Diversity in Functional Organization of Class I and Class II Biotin Protein Ligase

The cell envelope of Mycobacterium tuberculosis (M.tuberculosis) is composed of a variety of lipids including mycolic acids, sulpholipids, lipoarabinomannans, etc., which impart rigidity crucial for its survival and pathogenesis. Acyl CoA carboxylase (ACC) provides malonyl-CoA and methylmalonyl-CoA, committed precursors for fatty acid and essential for mycolic acid synthesis respectively. Biotin Protein Ligase (BPL/BirA) activates apo-biotin carboxyl carrier protein (BCCP) by biotinylating it to an active holo-BCCP. A minimal peptide (Schatz), an efficient substrate for Escherichia coli BirA, failed to serve as substrate for M. tuberculosis Biotin Protein Ligase (MtBPL). MtBPL specifically biotinylates homologous BCCP domain, MtBCCP87, but not EcBCCP87. This is a unique feature of MtBPL as EcBirA lacks such a stringent substrate specificity. This feature is also reflected in the lack of self/promiscuous biotinylation by MtBPL. The N-terminus/HTH domain of EcBirA has the self-biotinable lysine residue that is inhibited in the presence of Schatz peptide, a peptide designed to act as a universal acceptor for EcBirA. This suggests that when biotin is limiting, EcBirA preferentially catalyzes, biotinylation of BCCP over self-biotinylation. R118G mutant of EcBirA showed enhanced self and promiscuous biotinylation but its homologue, R69A MtBPL did not exhibit these properties. The catalytic domain of MtBPL was characterized further by limited proteolysis. Holo-MtBPL is protected from proteolysis by biotinyl-5′ AMP, an intermediate of MtBPL catalyzed reaction. In contrast, apo-MtBPL is completely digested by trypsin within 20 min of co-incubation. Substrate selectivity and inability to promote self biotinylation are exquisite features of MtBPL and are a consequence of the unique molecular mechanism of an enzyme adapted for the high turnover of fatty acid biosynthesis.


Introduction
Mycobacterium tuberculosis has become resistant to most drugs. The cell wall, composed of almost 60% lipids that are long chain, branched fatty acids, is highly hydrophobic and hence refractory to several components of human defense system. It also provides an effective permeability barrier against several anti-mycobacterial agents [1][2][3]. The rich diversity of lipids present in M. tuberculosis is reflected at the genomic level by a large repertoire of genes for lipid biosynthesis. M. tuberculosis, for example, has ,300 enzymes involved in lipid synthesis while E. coli has only about 50 [4][5][6][7].
Biotin-dependent enzymes are involved in carboxylation and decarboxylation reactions. Acyl CoA carboxylases (ACC) catalyze biotin-dependent carboxylation of nascent molecules such as acetyl-CoA, propionyl-CoA etc. These carboxylases are multisubunit, multi-domain proteins consisting of a and b subunits. M. tuberculosis has three copies of a-subunits which are composed of a N-terminus biotin carboxylase (BC) and a C-terminus biotin carboxyl carrier protein (BCCP). All biotinyl domains so far reported have a target lysine at 235 th residue from C-terminus for biotinylation [8]. Hence, a protein composed of C-terminus 87 amino acids of acc is an efficient substrate for Biotin Protein ligase [8]. The b-subunit has carboxyl transferase (CT) activity [8].
Biotinylation of BCCP is catalyzed by Biotin Protein Ligase (BPL) which promotes an amide linkage between the carboxyl group of biotin and the e-amino group of a specific lysine residue nestled within a conserved 'AMKM' sequence of BCCP. Biotinylation converts inactive apo-BCCP to functional holo-BCCP that participates in the transcarboxylation reaction [9,10]. Thus, BCCP has two functions -mechanistic by serving as carboxyl carrier in overall carboxylation reaction and structural, by swinging carboxybiotin to the carboxyl transferase component of ACC. BC carboxylates the ureido nitrogen atom of biotin covalently bound to BCCP which moves {CO 2 }-biotin to the active site of carboxyl transferase (CT), for the transfer of a carboxyl group to acetyl or propionyl CoA [11,12]. The entire sequence of carboxylation reaction and the key role played by BCCP is schematically represented in Figure 1.
In spite of a highly conserved function, BCCPs display unique features for their respective biotinylating enzymes. In solution, apo-BCCP (E. coli, Pyrococcus horikoshii) is a flattened bbarrel structure comprising of two four-stranded b sheets [12,13]. In most BCCPs, the biotinable lysine is nestled within the conserved tetrapeptide 'AMKM' sequence in an exposed b-turn of BCCP domain. However, in Sulfolobus tokodaii, the canonical lysine residue within the sequence 'AMKS' was not biotinylated by EcBirA [14,15]. In Aquifex aeolicus (AaBPL), the target lysine is within the 'ALKV' sequence [16]. BCCP of M. tuberculosis (MtBCCP) is part of a multi-domain enzyme, biotin carboxylase and this probably alters its dynamics with the cognate enzyme, MtBPL.
MtBPL belongs to class I BPLs which lack a DNA binding domain at their N-termini unlike the class II BPLs (e.g. EcBirA) hence are devoid of repressor function exhibited by class II BPLs [17][18][19]. Our previous study showed that the two enzymes differ in several ways from structural organization to ligand interactions [20]. EcBirA can biotinylate BCCPs of other species. MtBPL as shown in this study, in contrast, to EcBirA exhibits exquisite substrate specificity. The differences in their activities are correlated here with their intrinsic metabolic functions.

Protein purification
It has been reported that the C-terminus domain of BCCP (apo-BCCP 87 ), does not self-associate and was a good substrate for biotinylation reaction [11,12]. Hence EcBCCP 87 and MtBCCP 87 expressed in pET28a were used for avidin blot assays. The BCCP was purified by Ni-NTA column chromatography. The apo form was separated from the holo form using a Mono Q column preequilibrated with 10 mM Tris-HCl buffer (pH-8.0) prior to the elution of the protein with a salt gradient (0-100% 10 mM Tris-HCl pH-8.0, 1 M NaCl). Fractions containing apo-BCCP were checked on avidin blot, pooled and dialyzed against 10 mM Tris-HCl pH 8.0, 50 mM KCl, 2.5 mM MgCl 2 (standard buffer). Thus ,95% of the purified MtBCCP 87 /EcBCCP 87 was found to be in their apo form. The biotinylation reaction was found to be dependent on Mg 2+ , ATP and biotin. BCCP and BPL were dialyzed against the standard buffer prior to use.
For self-biotinylation assays, BL21 containing EcBirA construct was grown in M9 media supplemented with 2% glucose for 5 h and induced for 3 h to prevent endogenous self-biotinylation. The eluted protein was dialyzed, concentrated and dialyzed against standard buffer.

Domain architecture of Biotin protein ligase
The domain structure of MtBPL and EcBirA was obtained from pfam ( Figure 2) [21]. The different domains of BPL are: HTH. The helix-turn helix domain. BPL_LipA_ LipB. This family includes biotin protein ligase, lipoate-protein ligase A and B.
BPL C. The C-terminus domain has a SH3-like barrel fold, the function of which is unknown. BPL family is a member of clan TRB (Transcriptional repressor beta-barrel domain).This betabarrel domain is found at the C-terminus of a variety of transcriptional repressor proteins. As shown in the Figure 2, Biotin Protein Ligase of M.tuberculosis lacks the N-terminus HTH domain and hence does not function as a repressor.

Substrate specificity of MtBPL
The molecular behavior of MtBPL and EcBirA are different due to the presence of an additional repressor function in EcBirA. It has been documented that EcBirA biotinylates BCCPs from other species except the one from S. tokadii [15]. In fact, EcBirA efficiently biotinylated the synthetic biotinable minimal peptide of sequence 'GLNDIFEAQKIEWH' (Schatz peptide) which is known to be a good substrate for BPLs (Figure 3b). In contrast, MtBPL failed to biotinylate Schatz peptide (Figure 3a). Subsequently, we investigated the ability of MtBPL and EcBirA to cross biotinylate EcBCCP 87 and MtBCCP 87 . BCCPs (5 mM) were incubated with 500 mM biotin, 3 mM ATP, 100 nM EcBirA or MtBPL for 30 min at 37uC. EcBirA efficiently biotinylated both the BCCPs but MtBPL selectively biotinylated its cognate substrate (MtBCCP 87 ) alone and failed to biotinylate EcBCCP 87 (Figure 3c).

Self-biotinylation of EcBirA
When substrate specificity of BPLs was explored, at higher enzyme concentration, a protein with molecular weight corresponding to EcBirA was detected on avidin blot indicating that EcBirA undergoes self-biotinylation. This is consistent with the report of Choi-Rhee et al [22]. Therefore, we investigated if MtBPL was capable of self-biotinylation like its counterpart in E. coli. MtBPL or EcBirA (250-2000 nM) were subjected to biotinylation reaction for 1 h in the absence of BCCP. The biotinylation mixture was resolved on 12% SDS-PAGE, transferred onto nitrocellulose membrane and detected by streptavidin HRP. The control, EcBirA was self-biotinylated at concentration as low as 500 nM (Fig S1). In contrast, MtBPL did not undergo self-biotinylation even at 2000 nM (Lane 2-6, Fig S1). Hence, our focus was to study the implications of the lack of self-biotinylation in MtBPL.
EcBirA has an additional N-terminus HTH domain which contributes to the repressor function of the protein ( Figure 2). Earlier reports suggested that truncated EcBirA (D1-34) was enzymatically active but did not undergo self-biotinylation [22]. This suggested that the N-terminus probably carries the biotinable residues. So, the N-terminus domain (1-65 amino acids) was independently cloned in pGEX4T-1. The fused GST-HTH domain of EcBirA (pGEN1) was subjected to biotinylation using enzymatic concentration of 100 nM MtBPL/EcBirA. The fused protein was biotinylated by full length EcBirA (Figure 4). The control GST protein was not biotinylated by EcBirA. This confirms that the self-biotinable lysine is within the N-terminus/ HTH domain of EcBirA. It also suggests that the catalytic and self biotinable domain require no physical contiguity for the covalent modification. Hence, this construct was used to investigate if the lack of self biotinylation in MtBPL was because (i) MtBPL lacks the N-terminus domain or (ii) the enzyme was deficient in promoting self-biotinylation. MtBPL failed to biotinylate HTH-GST fusion protein (pGEN1) but EcBirA efficiently biotinylated the fusion protein. This suggests that the mere presence of self -biotinable residue does not confer MtBPL an ability to self biotinylate. Furthermore, non-specific proteins such as BSA was biotinylated by EcBirA but not by MtBPL (Fig S2). This clearly reinstates that MtBPL does not catalyze indiscriminate biotinylation. Thus, the inability of MtBPL to undergo self-biotinylation could be attributed to two factors: absence of an HTH domain and a stringent catalytic specificity of the enzyme. (Figure 4).

Competitive inhibition of self -biotinylation by Schatz peptide
The intermediate molecule, bio-59AMP, appears to play a central role in several processes. We investigated if bio-59AMP was preferentially used for self-biotinylation of HTH domain or biotinylation of biotin acceptor molecule. For this, self-biotinylation of EcBirA was performed in the presence of saturated concentration of Schatz peptide or BCCP 87 (5 mM). EcBirA failed to undergo self-biotinylation or promote biotinylation of heterologous HTH (pGEN1) domain in the presence of excess biotin acceptor molecule such as Schatz peptide ( Figure 5). Indeed, the bio-59AMP synthesized was used for biotinylation of biotin acceptor molecules, Schatz peptide and BCCP, rather than for self-biotinylation. Also to confirm that the covalently modified selfbiotinylated EcBirA was dialyzed to remove unbound biotin and ATP and then incubated with Schatz peptide. The covalently modified self-biotinylated EcBirA failed to endogenously biotinylate Schatz peptide. However, the addition of biotin and ATP to previously self-biotinylated EcBirA led to the conversion of apo-Schatz peptide to biotinylated form ( Fig S3).

Mutation analysis
Choi-Rhee et al have shown that the affinity of R118G mutant of EcBirA for biotin decreased by ,100 fold and the self-biotinylation increased several fold [22]. However, for the homologous R69A mutant of MtBPL the binding constant for biotin was nearly the same as that observed for the wild type protein (data not shown) . Also, the R69A mutant of MtBPL did not undergo self-botinylation (Lane 12, Figure S1). This highlights the differences in the structural and functional organization of EcBirA and MtBPL.

Limited proteolysis
Purified MtBPL was subjected to proteolytic digestion with protease trypsin for 20 min and the products were analyzed on 12% SDS PAGE in order to define the domain boundaries within the enzyme. The enzyme was subjected to limited proteolysis in the presence and absence of biotin and MgATP. Trypsin generated two fragments, one of about ,8.2 kDa and the other of ,21 kDa as determined by N-terminus sequencing and SDS-PAGE (Figure 6a, b). The ,8.2 kDa has an N terminus His-tag which was identified by its reactivity with the anti-His antibody. Also, the ,8.2 kDa fragment was susceptible to further proteolysis. The N-terminus sequencing of these products revealed the cleavage occurred between Arg-72 and Gly-73 for trypsin. Since these cleavage points are located around the conserved biotin binding site (GRGRHGR), MtBPL was subjected to proteolytic digestion in the presence of saturating amounts of the substrates, biotin and ATP as well as both of them together. Incubation with ATP did not alter the cleavage by trypsin with 83% of the protein being digested. Incubation with biotin did reduce the proteolysis with nearly 40% of the protein intact. However, incubation of MtBPL with both biotin and ATP completely protected nearly all the protein from proteolytic digestion by trypsin. This was also observed when the protein was pre-incubated with chemically synthesized bio-59AMP. In fact the intermediate molecule, biotinyl-59AMP protected the protein from proteolytic digestion for over 24 h. Thus, when biotin and ATP were pre-incubated with the enzyme, biotinyl-59AMP was synthesized and this intermediate molecule protected the protein from proteolysis by binding to the active site of the enzyme. MtBPL was incubated with saturating amounts of biotin and non-hydrolyzable ATP analogue AMPpNpp and then treated with trypsin. The protein showed reduced protection against the protease as the nonhydrolyzable ATP analog failed to synthesize biotinyl-59AMP. Taken together, these results suggest that the binding of the substrates and/or the formation of the intermediate, biotinyl-59AMP, protects BPL from protease cleavage. The subunit assembly of accA3 and accD6 complex in association with esubunits has been studied in detail [8]. The BCCP domain of heavier a (accA) subunit interacts with three distinct heterologous proteins; BCCP-BC, BCCP-CT and BCCP-BPL. Considering the complexity of the cell wall of M. tuberculosis, it is not surprising that the pathogen has so many of these enzymes with biotinyl domains.
BCCP is a key player in carboxylation and transcarboxylation reactions which shuttles carboxyl group from BC to CT of ACC to  initiate fatty acid elongation. As a prelude to carboxylation of biotin to transcarboxylation of acyl-CoA, BPLs must selectively interact with BCCP. Relating structure to function of a protein that participates in multiple interactions is fraught with difficulties [23,24]. From the crystal structures of PhBCCP and EcBCCP, it was evident that target lysine is located at the type 1 bturn [11,25]. In most post-translational modifications, the primary structure surrounding target residue is critical. But from the biotinylation results of this study it is apparent that while the motif is necessary it is not enough for biotinylation. Indeed failure of MtBPL to biotinylate EcBCCP 87 is consistent with this argument. Hence, specific conformational feature(s) around the motif are necessary for biotinylation of the acceptor domain.
We reported earlier that BCCP domain of accA1 was efficiently biotinylated and hence probably participates in the acetyl CoA carboxylase activity [20]. M. tuberculosis has three BCCP domains each one belonging to a biotin carboxylase paralog. Our interest was to study the specificity of MtBPL for the reactive biotinable lysine residue(s). This was of interest especially considering that EcBirA could biotinlate BCCP from S. cerevisiae. Our study clearly defines the substrate specificity of MtBPL. The gram positive protein ligase could not biotinylate Schatz peptide or EcBCCP at all the conditions tested. In contrast, EcBirA could biotinylate Schatz peptide and also MtBCCP showing broad substrate specificity. Association of BirA -BCCP is complex and in E. coli, a cysteine residue in the conserved hydrophobic patch (LCIV) of b4-b5 turn promotes dimerization of apo-EcBCCP. On biotinylation, the cysteine residue is buried contributing to monomerization of holo-EcBCCP [12]. However, PhBCCP and MtBCCP lack this crucial cysteine residue. The C-terminus of BPL undergoes relatively large conformational changes to accommodate BCCP [13]. The BCCP domains from different species have varied structural organization to interact with their homolgous enzyme(s) [26,27]. Display of a stringent specificity for its substrate is probably very critical for MtBPL due to the presence of different paralogs of BCCPs (accA1, accA2, accA3) in its genome.
While Choi Rhee et al showed that R118G mutant of EcBirA promotes self-biotinylation and also biotinylates BSA, we show that wild type EcBirA itself at higher concentration of ATP and biotin exhibited self-biotinylation. It also promoted promiscuous biotinylation of BSA. In contrast, MtBPL did not undergo selfbiotinylation nor promote appreciable promiscuous biotinylation of BSA ( Fig S1, S2, and S3).
Certain BPLs have a flexible active site domain that accommodates different substrates. Though EcBirA and MtBPL share considerable sequence homology they differ in their activities in a fundamental manner. A profound difference between EcBirA and MtBPL is self-biotinylation exhibited by the former enzyme. The N-terminus domain (HTH domain) of EcBirA is the site of self-biotinylation. D1-34 EcBirA failed to undergo self-biotinylation [22]. Also, biotinylation of heterologus pGEN1 (1-65 amino acid N-terminus domain of EcBirA) confirmed that HTH domain had the self-biotinable lysine residue. As mentioned earlier, MtBPL failed to undergo self-biotinylation probably because it lacks the HTH (repressor) domain. Sequence analysis showed that EcBirA, PhBPL and AaBPL have 18, 25 and 16 lysine residues compared to just 2 residues in MtBPL. The two lysine residues of MtBPL are within the conserved 'KWPND' and 'KIAGLEV'motifs and are probably part of the active site. The invariant lysine within the KIAGLEV plays an essential catalytic role during synthesis of bio-59AMP and the KWPND shares the motif with streptavidin. Thus, the lack of self-biotinylation in MtBPL is due to the absence of a biotinable lysine residues. The specific lysine residues involved in the self-biotinylation of the HTH domain are currently under investigation in our laboratory. In biotinylation of BCCP, an electrostatic interaction between negative phosphate group of bio-59AMP and positively charged lysine of BCCP are key elements. The uncharged lysine in BCCP is deprotonated by aspartate residues of EcBirA which promotes a nucleophilic attack on the electrophilic carbonyl group of bio-59AMP leading to covalent modification [27]. It is possible that a similar mechanism promotes self-biotinylation of HTH domain of EcBirA. However, the self-biotinable residues in BPLs may not have sufficient accessibility and reactivity for accepting biotin and hence require longer incubation which perhaps accounts for a lag period of 1 h. Intermolecular interaction of BPL and BCCP probably allows for a snug fit which in turn promotes a fast and efficient covalent modification of the acceptor target lysine in BCCPs. On the other hand intramolecular folding of BPL initiated by bio-59AMP may impart steric hindrance which probably restrains orientation of the adenylate towards the self-biotinable lysine.
Intramolecular folding in EcBirA enables deprotonation of selfbiotinable/promiscuous biotinable lysine residue leading to its covalent modification. However, the transition state of MtBPL probably selects the specific acceptor molecule which in turn explains its stringent specificity for its cognate BCCP. Studies reported show that MtBPL differs from EcBirA and probably other BPLs in many additional ways; (a) MtBPL is a monomer in both its apo and holo forms and has relatively lower affinity for biotin and bio-59AMP (b) In EcBirA, self-biotinylation was enhanced in R118G mutant which releases bio-59AMP leading to increased self-biotinylation of the mutant protein. The R69A mutant of MtBPL failed to undergo self-biotinylation suggesting that the proclivity of the enzyme for biotinylation was different from that of EcBirA. The R69A MtBPL has similar affinity for biotin as that of wild type in contrast to R118G EcBirA which exhibited reduced affinity for biotin. Self-biotinylation of EcBirA occurs only in the absence of a biotin acceptor molecule. This is of relevance to the repressor function of EcBirA which occurs only in the absence of biotin acceptor molecule. Limited proteolysis study further reveals that the folding of the ligases are different. Our studies show that MtBPL is cleaved at the N-terminus (72-73 amino acids) whereas EcBirA is known to be cleaved at the C-terminus (217-218 amino acids) [28]. MtBPL exhibit restricted cleavage in the presence of substrate suggesting that scissile site interacts with the substrate.
While the biotin binding site is constituted by the conserved 'GRGRHGR' in both the BPLs but binding of biotin/bio-59AMP promotes conformational change in EcBirA [28,29].
Self-biotinylation is intrinsic to the catalytic function of the given BPL as availability of the self biotinable domain of EcBirA (pGEN1) does not promote promiscuous biotinylation by MtBPL. In support, AaBPL which lacks the HTH domain undergoes selfbiotinylation at higher enzyme concentration (.500 nM) as observed with EcBirA [30,31]. In MtBPL, the lack of selfbiotinylation is due to both substrate stringency of the enzyme and also due to the lack of a target lysine residue. The absence of self-biotinylation in MtBPL is probably a desirable feature to facilitate the high demands of fatty acid biosynthesis in M. tuberculosis. However, in other biotin protein ligases with or without the HTH domain, self-biotinylation is seen to take place.

Proposed rationale for the diverse functional organization of BPLs
MtBPL. We reported earlier that MtBPL in spite of lower affinity for biotin had K m similar to that of EcBirA [20]. Deletion of N-terminus domain of EcBirA decreases binding affinity of the enzyme by ,100 fold [29]. This suggests that higher binding constant of EcBirA for biotin may be directed towards covalent modification of HTH domain. In MtBPL, fatty acid synthesis plays central role for its cell wall synthesis. As this is a rate limiting step, the enzyme avoids self/promiscuous biotinylation to conserve biotin, a scarce co-factor whose biosynthesis itself is an extremely slow process. This is due to the low turn over of BioB and its degradation under low iron concentration [32,33]. Additionally, uncoupling biotinylation and repressor functions would favor fatty acid biosynthesis [34]. Hence, the mycobacterium cell probably reserves all the biotin at its disposal for biotinylation of acc to meet the demands of cell wall biosynthesis (Figure 7a).
E. coli. The intermediate molecule, bio-59AMP can be utilized for any of the three function: biotinylation of BCCP, self-biotinylation or as a co-repressor depending on the cellular demands (Figure 7b). when the cell does not require biotin for biotinylation reaction, bio-59AMP [++] probably needs to functions as a co-repressor of biotin biosynthetic pathway and repress synthesis of biotin. However, this would be favored only if E.coli does not require immediate fatty acid biosynthesis to operate. But the bacterium during the transition, probably requires additional time to decide whether it wants to block the biotin biosynthetic pathway. Under such a situation, in the absence of BCCP, the bio-59AMP is directed towards selfbiotinylation. This prevents the bio-59AMP to be utilized as a co-repressor of biotin biosynthetic pathway. The selfbiotinylated EcBirA is enzymatically active to participate in the biotinylation of BCCP. This is primarily because transcription activation or repression has to be modulated according to the cellular requirements [34] 3) However, when the concentration of bio-59AMP [+++] is abundant it functions as a co-repressor and shuts the biotin biosynthetic pathway.
Our results support the proposed hypothesis as self-biotinylation is competitively inhibited by biotin acceptor molecule which is is increased in the presence of operator sequence of biotin biosynthetic pathway [18].
The preferred order of bio-59AMP utilization by EcBirA is:

BiotinylationwSelf-biotinylationwCo-repression
Thus the evolutionary process has devised different mechanism in EcBirA and MtBPL commensurate with the functional requirement of the organism. The biotin repressor function is separated from enzyme function in MtBPL as lipid biosynthesis is very critical in M. tuberculosis. As the repressor function is not coupled to the enzyme function the enzyme does not promote self-biotinylation. However, in E. coli during the evolutionary process, the enzyme has probably compromised its substrate specificity and has also acquired self as well as promiscuous biotinylation. Yao et al [35,36] suggested that though functionality and overall folding of biotinyl domains are conserved through evolution, the detailed structures of BPL-BCCP binding interface may vary among different species. The substrate stringency of MtBPL may add to its ability to regulate the acyl CoA carboxylases in M. tuberculosis.
In conclusion, our studies with MtBPL show that biotinylation process is not dependent merely on recognition of a target residue but involves an intricate play between the biotinyl acceptor (BCCP) and its cognate ligase. MtBPL plays an active role in substrate selection which occurs by an integration of an intricate series of events involved in BPL-BCCP interaction and biotin demands of the cell. The stringency exhibited by MtBPL makes it a suitable target for the development of anti-mycobacterials and vaccine.

Protein methods
M. tuberculosis BPL ( Rv3279c) was cloned into pET28a at NdeI/ HindIII sites and the protein purified as described by Purushothaman et al [20]. Mutant R69A was generated by site-directed mutagenesis and cloned into NdeI/HindIII sites and sequence analyzed. The procedure used for the purification of the mutant protein was identical to that of its wild type counter-part wild type [20]. EcBirA and (D1-65) EcBirA. M. tuberculosis has three acetyl-/propionyl coenzymeA carboxylase a subunit accA1 (Rv2501c), accA2 (Rv0973c), accA3 (Rv3285c), and a putative acetyl CoA carboxylase subunit BCCP TB7.3 (Rv3221c) and six b subunit, accD genes [7,8]. All biotinyl domains so far reported have target lysine at 235 th residue from C-terminus for biotinylation. Hence, we cloned the C-terminus 87 amino acid residues of accA1 as the substrate for MtBPL. MtBCCP 87 and EcBCCP 87 were cloned into pET28a. The PCR primers used for amplification reaction are listed in Table 1.
For self-biotinylation studies, BL21 expressing EcBirA was grown in M9 minimal media supplemented with 2% glucose for 4 h and then induced with 100 mM IPTG for 3 h. This was carried out to prevent autologous self-biotinylation.

Schatz minimal peptide
A minimal peptide, Schatz peptide, which is efficiently biotinylated by EcBirA GLNDIFEAQKIEWH (Genscript, USA) [37], was used for some of the experiments (37). The peptide (5 mM) was incubated with 100 nM of EcBirA/MtBPL, biotin (500 mM), ATP (3 mM) for 1 h at 37uC in standard buffer and the biotinylation was detected by MALDI-TOF.

Matrix-assisted laser desorption time of flight mass spectrometry
The molecular weight of Schatz and holo-Schatz peptides were determined by MALDI-TOF MS using a Ultraflex TOF/TOF , (Bruker Daltonics Germany) equipped with a N2 Laser, 337 nm, 50 Hz operating in the 25 KvA reflector mode. Samples were dialyzed against water and l ml of sample was mixed with equal volume of matrix solution on a stainless steel plate and air-dried prior to analysis. The matrix solution used was a-cyano-4hydroxycinnamic acid in 50% acetonitrile, 0.1% (v/v) trifluoroacetic acid. Mixture of appropriate standards was used for calibration and Schatz and holo-Schatz peptide analytes were analyzed as described above and calibration was performed using the known protonated molecular ion (MH1).

Biotinylation assay
Biotin acceptor molecule (BCCP or Schatz peptide) were incubated with 500 mM biotin, 3 mM ATP, 2.5 mM MgCl 2 and 100 nM MtBPL or EcBirA in standard buffer for 1 h at 37uC. The biotinylated proteins were detected by avidin blot and mass spectrometry..

Self-biotinylation reaction
To determine self-biotinylation, different concentration of EcBirA/MtBPL were incubated with 3 mM (ATP), biotin (500 mM) in standard buffer for 1 h at 37uC. The biotinylated protein was then detected by streptavidin blot.

Limited proteolysis
MtBPL (5 mM) was incubated with trypsin (1:200) dilution and incubated at 37uC for 20 min. The enzyme was pre-incubated with biotin (500 mM) and ATP (3 mM) at 37uC for 30 min prior to trypsin digestion. After protease treatment, the sample solubilizing dye was added to the protein , boiled and loaded on to a 10% SDS-PAGE. The resolved proteins were scanned and percentage of proteolysis determined. The digested product was sequenced from the N-terminus on an Applied Biosystems Precise 491 CLC Protein Sequencer.