A division of labor between two biotin protein ligase homologs

Abstract

Group I biotin protein ligases (BPLs) catalyze the covalent attachment of biotin to its cognate acceptor proteins. In contrast, Group II BPLs have an additional N-terminal DNA-binding domain and function not only in biotinylation but also in transcriptional regulation of genes of biotin biosynthesis and transport. Most bacteria contain only a single biotin protein ligase, whereas Clostridium acetobutylicum contains two biotin protein ligase homologs: BplA and BirA′. Sequence alignments showed that BplA is a typical group I BPL, whereas BirA′ lacked the C-terminal domain conserved throughout extant BPL proteins. This raised the questions of why two BPL homologs are needed and why the apparently defective BirA′ has been retained. We have used in vivo and in vitro assays to show that BplA is a functional BPL whereas BirA′ acts as a biotin sensor involved in transcriptional regulation of biotin transport. We also successfully converted BirA′ into a functional biotin protein ligase with regulatory activity by fusing it to the C-terminal domain from BplA. Finally, we provide evidence that BplA and BirA′ interact in vivo.

1 ∣. INTRODUCTION

Biotin (vitamin B7) is essential for all three domains of life where it acts as a cofactor for biotin-dependent enzymes, including key carboxylases, decarboxylases, and transcarboxylases of central metabolism (Satiaputra et al., 2016). Biotinylation refers to the extremely specific process whereby biotin becomes covalently attached to a single lysine residue of its cognate biotin receptor protein(s) (Chapman-Smith & Cronan, 1999). Protein biotinylation requires tight regulation since the biotinylation process is involved in protein signaling, localization, activation, and degradation (Sternicki et al., 2017). De novo biotin synthesis is a metabolically expensive process, requiring as many as 20 equivalents of ATP and at least 5 enzymes to assemble a molecule of biotin (Feng et al., 2013).

Biotin protein ligase (BPL) is the enzyme responsible for covalent attachment of biotin to its cognate enzymes and is essential for growth (Chapman-Smith & Cronan, 1999; Sirithanakorn & Cronan, 2021). Biotinylation proceeds in two discrete steps (Figure 1): biotin and ATP react to form biotinyl-5′AMP (Bio-5′-AMP). The protein-bound Bio-5′-AMP intermediate is then attacked by the ε-amino group of a unique lysine residue of the acceptor protein (Chapman-Smith & Cronan, 1999). This results in the amide-linked biotin-modified protein with release of AMP. To date, three classes of BPL enzymes have been reported (Satiaputra et al., 2016). Class I BPLs are monofunctional enzymes composed only of the conserved catalytic modules required for biotinylation. In contrast, Class II BPLs, also known as BirAs, are bifunctional enzymes since they contain an additional N-terminal DNA-binding domain. The N-terminal DNA-binding domain of BirA proteins allows them to function not only in biotinylation but also in transcriptional regulation of biotin synthesis and/or transport (Cronan, 1988, 1989). The most extensively studied BirA is that of Escherichia coli. When biotin is limiting or unbiotinylated biotin-accepting proteins accumulate, BirA is a monomeric biotinylation enzyme. When the cell is replete with biotin and all cognate proteins have been biotinylated, the BirA-Bio-5′-AMP complex dimerizes (Beckett, 2007). The BirA-Bio-5′-AMP dimers bind to specific operator sites within the promoters of the genes encoding biotin synthetic genes and biotin transporters to repress transcription (Chakravartty & Cronan, 2013; Cronan, 1989; Henke & Cronan, 2016). This guards against wasteful synthesis and transport of biotin. Class III BPLs are found in mammals and certain eukaryotes and contain a large extended N-terminal domain used for substrate proof reading (Beckett, 2018). Most organisms possess only a single BPL. An exception is Francisella novicida which expresses both a class I BPL (called BplA) and a BirA (Feng et al., 2015). Although both proteins have BPL activity, the class I BplA is the major enzyme and is necessary for virulence whereas the class II BirA homolog provides transcriptional regulation (Feng et al., 2015). Another example is Lactococcus lactis, a bacterium unable to synthesize biotin, which uses its BirA homolog to regulate one of the two BioY biotin transporters (Zhang et al., 2016). L. lactis also encodes an active BPL protein. The rationale proposed for the presence of two enzymes is that neither enzyme has sufficient ligase activity to support cellular growth and biotinylation (Zhang et al., 2016).

FIGURE 1.

The biotin protein ligase reaction and model of BirA′ and BplA regulation. (a) The protein biotinylation reaction proceeds in two steps. In the first step, BPLA binds biotin and ATP to synthesize Bio-5′-AMP (biotinoyl-5′-adenylate) with release of pyrophosphate. In the second step, the conserved lysine residue of the acceptor protein attacks the mixed anhydride bond to form biotinylated acceptor protein and AMP. (b) Model of BirA′ and BplA regulation. Under repression conditions, cells are replete with biotin and all acceptor proteins (AccB) are biotinylated. BirA′ then dimerizes and represses transcription of biotin synthesis and/or transport genes. Under derepression conditions, cells are starved for biotin and/or have an excess level of acceptor proteins (AccB). BirA′ is then unable to dimerize and exert repression. The dashed line posits that BplA and BirA′ may interact

In our work on fatty acid synthesis in the industrially important bacterium, Clostridium acetobutylicum, we noted that sequence alignments with E. coli BirA showed two putative BPLs (Figure S1). The first BPL, that encoded by gene Ca 0589, is a standard group I Bpl we call BplA. The second atypical BPL called BirA′ (where the prime means C-terminally truncated) is encoded by gene Ca 0212 and seemed a defective group II BPL that lacked the conserved and essential C-terminal domain (Chapman-Smith et al., 2001). This raised the question of why BirA′, a seemingly inactive ligase, has been preserved when the BplA ligase is present. Three predicted C. acetobutylicum, BirA′ DNA-binding sites are found adjacent to putative biotin biosynthetic or transport genes (Nölling et al., 2001; Rodionov et al., 2002). We report that BplA is the biotin ligase responsible for protein biotinylation whereas BirA′ is a biotin sensor involved in transcriptional regulation of biotin transporter operons (Figure 1b). BirA′ could be converted into a functional biotin protein ligase by fusing the C-terminal domain of BplA to the BirA′ C-terminus. We show that biotin-engendered dimerization of BirA′ is required for transcriptional regulation and that BplA and BirA′ interact in a two hybrid assay.

2 ∣. RESULTS

2.1 ∣. The BplA protein encoded by Ca0589 has biotinylation activity whereas the BirA′ protein encoded by Ca0212 lacks enzymatic activity

To test the ligase activities of BirA′ and BplA, we tested complementation of the E. coli b/rA(Ts) mutant strain BM4062 ara+ with the pBAD322-derived plasmids. Strain BM4062 ara+ was derived from the birA85 strain BM4062 of Barker and Campbell (1980) by transduction to ara+ (arabinose is toxic to araD strains such as BM4062). The birA85 allele (Arg 235 to Ser) results in defective biotin utilization at 29°C and a lack of growth and bio operon regulation at 42°C even in the presence of a great excess of biotin (Barker and Campbell, 1980). Hence, growth at 42°C of BM4062 ara+ carrying a recombinant plasmid indicates that the plasmid expresses a functional biotin ligase.

The BplA-encoding plasmid allowed growth of the E. coli b/rA(Ts) strain BM4062 activity at 4 nM biotin and 42°C the nonpermissive temperature after induction with 0.02% arabinose (Figure 2). In contrast, a parallel experiment with the BirA′ plasmid showed only barely detectable (background) growth after >16 hr of incubation indicating that expression of BirA′ could not functionally replace E. coli BirA ligase activity (Figure 2). This was the result expected since mutation in the E. coli BirA C-terminal domain resulted in loss of ligase activity (Chapman-Smith et al., 2001). To characterize the ligase activities in vitro, BplA and BirA′ were purified together with the B. subtilis AccB biotin-accepting domain (AccB-86). The reactions contained α-³²P labeled ATP, biotin and Mg²⁺. Labeled ATP allowed detection of Bio-5′ AMP upon separation of the products by thin layer chromatography. Upon addition of AccB-86 the biotin moiety was transferred to the protein resulting in accumulation of labeled AMP. In reactions with BplA, formation of the Bio-5′-AMP intermediate was detected upon biotin addition (Figure 3a). Upon addition of the AccB-86 biotin-accepting protein, Bio-5′-AMP disappeared and the label appeared in AMP indicating transfer of biotin to AccB-86. The BirA′ protein was completely inactive in this assay, no Bio-5′AMP was detected (Figure 3a). Note that the AccB biotin domain (essentially the C-terminal half of the AccB protein) was used as the biotin acceptor rather than the full length AccB which is prone to aggregation (Sirithanakorn & Cronan, 2021). Ligation of biotin to AccB was assayed by the increased rate of migration on nondenaturing gels of the biotinylated species. AccB-86 like other biotin domains is a small (9.4 kDa), acidic protein (pi of 4.5) and thus migrates toward the positive pole in nondenaturing gel electrophoresis. Biotinylation of the target lysine makes the protein more acidic (the pi becomes 4.37) resulting in increased mobility (Chapman-Smith et al., 1994). Purified AccB-86 was incubated with BplA with or without addition of ATP or biotin. The increase in AccB-86 mobility was observed only when both ATP and biotin were added (Figure 3b). BirA′ showed no activity (no gel shift) in this assay (Figure S2a). Finally, the product of the BplA ligase activity was confirmed by tandem mass spectroscopy (Figure S3).

FIGURE 2.

Complementation of E. coli birA(Ts) strain BM4062 strains by expression of BirA′ and BplA proteins. Strains are grown on M9 minimal medium containing indicated concentrations of biotin, 0.02% L-arabinose, and X-gal. Blue colonies indicate transcription of the bioF-lacZ fusion whereas white colonies indicate transcriptional repression of the biotin operon by BirA binding to bioO. The plates are incubated at 42°C

FIGURE 3.

In vitro characterization of BPL ligase function. (a) Thin layer chromatographic analysis of the BirA′ and BplA reactions. Biotin transfer to the B. subtilis acceptor protein AccB-86, is detected by the production of AMP and concomitant consumption of ATP. (b) Native Page AccB-86 gel shift analysis of BplA. Each lane contains a 20 μl reaction. For each reaction, 50 mM HEPES (pH 7.5), 15 mM MgCl₂, 0.3 mM TCEP, with or without 15 μM BplA, 150 μM AccB-86, 15 μM ATP, and 150 μM biotin are added. The gel is 18% polyacrylamide. The biotinylated form of AccB-86 is indicated by the arrow

2.2 ∣. The BirA′ protein has regulatory function

In E. coli and Bacillus subtilis, transcriptional regulation of the biotin synthetic and transporter genes depends not only on the concentration of biotin but also on the level of acceptor protein (Cronan, 1988; Henke & Cronan, 2014, 2016). Since C. acetobutylicum is an obligate anaerobic Firmicute that lacks genetic tools, we performed in vivo experiments using a related spore-forming Firmicute, B. subtilis, which has biotin operators (bioOs) that are highly similar to the three proposed C. acetobutylicum BirA′-binding sites (Rodionov et al., 2002) (Figure 4a), hence B. subtilis strains were used to monitor regulation of biotin operon transcription by BplA and BirA′ upon manipulation of biotin and AccB levels. Note that because C. acetobutylicum is a biotin auxotroph (Johnston & Goldfine, 1992) indicating that the putative biotin synthetic genes are inactive, we have focused our regulation assays on the operons that contain a bioY transporter gene. At least one of the biotin transporters, bioY1 or bioY2, must be functional because biotin concentrations of a few nM (or less since cells must be subcultured at least three times in the absence of biotin to attain starvation) suffice to support growth of C. acetobutylicum (Johnston & Goldfine, 1992).

FIGURE 4.

β-Galactosidase assays of biotin operon transcription. (a). Sequence alignment of the B. subtilis bioO and BirA′-binding sites. (b, c). B. subtilis strains are grown in defined medium supplemented with the indicated concentrations of biotin and 100 μM IPTG. The results are the average of three independent experiments and the error bars denote standard error of the mean. The effects of AccB induction is due to depletion of the intracellular biotin concentration by the N-terminally deleted B. subtilis BirA that supports growth of the strains since BirA′ lacks biotinylation activity and the two chimera protein transfer biotin from Bio-5′-AMP to AccB-86 poorly (Figure S2). BirA′ denotes strain XS171 which is B. subtilis 1A330 bioW-lacZ birA::Ca 0212 amyE::ΔN birA. BirA′+AccB denotes strain XS208 which is B. subtilis 1A330 bioW-lacZ birA::Ca 0212 amyE::AccB and ΔN birA. BirA′206D denotes strain XS172 which is B. subtilis 1A330 bioW-lacZ birA::birA′206D amyE::ΔN birA. BirA′222G denotes strain XS173 which is B. subtilis 1A330 bioW-lacZ birA::birA′222G amyE::ΔN birA

In construction of B. subtilis strain XS171 the chromosomal B. subtilis birA was replaced by BirA′. The strain is a biotin auxotroph (due to the bioB141 mutation) and contains a bioW::lacZ transcriptional fusion to allow measurement of biotin operon transcription (Henke & Cronan, 2014). The strain also contains an N-terminally deleted B. subtilis birA under the control of an ectopic IPTG-inducible promoter. This mutant BirA protein lacks its regulatory domain and was provided to allow growth of the strain. Strain XS171was grown in defined media containing the indicated biotin concentrations plus 100 μM IPTG. Upon induction, increasing biotin concentration resulted in decreased expression of the biotin operon was observed indicating that BirA′ can bind the B. subtilis biotin operon promoter (Figure 4b). Strains in which the two chimera genes (see below) replaced the host birA gene were also assayed and showed regulatory behavior (Figure 4c).

Strain XS211 contains an additional ectopic IPTG-inducible gene encoding AccB-86 which gives higher levels of the biotin acceptor protein upon induction. Upon AccB-86 induction, expression of the biotin operon was derepressed at biotin concentrations that normally give repression (Figure 4b); a behavior similar to that seen with E. coli (Cronan, 1988). However, rather than a direct effect as seen in E. coli and B. subtilis (precluded due to the inability of BirA′ to make Bio-5′-AMP) the increased AccB-86 expression is due to depletion of intracellular biotin concentrations by biotinylation of AccB-86 catalyzed by the N-terminally deleted B. subtilis BirA. Biotin depletion limited repression by BirA and probably also repression of the two chimera proteins due to their weak abilities to transfer biotin from Bio-5′AMP to AccB-86 (Figure S2).

We also used an E. coli birA regulation knockout strain VC618 (Chakravartty & Cronan, 2012) to test the regulatory activities of BirA′ (Figure S4). All plates were supplemented with X-gal plus the indicated biotin concentrations and incubated at 30°C. BirA′ showed regulatory function in plates supplemented with 20 nM biotin, whereas BplA showed no regulatory function (Figure S4). β-Galactosidase assays were used to quantitate the levels of repression. Upon induction, decreased β-galactosidase activities were observed with increasing concentrations of biotin (Figure S5a) despite the different binding sites of BirA′ and E. coli BirA (Figure S6).

2.3 ∣. Affinity of BirA′ binding to the bioY1 operator

To test the binding affinities of the predicted C. acetobutylicum BirA′-binding sites, we purified BirA′ and performed electrophoretic mobility shift assays (EMSAs) on DNA fragments containing the bioY1-binding site. Purified BirA′ bound the bioY1 operator without addition of biotin or ATP (Figure S7a). This seemed due to co-purification of biotin or Bio-AMP that was retained in the active site as previously seen with S. aureus BirA (Henke & Cronan, 2016). We attempted to clear the active site by hydroxylamine treatments as used previously (Henke & Cronan, 2016) but these were ineffective (Figure S8d) probably because they are targeted at Bio-5′-AMP which BirA′ is unable to synthesize. We then turned to expressing BirA′ in biotin starved cultures of E. coli strain ER47, a biotin auxotroph (Choi-Rhee & Cronan, 2005). The BirA′ protein samples purified from strain ER47 contained no detectable biotin (Figure S8d). Biotin-free BirA′ was retested for DNA binding of the bioY1 operator and binding was again observed in the absence of additional biotin or ATP. However, binding was enhanced by biotin addition (Figure 5). Although dimerization of the E. coli and B. subtilis BirAs requires Bio-5′-AMP, addition of only biotin to the Staphylococcus aureus BirA results in significant dimerization (Wang & Beckett, 2017) and operator binding (Henke & Cronan, 2016).

FIGURE 5.

Electrophoretic mobility shift assays of DNA binding by BirA′. EMSA showing biotin free BirA′ binding to C. acetobutylicum bioY1, B. subtilis bioO and E. coli bioO. blap is used as negative control

DNA-binding activity was not seen with the control DNA (blap, which encodes a small B. subtilis protein of unknown function). Indeed BirA′ bound to the B. subtilis bioO operator which led us to test BirA′ binding in vivo (Figure 4). Surprisingly, BirA′ showed weak binding affinity toward E. coli bioO (Figure 5). We also tested the binding affinity of BirA′ toward the C. acetobutylicum bioY1 and bioY2 operators and found that both bound BirA′ similarly to the cognate bioO (Figure S7c). Furthermore, BirA′ bound the C. acetobutylicum bioY1- and bioY2-binding sites in the presence or absence of ATP (Figure S7d), indicating that ATP plays no role in binding.

These data raised the question of why the EMSA analysis detected DNA-binding activity by BirA′ when no biotin was present. One possibility was that binding was an EMSA cage effect artifact (Cann, 1998). To test this possibility we turned to fluorescence anisotropy, an assay that allows DNA binding to be measured in free solution. Protein preparations purified from E. coli strains ER47 and BL21 STAR were used in the analyses. The DNA-binding activities of BirA′ were tested with a fluorescein 5′ end-labeled 33 base pair DNA containing either the C. acetobutylicum bioY1- or bioY2-binding sites. The changes in anisotropy were plotted against the protein concentrations to obtain binding curves using Graphpad Prism 9 (Figure S9) and the dissociation constants are listed in Table. 1. BirA′ purified from BL21 STAR showed similar binding affinities toward C. acetobutylicum bioY1 and bioY2 sites (Figure S9a,b), with K_d values of 15.4 ± 4.2 nM and 20.28 ± 4.73 nM in the presence of biotin, respectively. BirA′ purified from ER47 showed lower binding affinity toward C. acetobutylicum bioY1 with a K_d of 94.9 ± 28.5 nM (Figure S9d) probably due to the low activity of protein purified from the biotin-starved strain ER47. The binding affinity toward E. coli bioO was also tested and a fluorescein end-labeled 44 bp was used in the assay (Figure S9c). Compared with BirA′-binding affinity toward C. acetobutylicum bioY1, a nearly two-fold decrease of binding affinity with a K_d of 47.49 ± 9.69 nM was obtained. ATP did not aid binding: no significant difference was observed between biotin addition and addition of ATP plus biotin. Much higher K_d values were observed when biotin was absent, indicating that biotin is the regulatory ligand.

TABLE 1.

The dissociation constants of BirA′, BirA′206D, and BirA′222G obtained by fluorescence anisotropy with biotin addition or with addition of biotin plus ATP (±95% confidence intervals)

Protein	DNA binding site	Dissociation constant (nM)
		Protein	ATP	Biotin	ATP + Biotin
BirA′	C. acetobutylicum bioY1	897.85 ± 5.93	695.14 ± 14.33	15.43 ± 4.2	21.86 ± 5.66
BirA′a	C. acetobutylicum bioY1	868.21 ± 11.42	789.57 ± 16.5	94.9 ± 28.5	92.67 ± 23.06
BirA′	C. acetobutylicum bioY2	601.85 ± 10.54	544.6 ± 9.7	20.28 ± 4.73	24.53 ± 4.82
BirA′	E. coli bioO	2088.15 ± 6.11	2,180.47 ± 4.88	47.49 ± 9.69	43.14 ± 10.15
BirA′206D	C. acetobutylicum bioY1	1,385.74 ± 4.77	665.14 ± 11.56	49.05 ± 19.11	29.27 ± 11.56
BirA′222G	C. acetobutylicum bioY1	1,096.63 ± 2.15	897.85 ± 3.23	56.7 ± 24.3	29.71 ± 6.17

^aBiotin free BirA′ was purified from E. coli ER47.

2.4 ∣. Conversion of BirA′ into a functional BirA

Previous work in this lab reported the successful conversion of a group II BirA into a group I BPL (Henke & Cronan, 2014), indicating the possibility of interconversion between the two biotin protein ligase types. Since BirA′ lacks the C-terminal domain and has no ligase function, we hypothesized that we might be able to convert BirA′ into a functional BirA by attaching the C-terminal domain of BplA.

A series of chimeric BirA′ proteins were constructed by fusing C-terminal segments of BplA to BirA′, and two showed ligase function, BirA′206D and BirA′222G (where the prime means C-terminal truncation, and the number denotes the amino acid residue of BirA′ where the fusion junction was made). BirA′206D was constructed by fusing the coding region of the first 266 residues of BirA′ with that of the last 65 residues of BplA (starting from 206D in Ca 0589) whereas BirA′222G was constructed by fusing the coding region of the first 266 residues of BirA′ with that of the last 49 residues BplA (starting from 222G in Ca 0589). The ligase functions of the chimera BirA′ proteins were tested via complementation assays using the E. coli mutant strain BM4062 ara+. All plates were supplemented with X-gal, the indicated concentrations of biotin and incubated at 42°C. Both BirA′206D and BirA′222G expressed from pBAD322C and induced with 0.02% arabinose complemented E. coli BirA ligase activity and regulatory function in strain BM4062 ara+ at 4 nM biotin (Figure 2). We further used an E. coli birA regulation knockout strain VC618 to test the regulatory activities of the two chimera proteins. All plates were supplemented with X-gal, the indicated concentrations of biotin and incubated at 30°C. Both BirA′206D and BirA′222G showed regulatory function in plates supplemented with 20 nM biotin and 0.02% arabinose (Figure S4). The ligase functions of BirA′206D and BirA′222G were tested in vitro via native gel shift assays with AccB-86. Purified AccB-86 was incubated with BirA′206D or BirA′222G with or without ATP and biotin and modest shifts in AccB-86 migration were observed (Figure S2a,b). The magnitude of the gel shifts were not altered by increases in AccB-86 concentration (Figure S2b), consistent with a deficiency in biotin transfer rather than weak binding of the acceptor.

The regulatory functions of the fusion proteins were tested in the B. subtilis bioW::lacZ strain. The chromosomal B. subtilis birA was replaced with birA′206D (strain XS173) or birA′222G (strain XS172). The N-terminally deleted B. subtilis BirA under the control of an IPTG-inducible promoter was provided to ensure survival of the strain. The strain was grown in chemically defined media containing the indicated concentrations of biotin and 100 μM IPTG. Upon induction, decreased expression of the biotin operon was observed with increasing biotin concentrations. Compared with BirA′, however, weaker regulatory functions were observed for both chimera proteins (Figure 4c). We also compared the expression levels of biotin operon transcription with and without IPTG induction and only minor differences were seen (Figure 4c).

Since previous data showed that BirA′ had weak binding affinity toward E. coli bioO, we tested biotin operon expression in VC618, an E. coli birA regulation knockout strain (the yeast Bpl1 supplied ligase activity). All E. coli strains were grown in defined medium supplemented with the indicated concentrations of biotin and 0.02% L-arabinose. With increasing biotin concentrations, decreased levels of expression were observed in BirA′, BirA′206D, and BirA′222G (Figure S5a). The compromised binding activities of BirA′206D and BirA′222G were also tested by fluorescence anisotropy using a fluorescein end-labeled 33 bp C. acetobutylicum bioY1-binding site (Figure S9e,f) and an increased K_d was observed in both cases with biotin and/or ATP addition. However, in contrast to BirA′, the binding affinities of BirA′206D and BirA′222G were enhanced by ATP addition (Table 1).

2.5 ∣. Oligomerization states of BirA′, BplA, BirA′206D and BirA′222G.

Dimerization is required for DNA binding by E. coli BirA and dimerization is dependent on biotin and ATP binding to form Bio-5′-AMP (Beckett, 2007; Cronan, 1988). To determine the oligomerization state of BirA′, BplA, BirA′206D, and BirA′222G, purified proteins were incubated with the crosslinker ethylene glycol bis(succinimidyl succinate) in the presence or absence of biotin and ATP and the products were analyzed on SDS gels (Figure S10). Note, that because crosslinked proteins are a mixture of species having different shapes diffuse bands are formed and minor bands may not be detected by staining, depletion of the monomer bands provides the best estimate of crosslinking efficiency. However, the monomer bands are also diffuse due to intramolecular crosslinking. In these gels, the monomer bands of BirA′ and the two chimera proteins decreased markedly when crosslinked in the presence of biotin whereas crosslinking in the absence of biotin resulted in almost no depletion of the monomer bands. These data indicate that biotin is required for formation of BirA′ and chimera dimers. In contrast, BplA showed no depletion of the monomer bands upon crosslinking in either the presence or absence of biotin which demonstrated that BplA is a monomeric protein.

Size exclusion chromatography was also used to characterize the oligomerization state of BirA′, BplA, and the two chimera proteins. Biotin (10 μM) was added into the running buffer to stabilize the oligomerization state. Size exclusion chromatography indicted that BirA′, BirA′206D, and BirA′222G are dimers whereas BplA is a monomer (Table S4).

2.6 ∣. Testing for interaction between BirA′ and Bpl

Previous studies posited the possibility of Group I Bpls having interactions with other unknown proteins (Satiaputra etal., 2016) and we have established that domains can be exchanged between BirA′ and BplA. We were thus curious to know if BirA′ could interact with BplA. To explore the possibility, we used BplA to ensure the survival of B. subtilis strain in which the B. subtilis birA had been replaced by BirA′ and conducted β-galactosidase assays. Compared with the strain that expressed the N-terminal deleted BirA (strain XS171) (Henke & Cronan, 2014), the strain where BplA provided BPL activity (strain XS209) was less sensitive to the changes of biotin concentration (Figure S5b). The decreased sensitivity was also observed in strains with increased AccB levels. As shown in Figure S5b strains that have increased AccB levels showed derepression whereas the strain that had BplA as sole biotin protein ligase (strain XS210) was less sensitive to changes of biotin concentration compared to strain XS211. One possibility is that BplA is a more active biotin protein ligase than the N-terminal deleted B. subtilis BirA and more rapidly loads biotin onto its cognate acceptor protein. Another possibility is that interaction between BplA and BirA′ could occur and interfere with BirA′ binding.

2.7 ∣. Evidence for interactions between BIRA’and BPLA

To test for possible interactions between BirA′ and BplA, C-terminal His₆-tagged BirA′ and C-terminal S-tagged BplA were expressed under the control of T7 promoter using the pACYCDuet-1 vector, and procedures for purifying compound BirA′-Bpl are illustrated in Figure S11a. Cell lysates containing a mixture of dimer BirA′, monomer BplA and compound BirA′-BplA were first chromatographed through a Ni-NTA agarose column to bind the His₆-tagged proteins and remove S-tagged BirA′. The purified protein mixture was chromatographed on an S-protein agarose to remove any His₆-tagged BplA monomer and thus obtain mixtures containing both BirA′ and BplA. Western blotting against the S-tagged proteins were performed to test the presence of BirA′ in the His-tagged protein fraction and S-tagged BirA′ was observed in the protein aggregate (Figure S11b), an indication that BirA′ and BplA interact. We then subjected the S-tagged proteins to size exclusion chromatography but no complex was seen possibly due to transient and/or weak interactions between BirA′ and BplA and the dilution inherent in size exclusion chromatography (data not shown). We switched the tags between the two proteins (constructed His₆-tagged BplA and S-tagged BirA′) and repeated the purification and the size exclusion chromatography but again no complex was observed (data not shown).

Next, we tested for interaction between BirA′ and BplA by use of the bacterial adenylate cyclase-based two-hybrid system (Olson et al., 2018). Two pairs of vectors were used. In pair 1, T25 was fused into the N-terminal of BirA′ and T18 was fused into the N-terminus of BplA. In pair 2, T25 was fused into the N-terminal of BplA and T18 was fused into the N-terminus of BirA′. The two pairs, along with the positive control (which formed a leucine zipper), empty vectors and negative control were tested for β-galactosidase activity following the established protocol (Olson et al., 2018). Compared with the empty vectors and negative control, a 5-fold increase was detected in pairs 1 and 2, indicating interactions between the two proteins (Figure 6). Although the interactions were weaker than those seen in the leucine zipper positive control, note that these monomeric leucine zipper proteins rapidly form a very stable dimer (Holtzeret al., 2001) whereas the BplA-BirA′ interaction may appear weak because it is transient.

FIGURE 6.

Characterization of interactions between BirA′ and BplA via the bacterial adenylate cyclase-based two-hybrid assay detected by β-galactosidase activity. The results are the average of eight independent experiments and the error bars denote standard error of the mean

3 ∣. DISCUSSION

BirA′, the protein encoded by gene Ca 0212, has regulatory function despite lacking the canonical BPL C-terminal domain and thus the ability to synthesize Bio-5′-AMP. The BirA′ regulatory ligand is biotin rather than the canonical Bio-5′-AMP. This is readily rationalized in the context of C. acetobutylicum, a natural biotin auxotroph, which must obtain biotin from the environment. Biotin is a rare commodity in nature and thus high affinity transport systems are required for utilization. We believe that the physiological role of BirA′ is to control biotin uptake by directly regulating transcription of the transporter genes. When the cytosol becomes replete with biotin BirA′ would bind biotin and shut down transcription of the transporters. Since BioY proteins consume ATP (Sirithanakorn & Cronan, 2021) this avoids wasting energy. It is also possible that high levels of biotin are toxic. This scenario comes from the observation that, although the two half reactions of the key E. coli fatty acid synthesis enzyme, acetyl-CoA carboxylase, both function with free biotin, the overall reaction does not (Cronan & Waldrop, 2002). The overall reaction requires a biotinylated substrate protein (AccB) that couples the two half reactions. Hence, high biotin concentrations could uncouple the C. acetobutylicum acetyl-CoA carboxylase by competing with the biotinylated substrate protein in formation of the first intermediate, carboxybiotin. Note that acetyl-CoA carboxylase is the dominant C. acetobutylicum biotin-requiring enzyme as shown by the ability of fatty acid supplementation to replace biotin (Johnston & Goldfine, 1992).

The BplA protein encoded by the Ca0589 gene is a classical group I biotin protein ligase composed of the core catalytic and C-terminal domains. The core domain has a structure and sequence conserved throughout the BPL family. The C-terminal domain is essential for binding ATP and plays roles in acceptor protein substrate binding (Chapman-Smith et al., 2001) and dimerization (He et al., 2018) (Chakravartty & Cronan, 2012). The lack of the C-terminal domain in BirA′ resulted in the expected loss of the ability to synthesize Bio-5′-AMP (Figure 3a) and thus BirA′ is not a ligase but has retained function as a biotin sensor. However, upon fusion with the C-terminal domain of BplA, BirA′ became a functional biotin protein ligase, indicating the importance of the C-terminal domain for catalysis.

Group I BPLs catalyze biotinylation but lack the DNA-binding domain required to regulate genes involved in biotin synthesis and transport, which raises the questions of how biotin synthesis and transport are regulated in bacteria that lack BirA proteins. Comparative genomic analyses have identified two biotin operon transcriptional repressors, BioR and BioQ (Sirithanakorn & Cronan, 2021). However, the ligands that modulate repression by these proteins remain unknown.

4 ∣. CONCLUSIONS

We report the first example of biotin directly exerting transcriptional regulation of biotin transport. BirA′ is unable to convert biotin to Bio-5′-AMP, the canonical regulatory ligand and thus is an ATP-independent biotin sensor. The physiological role of BirA′ in C. acetobutylicum is to regulate biotin uptake from the environment which provides a rationale for its maintenance in the genome. It could play a role in regulating biotin synthesis, but this is moot because C. acetobutylicum is unable to synthesize biotin. The interaction between BirA′ and BplA may be a means to adjust BplA activity to biotin supply. A plausible scenario is that some progenitor of C. acetobutylicum had a functional biosynthesis pathway regulated by a classical BirA bifunctional ligase. When the ability to synthetize biotin was lost, duplication and subsequent remodeling of the bifunctional ligase gave the BplA and BioA′ proteins. BplA was required for biotinylation of acetyl-CoA carboxylase and pyruvate carboxylase whereas BioA′ regulated biotin transport.

BirA′ weakly interacts with Bpl. This is the first report of protein-protein interactions observed between a biotin sensor and a biotin protein ligase, indicating a possible new mechanism for regulation of biotin metabolism. The interaction could provide a means to coordinate ligation and bio operon regulation analogous to that of the Group II BirAs of E. coli and B. subtilis. BirA′ would provide transcriptional regulation in response to intracellular biotin concentrations whereas BplA would provide ligation and via interaction with BirA′ could link transcriptional regulation to the level of unbiotinylated acceptor protein. An attractive notion is that BplA could acquire biotin from the BirA′ binding site because under conditions of limited biotin the BirA′ binding site must be cleared to relieve repression of bioY. However, if biotin binding by BirA′ is significantly weaker than BplA biotin binding, direct interaction would not be needed: BplA could clear the BirA′ binding site by simple competition. Considering the apparent weakness of the protein-protein interaction detected via the two hybrid assay, other proteins or ligands might assist in the process.

5 ∣. EXPERIMENTAL PROCEDURES

5.1 ∣. Strains, chemicals and culture media

Reagents and chemicals were obtained from Research Products International 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. Oligonucleotides were purchased from Integrated DNA Technologies and are listed in Table S4. PCR amplification was performed using Q5 high fidelity DNA polymerase (New England BioLabs) according to manufacturer protocols. DNA constructs were sequenced by ACGT, Inc. The rich medium for growth of E. coli and B. subtilis was LB broth. The defined medium for E. coli was M9 salts supplemented with 0.5% glucose (or 0.5% glycerol) and 0.1% vitamin-free Casamino Acids (Difco). The defined medium for B. subtilis was Spizizen salts supplemented with 0.5% glycerol and 0.05% vitamin-free Casamino Acids (Difco) plus 0.01% each of tryptophan, tyrosine, isoleucine, and phenylalanine. The defined medium for protein purification in E. coli (protein purification medium) was M9-XS1 salts supplemented with 2 mM MgSO₄, 0.5x Trace Metals, 0.5%CAA, 0.4% glycerol, and 1% glucose. The M9-XS1 salts contains (in g/L) Na₂HPO₄, 7.1; K₂HPO₄, 3.4; NaCl, 0.5; NH₄Cl, 2.675 adjusted to pH 8.0–8.2. For 100 ml 1000X Trace Metals, contained 50 ml 0.1 M FeCl₃, 2 ml 1 M CaCl₂, 1 ml 1 M MnCl₂, 1 ml 1 M ZnSO₄, 1 ml 0.2 M CoCl, 2 ml 0.1 M CuCl₂, 2 ml 0.2 M NiCl₂, 2 ml 0.1 M Na₂MoO₄, 2 ml 0.1 M H₃BO₃ and 37 ml MilliQ H₂O. Antibiotics were used at the following concentrations (μg/ml): kanamycin sulfate, 50; chloramphenicol, 20; erythromycin sulfate, 12.5; lincomycin, 12.5; ampicillin 100, and spectinomycin 100.

The 6×-His Tag Monoclonal Antibody (4E3D10H2/E3) was purchased from Thermo Fisher Scientific, catalog # MA1-135, RRID AB_2536841. Goat anti-Mouse IgG (H + L) Secondary Antibody, HRP was purchased from Thermo Fisher Scientific, catalog # 31,430, RRID AB_228307. Gel filtration standards was purchased from BIO-RAD, catalog # 1,511,901. S-protein HRP Conjugate (catalog # 69,047), S-protein Agarose (catalog # 69,704), and S-Tag Monoclonal Antibody (catalog # 71,549) were purchased from Millipore Sigma. Ni-NTA agarose, (lot No. 139,290,134) was purchased from QIAGEN.

5.2 ∣. Plasmids And Plasmid Constructions

5.2.1 ∣. Complementation plasmids

All PCR amplified C. acetobutylicum genome segments were from strain ATCC 824 genomic DNA. The birA′206D gene contains the 5′ 1–798 bp segment of Ca 0212 and the last 198 bp of Ca 0589. The birA′222G gene contains the 5′ 1–798 bp segment of Ca 0212 and the last 150 bp of Ca 0589. The Ca 0212 and Ca 0589 genes were PCR amplified using primers SKH147 and SKH148, SKH149 and SKH150, respectively. The 5′ primers contained Eco RI sites whereas the 3′primers contained SalI sites and were ligated into pBAD322Cm cut with EcoRI and SalI to construct plasmids pSKH032 and pSKH033, respectively. The 5′ 1–798 bp segment of Ca 0212 was PCR amplified using primers SXJ125 and SXJ138. The last 150 bp of Ca 0589 was PCR amplified using primers SXJ139 and SXJ124. These two fragments overlap and were assembled by overlapping PCR and inserted into EcoRI and SalI cut pBAD322C to give plasmid pXS1. The last 198 bp of Ca 0589 was PCR amplified using primers SXJ144 and SXJ124. The Ca 0212 1–798 bp which overlapped with the last 198 bp of Ca 0589 were assembled by overlapping PCR and inserted into EcoRI and SalI cut pBAD322Cm at to create plasmid pXS2.

5.2.2 ∣. Protein expression/purification plasmids

The Ca 0212 gene was PCR amplified with primers SKH151 and SKH152 and ligated into NcoI and XhoI digested pET28b to give plasmid pSKH036. The Ca 0589 gene was PCR amplified with primers SKH155 and SKH156 and ligated into NcoI and XhoI digested pET28b to create plasmid pSKH037. DNA fragments containing Ca 0212 1–798 bp and the last 150 bp of Ca 0589 were PCR amplified from pXS1 using primers SXJ206 and SXJ208, and pET28b was PCR amplified using primers SXJ207 and SXJ209. These two fragments were assembled using Gibson Assembly to create pXS3 (Gibson et al. 2009). DNA fragments containing 1–798 bp of Ca 0212 and the last 198 bp of Ca 0589 were PCR amplified from pXS3 using primers SXJ206 and SXJ208, and pET28b was PCR amplified from pET28b using primers SXJ207 and SXJ209. These two fragments were assembled using Gibson Assembly to create plasmid pXS4.

The Ca 0589 gene was PCR amplified with primers SXJ269 and SXJ270 and ligated into pACYCDu-et-1 NcoI and HindIII sites to create plasmid pXS12. The Ca 0212 gene was PCR amplified with primers SXJ271 and SXJ272, and pXS12 was PCR amplified with primers SXJ274 and SXJ273. The two fragments were assembled using Gibson Assembly to create plasmid pXS13. The Ca 0589 gene was PCR amplified with primers SXJ296 and SXJ295 and ligated into pACYCDu-et-1 at NcoI and HindIII to create plasmid pXS15. The Ca 0212 gene was PCR amplified with primers SXJ271 and SXJ294, and pXS15 was PCR amplified with primers SXJ274 and SXJ293. The two fragments were assembled using Gibson Assembly to create plasmid pXS16.

5.2.3 ∣. Construction of B. subtilis plasmids

The Ca 0589 gene was PCR amplified using primers SXJ151 and SXJ152 and was ligated into pDR111 at HindIII and SphI to create plasmid pXS5. The B. subtilis genes were amplified from B. subtilis 168 genomic DNA. B. subtilis accB-86 was PCR amplified using primers SXJ236 and SXJ241 and Ca 0589 was PCR amplified using primers SXJ240 and SXJ242. Those two fragments were assembled and ligated into pDR111 cut with SalI and SphI to create plasmid pXS6. B. subtilis Δ5-63 birA was PCR amplified using primers SXJ103 and SXJ104 and was ligated into pDR111 cut with SalI and SphI to create plasmid pXS7. B. subtilis accB-86 was PCR amplified using primers SXJ236 and SXJ237 and B. subtilis Δ5-63 birA was PCR amplified using primers SXJ238 and SXJ239. Those two fragments were assembled and ligated into the SalI and SphI sites of pDR111 to create plasmid pXS8. The last 500 bp of the B.subtilis cca gene was PCR amplified using primers SKH165 and SKH166. B. subtilis Δ5-63 birA was PCR amplified using primers SKH167 and SKH168. These two fragments were assembled into pDG780-panB at BamHI and EcoRI using Gibson Assembly (NEB) to create plasmid pSKH039. The last 500 bp of B. subtilis cca was PCR amplified using primers SXJ194 and SXJ195. Ca 0212 was PCR amplified using primers SXJ196 and SXJ197. The birA′206D gene was PCR amplified from pXS2 using primers SXJ196 and SXJ198. The birA′222G gene was PCR amplified from pXS2 using primers SXJ196 and SXJ198. The cca500 and Ca 0212 were assembled into pDG780-panB at BamHI and EcoRI to create plasmid pXS9. The cca500 and birA′206D gene were assembled into pDG780-panB at BamHI and EcoRI to create plasmid pXS10. The cca500 and birA′222G genes were assembled into pDG780-panB at BamHI and EcoRI to create plasmid pXS11.

5.2.4 ∣. Bacterial adenylate cyclase-based two-hybrid plasmids

The Ca 0212 gene was PCR amplified with primers SXJ409 and SXJ410, and pKT25-zip was PCR amplified with primers SXJ411 and SXJ412. The two fragments were assembled using Gibson Assembly to create plasmid pXS17. The Ca 0589 gene was PCR amplified with primers SXJ413 and SXJ414, and pKT25-zip was PCR amplified with primers SXJ415 and SXJ416. The two fragments were assembled using Gibson Assembly to create plasmid pXS18. Ca 0212 was PCR amplified with primers SXJ417 and SXJ418, and pUT18C-zip was PCR amplified with primers SXJ419 and SXJ420. The two fragments were assembled using Gibson Assembly to create plasmid pXS19. The Ca 0589 gene was PCR amplified with primers SXJ421 and SXJ422, and pUT18C-zip was PCR amplified with primers SXJ423 and SXJ424. The two fragments were assembled using Gibson Assembly to create plasmid pXS20. pKT25-zip was PCR amplified with primers SXJ433 and SXJ414 to create plasmid pXS21, and pUT18C-zip was PCR amplified with primers SXJ436 and SXJ437 to create plasmid pXS22.

All plasmids were verified via sequencing and are listed in Table S3. DNA fragments cloned into pBAD322cm were verified using primers SXJ62 and SXJ63. DNA fragments cloned into pET28b were verified using primers SXJ192 and SXJ193. DNA fragments cloned into pACYCDuet-1 were verified using primers SXJ255 and SXJ256 for MCS1, and primers SXJ257 and SXJ258 for MCS2. DNA fragments cloned into pDR111 were verified using primers SXJ54 and SXJ55. DNA fragments cloned into pDG780 were verified using primers SXJ56 and SXJ57. DNA fragments cloned into pKT25 were verified using primer SXJ435. DNA fragments cloned into pUT18C were verified using primer SXJ438.

5.3 ∣. Bacillus subtilis strain constructions

B. subtilis competent cell preparation and transformation were carried out as described by Dubnau and Davidoff-Abelson (Dubnau & Davidoff-Abelson, 1971). Strain SKH001 was transformed with linearized pSKH039 to replace the B. subtilis birA with B. subtilis birA ΔNTD with a kanamycin cassette to give strain XS67. Strain SKH001 was transformed with linearized pXS7 to replace the B. subtilis amyE with B. subtilis birA ΔNTD with a kanamycin cassette to give strain XS61. Strain XS61 was transformed with linearized pXS9 to give strain XS171. Strain XS61 was transformed with linearized pXS10 to give strain XS172. Strain XS61 was transformed with linearized pXS11 to give strain XS173. Strain SKH001 was transformed with linearized pXS8 to replace the B. subtilis amyE with B. subtilis accB-86 and birA ΔNTD with a kanamycin cassette to give strain XS208. Strain XS208 was transformed with linearized pXS9 to give strain XS211. Strain SKH001 was transformed with linearized pXS5 to replace the B. subtilis amyE with C. acetobutylicum Ca 0589 with a kanamycin cassette to give strain XS109. Strain XS109 was transformed with linearized pXS9 to give strain XS209. Strain SKH001 was transformed with linearized pXS6 to replace the B. subtilis amyE with B. subtilis accB86 and C. acetobutylicum 0589 with a kanamycin cassette to give strain XS207. Strain XS207 was transformed with linearized pXS9 to give strain XS210. The integration events were verified by PCR and sequencing. The strains constructed are listed in Table S2.

5.4 ∣. Complementation analysis

The E. coli strain BM4062 was transformed with pBAD322Cm, pSKH015, pSKH032, pSKH033, pXS1, and pXS2. Transformants were selected on LB plates supplemented with streptomycin and chloramphenicol at 30°C. The strains were grown at 42°C on defined medium plates supplemented with chloramphenicol, streptomycin, 25 μM 5-bromo-4-chloro-indolyl-β-D-galactopyranoside (X-gal), indicated concentration of biotin and 0.02% arabinose. E.coli strain VC618 was transformed with pBAD322CM, pSKH015, pSKH032, pSKH033, pXS1, and pXS2. Transformants were selected on LB plates supplemented with ampicillin and chloramphenicol at 30°C. The strains were grown at 30°C on defined medium plates supplemented with chloramphenicol, streptomycin, 25 μM 5-bromo-4-chloro-indolyl-β-D-galactopyranoside (X-gal), indicated concentrations of biotin and 0.02% arabinose.

5.5 ∣. Bio-5′-AMP synthesis assay

The assays contained 50 mM Tris-HCl buffer (pH 8.0), 5.5 mM MgCl₂, 5 μM ATP, 25 μM biotin, 5 μM BirA, 0.1 μM [α-³²P] ATP and with or without 50 μM AccB-86 for a total reaction mixture of 20 μl. The reaction mixtures were incubated at room temperature for 30 min. A portion of each reaction mixture (1 μl) was spotted on cellulose thin-layer chromatography (TLC) plates and developed in isobutyric acid-NH₄OH-water (66:1:33) (Prakash & Eisenberg, 1979). The thin-layer chromatograms were dried for 10 hr and exposed to a phosphorimaging screen and visualized using a Fujifilm FLA-3000 Phosphor Imager and Fujifilm Image Gauge software.

5.6 ∣. Purification of BirA′, BplA, BirA′206D and BirA′222G

E. coli BL21 Star (DE3) was transformed with pSKH036 for BirA′ purification. E. coli BL21 Star (DE3) was transformed with pSKH037 for BplA purification. E. coli BL21 Rosseta was transformed with pXS2 for BirA′206D purification. E. coli BL21 Codon Plus was transformed with pXS1 for BirA′222G purification. The strain was grown at 37°C in protein purification medium with kanamycin to an OD₆₀₀ of 0.8 and expression was induced by addition of IPTG to 1 mM. Following growth for an additional 16 hr at room temperature (24°C), the cells were recovered by centrifugation and suspended in lysis buffer (50 mM (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 10 mM imidazole, 0.5 mM tris(2-carboxyethyl)phosphine (TCEP), 250 mM NaCl, 5% glycerol, pH 7.5. The cells were lysed by passage through a French pressure cell and the lysate was centrifuged to remove unbroken cells and cellular debris and the supernatant was added to Ni NTA beads (Qiagen) and incubated for 1 hr before the beads were added to a disposable 10 ml polypropylene column. The column was washed with 30 ml of wash buffer (lysis buffer containing 60 mM imidazole). Proteins were eluted in 1 ml fractions with elution buffer (lysis buffer containing 250 mM imidazole). The fractions were analyzed by SDS-PAGE to determine purity. Pure fractions were combined and dialyzed against storage buffer (lysis buffer lacking imidazole). Aliquots were flash frozen and stored at −80°C.

To express and purify BirA′ free of biotin, E. coli strain ER47 was transformed with pSKH036. The strain was grown overnight at 37°C in medium A (1× M9, 1 mM MgSO₄, 1% CAA, 0.2% glucose, 4 nM biotin) supplemented with kanamycin. Cells were collected by centrifugation and suspended in biotin free medium A at an OD₆₀₀ of 0.6 and grown for 2 hr at 37°C. Cells were collected by centrifugation and suspended in medium B (1× M9, 1 mM MgSO₄, 1% CAA, 0.2% glycerol, 1 mM TPTG) supplemented with kanamycin and grown for 3 hr at 37°C. The cells were then recovered, suspended, and the proteins were purified as above.

5.7 ∣. Purification of B. subtilis AccB-86

E. coli strain BL21 (DE3) was transformed with pSKH003. The strain was grown at 37°C in LB medium supplemented with kanamycin to an OD600 of 0.8 and induced by addition of IPTG to 1 mM for an additional 4 hr at 30°C. 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) (Chapman-Smith et al. 1994).

5.8 ∣. Tandem mass spectroscopy of AccB-86

LC-MS/MS analysis of the intact AccB proteins was performed on a Thermo Scientific Fusion Orbitrap mass spectrometer coupled to a Dionex Ultimate 3,000 RSLC nano UHPLC. The separation was performed on a Thermo Scientific Acclaim PepMap RSLC (75 μm × 15 cm) C18, 2 μm, 100 Å column at 300 nl/min. The gradient (A: 0.1% formic acid in water, B: 0.1% formic acid in acetonitrile) was run over 76 min with the following compositions: 0–6 min 1%–4% B, 6–51 min 4%–35% B, 51–56 min 35%–50% B, 56–60 min 50%–90% B, 60–64 min 90% B, 64–66 min 90%–2% B, 66–76 min 2% B. The Fusion Tribrid settings were in the Peptide mode and the Universal method with resolution of 120,000, precursor mass range 300–1600, RF lens 60%, AGC 2.0e5, Injection Time 100 ms, Intensity threshold 5.0 e3, Charge state 1–7, Exclude after 2 times in 30 s for 60 s, Mass tolerance ± 10 ppm, Data Dependent MS², IT CID, Isolation 1.6 m/z, CID activation, Collision energy 35%, IT detection in Rapid Scan mode, AGC 1.0e4, All parallelizable time, maximum Injection time 60 s, 3 s/cycle.

Raw data were processed into peak lists using Mascot Distiller 2.7.1.0 (Matrix Science) and searched using Mascot 2.7.0.1 against a custom database composed of the AccB-86 and the B subtilis AccB sequence. Mascot search settings included biotinylation, oxidation (Met) variable modifications, precursor ion mass tolerance ±10 ppm, fragment ion mass tolerance ±0.3 Da and threshold for accepting individual spectra of 13.

5.9 ∣. β-Galactosidase assays

B. subtilis strains were grown overnight in defined media containing 2 nM biotin, kanamycin, erythromycin, lincomycin, and spectinomycin. Cultures were diluted to OD₅₉₅ of 0.3 in defined media containing various concentrations of biotin (2, 4, 8, 20, 40, 80, 200 nM). Samples were grown with or without 0.1 μM IPTG for 6 hr. β-Galactosidase activity was determined as described by Harwood and Cutting (Harwood & Cutting, 1990) following permeabilization of the cells with lysozyme.

6 ∣. Electrophoretic mobility shift assay of DNA binding

The predicted C. acetobutylicum DNA-binding site upstream of bioY1 was PCR amplified with primers SKH153 and SKH154, the predicted C. acetobutylicum DNA-binding site upstream of bioY2 was PCR with primers SXJ129 and SXJ130 and the predicted C. acetobutylicum DNA-binding site downstream of bioY was PCR amplified with primers SXJ131 and SXJ132. The negative control DNA (blap) was amplified from B. subtilis 168 gene yngHB encoding the biotin/lipoyl attachment protein with primers SKH028 and SKH029. All DNA fragments were 125 bp in length. The PCR products were assayed on a 2.0% agarose gel and purified using a QIAquick PCR Purification Kit (Qiagen). DNA concentrations were determined at OD₂₆₀ by using a NanoDrop 2000c. The DNA-binding reaction contained 50 mM HEPES (pH 7.5), 50 mM NaCl, 10% glycerol, 40 nM DNA, indicated concentrations of BirA′, 1 μM ATP, 1 mM MgCl₂, and 1 μM biotin. The binding reactions were incubated at room temperature for 45 min and then loaded into a 6% DNA retardation gel (Invitrogen). The gel was run in 0.5× TBE at 100 V for 90 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.

6.1 ∣. Biotin bioassays

E. coli strain NRD25 (Choi-Rhee & Cronan, 2005) was grown overnight in defined media supplemented with chloramphenicol and 1 nM biotin. The cells were recovered by centrifugation at 6000 ×g for 3 min and washed 4× with 1 ml of M9 medium. The cells were re-suspended in 1 ml of glucose M9 medium and subcultured into 100 ml of glucose M9 minimal media lacking biotin and containing 5 units of avidin. These cultures were incubated at 37°C for 5 hr, centrifuged at 15,000 ×g for 5 min and the cell, pellets washed 5× with 1 ml of M9 medium. The cells were then suspended in 1 ml M9 and added to 150 ml of minimal media agar containing the redox indicator 2,3,5-triphenyl tetrazolium chloride (0.1%, w/v) (Lin et al., 2010) Six ml of the agar mixture was added to sectored petri dishes. A sterile 6 mm paper disc was applied to the top of the agar and spotted with 10 μl of biotin standards or denatured protein samples. Protein samples were denatured by heating to 99°C for 20 min and centrifuged at 15,000 ×g for 5 min. The supernatant was collected and used to spot the paper discs. The pmol values were calculated from the protein concentrations before denaturation. The plates were incubated at 30°C overnight. Growth of strain NRD25 was visualized as a deposit of red formazan.

6.2 ∣. Fluorescence anisotropy analyses

DNA fragments containing the 5′ -fluorescein-label, primers SXJ 181, 182 and 249, were synthesized by IDT. DNA-binding reactions contained 50 mM Tris-HCl (pH 8.0), 50 mM NaCl, 10% glycerol, 2 nM DNA, and indicated concentrations of BirAs. Buffer containing the fluorescein labeled DNA was added to a black 96 well microplate (Molecular Devices). The indicated amount of a BirA sample was added to the wells in a final reaction volume of 20 ul. Reactions were incubated at room temperature for 40 min. Anisotropy values were measured using the fluorescence polarization function of the Analyst HT Plate Reader (Molecular Devices) at the High-Throughput Screening Facility, School of Chemical Sciences at the University of Illinois.

6.3 ∣. Western blotting

Protein samples were loaded and separated on a 12% SDS-polyacrylamide gel and transferred by electrophoresis to Immobilon-P membranes (Millipore) for 60 min at 90 V. The membranes were preblocked with TBS buffer (20 mM Tris base and 150 mM NaCl, pH 7.5) containing 0.1% (v/v) Tween-20 and 5% non-fat dry milk. For His₆-tagged protein visualization, the membranes were probed for 1 hr with 6×-His Tag monoclonal antibody diluted 1:2000 in the above buffer and washed 4 times with TBS buffer. Following incubation with Goat anti-Mouse IgG (H + L) secondary antibody, the labeled proteins were detected using Quantity One software. For S-tagged protein visualization, the membranes were probed for 1 hr with S-Tag monoclonal antibody (diluted 1:5,000) in the above buffer, and washed 4 times with TBS buffer. Following incubation with S-protein HRP conjugate (diluted 1:5,000), the labeled proteins were detected using Quantity One software.

6.4 ∣. Size exclusion chromatography

The Superdex200 Increase 10/300 GL column was purchased from GE Life Sciences and run following user instructions. The running buffer was 0.05 M Phosphate buffer containing 0.15 M NaCl, 0.5 mM TCEP, and 10 μM biotin (pH 7.4) with a flow rate of 0.1 ml/min. Gel filtration standard was purchased from BIO-RAD (catalog # 1,511,901) and used following manufacturer instructions and protein samples were loaded as the manufacturer directed.

6.5 ∣. Chemical cross-linking of proteins

Protein samples were incubated with or without 100 μM biotin and/or ATP and indicated concentration of ethylene glycol bis[succinimidylsuccinate] (EGS) at room temperature for 30 min. SDS-loading dye was added and samples were heated to 99°C for 5 min. Samples were loaded on 4%–20% gradient SDS polyacrylamide gels (BioRad) and run at 110 v for 75 min.

6.6 ∣. Sequence alignments

Sequence alignments and the output was processed by Jalview 2 to generate the final figure (Waterhouse et al. 2009).

DATA AVAILABILITY STATEMENT

The data from this investigation are available from JEC upon request.