Successful Conversion of the Bacillus subtilis BirA Group II Biotin Protein Ligase into a Group I Ligase

Group II biotin protein ligases (BPLs) are characterized by the presence of an N-terminal DNA binding domain that allows transcriptional regulation of biotin biosynthetic and transport genes whereas Group I BPLs lack this N-terminal domain. The Bacillus subtilis BPL, BirA, is classified as a Group II BPL based on sequence predictions of an N-terminal helix-turn-helix motif and mutational alteration of its regulatory properties. We report evidence that B. subtilis BirA is a Group II BPL that regulates transcription at three genomic sites: bioWAFDBI, yuiG and yhfUTS. Moreover, unlike the paradigm Group II BPL, E. coli BirA, the N-terminal DNA binding domain can be deleted from Bacillus subtilis BirA without adverse effects on its ligase function. This is the first example of successful conversion of a Group II BPL to a Group I BPL with retention of full ligase activity.


Introduction
Biotin protein ligase (BPL) is required for the covalent attachment of biotin to biotin-dependent enzymes. This attachment proceeds in a two-step reaction. First, BPL binds both biotin and ATP to synthesize biotinoyl-59-AMP (Bio-59-AMP, also called biotinoyl-adenylate) with release of pyrophosphate [1]. The eamino group of the conserved lysine residue of the acceptor protein acts as a nucleophile to attack the Bio-59-AMP mixed anhydride bond to give covalently attached biotin plus AMP (Fig. 1). Microbial BPLs are readily placed into two groups [2]. Both groups have catalytic and C-terminal domains that show strong structural conservation [3][4][5][6][7] whereas Group II BPLs are characterized by addition of an N-terminal helix-turn-helix (HTH) DNA binding domain that permits transcriptional regulation of the biotin synthetic genes. E. coli BirA, the paradigm for regulation of biotin biosynthesis, is the best studied Group II BPL. Transcriptional repression of the E. coli biotin operon occurs when BirA accumulates Bio-59-AMP because all biotin acceptor proteins have been biotinylated [8][9][10]. Bio-59-AMP accumulation results in dimerization of BirA and subsequent DNA binding [6,11,12]. In all four E. coli BirA crystal structures [13][14][15] the HTH structure is spatially well removed from the other domains of the protein and thus deletion of the N-terminal DNA binding domain was expected to convert this Group II BPL into a fully functional Group I ligase. However, this was not the case: the resulting protein had severely compromised ligase activity [16]. This was also true for ligases having smaller N-terminal deletions [17].
The sequences of other Group II BPLs suggest such proteins are found in c-Proteobacteria, Bacilli, and Clostridii [2], although only the proteins from E. coli [6] and (very recently) Staphylococcus aureus [7] have been enzymatically characterized and crystallized. One of the Group II BPLs, B. subtilis BirA, has only 27% amino acid sequence identity to E. coli BirA (Fig. 2). Despite this low sequence identity, Bacillus subtilis birA has been shown to complement the ligase activity of a temperature-sensitive E. coli birA85 strain [18]. Moreover, B. subtilis birA mutants show constitutive expression of a bioW-lacZ fusion, suggesting that BirA regulates biotin operon transcription [18]. B. subtilis microarray data identified two additional transcripts, yuiG and yhfUST, regulated by biotin and BirA [19]. Both YuiG and YhfU have strong sequence similarity to the structurally characterized BioY biotin transporter of Lactococcus lactis [20] and other well characterized energy-coupling factor (ECF) biotin transporters [21]. All three transcripts have similar predicted BirA binding sites [2,19]. B. subtilis has two known biotinylated proteins, pyruvate carboxylase (PyC) and the biotin carboxyl carrier protein (AccB) subunit of acetyl-CoA carboxylase (http://genodb.pasteur.fr). A third B. subtilis protein, biotin/lipoyl attachment protein (BLAP) encoded by the yngHB gene was found to be biotinylated and lipoylated when expressed in E. coli [22] although subsequently this protein was found not to be lipoylated by the B. subtilis enzymes that modify the known cognate lipoic acid acceptor proteins [23].
A major shortcoming of our picture of the enzymatic and in vivo regulatory activities of Group II BPLs is that it is based on a single example, E. coli BirA. For this reason we decided to study the enzymatic and regulatory properties of the BirA of B. subtilis, a bacterium that is evolutionarily diverse from E. coli. Moreover the amino acid sequence of B. subtilis BirA differs markedly from that of the E. coli protein and in vivo B. subtilis BirA biotinylates multiple proteins whereas E. coli BirA modifies only a single protein. Here we report that B. subtilis BirA is a Group II BPL that follows the E. coli model of regulation, but unlike the E. coli protein, B. subtilis BirA can be successfully converted into a fully active Group I BPL.

Strains chemicals and culture media
The bacterial strains used were derivatives of B. subtilis 168 and E. coli K-12 (Table 1). The rich medium used to grow E. coli and B. subtilis was LB. The defined medium for E. coli was M9 salts supplemented with 0.5% glucose and 0.01% vitamin-free Casamino Acids (Difco) whereas the defined medium for B. subtilis was Spizizen salts [24] supplemented with 0.5% glycerol and 0.05% vitamin-free Casamino Acids (Difco) in addition to 0.01% each of tryptophan, tyrosine, isoleucine and phenylalanine. Antibiotics were used at the following concentrations (in mg ml 21 ): sodium ampicillin, 100; kanamycin sulfate, 50; chloramphenicol, 25; tetracycline HCl, 12; erythromycin, 100; lincomycin, 12.5; streptomycin sulfate, 50 and spectinomycin sulfate, 50. The 15:1 mixture of ticarcillin disodium salt and potassium clavulanate (Research Products International) was used at 25 mg ml 21 . Oligonucleotides were purchased from Integrated DNA Technologies. PCR amplification was performed using Taq polymerase (New England BioLabs) and Pfu polymerase (Stratagene) according to the manufacturer's specifications. DNA constructs were sequenced by ACGT, Inc. Reagents and chemicals were obtained from Sigma-Aldrich and Fisher, unless otherwise noted. New England BioLabs supplied restriction enzymes and T4 DNA ligase. Life Technologies provided SYBER Green I Nucleic Acid Gel stain and the 6% DNA Retardation Novex TBE Gels. Perkin Elmer provided [a-32 P]ATP (6,000 Ci/mmol). Analtech TLC Uniplates of microcrystalline cellulose matrix were purchased from Sigma-Aldrich.

Plasmids and plasmid constructions
The plasmids used and constructed are given in Table. 2. The B. subtilis birA gene was amplified by PCR from B. subtilis strain 168 genomic DNA with primers SKH001 and SKH002 (all primers are given in Table 3) that added NdeI and XhoI sites. The product was digested with NdeI and XhoI and ligated into the same sites of pET19b, with an N-terminal hexahistidine tag to give pSKH001.
The C-terminal 86 residues of B. subtilis AccB (AccB-86) were amplified by PCR from B. subtilis strain 168 genomic DNA with primers SKH005 and SKH006 that added NcoI and XhoI sites. The product was digested with NcoI and XhoI and ligated into the same sites of pET-28 to give pSKH003 which encoded the untagged protein. The C-terminal 77 residues of B. subtilis PyC (PyC-77) were amplified by PCR from B. subtilis strain 168 genomic DNA with primers SKH007 and SKH009 and inserted into pET-28 by the same procedures.
B. subtilis BirA, D2-64 BirA, D2-66 BirA, D2-74 BirA, and D1-81 BirA were amplified by PCR from B. subtilis strain 168 genomic DNA with forward primers SKH038, SKH036, SKH046, SKH048, SKH045, respectively, plus reverse primer SKH037. The primers added EcoRI and SalI sites. The products were digested with EcoRI and SalI and ligated into the same sites of pBAD322K, to give pSKH006, pSKH005, pSKH009, pSKH010 and pSKH011, respectively. The products were also ligated into the same sites of pBAD322Cm to give pSKH013, pSKH012, pSKH016, pSKH017 and pAKH018, respectively. E. coli BirA and D2-64 BirA were amplified by PCR from E. coli MG1655 genomic DNA with primers SKH049 and SKH043, and SKH042 and SKH043, respectively, thereby adding EcoRI and SalI sites. The products were digested with EcoRI and SalI and ligated into the same sites of pBAD322K, to give pSKH008 and pSKH007 respectively. The products were also ligated into the same sites of pBAD322Cm to give pSKH015 and pSKH014 respectively.
A 500 base pair internal fragment of B. subtilis bioW was PCR amplified from B. subtilis strain 168 genomic DNA with primers SKH057 and SKH058 that added EcoRI and BamHI sites. The product was digested with EcoRI and BamHI and ligated into the same sites of pMUTIN4 to give pSKH023.
B. subtilis accB-86 was PCR amplified from strain 168 genomic DNA with primer SKH069 which added a HindIII and a strong RBS site (AAGGAGGAAAAAATATG) plus primer SKH070 which contained an SphI site. The product was digested with HindIII and SphI and ligated into the same sites of pDR111 to give pSKH0024.

Bacillus subtilis strain construction
B. subtilis competent cell preparation and transformation were carried out as described by Dubnau and Davidoff-Abelson [25]. To create a bioW-lacZ chromosomal fusion, pSKH023 was transformed into strain 1A330 creating strain SKH001. Single crossover integration into bioW was verified by PCR and sequencing. To construct an IPTG inducible chromosomal copy of accB-86, strain SKH001 was transformed with linearized   pSKH024 creating strain SKH002. Double crossover integration into amyE was verified by the amylase production screen of Harwood and Cutting [26]. Integration was also verified by PCR and by sequencing with primers SKH065 and SKH066, SKH067 and SKH075 and SKH073 and SKH074.

Structural modeling and sequence alignment
The B. subtilis AccB biotin attachment domain was identified by InterProScan [27] and structural modeling to E. coli AccB-87 crystal structure (PDB 1A6X) using Swiss-Model automated mode [28][29][30]. The B. subtilis PyC biotin attachment domain was identified by InterProScan [27] and structural modeling to the Staphylococcus aureus pyruvate carboxylase biotin attachment domain crystal structure (3HBL) using Swiss-Model as above. B. subtilis BirA N-terminal deletions were determined by modeling to S. aureus BirA crystal structure (4DQ2) as a template using Swiss-Model automated mode as above. The final image was created using UCSF Chimera package [31]. Sequence alignments were created using the Clustal Omega (http://www.ebi.ac.uk/Tools/ msa/clustalo/) and the output was processed by ESPript 3.0 (http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi) to generate the final figure [32].

Complementation analyses
Two E. coli birA strains were tested. In the more straightforward test strain BM4092 [33] was transformed with either plasmid pSKH001 or pSKH002 followed by selection for transformants by plating on LB supplemented with ampicillin or kanamycin, respectively. These strains were grown at 37uC on M9 minimal plates [34] with 25 mM 5-bromo-4-chloro-indolyl-b-D-galactopyranoside (X-gal) and varying concentrations of biotin (1.6 nM, 4.1 nM, 41 nM, 4.1 mM, or 41 mM) [33]. Note that gene expression from T7 promoter based multi-copy plasmids in the absence of T7 RNA polymerase has been shown to be equivalent to that of a single copy plasmid [35]. Derivatives of strain BM4092 expressing the mutant BirAs encoded by kanamycin resistant plasmids pSKH005, pSKH006, pSKH007, pSKH008, pSKH009, pSKH010 or pSKH011 were similarly obtained and tested.
Strain VC618 which carries a deletion of the chromosomal birA gene was transformed with pSKH012, pSKH013, pSKH014, pSKH015, pAKH016, pSKH017, or pSKH018 and transfor-  mants were selected on LB plates supplemented with ampicillin and chloramphenicol at 30uC. These strains were cured of the temperature sensitive plasmid VC18 which expresses the Saccharomyces cerevisiae Bpl1 ligase by growth at 42uC on the defined medium supplemented with chloramphenicol and biotin. Loss of the temperature-sensitive plasmid was indicated by ampicillin sensitivity [36]. These strains were then grown at 37uC on M9 minimal plates [34] with 25 mM X-gal and varying concentrations of biotin as above [33].

Protein purification
For purification of the wild type and B. subtilis N-terminally deleted BirA Proteins E. coli strain BL21 (l DE3) was transformed with pSKH001, pSKH019, pSKH020, pSKH021, or pSKH022. The strains were grown at 37uC in LB medium supplemented with ticarcillin-clavulanate to an OD 600 of 0.8 and induced by addition of IPTG to 1 mM for an additional 6 h at 37uC. Cells were centrifuged and resuspended in lysis buffer which was 50 mM Tris-HCl, 500 mM NaCl, 0.1 mM tris(2-carboxyethyl)phosphine (TCEP), 10 mM imidazole, 5% glycerol, pH 8.0). The cells were lysed by passage through a French pressure cell and the lysate was centrifuged. The supernatant was added to Ni NTA beads (Qiagen) and incubated for 30 min. The mixture was added to a disposable 10 ml polypropylene column (Pierce) and washed with three column volumes of wash buffer (50 mM Tris-HCl, 500 mM NaCl, 0.1 mM TCEP, 60 mM imidazole, 5% glycerol, pH 8.0). BirA was eluted in 1 ml fractions with elution buffer (50 mM Tris-HCl, 500 mM NaCl, 0.1 mM TCEP, 250 mM imidazole, 5% glycerol, pH 8.0). Fractions were subjected to SDS-PAGE to determine purity. Pure fractions were combined and dialyzed against storage buffer (50 mM Tris-HCl, 500 mM NaCl, 0.1 mM TCEP, 5% glycerol, pH 8.0). Aliquots were flash frozen and stored at 280uC.
The B. subtilis acetyl-CoA carboxylase biotin carboxyl carrier protein biotin attachment domain (AccB-86) was purified from E. coli strain BL21 (l DE3) transformed with pSKH003. The strain was grown at 37uC in LB medium supplemented with kanamycin to an OD 600 of 0.8 and induced by addition of IPTG to 1 mM for an additional 4 h at 30uC. The cells were centrifuged and resuspended in starting buffer (20 mM Tris-HCl, 1 mM NaCl, 0.1 mM TCEP, 5% glycerol, pH 8.0). The cells were lysed by passage through a French pressure cell. The lysates were centrifuged and the supernatant was subjected to 60% isopropanol precipitation followed by anion exchange chromatography using a HiTrap Q FF column (GE Healthcare) and fast liquid chromatography (AKTA) [37,38].
The B. subtilis pyruvate carboxylase biotin attachment domain (PyC-77) was purified E. coli strain BL21 (l DE3) transformed with pSKH004. The strain was grown at 37uC in LB medium supplemented with kanamycin to an OD 600 of 0.8 and induced by addition of IPTG to 1 mM for an additional 4 h at 30uC. The cell were harvested by centrifugation and resuspended in starting buffer (50 mM sodium acetate, 100 mM NaCl, 0.1 mM TCEP, 5% glycerol, pH 4.9). The cells were lysed by passage though a French pressure cell and centrifuged. The supernatant was subjected to cation exchange chromatography using HiTrap SP FF column (GE Healthcare) and fast liquid chromatography (AKTA). PyC-77 was eluted by an NaCl gradient. Fractions were subjected to SDS-PAGE and fractions containing PyC-77 were combined and dialyzed against 50 mM 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES) buffer (pH 7.5) containing 500 mM NaCl and 0.1 mM TCEP. To further purify PyC-77 the protein was subjected to Superdex-75 size exclusion chromatography. Eluted fractions were analyzed by SDS-PAGE to determine purity. Fractions containing pure PyC-77 were combined, dialyzed against 50 mM HEPES buffer (pH 7.5) containing 500 mM NaCl, 0.1 mM TCEP, and 5% glycerol, flash frozen and stored at 280uC.
Plasmid pQC026C a derivative of vector pET28b encoding a terminal hexahistidine tagged BLAP was transformed into E. coli strain BL21 (l DE3). The strain was grown in LB medium to an OD 600 of 0.8 and induced by addition of IPTG to 1 mM for an additional 4 h. BLAP was purified by Ni +2 affinity chromatography as previously described [23]. The C-terminal hexahistidine tagged D2-65 BirA was purified as previously described [17].

Electrophoretic Mobility Shift Assays (EMSA) of DNA binding
The predicted B. subtilis BirA binding sites upstream of the coding sequences of bioO, yhfU and yuiG were PCR amplified from B. subtilis 168 genomic DNA with primers SKH014 and SKH015, SKH016 and SKH017 and SKH018 and SKH019, respectively. A DNA fragment containing one half-site of B. subtilis bioO was similarly amplified with primers SKH034 and SKH035. E. coli bioO was amplified with primers SKH026 and SKH027 from MG1655 genomic DNA. Negative control DNA (blap) was amplified from B. subtilis 168 genomic DNA with primers SKH028 and SKH029. All DNA fragments were 125 bp in length. The PCR products were sized on a 1.8% agarose gel and purified using a QIAquick PCR Purification Kit (Qiagen). DNA concentrations were determined at OD 260 by using a NanoDrop 2000c. Purified BirA was incubated with the small nucleophile hydroxylamine at neutral pH to cleave any Bio-59-AMP bound in the active site [39] and then dialyzed against storage buffer. The reaction contained 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 50 mM NaCl, 10% glycerol, 40 nM DNA and various concen-  [36]. The binding reactions were incubated at room temperature for 30 min and then loaded into a 6% DNA retardation gel (Invitrogen). The gel was run in 0.5X TBE at 100 V for 85 min. The gel was stained with SYBR Green I nucleic acid gel stain (Invitrogen) and visualized using Bio-Rad Chemidoc XRS and Quantity One software.

b-Galactosidase Assays
Cultures were grown overnight in defined medium containing 1.6 nM biotin. The cultures were diluted to an OD 595 of 0.2 in defined media containing various concentrations of biotin (1.6, 4, 20, 40, 80, 400 nM, 4 mM) and grown to OD 595 of 0.8 and induced with 1 mM IPTG for an additional 2 h. b-Galactosidase activity was determined as described by Harwood and Cutting for Bacillus following permeabilization with lysozyme [26].

B. subtilis BirA is a Group II BPL that binds three operator sites
To test the relative affinities of the predicted B. subtilis BirA binding sites we purified the protein (Fig. 3) and performed electrophoretic mobility shift assays (EMSAs) on DNA fragments containing the three sites (Fig. 4A). Full dependence of binding on ATP and biotin required treatment of the protein with neutral hydroxylamine to remove Bio-59-AMP accumulated in the active site during expression in E. coli. With the treated BirA binding of the biotin biosynthetic operator (bioO) was observed only in the presence of both biotin and ATP (Fig. 4B). BirA also bound the yhfU and yuiG operators only in the presence of biotin and ATP (Fig. 4C, D). Although the three binding sites have slightly different DNA sequences (Fig. 4A), analysis of binding over a range of BirA concentrations showed that the three sites had very similar binding affinities (Fig. 4E). B. subtilis BirA failed to show non-specific DNA binding (Fig. 4F) as assayed by use of a fragment from the coding sequence of the yngHB gene. BirA preparations that had not undergone hydroxylamine treatment showed some interaction with bioO in the absence of biotin and ATP (Fig. 4G). BirA did not interact with a site that was composed of only one of the B. subtilis bioO inverted repeats suggesting that the form of BirA active in DNA binding is a dimer (Fig. 4H). B. subtilis BirA interacted only very weakly with the E. coli bioO DNA site (Fig 4I).

B. subtilis BirA biotinylates three cognate proteins and the reactions proceed via Bio-59-AMP
Each of the acceptor proteins, AccB-86, PyC-77 and BLAP (Fig. 5A), was purified after high-level expression in E. coli (Fig. 3). In the first two cases the N-terminal halves of the proteins were deleted to avoid protein aggregation during purification (the deleted segments are responsible for interaction with other proteins of the enzyme complexes and play no role in biotinyla- tion). Electrospray ionization mass spectrometry results matched the theoretical masses of the apo forms of the three proteins and showed that the preparations were free of the biotinylated forms (Fig. 5B). When the apo forms of AccB-86, PyC-77 and BLAP were incubated with ATP, biotin and purified B. subtilis BirA and subsequently analyzed by mass spectrometry, mass values very similar to the theoretical values for biotinylated forms of all three acceptor protein were obtained (Fig. 5B). In the presence of a-32 Plabeled ATP and biotin B. subtilis BirA formed labeled Bio-59-AMP. Upon addition of any of the three acceptor proteins (AccB- The in vitro data presented above indicate that B. subtilis BirA behaves in a manner that strongly parallels that of E. coli BirA despite the low sequence similarity of the two proteins. However, E. coli BirA has recently been shown to undergo extensive interdomain communication that is required for full ligase activity [17]. These observations confirm and extend those of Xu and Beckett [16] and demonstrate that the E. coli BirA N-terminal domain plays a role in organizing the active site of BirA [17]. Specifically, the wing of the winged HTH structure interacts with the ligase active site biotin binding loop and acts to organize the active site to give high affinity binding of biotin and Bio-59-AMP [17]. To determine if another Group II BPL, that of B. subtilis, requires such communication of the catalytic and N-terminal domains for full ligase activity, we constructed genes encoding several N-terminal B. subtilis BirA deletion proteins. The deletion endpoints were based on structural modeling of B. subtilis BirA based on the crystal structure of S. aureus BirA (PDB 4DQ2) [7] (Fig. 6). BirA deletions D2-63 and D2-65 eliminated the predicted N-terminal domain whereas BirA deletions D2-74 and D1-81 also cut into the predicted central catalytic domain. Complementation assays using the E. coli birA1 mutant strain BM4092 and the E. coli DbirA deletion strain VC618 were used to test the ligase activities of the B. subtilis BirA N-terminally deleted proteins. These assays showed that upon expression of the B. subtilis D2-63 BirA, D2-65 BirA and wild type proteins in strain BM4092 all three BPLs supported growth equally well on medium containing 1.6 nM biotin (the minimal level allowing growth of E. coli) (Fig. 7A). These results indicated that the ligase activities of these two deletion proteins were essentially normal. In contrast, upon expression of the E. coli BirA lacking its amino terminus (D2-65 BirA), growth of the transformed strain required a biotin concentration that was 1000-fold greater. The complementation activities of the other two B. subtilis deletion proteins, BirA D2-74 and BirA D1-81 were either partially (BirA D2-74) or totally (BirA D1-81) compromised (Fig. 7A). To ensure that residual activity of the host BirA1 protein did not play a role in growth restoration we also performed complementation of the DbirA E. coli strain VC618 and obtained a similar complementation pattern (Fig. 7B).
The ligase activities of the purified B. subtilis BirA N-terminal deletion proteins (Fig. 3) were also tested by in vitro biotinylation assays. As expected from the complementation results the D2-63 and D2-65 BirA proteins showed Bio-59-AMP synthesis and biotin transfer activities indistinguishable from those of the wild type protein (Fig. 8). In contrast the Bio-59-AMP synthetic abilities of the D2-74 and D1-81 BirAs were significantly reduced relative to wild type BirA. However, the biotin transfer activity of the D2-74 BirA was comparable to wild type levels. Biotin transfer by the D1-81 BirA was significantly reduced compared to the wild type protein (Fig. 9). In agreement with prior work [16,17] the E. coli D2-65 BirA was significantly reduced in Bio-59-AMP synthesis and in biotin transfer (Fig. 8).

Bacillus subtilis BirA dimerizes in the presence of biotin and ATP
Dimerization is required for DNA binding of E. coli BirA and dimerization is dependent on biotin and ATP binding. To determine the oligomerization state of B. subtilis BirA, purified  BirA was incubated with EGS in the presence or absence of biotin and ATP. Chemical crosslinking indicated that BirA forms a dimer only in the presence of both biotin and ATP (Fig. 9A). Dimerization was also observed for the N-terminally truncated D2-65 BirA (Fig. 9B). Note that the characterized Group I BPLs can either be dimeric as is the Pyrococcus horikoshii protein [3,41] or monomeric as are the Mycobacterium tuberculosis [4], Aquifex aeolicus [5] and Propionibacterium freudenreichii subsp. shermanii [42] BPLs.
Bacillus subtilis BirA follows the E. coli BirA regulatory model In E. coli regulation of the biotin biosynthetic gene transcription depends not only on the concentration of biotin but also the levels of unbiotinylated AccB [8][9][10]. This is an important attribute because biotin is only active in central metabolism when it is protein bound. Hence, the rate of biotin operon transcription is sensitive not only to the intracellular concentration of biotin, but also to the supply of the proteins to which the biotin must be attached. Moreover, accumulation of the unmodified protein increases the rate of biotin synthesis thus ensuring replacement of the biotin consumed in protein modification.
To determine if this regulatory facet also applies to another Group II BPL, we constructed a B. subtilis strain that could be used to monitor regulation of biotin operon transcription upon alteration of the levels of the AccB-86 acceptor protein. This strain contains the bioB141 mutation rendering it a biotin auxotroph, a bioW-lacZ transcriptional fusion and an ectopic IPTG-inducible gene encoding AccB-86. This strain was grown in chemically defined media containing various concentrations of biotin. Upon induction of AccB-86 expression with IPTG bgalactosidase assays showed that expression of the biotin operon was dereprepressed at biotin concentrations that normally cause repression (Fig. 10) in a manner similar to that seen in E. coli [8][9][10].

Discussion
B. subtilis BirA complements the ligase activity of E. coli BirA mutant strain BM4092, but fails to complement the regulatory function of E. coli BirA. This latter result is not surprising since E. coli and B. subtilis BirAs bind different operator sequences and have different spacing of the palindromic elements. We have shown that B. subtilis BirA is a Group II BPL that binds equally well to the three operators predicted by others (Fig. 4). Binding is observed only in the presence of both biotin and ATP indicating that Bio-59-AMP is the regulatory ligand. Interestingly, EMSA showed that B. subtilis BirA weakly binds E. coli bioO (Fig. 4I), suggesting some distant relationship between the operators. Similar to E. coli, dimerization of B. subtilis BirA occurs only in the presence of biotin and ATP, a further indication that Bio-59-AMP is the regulatory ligand (Fig. 9A). The B. subtilis D2-65 BirA also formed dimers indicating that the N-terminal domain is not required for dimerization (Fig. 9B).
Deletions of the N-terminal DNA binding domain of E. coli BirA result in proteins having only very weak ligase activities indicating inter-domain interactions are required for full ligase activity [16,17]. These interactions have been shown to occur between the wing of the winged HTH domain and the biotin-binding loop of the catalytic domain. Deletion of only the fourteen-residue wing has as drastic an effect as deletion of the entire N-terminal domain, but was largely restored by insertion of a foreign wing of similar structure [17]. Moreover, a mutation within the wing can restore function to proteins having mutant biotin binding loops [17].
Although the HTH domain of our modeled B. subtilis BirA structure includes a wing, the structure is not required for full ligase activity. Both the D2-63 and D2-65 BirAs lack the entire Nterminal domain but performed the ligase partial reactions as well as the wild type protein both in vivo and in vitro (Fig. 8). Deletions that entered the predicted catalytic core of the protein resulted in either compromised (D2-74 BirA) or highly defective (D1-81 BirA) proteins (Fig. 8). The Bio-59-AMP synthetic ability of the D2-74 BirA was significantly decreased, although the protein transferred the biotin moiety normally suggesting that the putative a-helix near residue 74 may be involved in stabilizing biotin and ATP or Bio-59-AMP binding. Although B. subtilis D1-81 BirA failed to replace the E. coli ligase in vivo (Fig. 7), it retained weak Bio-59-AMP synthesis and biotin transfer activities (Fig. 8) suggesting that the loop predicted near residue 81 is required for efficient binding of biotin and ATP. These data indicate that unlike E. coli, B. subtilis BirA does not require an intact N-terminal DNA binding domain for full ligase activity. Instead, the first a-helix and loop of the modeled central domain seem to be important in binding biotin and ATP.
E. coli BirA is a highly dynamic protein as shown both by physical [43][44][45] and mutational [36] analyses. In the latter case mutations in the catalytic and C-terminal domains as well as in the HTH domain can result in super-repressor phenotypes. These proteins are BirAs that repress bio operon expression even at biotin concentrations that normally fail to repress transcription. Although E. coli BirA has been studied for over 30 years [40,46] and four different crystal structures have been solved, the details of how the protein binds its operator site remain unknown. Despite much effort in several laboratories no diffraction grade crystals of the operator DNA with liganded E. coli BirA have been obtained. This could be due to competing interactions of the N-terminal domain with the catalytic domain and operator giving a mixture of molecular species that preclude crystallization. If so, the apparent lack of such interactions in B. subtilis BirA may allow crystallization of this protein with its operator DNA.
The BLAP protein was biotinylated by B. subtilis BirA in vitro suggesting that BLAP may be the biotin carboxyl carrier protein for a biotin-dependent enzyme. The genes that neighbor yngHB, the BLAP encoding gene, are annotated as propionyl-CoA carboxylase subunits. However, a recent report suggests that these genes may comprise a methylcrotonyl-CoA carboxylase involved in leucine degradation [47].