1
Brucella BioR regulator defines a complex regulatory mechanism for 1
bacterial biotin metabolism 2
3
Youjun Feng1,2*, Jie Xu3, Huimin Zhang4, Zeliang Chen3* and Swaminath 4
Srinivas5 5
6
1 Department of Microbiology, University of Illinois, Urbana, IL 61801, the 7
United States 8
9
2Institute of Microbiology, College of Life Science, Zhejiang Uninversity, 10
Hangzhou 310058, Zhejiang, P. R. China 11
12
3Department of Infectious Disease Control, Beijing Institute of Disease Control 13
and Prevention, Beijing 100071, P.R. China 14
15
4Institute for Genomic Biology, University of Illinois, Urbana, IL 61801, the 16
United States 17
18
5Department of Biochemistry, University of Illinois, IL 61801, the United States 19
20
21
22
23
*Correspondence to: Youjun Feng ([email protected]), Department of 24
Microbiology, University of Illinois, B103 Chemical and Life Sciences 25
Laboratory, 601 S. Goodwin Ave, Urbana, IL 61801, Phone: (217) 333-7919; 26
Fax: (217) 244-6697; Zeliang Chen ([email protected]), Institute of 27
Disease Control and Prevention, Academy of Military Medical Sciences, No. 28
20, Dongdajie, Fengtai District, Beijing 100071, P. R. China. Tel: 29
86-10-66948434; Fax: 86-10-6694843430
Copyright © 2013, American Society for Microbiology. All Rights Reserved.J. Bacteriol. doi:10.1128/JB.00378-13 JB Accepts, published online ahead of print on 31 May 2013
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
2
Abstract 31
The enzyme cofactor biotin (Vitamin H or B7) is an energetically expensive 32
molecule whose de novo biosynthesis requires 20 ATP equivalents. It seems 33
quite likely that diverse mechanisms have been evolved to tightly regulate its 34
biosynthesis. Unlike the model regulator BirA, a bi-functional biotin protein 35
ligase with a capability of repressing biotin biosynthetic pathway, we recently 36
reported an alternative machinery, BioR, a new type of GntR family 37
transcriptional factor that can repress the expression of bioBFDAZ operon in 38
the plant pathogen Agrobacterium tumefaciens (A. tumefaciens). However, 39
quite unusually, a closely related human pathogen, Brucella melitensis (B. 40
melitensis) has four putative BioR binding sites (both bioR and bioY possesses 41
one site in the promoter region, whereas the bioBFDAZ (bio) operon contains 42
two tandem BioR boxes). This raised the question of whether BioR mediates 43
the complex regulatory network of biotin metabolism. Here we report that this 44
is the case. The B. melitensis BioR ortholog was over-expressed and purified 45
to homogeneity, and its solution structure was found to be dimeric. Functional 46
complementation in a bioR isogenic mutant of A. tumefaciens elucidated that 47
Brucella BioR is a functional repressor. Electrophoretic mobility shift assays 48
demonstrated that the four predicted BioR sites of Brucella plus the BioR site 49
of A. tumefaciens can all interact with the Brucella BioR protein. In a reporter 50
strain that we developed on the basis of a double mutant of A. tumefaciens 51
(ΔbioR ΔbioBFDA), the β-gal activity of three plasmid-borne transcriptional 52
fusions (bioBbme-lacZ, bioYbme-lacZ and bioRbme-lacZ) were dramatically 53
decreased upon overexpression of Brucella bioR. Real-time quantitative PCR 54
analyses showed that expressions of bioBFDA and bioY are significantly 55
elevated upon removal of bioR from B. melitensis. Together, we conclude that 56
not only is Brucella BioR a negative auto-regulator, but also a repressor of 57
expression of bioY and bio operon that separately function in biotin transport 58
and biosynthesis pathway.59
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
3
Introduction 60
Biotin (vitamin H or B7) is a covalently-bound enzyme cofactor required 61
by three domains of life (1). This requirement of biotin has long been 62
recognized in a class of important biotin-dependent enzymes involved in 63
central metabolism, such as carboxylases and decarboxylases (2). Generally, 64
bacterial biotin metabolism encompasses the following three processes: 65
transport/uptake, de novo synthesis, and utilization (1, 3, 4). In 66
microorganisms with a full capability of synthesizing biotin, four universal 67
genes (bioF, bioA, bioD, and bioB) appear to constitute the majority of the 68
biotin biosynthetic pathway (1, 5). In E. coli, which defines the paradigm, biotin 69
synthesis proceeds via a four step path with pimeloyl-ACP as precursor (6). 70
Recently, we found that the generation of pimeloyl moiety in earlier steps of 71
biotin synthesis is involved in a modified Type II fatty acid biosynthetic pathway 72
in E. coli (6-8). In the BioC-BioH pathway of pimelate synthesis, BioC 73
methylates malonyl-CoA (or ACP) and gives a methyl malonyl-thioester 74
destined to fatty acid biosynthesis to act as a primer (6, 7, 9), whereas the bioH 75
gene product demethylates the pimeloyl-ACP methyl ester to form pimeloyl- 76
ACP after two rounds of the fatty acid elongation cycle (6, 7, 10). The 77
subsequent four step pathway functions in assembling the double rings in the 78
biotin molecule (6-8) (seen in Fig. 1A). First, BioF (7-keto-8-aminopelargonic 79
acid synthase) condenses the activated form of pimelic acid (pimeloyl-ACP) 80
with L-alanine. Second, BioA (7, 8-diaminopelargonic acid synthase) catalyzes 81
the generation of 7, 8-diaminopelargonic acid that is followed by the BioD 82
(dethiobiotin synthase) mediated formation of ureido ring in this molecule. 83
Finally, BioB (biotin synthase) converts dethiobiotin into biotin. In contrast, the 84
microorganisms that only possesses an incomplete biotin synthesis pathway 85
(e.g., probiotic bacteria Lactococcus and human pathogen Streptococcus 86
species) seem likely to have evolved an alternative mechanism of biotin 87
scavenging to fulfill their metabolic requirements (2, 3, 11) (Fig. 1A). Energy 88
coupling factor (ECF)-type transporters have been identified for vitamin uptake 89
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
4
in prokaryotes (12-14). This type of tripartite ECF-type transporter contains 90
three elements: S component (a membrane-embedded, substrate-binding 91
protein), A/A’ component (an energy-coupling module that comprises two 92
ATP-binding proteins) and T component (a trans-membrane protein) (14-16). 93
Recently, Hebbeln et al. (11) reported biochemical evidence that Rhodobacter 94
capsulatus ECF transporter (BioMNY) and the single S component, bioY gene 95
product can both actively function in biotin uptake. 96
Most of our current knowledge on regulation of bacterial biotin 97
metabolism comes from studies with E. coli (1, 17). Biotin protein ligase (BPL) 98
is referred to as a universal enzyme widespread throughout all forms of life 99
which covalently attaches biotin to its cognate acceptor proteins such as the 100
AccB subunit (also called biotin carboxyl carrier protein, BCCP) of acetyl-CoA 101
carboxylase (ACC), which catalyzes the first committed step of fatty acid 102
biosynthesis (17). The E. coli birA gene product is a model BPL that has 103
undergone extensive investigations ranging from genetics (18-20), 104
biochemistry (21-25), to structural biology/biophysics (21, 26, 27). Unlike the 105
Group I BPL (also called mono-functional BirA) that lacks N-terminal DNA 106
binding domain (1, 28), the E. coli BirA is an unusual bi-functional BPL in that it 107
also acts as a repressor for the biotin operon (1, 17, 29). The fact that E. coli 108
BirA enzyme possesses these two divergent functions allows it to 109
physiologically sense the intracellular levels of both biotin and un-biotinylated 110
biotin accepting protein BCCP (1, 18, 30). The ligand of BirA repressor is 111
biotinoyl-5’-AMP (also called biotinyl-adenylate) that is the intermediate of 112
BirA-catalyzed ligation (31). Upon excess biotin being present, the 113
biotinoyl-5’-AMP at high levels promotes accumulation of a ligand-bound form 114
of BirA, the functional repressor complexes at the bio operator by triggering 115
BirA dimerization, which consequently leads to the repression of bio operon (1). 116
Similarly, a high level of ligand-bound BirA complex can accumulate when 117
AccC is overproduced, because that tightly ties up the apo-form of AccB in a 118
complex with the poor biotinylation substrate (1, 29). In contrast, the 119
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
5
transcription of bio operon can be de-repressed once the concentration of 120
liganded-BirA is significantly decreased (1, 17). In fact, this kind of situation 121
might be easily triggered by either an inhibition of biotinoyl-5’-AMP formation 122
by intracellular biotin limitation or by increased consumption of 123
biotinyl-adenylate due to overproduction of un-modified acceptor protein AccB 124
(1, 29). Earlier analyses of comparative genomics characterized a GntR-type 125
transcription factor BioR and its recognition signals (referred to as BioR box) 126
upstream of biotin-related genes in a group of α-proteobacteria like A. 127
tumefaciens and it was proposed that BioR was a new regulator of bacterial 128
biotin sensing, but required further experimental demonstration (32). Very 129
recently, we reported that the plant pathogen A. tumefaciens has evolved a 130
new mechanism, a two-protein paradigm of BirA and BioR to sense the 131
demand of biotin, validating the above hypothesis of Rodionov and coworkers 132
(33). 133
Brucellosis is an endemic disease with an estimated 500,000 cases 134
globally each year, and pathogenic species of Brucella are the causative 135
agents of such kind of zoonotic infectious disease (34). According to the 136
difference of their phenotypes and host habitats, the genus of Brucella have 137
been classified into 10 known species plus one new strain Brucella sp. NVSL 138
07-0026 (35-38) (http://en.wikipedia.org/wiki/Brucella). Among them, most of 139
Brucella genomes have been decoded (39). In addition to the six 140
earlier-recognized species, B. melitensis (sheep and goats), B. abortus (cattle), 141
B. canis (dogs), B. neomoate (dessert wood rats), B. ovis (sheep), and B. suis 142
(pigs), they are five newly-added species, namely B. ceti from cetaceans, B. 143
microti from voles, B. pinnipediae from pinnipeds, B. inopinata with unknown 144
host and Brucella sp. NVSL 07-0026 from baboon 145
(http://en.wikipedia.org/wiki/Brucella) (39). Although bacterial biotin 146
metabolism can be regarded as a promising/potential antibacterial drug target 147
(40), the regulatory mechanism in these pathogenic Brucella species still 148
remains poorly known. Surprisingly, our genomic sequence-based analyses of 149
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
6
multiple Brucella species revealed that their bio operon-encoding sequences 150
(bioBFDAZ and bioY) and their promoter regulatory regions were almost 100% 151
identical. This extreme conservation was also observed in the bioR orthologs. 152
Unlike the scenario seen with A. tumefaciens that features only one BioR box 153
located in bioB locus (Fig. 1B) (32, 33), four putative BioR palindromes in total 154
were detected in each of Brucella species (both bioR and bioY possesses one 155
site in the promoter region, whereas the bioBFDAZ operon contains two 156
tandem BioR boxes), indicating a complicated regulatory network underlying 157
their biotin sensing machineries. 158
In this paper, we are the first to report a complex regulatory network of 159
biotin metabolism found throughout all the species of Brucella. Using a LacZ 160
reporter system that we developed on the basis of some engineered A. 161
tumefaciens strains, we demonstrated that BioR_bme acts as a functional 162
repressor for bio operon transcription (Fig. 1C). Not only does BioR_bme 163
negatively auto-regulate itself, it also exerts repression on the expression of 164
bioY that encodes a major player involved in biotin scavenging from the host 165
and/or its inhabiting environment.166
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
7
Materials and Methods 167
Bacterial strains and growth conditions 168
E. coli strains are K-12 derivatives (Table 1) and grown at 37oC. A. 169
tumefaciens used here were all engineered lacZ-fusion reporter strains that we 170
had recently developed (Table 1) (33). The three kinds of media used for the 171
maintenance of both E. coli and A. tumefaciens include Luria-Bertani (LB) 172
medium (tryptone, 10g/L; yeast extract, 5 g/L; NaCl, 10 g/L; pH 7.5), rich broth 173
(RB medium; 10 g of tryptone, 1 g of yeast extract, and 5 g of NaCl per liter), 174
and a defined M9 minimal medium supplemented with 0.4% glucose or 175
another carbon source, 0.1% Vitamin-Free Casamino Acids and 0.001% 176
thiamine. Mannitol-Glutamate/Luria Medium (MG/L medium, 5 g of Mannitol, 177
1.16 g of Monosodium glutamate, 12.5 g of LB broth, 0.1 g of MgSO4 and 1 mg 178
biotin per liter, pH 7.0) was utilized to prepare the competent cells of A. 179
tumefaciens. The strain B. melitensis 16M was maintained in BD™ Trypic Soy 180
Broth medium (TSB, 17 g of tryptone (pancreatic digest of casein), 3 g of 181
peptone (soybean digest), 2.5 g of glucose, 5 g of NaCl and 2.5 K2HPO4 per 182
liter, pH7.2). Both A. tumefaciens and B. melitensis were cultivated at 30oC. If 183
required, antibiotics were added as follows (in mg/liter): sodium ampicillin, 100 184
for E. coli; kanamycin sulfate, 25 for E. coli and 50 for A. tumefaciens; 185
gentamycin sulfate, 50 for E. coli and 30 for A. tumafaciens and spectinomycin, 186
100 for both E. coli and A. tumefaciens. 187
188
Plasmids and molecular techniques 189
B. melintensis bioR gene was synthesized in vitro here using an 190
overlap-PCR method. First, we amplified four pieces of overlapping DNA 191
fragments using four sets of combined oligo-nucleotide primers (Table 3): 1) 192
bioRBME-F1, bioRBME-F2, bioRBME-F3 plus bioRBME-F4(r); 2) bioRBME-F4, 193
bioRBME-F5, bioRBME-F6, bioRBME-F7, bioRBME-F8 plus bioRBME-F9(r); 3) 194
bioRBME-F9, bioRBME-F10, bioRBME-F11, bioRBME-F12, bioRBME-F13 plus 195
bioRBME-F14(r); 4) bioRBME-F14, bioRBME-F15, bioRBME-F16, bioRBME-R2 plus 196
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
8
bioRBME-R1. The resulting four PCR products were mixed in an equal ratio and 197
used as the template for subsequent rounds of PCR reaction in which the 198
specific primers bioRBME-F plus bioRBME-R1 were used (Table 1). The 199
resultant PCR product with expected size of ~0.7 kb was purified and 200
subjected to direct DNA sequencing. Finally, the bioRbme gene was cloned into 201
pET28a expression vector via BamHI and XhoI cuts, giving the recombinant 202
plasmid pET28-bioRbme (Table 1). The IPTG-inducible, tightly regulated 203
expression vector pSRKGm (33, 41) was used for functional assays in A. 204
tumefaciens. The bioRbme was amplified with the primers of bioRBME-CF plus 205
bioRBME-CR (Table 2) and inserted into pSRKGm via the sites of NdeI and 206
NheI, giving pSRK-bioRbme (Table 1). There promoter sequences of bioBbme 207
(396 bp), bioYbme (372 bp) and bioRbme (377 bp) were synthesized by 208
Integrated DNA Technology (IDT) and encoded by the following three plasmids 209
pIDT-PbioBbme, pIDT-PbioYbme and pIDT-PbioRbme (Table 1). Subsequently, 210
the promoter regions were cut from the pIDT vector with SmaI and BamHI, and 211
cloned into pRG970, a low-copy lacZ reporter vector (42, 43) via the same cuts, 212
yielding pRG-PbioBbme, pRG-PbioYbme, and pRG-PbioRbme, respectively 213
(Table 1). All the required recombinant plasmids were verified by either PCR 214
detection, digestion of restriction enzymes, and direct DNA sequencing. 215
216
Expression, purification and characterization of BioR proteins 217
The two versions of bacterial BioR proteins (referred to as BioR_bme 218
and BioR_at) were produced in E. coli BL21 (DE3) carrying the appropriate 219
expression plasmids (pET28a-bioRbme and pET28a-bioRat, Table 1) and 220
purified as described recently (33) with minor changes. The purified 221
recombinant BioR_bme with 6x His-tag at N-terminus was separated by12% 222
SDS-PAGE and then confirmed using Western blotting assay with anti-6x His 223
primary antibody. The identity of BioR_bme was determined by liquid 224
chromatography quadrupole time-of-flight mass spectrometry of tryptic 225
peptides (44, 45). 226
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
9
Electrophoretic mobility shift assays 227
Gel shift assays were performed to test the binding of the BioR_bme to 228
its cognate DNA sequences of B. melitensis as described by us earlier with 229
minor changes (33, 45-47). The digoxigenin-labeled DNA probes (5 in total) 230
were synthesized in vitro by annealing two complementary primers (Table 2) in 231
TEN buffer (10 mM Tris-HCl, 1 mM EDTA, 100 mM NaCl, pH 8.0) plus the 232
terminal transferase-aided labeling with DIG-ddUTP (Roche). Native PAGE 233
(8%) was utilized to detect the DNA-protein complex formed by an incubation 234
of the DIG-labeled DNA probes (0.2 pmol) with/without BioR_bme (or BioR_at) 235
in binding buffer (Roche) at room temperature for 20 min. Finally, signal 236
capture was done by exposure to ECL film (Amersham) (44, 46). 237
238
Biotin bioassay-based cross-feeding 239
To visualize the effect on biotin production exerted by bioRbme 240
expression in A. tumefaciens, we designed a biotin bioassay-based 241
cross-feeding experiment in which the biotin auxotrophic strain of E. coli, ER90 242
(ΔbioF bioC bioD) is cross-fed by A. tumefaciens species. The biotin-free M9 243
minimal medium plates were made as previously described (6, 33, 48), 244
containing 0.01% (w/v) the redox indicator 2, 3, 5-triphenyl tetrazolium chloride 245
(TTC). Of particular note, prior to pouring the plates, the indicator ER90 strain 246
was suspended into the melted agar medium (~50 oC) at the final bacterial 247
concentration (optical density at 600 nM (OD600) is around 0.05, i.e., ~5.0x107 248
CFU/ml). 249
The four strains of A. tumefaciens used here included NTL4 (WT), 250
FYJ283 (ΔbioBFDA), FYJ212 (ΔbioRat) and FYJ341 (ΔbioR::Km+bioRbme). 251
Except that the biotin auxotroph strain FYJ283 was maintained in 5 ml of M9 252
medium supplemented with 1 nM biotin, all the other three strains were 253
cultivated in biotin-free M9 minimal media of 5 ml overnight. Subsequently, 254
these overnight cultures were collected by centrifugation (3000 rpm, 10min), 255
washed three times using the M9 liquid medium, and transferred into 100 ml of 256
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
10
biotin-free M9 media for 6 more hours of growth at 30oC to exhaust trace 257
amount of intracellular biotin in the biotin auxotroph strain FYJ283. Following 258
three rounds of washing with same media, all the bacteria were suspended in 259
M9 media and their optical densities at 600 nM were adjusted to 1.5. 20 ul of A. 260
tumefaciens culture (OD600=1.0) was spotted on the paper disc, and 261
maintained at 30 oC for overnight. A red deposit of formazan suggested that the 262
indicator strain ER90 is fed by the A. tumefaciens strains, and the “red” area of 263
the growth circle (square centimeters, cm2) represents the level of biotin 264
produced by the different feeder strains. 265
266
Genetic manipulations 267
Three plasmid-borne lacZ transcriptional fusions (pRG-PbioBbme, 268
pRG-PbioYbme and pRG-PbioRbme, in Table 1) were separately 269
electroporated into the double mutant strain of A. tumefaciens, FYJ284 270
(ΔbioR::Km, ΔbioBFDA), giving reporter strains FYJ319 (ΔbioR::Km, 271
pRG-PbioBbme), FYJ321 (ΔbioR::Km, pRG-PbioYbme) and FYJ344 272
(ΔbioR::Km, pRG-PbioRbme), respectively (Table 1). In addition to the 273
well-established reporter strain FYJ291 that carries the plasmid-borne 274
PbioBAT-lacZ transcriptional fusion (33) (Table 1), we introduced the low-copy 275
expression plasmid (pSRK-bioRbme) into the above three reporter strains for 276
function analyses of Brucella bioR. 277
As we earlier reported (49) with little modifications, we deleted bioR 278
(BMEI0320) from B. melitensis 16M using the strategy of homologous 279
recombination. Both multiplex-PCR assays and direct DNA sequencing of PCR 280
products were employed to confirm the acquired ΔbioR mutant of B. melitensis 281
(Table 1). 282
283
RNA isolation and Real-time quantitative RT-PCR 284
Bacterial total RNAs were isolated from the log-phase cells of B. 285
melitensis 16M and its bioR isogenic mutant grown in TSB medium (OD600 is 286
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
11
around 1.0), using the RNeasy bacterial RNA isolation kit (Qiagen). The 287
validated RNA samples in which contaminated DNA is not detected by general 288
PCR with primers 16S-F plus 16S-R (Table 2) were subjected to SYBR® Green 289
single-step real-time reverse transcription (RT)-PCR experiments. 16S rDNA 290
acted as an internal reference gene, and other five genes of interest included 291
bioB, bioF, bioD, bioA and bioY (Table 2). The relative expression levels were 292
determined using the 2-ΔΔct method reported by Livak et al. (50). 293
294
β-Galactosidase assays 295
Two different methods were adopted here to assay the β-galactosidase 296
activity. Firstly, the engineered A. tumefaciens reporter strains containing 297
appropriate plasmids (Table 1 & 2) were inoculated onto MacConkey agar 298
plates with 0.4% lactose as sole carbon source (Thermo Scientific) to initially 299
visualize the differences in their LacZ activity. Secondly, we subjected the 300
bacterial lysates sampled from log-phase of culture to a treatment with sodium 301
dodecyl sulfate-chloroform (51) to quantify their β-galactosidase activities in a 302
manner similar to before (48). 303
304
Bioinformatic analyses 305
The multiple alignments of either BioR protein or BioR-binding sites 306
were conducted using the program of ClustalW2 307
(http://www.ebi.ac.uk/Tools/clustalw2/index.html), and final output was 308
processed by the server of ESPript 2.2 309
(http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi). Transcription start site was 310
predicted using the method of Neutral Network Promoter Prediction 311
(http://www.fruitfly.org/seq_tools/promoter.html).312
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
12
Results 313
Characterization of B. melitensis BioR 314
B. melitensis, a member of α-proteobacteria, contains two circular 315
chromosomes: Ch I (Accession no.: AE008917.1) is 2,117,144 bp long, while 316
Ch II (Accession no.: AE008918.1) is 1,177,787 bp in length (52). A gene 317
orthologous to bioR (BMEI0320) is located on Chromosome I and encodes a 318
230 amino acid of polypeptide that is four residues longer than that of A. 319
tumefaciens BioR (Fig. 2A). Sequence alignment of these two BioR proteins 320
(BioR_bme and BioR_at) showed that they share 76.5% identity and 65.5% 321
similarity, respectively (Fig. 2A). As predicted by Rodionov and coworker (32), 322
these two proteins have a conserved N-terminal DNA-binding motif with 323
helix-turn-helix structure (Fig. 2A and F). To test its putative function, we 324
over-expressed the recombinant BioR_bme protein in E. coli and purified it to 325
homogeneity (Fig. 2B). Although the prevalent form (~90%) of the purified 326
BioR_bme protein when loaded on SDS-PAGE occurs at the position of 327
~26kDa (which is consistent with estimated molecular weight of its monomer), 328
a small amount of protein (~10%) is consistently present at the position of ~52 329
kDa, implying that BioR_bme can from a dimer in solution. To rule out the 330
possibility for the impurities in the BioR_bme sample, we carried out Western 331
blotting analyses with anti-6X His tag primary antibody to address this issue. 332
As expected, both forms of protein were recombinant forms of the protein with 333
6X His tag (Fig. 2C). Using the other approach, chemical crosslinking assay, 334
we also proved that the dimer form of BioR_bme protein is appreciably 335
increased upon addition of chemical crosslinker EGS, which is similar to our 336
observation with A. tumefaciens BioR (Fig. 2D) (33). Finally, liquid 337
chromatography mass spectrometry of tryptic peptides demonstrated that the 338
two protein bands cut from the gel indeed come from the same B. melitensis 339
BioR protein with corresponding coverage score of 81% for the monomer form 340
of 26kDa (not shown), and 80% for the form of ~52kDa (Fig. 2E). It can be 341
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
13
concluded that BioR_bme can exhibit solution structure of dimer, which is in 342
consistent with scenario seen with BioR_at (33). 343
344
B. melitensis bioR homologue is functional in vivo 345
We recently engineered a reporter strain FYJ291 that is a double 346
mutant of A. tumefaciens (ΔbioR ΔbioBFDA) carrying the low-copy 347
plasmid-borne bioBat-lacZ transcriptional fusion (33). This reporter strain can 348
indicate whether the bioR orthologue is functional or not. Upon growth on a 349
MacConkey agar plate with 0.4% lactose as a sole carbon source, the reporter 350
strain FYJ291 gave purple colonies, indicating that the bioBat-lacZ fusion has 351
strong β-gal activity upon removal of BioR_at (Fig. 3A). The introduction of 352
BioR_at into this reporter strain resulted in the formation of yellow colonies, 353
implying that extremely low β-gal activity of bioBat-lacZ is due to efficient 354
repression by the expression of BioR_at (Fig. 3A). In general agreement with 355
an observation with BioR_at, functional complementation of FYJ291 indicator 356
strain with plasmid-borne bioRbme also gave the similar phenotype of yellow 357
colonies (Fig. 3A). Assays for LacZ activities further revealed that expressions 358
of both bioRbme and bioRat leads to a 8 to 10-fold decrement of the bioBat 359
transcription level relative to that of FYJ291 reporter strain with de-repression 360
of BioR (Fig. 3B). Therefore, bioRbme is believed to be a functional 361
orthologous gene in vivo. It seems quite likely that BioR palindrome of A. 362
tumefaciens bioB can be bound by both BioR_at and BioR_bme. 363
364
B. melitensis BioR binds cognate palindromes 365
In the chromosome I of B. melItensis 16M (Accession no.: NC_003317), 366
the biotin transporter locus bioY (BMEI0319) that encodes a 191 residues of 367
polypeptide neighbors the gene bioR (BMEI0320). Each of them has a 368
predictive BioR-binding site (Fig. 4A and B) (32). By contrast, in its 369
closely-relative organism A. tumefaciens, these two genes do not have any 370
detectable BioR-binding sites in front of their coding sequences (32, 33). In 371
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
14
Chromosome II of B. melitensis 16M (NC_003318), the genes constituting of 372
biotin biosynthetic pathway (BMEII0775, BMEII0776, BMEII0777, BMEII0778 373
and BMEII0779) are organized into the bioBFDAZ operon (Fig. 4A). Unlike the 374
scenario that only one BioR palindrome (TTATCTATAA) is determined to be in 375
the very beginning of the bioB coding sequence from A. tumefaciens 376
bioBFDAZ operon, there are two discontinuous BioR-recognized sites 377
localized upstream of the translation start site of B. melitensis bioBFDAZ 378
operon (Fig. 4A and C). 379
Systematic bioinformatics analyses by The Neutral Network Program of 380
Promoter Prediction (http://www.fruitfly.org/seq_tools/promoter.html) 381
suggested that 1) the predicted transcription start site of bioR is a “T” 40 bp 382
upstream of the “ATG” translation initiation site and the putative BioR binding 383
site (TTATCTATAA) centered at the position 31 (Fig. 4C); the transcription of 384
bioY gene begins at a “T” 17 bp that is upstream of the “ATG” translation start 385
site, and that is also nearly in the center of the predicted BioR palindrome 386
(TTATCTATAA)(Fig. 4C); the bioBFDAZ operon can be transcribed in the 387
beginning of an “A” (37 bp upstream of “ATG” the translation start site) that 388
happens to separate the two BioR sites (site 1: TTATCTATTA and site 2: 389
TTATCTACAA) (Fig. 4C). Of note, the above four candidate 390
BioR-recognizable sites required further experimental validation and the 391
diversity in their positions also raised the possibility that a complex regulatory 392
network for biotin metabolism might exist. 393
Gel shift assays confirmed that BioR_bme can efficiently bind A. 394
tumefaciens bioB promoter (Fig. 5A and B), which generally validates our 395
above observation that BioR_bme represses the expression of A. tumefaciens 396
bioB in vivo (Fig. 3). The bioRbme gene’s own BioR-binding site was also 397
demonstrated to be functional by in vitro EMSA test (Fig. 5C and D). 398
Considering the fact that this BioR-binding site is located downstream of the 399
predicted transcription start site (Fig. 4C), we hypothesized that BioR_bme 400
might be an auto-repressor. Of particular note, our EMSA results proved the 401
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
15
earlier prediction by Rodionov and coauthors (32) is correct in this case of B. 402
melitensis bioBFDAZ operon (Fig. 4), i.e., the two tandem putative BioR 403
palindromes in the promoter region of this operon both exhibit abilities to 404
interact with BioR_bme protein (Fig. 5E-H). This is somewhat unexpected, but 405
not without precedent. A scenario similar to what we have encountered is that 406
two functional FadR-binding sites are present in the promoters of fadL and 407
fadD, of which the protein products are constitute the long chain fatty acids 408
(LCFA) transporter system in E. coli (48). We thereby speculated that the two 409
sites determined this operon under negative regulation by BioR in B. melitensis, 410
but this hypothesis required further in vivo evidence. Additionally, the promoter 411
region of the bioYbme gene that encodes the S component of ECF (energy 412
coupling factor)-type biotin transporter (11, 13, 14) was visualized to bind 413
BioR_bme protein in the in vitro assay (Fig. 5I and J). This might represent the 414
first example of the BioR-regulated transport/scavenge of biotin in bacteria. 415
To probe possible physiological ligands/effectors for BioR_bme binding, 416
we systematically tested the precursor (pimeloyl-ACP), intermediates (KAPA, 417
DAPA and DTB), and final product (biotin) of biotin synthesis pathway (Fig. 6A) 418
by employing EMSA approach. In much consistency with our recent 419
observation with A. tumefaciens BioR, these biotin-related metabolites 420
seemed to not interfere DNA-binding activity of BioR_bme even addition of 421
excessive metabolites (such as 500 pmol KAPA, DAPA, DTB & biotin) (Fig. 6B 422
and C). Also, we observed that excess of cold bioRbme DNA probe can 423
efficiently/competitively impaired interaction between DIG-labeled bioRbme 424
DNA probe and BioR_bme protein (Fig. 6C), validating that such kind of 425
DNA-protein binding is a specific physical interaction. 426
427
Complex regulation of biotin metabolism by BioR in B. melitensis 428
We applied two different approaches to dissect the in vivo role of B. 429
melitensis bioR in biotin metabolism. One is an assay for LacZ activity in 430
reporter strains, and the other is a qPCR-based comparison of the bioRbme 431
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
16
isogenic mutant with its parental strain of B. melitensis 16M. 432
First, we constructed three versions of plasmid-borne transcriptional 433
lacZ fusions (PbioBbme-lacZ, PbioYbme-lacZ and PbioRbme-lacZ) and then 434
introduced them into the engineered strains of A. tumefaciens (ΔbioR 435
ΔbioBFDA) that we recently developed, giving three reporter strains FYJ319 436
(ΔbioRat plus bioBbme-lacZ), FYJ321 (ΔbioRat plus bioYbme-lacZ) and FYJ344 437
(ΔbioRat plus bioRbme-lacZ), respectively (Table 1 and Fig. 7). When grown 438
on MacConkey agar plates with 0.4% lactose as sole carbon source, all the 439
three reporter strains exhibited a similar phenotype of purple colonies (Fig. 7A, 440
C and E), indicating this A. tumefaciens-based reporter system works well for 441
our purpose. Upon in trans complementation of pSRKGm-borne bioRbme 442
gene separately into the above three reporter strains, all the colonies with 443
yellow color were consistently observed to grow on the MacConkey indicator 444
plates (Fig. 7A, C and E). Such a dramatic change in colonial color clearly 445
illustrated the in vivo effect of repression by BioR_bme on these target genes. 446
In particular note, bioRbme seemed to be negatively auto-regulated, which 447
contrasts its counterpart in A. tumefaciens (32, 33). Subsequent analyses of 448
β-gal activities revealed that 1) over-expression of bioRbme gave a 8 to 12-fold 449
decrease of bioBFDAZ operon expression (Fig. 7B) , which is generally 450
consistent with our recent observation with bioRat (33); 2) the expression level 451
of bioYbme transporter gene was reduced 3 to 5-fold in the presence of 452
bioRbme expression (Fig. 7D); 3) the amplitude for auto-repression of BioRbme 453
itself was around 4 to 6-fold (Fig. 7F). 454
Second, we carried out qPCR assays for further addressing the 455
accumulated transcript level of the representative target genes (bioB, bioF, 456
bioD, bioA and bioY) in the bioR disrupted mutant of B. melitensis 16M. The 457
expression level of bio operon in the ΔbioRbme mutant was increased 2 to 458
3-fold relative to the wild type strain 16M (Fig. 7G). Also, removal of the bioR 459
gene from B. melitensis increased the bioY transcription to nearly 5-fold (Fig. 460
7G). In general agreement with the data from LacZ assays of reporter strains 461
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
17
(Fig. 7), real-time PCR analyses validated that BioR_bme functions as a 462
repressor in vivo. Together, we concluded that it is very different from the 463
scenario in A. tumefaciens and that a complex regulation network of biotin 464
metabolism by BioR exists in B. melitensis, i.e., Not only does BioR act as an 465
auto-repressor that negatively modulates bio operon of biotin biosynthetic 466
pathway, it also represses bioY transporter system (Fig.1). Although the bioY 467
gene had been proposed for years (2, 53), its regulated expression was very 468
recently determined in Corynebacteria, i.e., it can be repressed by a TetR-like 469
transcription factor BioQ (54). Therefore, our observation with bioY of B. 470
melitensis is in generally consistency with the scenario in Corynebacteria. 471
472
Physiological relevance of BioR regulation to biotin production 473
To gain a glimpse of the physiological consequence BioRbme-mediated 474
regulation, we established a cross-feeding system in which the feeder strains 475
are genetically modified A. tumefaciens species that are supposed to have 476
different ability to produce various levels of biotin (such as FYJ341, whose 477
bioRat is inactivated from chromosome, but carrying plasmid-borne bioRbme), 478
and recipient strain is biotin auxotrophic strain of E. coli, ER90 (ΔbioF bioC 479
bioD) (Table 1). The growth medium we used here is biotin-free M9 minimal 480
medium plates supplementing 0.1% (w/v) the redox indicator 2, 3, 5-triphenyl 481
tetrazolium chloride (TTC) (6, 33, 48). Since it is known that the wild type strain 482
NTL4 of A. tumefaciens can produce high levels of biotin, most of which are 483
secreted out of cells (33), the ER90 indicator strain can be cross-fed featured 484
by a circle of a red deposit of formazan (Fig. 8). In contrast, the strain FYJ283 485
(ΔbioBFDA) acts as negative control and can’t support ER90 growth, given 486
that both are biotin auxotrophic strains (Fig. 8). In relation to the wild type 487
strain NTL4, the ΔbioRat mutant FYJ212 seemed to produce appreciably more 488
biotin, and in turn trigger better growth of the indicator strain ER90 exhibiting 489
bigger colony size (Fig. 8). As we expected, the ER90 strain grew poorly upon 490
expression of plasmid-borne bioRbme in the feeder strain FYJ212, indicating 491
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
18
that BioRbme efficiently exerts negative regulation on biotin biosynthetic 492
pathway (Fig. 8). To the best of our knowledge, this is first report of 493
physiological relevance of BioR-mediated repression in bacterial biotin 494
production.495
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
19
Discussion 496
BioR, a class of GntR-type transcription factor, is exclusively restricted 497
to most of α-proteobacteria (32), seems to compensate for the loss of 498
regulatory function of BirA, a mono-functional biotin protein ligase (33). We 499
favored a two protein model of BirA and BioR that might represent an 500
alternative mechanism for bacterial biotin sensing. Given the fact that bioR 501
duplication exists (e.g., Paracoccus denitrificans, Fig. 9B and the number of 502
BioR boxes varies greatly in different species (Fig. 9A and B), we believe that 503
BioR-mediated regulation is definitely complex and diverse. We also noted that 504
the BioR signal (BioR-binding palindromes) seems to be very conserved in the 505
α-proteobacteria we examined (Fig. 9C and Table 4), indicating it might 506
represent a common regulatory mechanism present in these organisms. Given 507
the fact that all the four Brucella BioR signals (including bioRbme probe in Fig. 508
10A and B, bioYbme probe in Fig. 10C and D, bioBbme probe 1&2 in Fig. 509
10E-H) can efficiently bind to A. tumefaciens homologue of BioR, we are more 510
confident to believe that this type of diversified regulatory mechanism is 511
mediated by a relatively conserved DNA-protein interaction (Fig. 10). 512
Here, we prove that Brucella BioR is a functional member using our 513
newly-engineered A. tumefaciens strain-based reporter system, suggesting 514
that this approach serves as a useful tool with potent implications for functional 515
assays of other bioR homologues. An unanswered question about this 516
mechanism lies in the physiological ligand of BioR, which still remains 517
enigmatic. We had no success in identifying if even after three years of testing 518
a series of metabolite intermediates of biotin biosynthesis (Fig. 6) as well as 519
probing for the possible chemical modification of BioR by biotin (33). We 520
anticipated that dissecting crystal structures of BioR protein alone and its 521
complex with DNA target might be helpful in obtaining clues regarding this 522
question in the future. However the unusual performance of this wield BioR 523
protein in vitro (Not only does it easily precipitate on Superdex 200 column 524
during the process of gel filtration, but also quickly deposits on the Millipore 525
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
20
centrifugal filter units upon being concentrated) probably hampers the 526
structure-based approach to this question (not shown). Recently, a TetR-type 527
transcription factor BioQ was reported to regulate biotin metabolism in 528
Corynebacterium glutamicum (54), which generally supports our hypothesis, 529
i.e. diversified mechanisms have been evolving/developing for different 530
bacteria to sense fluctuant biotin demands upon occupying varying natural 531
reservoirs or host environments. 532
Brucella, the genus of Gram-negative facultative intracellular bacteria, 533
consists of a group of heterogeneous populations with 10 classified species. 534
Although the fact that the sequenced genomes of most Brucella species (9 out 535
of 10 in total) vary in size, encoding 3,200 to 3500 open reading frames 536
(ORFs), the BioR regulator and its regulated-bio operon plus the BioR binding 537
sites are extremely similar (not shown), indicating that it is a conserved 538
regulatory mechanism is widespread in Brucella. Although the annotated locus 539
of bioC (BMEI0182) that presumably encodes a putative O-methyltransferase, 540
is present in Chromosome I of Brucella, no BioR-binding palindrome is 541
detected (32), suggesting that earlier step of biotin synthesis is not the 542
rate-limited step controlled by BioR repressor. Intriguingly, two copies of bioY 543
transporter genes (BMEI1431 and BMEI0319) with 52.7% similarity and 33.3% 544
identity appeared on the same Chromosome I, however only the latter 545
BMEI0319 adjacent to the locus bioR (BMEI0320) evolved to possess a 546
BioR-recognizable site ahead of its coding sequence (Fig. 4A and C). We tend 547
to believe that its physiological advantage for this biotin scavenge machinery 548
(consisting of one regulated BioY transporter and one more unregulated one) 549
is that of ensuring that the regulation of biotin uptake is not tightly-controlled, in 550
case it is encountering/inhabiting a host mileu with limited biotin availability. 551
Retrospectively, such kind of regulated expression of biotin transporter bioY 552
somewhat is similar to scenario observed with longh chain fatty acid 553
transporter fadL and fadD expression (48). The fact that two BioR boxes are 554
present in front of the bioBFDAZ gene cluster encoding protein products 555
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
21
responsible for formation of double rings in biotin molecule allows bacteria to 556
finely/effectively modulate late steps of biotin synthesis, which is generally 557
consistent with the scenario in E. coli, the paradigm organism (17, 55). 558
Differing from the simplified regulatory network of BioR in A. tumefaciens (33), 559
a functional dissection of Brucella BioR revealed quite a bit of the complex 560
regulatory architecture of biotin metabolism (biotin transport system and biotin 561
biosynthetic pathway), which might be a selective/adaptive consequence of 562
the long-term co-evolution of Brucella species to their unique inhabiting 563
environment.564
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
22
Acknowledgements 565
This work was supported by National Institutes of Health (NIH) Grant 566
AI15650 from National Institute of Allergy and Infectious Diseases (NIAID), 567
The National Natural Science Foundation of China (81071320, 31000041 & 568
81171530) and National Basic Research Program of China (Grant No. 569
2009CB522600). We are grateful to three anonymous reviewers for a series of 570
constructive suggestions to improve this manuscript, and we would like to 571
thank Dr. Peter Yau (from the Biotechnology Center, University of Illinois at 572
Urbana-Champaign) for technical assistance in Q-TOF. 573
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
23
References 574
1. Beckett D. 2007. Biotin sensing: universal influence of biotin status on transcription. Annu 575
Rev Genet 41:443-464. 576
2. Rodionov DA, Mironov AA, Gelfand MS. 2002. Conservation of the biotin regulon and the 577
BirA regulatory signal in Eubacteria and Archaea. Genome Res 12:1507-1516. 578
3. Guillen-Navarro K, Encarnacion S, Dunn MF. 2005. Biotin biosynthesis, transport and 579
utilization in Rhizobia. FEMS Microbiol Lett 246:159-165. 580
4. Streit WR, Entcheva P. 2003. Biotin in microbes, the genes involved in its biosynthesis, its 581
biochemical role and perspectives for biotechnological production. Appl Microbiol Biotechnol 582
61:21-31. 583
5. Beckett D. 2009. Biotin sensing at the molecular level. J Nutr 139:167-170. 584
6. Lin S, Hanson RE, Cronan JE. 2010. Biotin synthesis begins by hijacking the fatty acid 585
synthetic pathway. Nat Chem Biol 6:682-688. 586
7. Lin S, Cronan JE. 2011. Closing in on complete pathways of biotin biosynthesis. Mol 587
Biosyst 7:1811-1821. 588
8. Cronan JE, Lin S. 2011. Synthesis of the alpha,omega-dicarboxylic acid precursor of biotin 589
by the canonical fatty acid biosynthetic pathway. Curr Opin Chem Biol 15:407-413. 590
9. Lin S, Cronan JE. 2012. The BioC O-methyltransferase catalyzes methyl esterification of 591
malonyl-acyl carrier protein, an essential step in biotin synthesis. J Biol Chem 592
287:37010-37020. 593
10. Agarwal V, Lin S, Lukk T, Nair SK, Cronan JE. 2012. Structure of the enzyme-acyl carrier 594
protein (ACP) substrate gatekeeper complex required for biotin synthesis. Proc Natl Acad Sci 595
U S A 109:17406-17411. 596
11. Hebbeln P, Rodionov DA, Alfandega A, Eitinger T. 2007. Biotin uptake in prokaryotes by 597
solute transporters with an optional ATP-binding cassette-containing module. Proc Natl Acad 598
Sci U S A 104:2909-2914. 599
12. Erkens GB, Berntsson RP, Fulyani F, Majsnerowska M, Vujicic-Zagar A, Ter Beek J, 600
Poolman B, Slotboom DJ. 2011. The structural basis of modularity in ECF-type ABC 601
transporters. Nat Struct Mol Biol 18:755-760. 602
13. Zhang P, Wang J, Shi Y. 2010. Structure and mechanism of the S component of a bacterial 603
ECF transporter. Nature 468:717-720. 604
14. Rodionov DA, Hebbeln P, Eudes A, ter Beek J, Rodionova IA, Erkens GB, Slotboom DJ, 605
Gelfand MS, Osterman AL, Hanson AD, Eitinger T. 2009. A novel class of modular 606
transporters for vitamins in prokaryotes. J Bacteriol 191:42-51. 607
15. Fisher DJ, Fernandez RE, Adams NE, Maurelli AT. 2012. Uptake of biotin by Chlamydia 608
Spp. through the use of a bacterial transporter (BioY) and a host-cell transporter (SMVT). 609
PLoS One 7:e46052. 610
16. Berntsson RP, ter Beek J, Majsnerowska M, Duurkens RH, Puri P, Poolman B, Slotboom 611
DJ. 2012. Structural divergence of paralogous S components from ECF-type ABC transporters. 612
Proc Natl Acad Sci U S A 109:13990-13995. 613
17. Chakravartty V, Cronan JE. 2012. Altered regulation of Escherichia coli biotin biosynthesis 614
in birA superrepressor mutant strains. J Bacteriol 194:1113-1126. 615
18. Cronan JE, Jr. 1989. The E. coli bio operon: transcriptional repression by an essential protein 616
modification enzyme. Cell 58:427-429. 617
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
24
19. Barker DF, Campbell AM. 1981. The birA gene of Escherichia coli encodes a biotin 618
holoenzyme synthetase. J Mol Biol 146:451-467. 619
20. Barker DF, Campbell AM. 1980. Use of bio-lac fusion strains to study regulation of biotin 620
biosynthesis in Escherichia coli. J Bacteriol 143:789-800. 621
21. Brown PH, Cronan JE, Grotli M, Beckett D. 2004. The biotin repressor: modulation of 622
allostery by corepressor analogs. J Mol Biol 337:857-869. 623
22. Choi-Rhee E, Schulman H, Cronan JE. 2004. Promiscuous protein biotinylation by 624
Escherichia coli biotin protein ligase. Protein Sci 13:3043-3050. 625
23. Chapman-Smith A, Mulhern TD, Whelan F, Cronan JE, Jr., Wallace JC. 2001. The 626
C-terminal domain of biotin protein ligase from E. coli is required for catalytic activity. 627
Protein Sci 10:2608-2617. 628
24. Barker DF, Campbell AM. 1981. Genetic and biochemical characterization of the birA gene 629
and its product: evidence for a direct role of biotin holoenzyme synthetase in repression of the 630
biotin operon in Escherichia coli. J Mol Biol 146:469-492. 631
25. Streaker ED, Gupta A, Beckett D. 2002. The biotin repressor: thermodynamic coupling of 632
corepressor binding, protein assembly, and sequence-specific DNA binding. Biochemistry 633
41:14263-14271. 634
26. Weaver LH, Kwon K, Beckett D, Matthews BW. 2001. Corepressor-induced organization 635
and assembly of the biotin repressor: a model for allosteric activation of a transcriptional 636
regulator. Proc Natl Acad Sci U S A 98:6045-6050. 637
27. Wilson KP, Shewchuk LM, Brennan RG, Otsuka AJ, Matthews BW. 1992. Escherichia 638
coli biotin holoenzyme synthetase/bio repressor crystal structure delineates the biotin- and 639
DNA-binding domains. Proc Natl Acad Sci U S A 89:9257-9261. 640
28. Purushothaman S, Gupta G, Srivastava R, Ramu VG, Surolia A. 2008. Ligand specificity 641
of group I biotin protein ligase of Mycobacterium tuberculosis. PLoS One 3:e2320. 642
29. Abdel-Hamid AM, Cronan JE. 2007. Coordinate expression of the acetyl coenzyme A 643
carboxylase genes, accB and accC, is necessary for normal regulation of biotin synthesis in 644
Escherichia coli. J Bacteriol 189:369-376. 645
30. Beckett D. 2005. The Escherichia coli biotin regulatory system: a transcriptional switch. J 646
Nutr Biochem 16:411-415. 647
31. Ke Y, Yuan X, Wang Y, Bai Y, Xu J, Song H, Huang L, Chen Z. 2012. Genome sequences 648
of Brucella melitensis 16M and its two derivatives 16M1w and 16M13w, which evolved in 649
vivo. J Bacteriol 194:5489. 650
32. Rodionov DA, Gelfand MS. 2006. Computational identification of BioR, a transcriptional 651
regulator of biotin metabolism in Alphaproteobacteria, and of its binding signal. FEMS 652
Microbiol Lett 255:102-107. 653
33. Feng Y, Zhang H, Cronan JE. 2013. Profligate biotin synthesis in α-Proteobacteria - A 654
develoing or degenerating regulatory system? Mol Microbiol 88:77-92. 655
34. Franco MP, Mulder M, Gilman RH, Smits HL. 2007. Human brucellosis. Lancet Infect Dis 656
7:775-786. 657
35. O'Callaghan D, Whatmore AM. 2011. Brucella genomics as we enter the multi-genome era. 658
Brief Funct Genomics 10:334-341. 659
36. Ficht T. 2010. Brucella taxonomy and evolution. Future Microbiol 5:859-866. 660
37. Moreno E, Cloeckaert A, Moriyon I. 2002. Brucella evolution and taxonomy. Vet Microbiol 661
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
25
90:209-227. 662
38. Corbel MJ. 1997. Brucellosis: an overview. Emerg Infect Dis 3:213-221. 663
39. Wattam AR, Williams KP, Snyder EE, Almeida NF, Jr., Shukla M, Dickerman AW, 664
Crasta OR, Kenyon R, Lu J, Shallom JM, Yoo H, Ficht TA, Tsolis RM, Munk C, Tapia R, 665
Han CS, Detter JC, Bruce D, Brettin TS, Sobral BW, Boyle SM, Setubal JC. 2009. 666
Analysis of ten Brucella genomes reveals evidence for horizontal gene transfer despite a 667
preferred intracellular lifestyle. J Bacteriol 191:3569-3579. 668
40. Woong Park S, Klotzsche M, Wilson DJ, Boshoff HI, Eoh H, Manjunatha U, Blumenthal 669
A, Rhee K, Barry CE, 3rd, Aldrich CC, Ehrt S, Schnappinger D. 2011. Evaluating the 670
sensitivity of Mycobacterium tuberculosis to biotin deprivation using regulated gene 671
expression. PLoS Pathog 7:e1002264. 672
41. Khan SR, Gaines J, Roop RM, 2nd, Farrand SK. 2008. Broad-host-range expression 673
vectors with tightly regulated promoters and their use to examine the influence of TraR and 674
TraM expression on Ti plasmid quorum sensing. Appl Environ Microbiol 74:5053-5062. 675
42. van Dillewijn P, Soto MJ, Villadas PJ, Toro N. 2001. Construction and environmental 676
release of a Sinorhizobium meliloti strain genetically modified to be more competitive for 677
alfalfa nodulation. Appl Environ Microbiol 67:3860-3865. 678
43. Van den Eede G, Deblaere R, Goethals K, Van Montagu M, Holsters M. 1992. Broad host 679
range and promoter selection vectors for bacteria that interact with plants. Mol Plant Microbe 680
Interact 5:228-234. 681
44. Feng Y, Cronan JE. 2011. Complex binding of the FabR repressor of bacterial unsaturated 682
fatty acid biosynthesis to its cognate promoters. Mol Microbiol 80:195-218. 683
45. Feng Y, Cronan JE. 2011. The Vibrio cholerae fatty acid regulatory protein, FadR, represses 684
transcription of plsB, the gene encoding the first enzyme of membrane phospholipid 685
biosynthesis. Mol Microbiol 81:1020-1033. 686
46. Feng Y, Cronan JE. 2010. Overlapping repressor binding sites result in additive regulation of 687
Escherichia coli FadH by FadR and ArcA. J Bacteriol 192:4289-4299. 688
47. Feng Y, Cronan JE. 2009. A new member of the Escherichia coli fad regulon: transcriptional 689
regulation of fadM (ybaW). J Bacteriol 191:6320-6328. 690
48. Feng Y, Cronan JE. 2012. Crosstalk of Escherichia coli FadR with global regulators in 691
expression of fatty acid transport genes. PLoS One 7:e46275. 692
49. Wang Y, Chen Z, Qiao F, Ying T, Yuan J, Zhong Z, Zhou L, Du X, Wang Z, Zhao J, Dong 693
S, Jia L, Yuan X, Yang R, Sun Y, Huang L. 2009. Comparative proteomics analyses reveal 694
the virB of B. melitensis affects expression of intracellular survival related proteins. PLoS One 695
4:e5368. 696
50. Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression data using real-time 697
quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402-408. 698
51. Miller J. 1972. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. . 699
52. DelVecchio VG, Kapatral V, Redkar RJ, Patra G, Mujer C, Los T, Ivanova N, Anderson I, 700
Bhattacharyya A, Lykidis A, Reznik G, Jablonski L, Larsen N, D'Souza M, Bernal A, 701
Mazur M, Goltsman E, Selkov E, Elzer PH, Hagius S, O'Callaghan D, Letesson JJ, 702
Haselkorn R, Kyrpides N, Overbeek R. 2002. The genome sequence of the facultative 703
intracellular pathogen Brucella melitensis. Proc Natl Acad Sci U S A 99:443-448. 704
53. Gloeckler R, Ohsawa I, Speck D, Ledoux C, Bernard S, Zinsius M, Villeval D, Kisou T, 705
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
26
Kamogawa K, Lemoine Y. 1990. Cloning and characterization of the Bacillus sphaericus 706
genes controlling the bioconversion of pimelate into dethiobiotin. Gene 87:63-70. 707
54. Brune I, Gotker S, Schneider J, Rodionov DA, Tauch A. 2012. Negative transcriptional 708
control of biotin metabolism genes by the TetR-type regulator BioQ in biotin-auxotrophic 709
Corynebacterium glutamicum ATCC 13032. J Biotechnol 159:225-234. 710
55. Solbiati J, Cronan JE. 2010. The switch regulating transcription of the Escherichia coli 711
biotin operon does not require extensive protein-protein interactions. Chem Biol 17:11-17. 712
56. Luo ZQ, Clemente TE, Farrand SK. 2001. Construction of a derivative of Agrobacterium 713
tumefaciens C58 that does not mutate to tetracycline resistance. Mol Plant Microbe Interact 714
14:98-103. 715
57. Goodner B, Hinkle G, Gattung S, Miller N, Blanchard M, Qurollo B, Goldman BS, Cao Y, 716
Askenazi M, Halling C, Mullin L, Houmiel K, Gordon J, Vaudin M, Iartchouk O, Epp A, 717
Liu F, Wollam C, Allinger M, Doughty D, Scott C, Lappas C, Markelz B, Flanagan C, 718
Crowell C, Gurson J, Lomo C, Sear C, Strub G, Cielo C, Slater S. 2001. Genome sequence 719
of the plant pathogen and biotechnology agent Agrobacterium tumefaciens C58. Science 720
294:2323-2328. 721
58. Wood DW, Setubal JC, Kaul R, Monks DE, Kitajima JP, Okura VK, Zhou Y, Chen L, 722
Wood GE, Almeida NF, Jr., Woo L, Chen Y, Paulsen IT, Eisen JA, Karp PD, Bovee D, Sr., 723
Chapman P, Clendenning J, Deatherage G, Gillet W, Grant C, Kutyavin T, Levy R, Li 724
MJ, McClelland E, Palmieri A, Raymond C, Rouse G, Saenphimmachak C, Wu Z, 725
Romero P, Gordon D, Zhang S, Yoo H, Tao Y, Biddle P, Jung M, Krespan W, Perry M, 726
Gordon-Kamm B, Liao L, Kim S, Hendrick C, Zhao ZY, Dolan M, Chumley F, Tingey 727
SV, Tomb JF, Gordon MP, Olson MV, Nester EW. 2001. The genome of the natural genetic 728
engineer Agrobacterium tumefaciens C58. Science 294:2317-2323. 729
59. Choi-Rhee E, Cronan JE. 2005. A nucleosidase required for in vivo function of the 730
S-adenosyl-L-methionine radical enzyme, biotin synthase. Chem Biol 12:589-593. 731
60. Choi-Rhee E, Cronan JE. 2005. Biotin synthase is catalytic in vivo, but catalysis engenders 732
destruction of the protein. Chem Biol 12:461-468. 733
61. Novichkov PS, Rodionov DA, Stavrovskaya ED, Novichkova ES, Kazakov AE, Gelfand 734
MS, Arkin AP, Mironov AA, Dubchak I. 2010. RegPredict: an integrated system for regulon 735
inference in prokaryotes by comparative genomics approach. Nucleic Acids Res 736
38:W299-307. 737
62. Novichkov PS, Laikova ON, Novichkova ES, Gelfand MS, Arkin AP, Dubchak I, 738
Rodionov DA. 2010. RegPrecise: a database of curated genomic inferences of transcriptional 739
regulatory interactions in prokaryotes. Nucleic Acids Res 38:D111-118. 740
741
742
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
27
Tables 743
Table 1 Bacterial strains and plasmids used in this study 744
Bacteria or plasmids
Relevant characteristics Origins
Bacterial strains NTL4 A derivative of A. tumefaciens C58
mutated to be sensitive of tetracycline
(56-58)
16M A virulent strain of B. melitensis Chen’s stock ΔbioR-16M The isogenic bioR mutant of B.
melitensis 16M Chen’s stock
Topo10 F-, ΔlacX74, a cloning host for recombinant plasmids
Invitrogen, (46, 47)
DH5α The E. coli host for DNA cloning Lab stock BL21(DE3) The engineered E. coli host for
protein expression Lab stock, (46, 47)
DH10b The E. coli host for large plasmid and BAC cloning
New England Biolabs (NEB)
ER90 MG1655, ΔbioF::cat ΔbioC ΔbioD (6, 59, 60) FYJ178 BL21(DE3) carrying 28a-bioRat This work FYJ212 A. tumefaciens NTL4, ΔbioR::Km This work FYJ218 BL21(DE3) carrying 28a-bioRbme This work FYJ283 A. tumefaciens NTL4, ΔbioBFDA This work FYJ284 FYJ212, ΔbioR::Km, ΔbioBFDA This work FYJ288 FYJ283 (NTL4, ΔbioBFDA) carrying
pRG970 This work
FYJ289 FYJ284 (NTL4, ΔbioR::Km, ΔbioBFDA) carrying pRG970
This work
FYJ290 FYJ283 (NTL4, ΔbioBFDA) carrying pRG-PbioBat
This work
FYJ291 FYJ284 (NTL4, ΔbioR::Km, ΔbioBFDA) carrying pRG-PbioBat
This work
FYJ292 FYJ291 (NTL4, ΔbioR::Km, ΔbioBFDA) carrying two plasmids pRG-PbioBat plus pSRKGm
This work
FYJ293 FYJ291 (NTL4, ΔbioR::Km, ΔbioBFDA) carrying two plasmids pRG-PbioBat plus pSRK-bioRat
This work
FYJ307 Topo10 carrying pSRK-bioRbme This work FYJ308 FYJ291 (NTL4, ΔbioR::Km,
ΔbioBFDA) carrying two plasmids pRG-PbioBat plus pSRK-bioRbme
This work
FYJ310 DH5α carrying pIDT-PbioBbme This work FYJ311 DH5α carrying pIDT-PbioYbme This work
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
28
FYJ312 DH5α carrying pIDT-PbioRbme This work FYJ313 DH10b carrying pRG-PbioBbme This work FYJ314 DH10b carrying pRG-PbioYbme This work FYJ315 DH10b carrying pRG-PbioRbme This work FYJ319 FYJ284 (NTL4, ΔbioR::km,
ΔbioBFDA) carrying pRG-PbioBbme This work
FYJ321 FYJ284 (NTL4, ΔbioR::km, ΔbioBFDA) carrying pRG-PbioYbme
This work
FYJ337 FYJ319 carrying pSRKGm-bioRbme This work FYJ339 FYJ321 carrying pSRKGm-bioRbme This work FYJ341 FYJ212 (NTL4, ΔbioR::Km) carrying
pSRKGm-bioRbme This work
FYJ344 FYJ284 (NTL4, ΔbioR::km, ΔbioBFDA) carrying pRG-PbioRbme
This work
FYJ346 FYJ344 carrying pSRKGm-bioRbme This work Plasmids pET28(a) Commercial T7-driven expression
vector, KmR Novagen
pSRKGm Broad-host-range expression vector with the tightly regulated promoter
(41)
pRG970 Low copy transcriptional promoter-less lacZ/Gus bi-directional fusion vector, SpcR
(42, 43)
28a-bioRat pET28(a) carrying A. tumefaciens bioR gene, KmR
This work
28a-bioRbme pET28(a) carrying B. melitensis bioR gene, KmR
This work
pSRK-bioRat pSRKGm encoding BioR_at, GmR This work pSRK-bioRbme pSRKGm encoding BioR_bme, GmR This work pRG-PbioBat pRG970 carrying A. tumefaciens
bioB promoter region, SpcR This work
pIDT-PbioBbme pIDT carrying B. melitensis bioB promoter sequence, AmpR
Integrated DNA Technologies (IDT)
pIDT-PbioYbme pIDT carrying B. melitensis bioY promoter sequence, AmpR
IDT
pIDT-PbioRbme pIDT carrying B. melitensis bioR promoter sequence, AmpR
IDT
pRG-PbioBbme pRG970 carrying B. melitensis bioB promoter sequence, SpcR
This work
pRG-PbioYbme pRG970 carrying B. melitensis bioY promoter sequence, SpcR
This work
pRG-PbioRbme pRG970 carrying B. melitensis bioR promoter sequence, SpcR
This work
745
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
29
Table 2 DNA primers used in this study 746
Primers Primer sequences BioRbioB_at site-Fb 5'-CTC TCT TGA GGA GGC AAA AAT TAT CTA TAA
TTT GCC ATT TAA CGA CCT GC-3’ BioRbioB_at site-Rb 5'-GCA GGT CGT TAA ATG GCA AAT TAT AGA TAA
TTT TTG CCT CCT CAA GAG AG-3’ BioRbioR_BME site-Fb 5'-GAC CGC CGG GGC AAA GAA CAT TAT CTA
TAA AAC CAT GAA TCA GAA TGT CC-3’ BioRbioR_BME site-Rb 5'-GGA CAT TCT GAT TCA TGG TTT TAT AGA TAA
TGT TCT TTG CCC CGG CGG TC-3’ BioRbioY_BME site-Fb 5'-GAA TAG ATC GAG TCT CGA TTT TAT CTA TAA
TTT GAG GAA ACC ATG GCG AC-3’ BioRbioY_BME site-Rb 5'-GTC GCC ATG GTT TCC TCA AAT TAT AGA TAA
AAT CGA GAC TCG ATC TAT TC-3’ BioRbioB_BME site1-Fb 5'-CAC AAC TTC CCC CAT CAA AAT TAT CTA TTA
TAT TAT CAT TTG TGG ATT CA-3’ BioRbioB_BME site1-Rb 5'-TGA ATC CAC AAA TGA TAA TAT AAT AGA TAA
TTT TGA TGG GGG AAG TTG TG-3’ BioRbioB_BME site2-Fb 5'-TTT GCA GAT TGA TTC TGT TTT TAT CTA CAA
TTT GGA GGA AGA ATG CCC TG-3’ BioRbioB_BME site2-Rb 5'-CAG GGC ATT CTT CCT CCA AAT TGT AGA TAA
AAA CAG AAT CAA TCT GCA AA-3’ bioRBME-F (BamHI) 5’-CG GGATCC ATG AAT CAG AAT GTC CCA
GCC-3' bioRBME-R1 (XhoI) 5’-CCG CTCGAG CTA CCC CAC AAT GGC GAA
GGA AT-3’ bioRBME-CF (NdeI) 5'-GGAATTC CATATG ATG AAT CAG AAT GTC CCA
GCC-3' bioRBME-CR (NheI) 5'-CTA GCTAGC CTA CCC CAC AAT GGC GAA G-3' PbioBBME-check 5'-CTG GAG CAG TTT CGC TTA AC-3' PbioYBME-check 5'-ACT GCG ACA AAA CGA TAT TG-3' PbioRBME-check 5'-CAC GCA TGA ATG GAA ACA GG-3' lacZ-rev 5'-GAC CAT TTT CAA TCC GCA-3’ 16S-F (371 bp) 5’-GTG GAA TTC CGA GTG TAG AGG-3’ 16S-R 5’-GTC CAG CCT AAC TGA AGG ATA G-3’ bioY-F (43-64) 5’-GTC TCT TCC CAG ATC GAA GTT C-3’ bioY-R (373-393) 5’-CAC GCT TGA AAG CCA TAG TGC-3’ bioB-F (14-34) 5’-GTG GAA AAG CAC GAG AAA CCG-3’ bioB-R (294-314) 5’-CCA TGA GTT TGG AGG CTT TCA-3’ bioF-F (1-21) 5’-GTG AAA CTC GAC ACC TAC CTG-3’ bioF-R (294-314) 5’-GTG GAA AGT GCT GCC AGA TTG-3’ bioD-F (120-140) 5’-GAA GAA ACC GAC AGC GAG ATC-3’ bioD-R (396-416): 5’-TGA TTG ATG GTG CCA AGG GCT-3’ bioA-F (205-225) 5’-GAT CTG GAC CAG ATC ATC TTC-3’
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
30
bioA-R (536-556) 5’-CTT CAA GAC GGT CGA GCA TAG-3’ a The underlined italic sequences are the introduced restriction sites. 747 b The bold letters are predicted core palindromes for BioR binding. 748
749
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
31
Table 3 A list of oligo DNA designed to in vitro synthesize B.melitensis bioR 750
gene 751
Oligos Sequences of oligo DNAsb bioRBME-F1 (1-52)a CG GGATCC ATG AAT CAG AAT GTC CCA
GCC TCC CGG CCC GCT CCC CGG AGC GAA ACA ATT G
bioRBME-F2 (32-91)a CTC CCC GGA GCG AAA CAA TTG CCG CCC GCA TCA GCC GCA TTC TCG CGG ATC GTA TCA TTG
bioRBME-F3 (71-130)a TTC TCG CGG ATC GTA TCA TTG CGG GCG AGA TCG AGC CCG GCA CCA AAC TGC GCC AGG ATC
bioRBME-F4 (111-170)a CAC CAA ACT GCG CCA GGA TCA TAT TGC CGA GGA ATT CCA GAC CAG CCA TGT GCC GGT GCG
bioRBME-F5 (151-210)a ACC AGC CAT GTG CCG GTG CGT GAA GCC TTC CGG CGG CTG GAG GCA CAG GGC CTC GCC GTT
bioRBME-F6 (191-250)a AGG CAC AGG GCC TCG CCG TTT CCG AAC CGC GGC GCG GCG TAC GCG TTG CCT CCT TCG ACA
bioRBME-F7 (231-290)a ACG CGT TGC CTC CTT CGA CAT TGG CGA AAT TCG CGA AGT GGC CGA AAT GCG CGC CGC GCT
bioRBME-F8 (271-330)a GCC GAA ATG CGC GCC GCG CTT GAG GTG CTT GCA CTG CGC CAT GCG GCC CCC CAC ATC ACC
bioRBME-F9 (311-370)a ATG CGG CCC CCC ACA TCA CCC GTG CCG TGC TGG ATG CCG CCG AAC AGG CCA CGC TGG AGG
bioRBME-F10 (351-410)a CGA ACA GGC CAC GCT GGA GGG CGA CAA GTC CCG CGA TGT GCG CAG TTG GGA AGA TGC GAA
bioRBME-F11 (391-450)a CGC AGT TGG GAA GAT GCG AAC CGG CGC TTC CAC CGT CTC ATT CTC ACC CCC TGC AAG ATG
bioRBME-F12 (431-490)a TTC TCA CCC CCT GCA AGA TGC CGC GCC TGC TCG CCG CCA TCG ACG ATC TTC ATG CGG CAA
bioRBME-F13 (471-530)a CGA CGA TCT TCA TGC GGC AAG CGC CCG TTT TCT CTT CGC CAC CTG GCG CTC GGC ATG GGA
bioRBME-F14 (511-570)a ACC TGG CGC TCG GCA TGG GAA GCA CGC ACC GAC CAC GAC CAC CGC GCA ATC CTC
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
32
GCC GCA bioRBME-F15 (551-610)a ACC GCG CAA TCC TCG CCG CAC TGC GAC
AAA ACG ATA TTG AAA GTG CGG CCA CCA TTC TCG
bioRBME-F16 (591-650)a AAG TGC GGC CAC CAT TCT CGC CCG CCA TGT GCA ATG GAT CGG CCA TCG CCC GGT CAA GAC
bioRBME-R2 (631-690)a CCC CAC AAT GGC GAA GGA ATC GCG CAC CTT TCC CGA AGC CGT CTT GAC CGG GCG ATG GCC
bioRBME-R1 (671-693)a CCG CTCGAG CTA CCC CAC AAT GGC GAA GGA AT
bioRBME-F4(r) CGC ACC GGC ACA TGG CTG GTC TGG AAT TCC TCG GCA ATA TGA TCC TGG CGC AGT TTG GTG
bioRBME-F9(r)
CCT CCA GCG TGG CCT GTT CGG CGG CAT CCA GCA CGG CAC GGG TGA TGT GGG GGG CCG CAT
bioRBME-F14(r) TGC GGC GAG GAT TGC GCG GTG GTC GTG GTC GGT GCG TGC TTC CCA TGC CGA GCG CCA GGT
a the numbers in bracket denote the position of overlapping PCR primers in 752
relative to the coding sequence of bioRBME. 753
b the underlined italic sequences are the introduced restriction sites. 754
755
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
33
Table 4 A collection of the predicted BioR-binding sites from α-proteobacteria 756
Organisms Genetic loci Sequences Positions Scores Rhizobiales Bradyrhizobium
japonicum blr2095 CTATAGATAA -46 4.39 blr2095 TTATCTACAA -27 4.32 bll2094 TTGTAGATAA -97 4.32 bll2094 TTATCTATAG -78 4.39
Bradyrhizobium
sp. BTAi1 BBta_2775 TAATCTATAA -83 4.23 BBta_2776 TTATAGATTA -49 4.23
Azorhizobium
caulinodans AZC_0363 TTATCTATAA -13 4.83 AZC_0362 TTATCTATAA 28 4.83
B. melitensis BMEI0319 TTATCTATAA -22 4.83 BMEI0320 TTATCTATAA -14 4.83 BMEII0775 TTATCTATTA -92 4.23 BMEII0775 TTATCTACAA -22 4.32
Sinorhizobium
fredii NGR234 NGR_c25140 CTATAGATAA -60 4.39
A. tumefaciens Atu3997 TTATCTATAA 85 4.83 Mesorhizobium
loti mlr7428 TTATCTATAA -13 4.83
Rhodobacterales Rhodobacter sphaeroides
RSP_1924 TTATCTATAA -203 4.83 RSP_1924 TTATAGATAG -159 4.39
Paracoccus denitrificans
Pden_2916 TAATAGATAA -80 4.23 Pden_2916 TTATAGATAC -37 4.08 Pden_1432 TTATCTATAA -84 4.83 Pden_1432 TTATAGATAG -40 4.39
Silicibacter pomeroyi
SPO3339 TTATAGATAG -64 4.39 SPO3339 TTATCTATAA -23 4.83
Sulfitobacter sp. EE-36
EE36_13898 TTATCTATAA -83 4.83
Rhodobacter sphaeroides
RSP_1925 TTATCTATAA -13 4.83
Silicibacter pomeroyi
SPO3340 TTATCTATAA -13 4.83
Sulfitobacter sp. EE-36
EE36_13903 CTATCTATAA -61 4.39
Silicibacter sp. TM1040
TM1040_3661 TTATAGATAA -94 4.83
BioR regulon was analyzed using Regpredict software (61) in 2 subdivisions of 757
α-proteobacteria (Rhizobiales and Rhodobacterales), which also can be in 758
much details accessed by logging Regprecise database (62). Position here is 759
relative to the translation start site, and the score is measured using the 760
recognition profile (position weight matrix). 761
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
34
Figure Legends 762
Fig. 1 A working model proposed for bacterial biotin metabolism and its 763
regulatory mechanism by BioR 764
A. Schematic diagram for bacterial biotin de novo biosynthetic pathway and its 765
alternative scavenging route 766
B. BioR represses biotin biosynthesis pathway in A. tumefaciens 767
C. Negative auto-regulation of BioR and its repression of both biotin 768
biosynthesis pathway and biotin transport system in Brucella 769
770
Fig. 2 Characterization of Brucella BioR 771
A. Sequence comparison of BioR protein from Brucella melitensis and A. 772
tumefaciens 773
The multiple alignment was conducted using ClustalW2 774
(http://www.ebi.ac.uk/Tools/clustalw2/index.html), and the final output was 775
given via data processing by program ESPript 2.2 776
(http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi). Identical residues are in 777
white letters with red background, similar residues are in black letters with 778
yellow background, varied residues are in black letters, and dots represent 779
gaps. The predicted secondary structure was shown in top. α: α-helix; β: 780
β-sheet; T: β-turns/coils. 781
782
B. SDS-PAGE profile of the purified Brucella BioR protein 783
784
C. Western blot analyses of the purified BioR protein using Anti-6xHis primary 785
antibody 786
The expected size of BioR protein in monomer is around 26 kDa, whereas the 787
dimeric size is about 52 kDa. The protein sample was separated with 4-20% 788
gradient Mini-PROTEAN@ TGXTM Gel (Bio-Rad). 789
790
D. Chemical cross-linking assays for the solution structure/state of BioR 791
protein 792
The minus sign decodes no addition of the chemical cross-linker EGS, 793
whereas the plus sign indicates addition of 20μM EGS. The cross-linking 794
reaction mixtures (20 μl in total) were loaded on the gradient SDS-PAGE as 795
above. 796
797
E. MS identification the recombinant Brucella BioR protein 798
The peptide fragments matching database sequences are given in bold and 799
under-lined type. 800
801
F. Modeled structure of Brucella BioR protein 802
Structure modeling was proceded by the software of SPDBV_4.01 using a 803
GntR regulator with known structure (PDB: 3C7J) of P. syringae pv., tomato str. 804
DC3000 (Accession no., AAO58874) as structural template. N: N-terminus, C: 805
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
35
C-terminus. 806
807
Fig. 3 Evidence that Brucella bioR is functional in vivo 808
A. Expression of Brucella bioR homologue represses transcription of bioBFDA 809
operon in A. tumefaciens 810
To visualize bioB-lacZ expression, we used MacConkey agar plate with 0.4% 811
lactose as a sole carbon source. The bacteria were maintained at 30 oC for 812
around 36 hours. Purple indicates strong β-gal activity, whereas yellow 813
denotes no/low β-gal activity. 814
815
B. Use of β-gal activity assay to test regulation of bioBFDA operon expression 816
by two BioR orthologues (one is from A. tumefaciens, and the other is from 817
Brucella) 818
Mid-log phase cultures in RB media were sampled for assays of β–gal activity. 819
The data is expressed in Average ± standard deviation (SD), and error bars 820
indicate SD. More than five independent experiments were performed. 821
The three strains included FYJ291 (ΔbioRat), FYJ293 (ΔbioRat + ΔbioRat), 822
and FYJ308 (ΔbioRat + ΔbioRbme), respectively. All the strains used here 823
carry the PbioBat-lacZ transcriptional fusion (Table 1). 0.3 mM IPTG was used 824
to induce expression of plasmid-borne bioRat (and/or bioRbme). 825
826
Fig. 4 BioR signals and promoters of three biotin-related loci, bioR, bioY and 827
bioBFDA operon of Brucella 828
A. Schematic diagram for bioR and its regulated genes 829
The oval symbol denotes BioR binding site 830
831
B. Multiple sequence alignment of the predicted BioR binding sites from a 832
collection of α-proteobacteria and the resulting sequence logo 833
In the top panel, the identical residues are white letters in red background, 834
similar residues are black letters in yellow background, and varied residues are 835
in black letters. In the bottom panel, the sequence logo is generated using 836
WebLogo (http://weblogo.berkeley.edu/logo.cgi). Designations: pd, 837
Paracoccus denitrificans; Rsph: Rhodobacter sphaeroides; bme: Brucella 838
melitensis; bj: Bradyrhizobium japonicum; at: Agrobacterium tumefaciens; ml: 839
Mesorhizobium loti; Ssp: Silicibacter sp. TM1040. 840
841
C. The promoters of bioR, bioBFDA operon and bioY in Brucella 842
The predicted BioR site is given in red and underlined letter, and the possible 843
ribosome binding site (RBS) is shown in purple and underlined type. The 844
anticipated -10 and -35 regions are underlined in yellow. Abbreviations: S 845
denotes transcription initiation site, and M denotes translation start site. 846
847
Fig. 5 Binding of Brucella BioR to promoters of series of target genes 848
A and B. Binding of the bioRat probe by various concentrations of Brucella 849
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
36
BioR. 850
C and D. Binding of the bioRbme probe by various concentrations of Brucella 851
BioR. 852
Binding of both bioRbme probe 1 (E and F) and probe 2 (G and H) by various 853
concentrations of Brucella BioR 854
I and J. Binding of bioYbme probe by various concentrations of Brucella BioR 855
The minus sign denotes no addition of protein. The protein levels of BioR (in 856
the right hand four lanes of each panel (left to right)) were 2, 4, 10 and 20 pmol. 857
The protein samples were incubated with 0.2 pmol of DIG-labeled probe in a 858
total volume of 15 µl. A representative result is given. The data were derived 859
from three independent gel shift assays (7.5% native PAGE). The abbreviation 860
of at and bme denotes A. tumefaciens and B. melitensis, respectively. 861
862
Fig. 6 Evaluation of possible effects of Brucella BioR binding DNA by 863
biotin-related metabolites 864
A. Schematic diagram for the four-step pathway of bacterial biotin biosynthesis 865
866
B. EMSA-based visualization for effects on DNA-BioR interplay excerted by 867
four kinds of biotin metabolites (Pimeloyl-ACP, KAPA, DAPA and DTB) 868
869
C. Binding of BioR_bme to the target DIG-labeled DNA probe is not affected by 870
the presence of biotin (or DTB), but can be fully/specifically impaired by the 871
excessive presence of the corresponding cold DNA probe 872
873
The signs of minus and plus separately denotes no addition and addition of 874
BioR_bme protein (10-20 pmol). The protein samples were incubated with 875
0.25 pmol of DIG-labeled bioRbme probe in a total volume of 20 µl. When 876
necessary, the cold probe (bioRbme probe, 15 (or 25) pmol) is supplemented. 877
The gel shift experiments are conducted with 7% native PAGE. A 878
representative photograph is given. Designations: KAPA, 879
7-keto-8-aminopelargonic acid; DAPA, 7, 8-diaminopelargonic acid; DTB, 880
dethiobiotin; bme, B. melitensis. 881
882
Fig. 7 In vivo evidence for complex regulation of biotin sensing in Brucella 883
A. MacConkey agar plate-based visualization of β-gal activity for altered 884
expression of Brucella BioBFDAZ operon in A. tumefaciens with/without the 885
presence of Brucella BioR 886
887
B. Assays for β-gal activity of bioBbme-lacZ transcriptional fusion in the 888
ΔbioRat mutant of A. tumefaciens in relative to the complemented strain 889
carrying plasmid-borne bioRbme 890
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
37
Two strains used here included FYJ319 (ΔbioRat) and FYJ337 (ΔbioRat+ 891
bioRbme) (Table 1). 892
893
C. MacConkey agar plate-based differentiation for altered expression level of 894
Brucella BioY in A. tumefaciens with/without the presence of Brucella BioR 895
896
D. Assays for β-gal activity of bioYbme-lacZ transcriptional fusion in the 897
ΔbioRat mutant of A. tumefaciens in comparison with the complemented strain 898
Two strains assayed here are FYJ321 (ΔbioRat) and FYJ339 899
(ΔbioRat+bioRbme), respectively (Table 1). 900
901
E. MacConkey agar plate-based observation revealed that expression of 902
Brucella bioR was repressed by its own protein product 903
904
F. Comparative analyses of β-gal activity of bioRbme-lacZ fusion in the 905
ΔbioRat mutant and the complemented strain 906
Two strains assayed here are FYJ344 (ΔbioRat) and FYJ346 907
(ΔbioRat+bioRbme), respectively (Table 1). 0.3 mM IPTG was used to induce 908
expression of plasmid-borne bioRbme. Mid-log phase cultures in RB media 909
were sampled for assays of β–gal activity. The data is expressed in Average ± 910
standard deviation (SD), and error bars indicate SD. More than five 911
independent experiments were performed. 912
913
G. Real-time qPCR-based visualization for effect of bioBFDA and bioY 914
expression exerted by Brucella BioR 915
Log-phase culture of B. melitensis 16M grown in RSB media were subjected to 916
total RNA isolation. In the real-time qPCR experiment, each gene was assayed 917
in triplicate. A representative result is given here. 918
919
Fig. 8 Physiological relevance of Brucella BioR-mediated repression to biotin 920
biosyntheis of A. tumefaciens 921
922
Four A. tumefaciens strains that are used to cross-feed E. coli strain ER90 923
lacking full bio operon included NTL4 (WT), FYJ283 (ΔbioBFDA), FYJ212 924
(ΔbioRat), and FYJ 341 (ΔbioRat+bioRbme), respectively. 925
926
The agar plates of M9 minimal medium without supplementing any vitamin are 927
prepared routinely with an exception of adding the biotin indicator strain ER90 928
(an engineered E. coli (ΔbioF::cat ΔbioC ΔbioD) whose growth is supported by 929
exogenously-feeding of biotin). 20 ul of A. tumefaciens strain (mid-log phase 930
culture whose optical density OD600 is adjusted to around 1.2) is spotted on 931
the paper disc, and maintained at 30oC for overnight. Red circles suggest that 932
indicator strain ER90 is fed by the A. tumefaciens strains, and the “red” area of 933
the growth circles (square centimeters, cm2) might represent the level of biotin 934
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from
38
produced by the different feeder strains. 935
936
Fig. 9 Unexpected diversity of BioR regulator, bio operon and its BioR signals 937
in α-proteobacteria 938
A. Diversified organization of bioR regulator, bio operon and its BioR signals in 939
Rhizobials 940
941
B. Complexity of genetic localizations of bioR regulator, bio operon and its 942
BioR signals in Rhodobacterales 943
Except that panD encodes an aspartate α-decarboxylase (seen in 944
Bradyrhizobium japonicum in Panel A), all the other genes listed above are 945
involved in biotin metabolism. Designations: B, bioB; F, bioF; D, bioD; A, bioA; 946
Z, bioZ; R, bioR; Y, bioY; M, bioM; N, bioN. 947
948
C. Sequence logo for the conservative BioR signals 949
BioR regulon from 2 subdivisions of α-proteobacteria (Rhizobiales and 950
Rhodobacterales) sampled from Regprecise database (62) was systematically 951
analyzed using Regpredict software (61). 952
All the proposed BioR-binding palindromes of were listed in Table 4 were 953
analyzed through WebLogo (http://weblogo.berkeley.edu/logo.cgi), giving the 954
sequence logo. 955
956
Fig. 10 A. tumefaciens BioR is functionally equivalent to Brucella BioR in that it 957
can bind to the promoter regions of all the four target genes 958
A and B. Binding of the bioRbme probe by various concentrations of A. 959
tumefaciens BioR 960
961
C and D. Binding of the bioYbme probe by various concentrations of A. 962
tumefaciens BioR 963
964
Binding of both bioRbme probe 1 (E and F) and probe 2 (G and H) by various 965
concentrations of A. tumefaciens BioR 966
967
The minus sign denotes no addition of BioR protein. The protein levels of 968
BioR_at (in the right hand six lanes of each panel (left to right)) were 0.05, 0.1, 969
0.25, 0.5, 1 and 2 pmol. The protein samples were incubated with 0.2 pmol of 970
DIG-labeled probe in a total volume of 15 µl. A representative result is given. 971
The data were derived from three independent gel shift assays. The 972
abbreviation of at and bme denotes A. tumefaciens and B. melitensis, 973
respectively. 974
on August 17, 2019 by guest
http://jb.asm.org/
Dow
nloaded from