c81
7mazEF3MazF3Rv 1102c103MazE3Unannotated77
8mazEF4MazF4Rv 1495105MazE4Rv 1494100
9mazEF5MazF5Rv 1942c109MazE5Unannotated78
10mazEF6MazF6Rv 1991c114MazE6Unannotated92
12mazEF8MazF8Rv 2274c105MazE8Unannotated82
13mazEF9MazF9Rv 2801c118MazE9Unannotated76
14vapBC1VapC1Rv 0065133VapB1Unannotated83
15vapBC2VapC2Rv 0301141VapB2Rv 030073
16vapBC3VapC3Rv 0549c137VapB3Rv 0550c88
17vapBC4VapC4Rv0595c130VapB4Rv 0596c85
18vapBC5VapC5Rv 0627135VapB5Rv 062686
19vapBC6VapC6Rv 0656c127VapB6Rv 0657c51
20vapBC7VapC7Rv 0661c145VapB7Rv 0662c122
21vapBC8VapC8Rv 0665112VapB8Rv 066490
22vapBC9VapC9Rv 0960127VapB9Unannotated73
23vapBC10VapC10Rv 1397c133VapB10Rv 1398c85
24vapBC11VapC11Rv 1561134VapB11Rv 156072
25vapBC12VapC12Rv 1720c129VapB12Rv 1721c72
26vapBC13VapC13Rv 1838c131VapB13Rv 1839c87
27vapBC14VapC14Rv 1953103VapB14Rv 195271
28vapBC15VapC15Rv 2010132VapB15Rv 200980
30vapBC17VapC17Rv 2527133VapB17Rv 252675
31vapBC18VapC18Rv 2546137VapB18Rv 254592
32vapBC19VapC19Rv 2548125VapB19Rv 254785
33vapBC20VapC20Rv 2549c131VapB20Rv 2550c81
34vapBC21VapC21Rv 2757c138VapB21Rv 2758c88
35vapBC22VapC22Rv 2829c130VapB22Rv 2830c71
36vapBC23VapC23Rv 2863126VapB23Unannotated82
37parDE1ParE1Rv 1959c98ParD1Rv 1960c83
38parDE2ParE2Rv 2142c105ParD2Unannotated71
  • * Number of amino acid residues.

TA loci have been demonstrated to exhibit activity in heterologous systems and such toxins could help identify targets for development of wide-spectrum antibiotics. In a recent study, two MazF homologues from M. tuberculosis have been shown to possess mRNA interferase activity and elicit toxicity in Escherichia coli (Zhu et al., 2006).

Here, we report the characterization of heterologous toxicity of all the 38 toxins encoded by M. tuberculosis and also the rescue activity of the cognate antitoxins in E. coli. We have cloned the toxin and antitoxin genes of all the 38 TA loci of M. tuberculosis and studied the effects of their expression on cellular growth and viability in E. coli. Results show that not all the toxins encoded by M. tuberculosis are functional in E. coli. This might also hold true for their activity in the native host.

Materials and methods

Growth condition and media

Cells were grown in Luria–Bertani (LB) broth. The medium was supplemented with ampicillin (100 μg mL−1), kanamycin (30 μg mL−1), arabinose (0.2%) or anhydrotetracycline (50 ng mL−1) as the per requirement. For all studies, E. coli BL21 [FompT hsdSB (rB mB) gal dcm] (Studier & Moffatt, 1986) was used and was grown with aeration at 37 °C.

PCR and other molecular biology procedures

PCR was performed in a final volume of 50 μL using 25 pmol of each primer, 1.5 U of expand high fidelity DNA polymerase (Roche) and 1 ng of template DNA. The annealing temperature was 55 °C and the polymerization temperature was 72 °C. The polymerization time was 1 min for amplification of a 1 kb fragment and an extension of 2 s was added to the polymerization time in each cycle. PCR products were purified using the PCR Purification Kit (Qaigen) and eluted in 50 μL of 10 mM Tris, pH 8.0. Restriction enzymes were procured from New England Biolabs and Roche. T4 DNA ligase (Roche) was used in all ligation reactions. Plasmid DNA was prepared using the Quick Spin Plasmid Purification Kit (Qiagen). All cloning procedures were performed using standard protocols (Sambrook & Russell, 2001).

Cloning of toxin and antitoxin encoding sequences

Antitoxin (At) encoding sequence was amplified from chromosomal DNA of M. tuberculosis H37Rv using suitable primers. The amplified product was treated with T4 DNA polymerase in the presence of dATP to create BsaI-compatible ends and inserted into BsaI-digested pLTAExp20 (Fig. 1) to generate pLTA(At)20.

Figure 1

Escherichia coli expression vectors for cloning of (a) antitoxin genes and (b) toxin genes. The maps are not to scale. β-Lactamase, ampicillin-resistance gene; TetR, tetracycline repressor; ColE1, ColE1 origin of replication with rop gene; F1, bacteriophage M13 origin of replication; TetP, tetracycline promoter-operator; H, sequence encoding 10 residues of histidine; stuffer, a noncoding 1.8 kb sequence flanked by multiple cloning sites (the foreign gene sequence is inserted in place of the stuffer sequence); AraC, sequence encoding AraC, which regulates transcription from the arabinose promoter; Kan r, kanamycin resistance gene; COLA, origin of replication from plasmid CloA; AraP, arabinose promoter-operator; St, eight-amino-acid strep tag-encoding sequence; NcoI, NheI, BsaI, Bsu36I, SpeI and EcoRI denote the restriction endonuclease sites available for cloning the foreign gene sequence.

Toxin (Tx) encoding sequence was amplified from chromosomal DNA of M. tuberculosis H37Rv using suitable primers. The amplified product was treated with T4 DNA polymerase in the presence of dATP to create BsaI-compatible ends and inserted into BsaI-digested pCAKExp10 (Fig. 1) to generate pCAK(Tx)10.

Plasmid construction was confirmed by DNA sequencing. All plasmids were introduced into E. coli by electroporation.

Toxin activity on solid medium

BL21 were transformed with pCAK(Tx)10 and grown in LB with kanamycin to an OD600 nm of 1.0. The culture was then streaked on LB agar supplemented with kanamycin with or without arabinose. The plates were incubated overnight at 37 °C and growth was observed.

Toxin activity assay in liquid medium

BL21 were transformed with pCAK(Tx)10 and grown in LB with kanamycin to an OD600 nm of 0.2. The cultures were then split and one part was supplemented with arabinose. Aliquots were removed every 60 min for assessment of OD600 nm and for plating serial dilutions on selective plates containing 1.0% glucose to determine CFU.

Antitoxin activity experiments

BL21 were transformed with pLTA(At)20 and pCAK(Tx)10 and grown in LB with ampicillin and kanamycin to an OD600 nm of 0.2. The cultures were then split and arabinose and/or anhydrotetracycline were added to some cultures. Culture aliquots were taken every 60 min for assessment of OD600 nm and for plating serial dilutions on selective plates containing 1% glucose to determine CFU.


Vector system for functional studies of TA systems

A pair of compatible, differentially regulated expression systems were constructed to study toxin and antitoxin activity in E. coli.

pLTAExp20 is a low copy vector carrying the ampicillin-resistance gene as a selectable marker and ColE1 origin of replication along with the rop gene. Expression of the foreign sequence is driven by the Tet promoter (TetP) system (Skerra, 1994) that is induced by anhydrotetracycline and repressed by the Tet repressor constitutively expressed from the β-lactamase promoter in the plasmid (Fig. 1a). The cloned sequence can be expressed as fusion with an N-terminal His10 tag. The plasmid also carries the bacteriophage M13 origin of replication to enable production of single-stranded DNA for site-directed mutagenesis of cloned sequences.

pCAKExp10 is a low copy vector carrying the origin of replication derived from plasmid ColA (Zverev & Khmel, 1985) and the kanamycin-resistance gene. Expression of the cloned gene sequence is driven by the promoter of the BAD arabinose-utilization operon in E. coli (AraP) that is induced by addition of arabinose to the medium and is subjected to catabolite repression by glucose (Guzman et al., 1995). The arabinose promoter is further regulated by the Ara C gene product encoded by the plasmid (Fig. 1b). The foreign gene can be expressed as a fusion with an N-terminal eight-amino-acid Strep tag (Schmidt & Skerra, 2007).

Both vectors contain a stuffer sequence flanked by noncompatible BsaI sites. Directional cloning of foreign gene sequences is carried out using a BsaI-digested vector. In this study, the antitoxin genes are cloned in the TetP-based system, while the toxin genes are cloned in the arabinose promoter-based system. Cotransformants can be separately induced for antitoxin or toxin production by addition of anhydrotetracycline or arabinose or coinduced by addition of both anhydrotetracycline and arabinose.

Cloning of Mycobacteria-encoded toxin and antitoxin gene sequences

Sequences encoding the 38 mycobacterial toxins (Table 1) were amplified from M. tuberculosis H37Rv genomic DNA using suitable primers and the amplified products were ligated into pCAKExp10 to obtain pCAK(Tx)10. The antitoxin-encoding sequences were also amplified similarly and ligated into pLTAExp20 to obtain pLTA(At)20. The ligation products were electroporated into BL21 cells. Recombinants expressing full-length proteins with no mutations were obtained for the toxin and antitoxin gene sequences. However, no recombinants were obtained for parE1 and parE2. To check the integrity of the parE1 and parE2 gene sequences in the M. tuberculosis genomic DNA used as a template in PCR, the complete parDE1 and parDE2 operons amplified from the same genome template were cloned in pLTAExp20. Full-length sequences with no mutation were obtained for both the toxin and the antitoxin gene pairs, confirming that the sequence of the genomic template was fine. Presumably, parE1 and parE2 were highly toxic to E. coli in the absence of parD1 and parD2 antitoxin sequences and therefore no recombinants were obtained for these two toxins.

To perform functional studies, it was important to have the parDE toxin and antitoxin partners cloned under differentially inducible promoter systems. Therefore, BL21 cells were transformed with pLTA(parD1)20 and pLTA(parD2)20 separately and then their competent cells were prepared. These BL21 competent cells contained an expression cassette producing ParD1 and ParD2, respectively, upon induction with anhydrotetracycline. The parE1 and parE2 ligation samples (in pCAKExp10) were then electroporated in cognate antitoxin-expressing BL21 competent cells. The transformants obtained on LB plates supplemented with ampicillin and kanamycin were screened for full-length toxin sequences. Recombinants containing full-length toxin and cognate antitoxin-encoding sequences with no mutation were obtained for parDE1 and parDE2 loci. This shows that ParE1 and ParE2 are highly toxic to E. coli and their expression plasmids can only be introduced into cells expressing cognate antitoxin.

Mycobacterium tuberculosis-encoded toxins inhibit E. coli cell growth and colony formation

Escherichia coli BL21 harbouring pCAK(Tx)10 were analysed for growth in the presence and absence of arabinose. For cells expressing HigB1, MazF3, MazF6, MazF7, MazF9, VapC4, VapC11, VapC20, ParE1 and ParE2 toxins, growth was completely abrogated on LB agar containing arabinose (Fig. 2) in comparison with LB agar without arabinose. This shows that these 10 toxins produced by induction on arabinose-containing medium were toxic to E. coli. For all the remaining 28 toxins, there was no difference in growth on medium with or without arabinose, indicating that these toxins did not affect growth of E. coli.

Figure 2

Toxicity of Mycobacterium tuberculosis TA loci. BL21 cultures grown at 37°C were streaked on LB plates containing kanamycin and supplemented (+ara) or not (−ara), as indicated, with 0.2% arabinose. Twelve clones were streaked on each plate. Plates were incubated overnight at 37°C. Cells harboured the pCAK(Tx)10 plasmid as per the numbering in Table 2. For the parDE locus analysis, BL21 harbouring pLTA(At)20 and pCAK(Tx)10 were grown. The clones showing reduced growth in arabinose-containing medium are indicated by their number as per Table 2.

To analyse the effects of toxin expression in a quantitative manner, the kinetics of cell growth was studied in a liquid culture. Arabinose induction was carried out for all the 38 toxin-expressing transformants. Culture density and viable cell count were checked at different time points postinduction. As shown in Table 2, 10 of the 38 toxins when expressed in E. coli resulted in complete cessation of growth within 3 h (Supporting Information, Fig. S1) and this was accompanied by a >105-fold decrease in the viable cell count. These 10 toxins had also caused complete abrogation of cell growth on arabinose-containing plates in the earlier experiment (Fig. 2). Expression of the RelE3 toxin resulted in a marginal reduction in culture density and a 10-fold decrease in the cell count. However, this difference was not seen in cell growth on solid medium containing arabinose, indicating that the toxic effect of RelE3 was minimal. No difference in growth in the presence or absence of arabinose was observed for the other 27 toxin-expressing cultures (Table 2, Fig. S1). Western blot analysis of the induced cultures was performed to check for expression of the toxin protein. All the 38 toxins were successfully expressed in E. coli upon induction of cultures with arabinose (Fig. S2). This proves that the difference in the activity of various toxins was not related to differences in protein expression levels in E. coli.

View this table:
Table 2

Killing effect of Mtb TA loci

Sl. no.LocusProteinOD after 3 h of growth (−ara/+ara)Fold change in CFU
  • * A600 nm of cultures grown in absence and presence of 0.2% arabinose.

  • Change in CFU of cultures grown in absence and presence of 0.2% arabinose for 3 h.

  • C1, clone expressing Mtb gyraseA protein.

  • C2, clone carrying empty vector pCAKExp10.

These results show that HigB1, MazF3, MazF6, MazF7, MazF9, VapC4, VapC11, VapC20, ParE1 and ParE2 are highly potent mycobacteria-encoded toxins that inhibit E. coli growth. Also, the toxic action of these proteins is not dependent on the presence of any other mycobacteria-specific factor.

Alleviation of toxin activity by cognate antitoxin

It was then decided to check whether the expression of putative cognate antitoxin can suppress the growth-inhibitory action of the mycobacterial toxin in E. coli. For this, BL21 cells were cotransformed with the toxin- and antitoxin-expressing plasmids and then studied for growth under different induction conditions. Induction of antitoxin production along with toxin production led to alleviation of growth inhibition (Fig. 3) and complete rescue of viable cell counts (data not shown). The antitoxin in all the 10 cases was able to neutralize the toxin and its killing effect on E. coli cells. Hence, all the 10 loci, namely, higBA1, mazEF3, mazEF6, mazEF7, mazEF9, vapBC4, vapBC11, vapBC20, parDE1 and parDE2, encode bonafide TA pairs.

Figure 3

Maintenance of cell viability by antitoxin expression. BL21 harbouring pLTA(At)20 and pCAK(Tx)10 for each TA locus were grown in LB medium supplemented with ampicillin and kanamycin to an OD0.2 nm. The culture was then split and supplemented or not with 0.2% arabinose and 50 ng mL−1 anhydrotetracycline at time zero. OD600 nm were taken at different times after induction and plotted. The experiments were performed in duplicate and average values were plotted. Uninduced cultures (empty triangles); arabinose-induced cultures (filled squares); and arabinose plus anhydrotetracycline-induced cultures (filled triangles). The number of the TA loci is as per Table 2.


In this paper, we have characterized the heterologous activity of all the 38 chromosomally encoded TA loci identified in M. tuberculosis by homology searches.

The demonstrated role of TA loci in stress responses and development of a persistent phenotype has made the TA locus biology an important area for research. The presence of 38 TA loci in the chromosome of M. tuberculosis H37Rv, the second highest in any genome after Nitrosomonas europea, has raised intriguing questions about their role in the M. tuberculosis life cycle. While several of the TA loci-encoded products have been annotated and assigned an Rv number, others are still unannotated and have been defined as a putative toxin or antitoxin based on their genetic organization and sequence homology (Table 1). In this study, we have provided data on the toxin and antitoxin activities of all the 38 loci. Our data show that the unannotated, putative sequences of parD2, higA1, mazE3, mazE6 and mazE9 indeed encode functional proteins. We also show that the putative mazEF7 is a bonafide TA locus coding for a TA pair that is functional in E. coli.

Regulated expression studies showed that 10 of the 38 M. tuberculosis-encoded toxins, namely HigB1, MazF3, MazF6, MazF7, MazF9, VapC4, VapC11, VapC20, ParE1 and ParE2, when overexpressed in E. coli, completely blocked bacterial growth and reduced the viable cell count by more than 5 log. This toxicity could be prevented by expression of cognate antitoxins namely, HigA1, MazE3, MazE6, MazE7, MazE9, VapB4, VapB11, VapB20, ParD1 and ParD2, respectively. Seven of the nine MazF toxins of M. tuberculosis have been studied previously for activity in E. coli (Zhu et al., 2006). There also, MazF3, MazF6 and MazF9 were found to be toxic in E. coli. MazF2 was reported to have weak toxicity; however, in our study MazF2 has not been found to affect E. coli growth. Zhu and colleagues had not studied mazEF7 and mazEF8 loci probably because the gene products of these loci are not annotated. Our results show that MazF7 is toxic in E. coli and this toxicity can be prevented by coexpression of MazE7. Thus, mazEF7 encodes a functional TA pair.

Whether all the TA loci identified by homology searches in M. tuberculosis encode functional proteins needs to be confirmed. In our study, less than one-third of the putative M. tuberculosis toxins have been found to be active in E. coli. Either several of these putative M. tuberculosis toxins are nonfunctional proteins; otherwise, the cellular targets of these toxins are absent in E. coli. It is important to note that the M. tuberculosis toxins found to be functional in E. coli belong to different toxin families. Moreover, toxins belonging to the same family exhibit different activity in E. coli, with some being toxic and others being nontoxic. Therefore, it remains to be studied whether all toxins belonging to one family have a similar mechanism of action and the same target or whether there are differences in cellular action within a family of toxins. Also, the extreme toxicity of M. tuberculosis-encoded ParE toxins vis-à-vis their plasmid encoded counterparts reported in the literature (Roberts et al., 1994) suggests that similar toxins might have differences in potency and or target recognition. This hypothesis gains strength from the recent data (Zhu et al., 2008) wherein M. tuberculosis-encoded MazF toxins exhibit a different sequence specificity in their activity as mRNA interferases and will therefore have differential target availability in different organisms. Also, MazF4, which is nontoxic in E. coli, shows RNA cleavage activity in vitro, suggesting that it's substrates are probably absent in E. coli. Whether it is toxic in the native host remains to be investigated. Similarly, MazF9, which is highly toxic and is not expressed well in E. coli (pers. commun.), was expressed well in M. smegmatis (Zhu et al., 2006), indicating that either it is nontoxic in M. smegmatis; otherwise, a neutralizing antitoxin counterpart is available for MazF9 in the bacterium.

Regulated expression studies of all the 38 M. tuberculosis-encoded toxins in the native host will be important to understand their functionality. It is possible that there may be a division of labour in TA loci whereby different loci become activated by different environmental cues and at different phases of the M. tuberculosis life cycle. There might also be redundancy in the M. tuberculosis genome, and some of these loci might be pseudogenes. Changes in the expression levels of some toxin genes have been observed in M. tuberculosis transcriptome studies (Betts et al., 2002; Dubnau et al., 2002; Dahl et al., 2003; Sassetti et al., 2003). However, the net effect of these changes on cellular metabolism and viability will depend on corresponding changes in the expression levels of cognate antitoxins. Because the TA loci are organized as an operon, RNA-level profiling alone may not be sufficient to infer the change in expression levels of these proteins. Expression profiling of these loci at both the RNA and the protein level under diverse M. tuberculosis growth conditions will be important to understand their activation.

TA loci-encoded toxins hold great potential for antibiotic action; however, they need to be studied in detail in diverse organism backgrounds before they can be useful for development of successful broad-spectrum antibacterial strategies.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Effect of toxin production on culture density and CFU of Escherichia coli.

Fig. S2. Western blot analysis of toxin expression in Escherichia coli.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.


The author gratefully acknowledges Professor V.K. Chaudhary for stimulating discussions during the course of this study and for critical evaluation of the manuscript. Help from Ms Shilpi Das in DNA sequencing and Ms Sunita Sharma in cloning experiments is acknowledged. This work was financially supported by the Department of Biotechnology, Government of India.


  • Editor: Roger Buxton


View Abstract