Biotin Synthesis in Ralstonia eutropha H16 Utilizes Pimeloyl Coenzyme A and Can Be Regulated by the Amount of Acceptor Protein

Ralstonia eutropha is applied in industry for the production of biopolymers and serves as a research platform for the production of diverse fine chemicals. Due to its ability to grow on hydrogen and carbon dioxide as the sole carbon and energy source, R. eutropha is often utilized for metabolic engineering to convert inexpensive resources into value-added products. The understanding of the metabolic pathways in this bacterium is mandatory for further bioengineering of the strain and for the development of new strategies for biotechnological production.

ABSTRACT

The biotin metabolism of the Gram-negative facultative chemolithoautotrophic bacterium Ralstonia eutropha (syn. Cupriavidus necator), which is used for biopolymer production in industry, was investigated. A biotin auxotroph mutant lacking bioF was generated, and biotin depletion in the cells and the minimal biotin demand of a biotin-auxotrophic R. eutropha strain were determined. Three consecutive cultivations in biotin-free medium were necessary to prevent growth of the auxotrophic mutant, and 40 ng/ml biotin was sufficient to promote cell growth. Nevertheless, 200 ng/ml biotin was necessary to ensure growth comparable to that of the wild type, which is similar to the demand of biotin-auxotrophic mutants among other prokaryotic and eukaryotic microbes. A phenotypic complementation of the R. eutropha ΔbioF mutant was only achieved by homologous expression of bioF of R. eutropha or heterologous expression of bioF of Bacillus subtilis but not by bioF of Escherichia coli. Together with the results from bioinformatic analysis of BioFs, this leads to the assumption that the intermediate of biotin synthesis in R. eutropha is pimeloyl-CoA instead of pimeloyl-acyl carrier protein (ACP) like in the Gram-positive B. subtilis. Internal biotin content was enhanced by homologous expression of accB, whereas homologous expression of accB and accC2 in combination led to decreased biotin concentrations in the cells. Although a DNA-binding domain of the regulator protein BirA is missing, biotin synthesis seemed to be influenced by the amount of acceptor protein present.

IMPORTANCE Ralstonia eutropha is applied in industry for the production of biopolymers and serves as a research platform for the production of diverse fine chemicals. Due to its ability to grow on hydrogen and carbon dioxide as the sole carbon and energy source, R. eutropha is often utilized for metabolic engineering to convert inexpensive resources into value-added products. The understanding of the metabolic pathways in this bacterium is mandatory for further bioengineering of the strain and for the development of new strategies for biotechnological production.

INTRODUCTION

The vitamin biotin (vitamin H or B₇) is essential for bacteria, as it acts as a cofactor in carboxylation reactions, for example, during fatty acid biosynthesis (1, 2). Biotin is composed of a ureido ring and a tetrahydrothiophene ring with an attached valeric acid chain. The late steps of biotin synthesis, which are involved in the formation of the ring structures, are well conserved among different species of bacteria, whereas the early steps of synthesis differ and are only poorly understood. The biotin synthesis pathway is best studied in Escherichia coli (3). In this bacterium, S-adenosyl-l-methionine (SAM)-dependent methyltransferase BioC adds a methyl moiety to malonyl-acyl carrier protein (ACP) to mask the terminal carboxyl group. This step ensures that the malonyl-ACP methyl ester is elongated by the enzymes of the fatty acid biosynthesis pathway, which do not act on carboxylic acids. In the following, the malonyl-ACP methyl ester is elongated by two cycles of fatty acid synthesis during which two molecules of malonyl-ACP were added to yield the pimeloyl-ACP methyl ester. To prevent further elongation of the fatty acid chain, the pimeloyl-ACP methyl ester esterase BioH removes the methyl group, and pimeloyl-ACP is formed. The late steps of biotin synthesis start with the formation of 7-keto-8-amino pelargonic acid (KAPA). KAPA synthase (BioF) condenses alanine with pimeloyl-ACP, whereby CO₂ and ACP are released. The aminotransferase BioA transfers the amino group from S-adenosyl-l-methionine (SAM) to KAPA, and the resulting 7,8-diaminononanoate (DAPA) is converted to dethiobiotin by addition of CO₂ catalyzed by DTB synthase (BioD). The thiophane ring is then formed by biotin synthase (BioB) employing a SAM molecule that serves not as a sulfur donor but as a radical starter for the reaction (4, 5).

In Bacillus subtilis, the substrate for BioF is formed by a BioI-BioW pathway. In previous studies it was suggested that cytochrome P450 (BioI) oxidatively cleaves acyl-ACP of the fatty acid synthesis pathway to pimeloyl-ACP or to free pimelic acid, which is activated by BioW to pimeloyl-CoA (6–9). Recent studies showed that BioF of B. subtilis is unable to convert pimeloyl-ACP to KAPA and that B. subtilis mutants lacking bioI are biotin prototrophic (10, 11). The current theory about the role of BioI and BioW in biotin synthesis in B. subtilis is that the non-oxygen-dependent BioW was acquired during evolution and replaced BioI in biotin synthesis. Nowadays, BioI has no major effect on biotin synthesis in B. subtilis (10). The pimelic acid precursor for biotin synthesis is probably provided by cleavage of intermediates of the fatty acid cycle catalyzed by an unknown thioesterase (10, 11). No homologous genes to bioC and bioH of E. coli are present in the genome of B. subtilis. Nevertheless, the late steps of biotin synthesis starting from pimeloyl-ACP or -CoA are the same in both organisms (6; Fig. 1).

FIG 1.

Biotin synthesis pathway in E. coli and B. subtilis. Question marks indicate unknown or undefined steps in the pathway. In E. coli, two molecules of acetyl-CoA are condensed to malonyl-CoA by acetyl-CoA carboxylase (1). To enable the entrance into the fatty acid biosynthesis (FAS) pathway that is catalyzed by FabH (2), BioC (3) adds a methyl-moiety to malonyl-CoA. After two cycles of elongation, BioH (4) removes the methyl moiety, and pimeloyl-ACP is formed. BioF (5) condenses l-alanine with pimeloyl-ACP to form 7-keto-8-amino pelargonic acid (KAPA). KAPA is converted to 7,8-diaminopelargonic acid by BioA (6), and BioD (7) forms the ureido ring by addition of CO₂. In the last step, BioB (8) catalyzes the conversion of dethiobiotin (DTB) to biotin. In B. subtilis, biotin is synthesized from pimelate as a substrate and not directly from pimeloyl-ACP. The source of pimelate is still unknown. It is predicted that BioI (9) cleaves intermediates of the FAS to form pimeloyl-ACP. As BioF of B. subtilis cannot convert ACP ester, a thioesterase (10) cleaves pimeloyl-ACP to generate pimelate. Pimelate is then activated by BioW (11) with CoA to pimeloyl-CoA. BioF of B. subtilis (12) adds l-alanine to pimeloyl-CoA and KAPA is formed. The following steps for biotin synthesis are the same as in E. coli. Additional abbreviations: ACP, acyl carrier protein; CoA, coenzyme A; SAM, S-adenosyl-l-methionine; SAH, S-adenosyl-l-homocystein; AMTOB, S-adenosyl-2-oxo-4-thiomethylbutyrate; 5′-DOA, 5′-deoxyadenosine.

The de novo synthesis of biotin is highly cost-intensive, as it requires 20 ATP equivalents (12), and it is therefore strictly regulated. In E. coli, the bifunctional protein BirA acts as a transcriptional regulator and as a biotin-protein ligase that attaches biotin to the AccB subunit of the acetyl-CoA carboxylase (ACC). ACC catalyzes the first step of the fatty acid biosynthesis, providing the precursor malonyl-ACP.

BirA activates biotin with ATP to biotinoyl-5′-AMP (bio-AMP), and in the absence of an acceptor protein, the bio-AMP intermediate remains bound to BirA. Accumulation of BirA:bio-AMP complexes result in formation of BirA dimers, which act as transcriptional regulators for the bio operon. If biotin is present in low concentrations, BirA catalyzes the attachment of bio-AMP to AccB, and no accumulation of BirA:bio-AMP occurs. In the absence of BirA dimers, bio operon transcription is derepressed. When the biotin concentration exceeds the amount of the acceptor protein AccB, bio-AMP remains bound to BirA, and active transcriptional regulator is formed by dimerization. Studies with E. coli showed that homologous expression of accB counteracted the regulation and resulted in increased biotin formation (13).

Another subunit of ACC, the biotin carboxylase AccC, forms unstable complexes with AccB. These complexes are poor substrates for biotinylation, and therefore increased amounts of AccC can negatively influence the derepression of bio operon transcription (13).

The chemolithoautotrophic β-proteobacterium Ralstonia eutropha is biotin prototrophic and ubiquitously found in soil and fresh water. The organism was reclassified and renamed several times in recent years and was finally designated Cupriavidus necator in 2004 (14–17). However, R. eutropha remained the most accepted designation and is used throughout this study. R. eutropha is well known as a production strain of biopolymers belonging to the class of poly(3-hydroxyalcanoates) (PHAs) and as a model organism for autotrophic growth with molecular hydrogen and carbon dioxide (18). The synthesis pathway of biotin and its regulation are unknown in R. eutropha. Several bio genes were annotated in the genome of this bacterium, but biotin formation in R. eutropha has not been analyzed so far. In the past years, R. eutropha has been established as a biotechnological production platform far beyond PHA, and the understanding of metabolic pathways of this bacterium has become crucial to exploit them by bioengineering for the production of fine chemicals or industrial intermediates (19–22). In this study, the early and the late steps of the biotin synthesis pathway of R. eutropha were investigated and compared to the well-studied pathways for biotin synthesis in E. coli and B. subtilis.

RESULTS

Identification of genes involved in biotin metabolism by in silico analysis and Tn5 mutagenesis.

In Escherichia coli, the bio operon comprises the genes bioA, bioB, bioF, bioC, and bioD, whereas bioH is located at a distinct location in the genome (23, 24). The gene bioA is located in the opposite direction from the other four genes of this cluster. Expression of these five genes is regulated by BirA, which attaches to an operator that is located between bioA and bioB. In silico analysis of the genome of Ralstonia eutropha revealed homologous genes to every gene of the biotin pathway of E. coli except for bioH. In R. eutropha, the genes bioA, bioF, bioD, and bioB form an operon together with accC1, coding for subunit C of acetyl-CoA carboxylase (gene loci H16_A0180 to H16_A0184). The bioC gene of R. eutropha shows only weak similarity to the bioC gene of E. coli and is not part of the operon (Fig. 2). In addition, deletion of the annotated bioC gene (gene locus H16_A0340) from the genome of R. eutropha does not result in a biotin-auxotrophic phenotype. Growth of the mutant was reduced neither on solid nor in liquid minimal medium without addition of biotin (data not shown). It is therefore possible that another gene product compensates for the lack of bioC or that biotin is not synthesized by a BioC-BioH pathway. Although a birA gene is also present in the genome of R. eutropha, its corresponding enzyme lacks a DNA-binding domain, and it is not known if it also functions as a regulator. The birA gene overlaps a downstream gene that is annotated as pantothenate kinase. This kinase shares no similarity with BirA of E. coli or with its DNA-binding domain.

FIG 2.

Genomic organization of biotin synthesis genes in E. coli, B. subtilis, and R. eutropha and sequence alignment of BioCs of E. coli and R. eutropha. (A) Comparison of the genomic organization of genes involved in biotin synthesis in E. coli, B. subtilis, and R. eutropha. Genes are shown as arrows, which indicate the direction of transcription. Gene names are noted, and those genes with a remote position are separated by a hatched line. The birA gene of R. eutropha overlaps a gene with the locus tag A0136, which is annotated as pantothenate kinase. (B) Sequence alignment of BioC of E. coli with BioC (locus tag A0340) of R. eutropha. The alignment was created using NCBI BLASTp, and identical amino acids are shown in black. A sequence identity of 30% was calculated.

Tn5 mutagenesis was performed to identify genes essential for biotin synthesis in R. eutropha. As Tn5 integrates randomly in the genome and leads to single-gene mutations (25), a biotin-negative phenotype is expected for those mutants that possesses a Tn5 integration in an essential gene for biotin synthesis. In total, 14,600 mutants were screened for biotin auxotrophy, and three independent mutants were obtained which are not able to grow without an external biotin supply. In all of these three mutants, the Tn5::mob had integrated into the bioB gene at different positions. No other biotin-negative mutants were obtained by Tn5 mutagenesis, although only single copies of bioA, bioD, and bioF exist in the genome of R. eutropha, and their disruptions were supposed to result in a biotin-negative phenotype.

Determination of biotin storage in and biotin requirement of an auxotroph mutant of R. eutropha.

A marker-free deletion mutant, R. eutropha ΔbioF, was generated using the suicide plasmid technique; the reintegration of bioF into the genome was excluded by PCR with internal primer pairs. The mutant was analyzed concerning growth on mineral salts medium (MSM) agar plates with 1% (wt/vol) gluconate and with or without supplementation of biotin. A clear negative phenotype without an external biotin source was only observed for the mutant after a second transfer to fresh plates without supplementation of biotin (Fig. 3). It is known from the literature that biotin is only present in trace amounts in bacterial cells. For E. coli, 100 to 200 molecules of biotin per cell are sufficient to promote cell growth (26, 27). This is probably also due to the few biotin-dependent enzymes. The biotin that was taken up from complex medium or minimal medium with supplementation of biotin could probably be stored in the cells and ensures cell growth during the next generations when biotin is absent from the medium.

FIG 3.

Biotin depletion in auxotroph mutant R. eutropha ΔbioF. (Left) Growth behavior in liquid mineral salts medium (MSM) with 1% (wt/vol) gluconate and with or without supplementation of 200 ng/ml biotin or with biotin after previous cultivation in the absence of biotin. (Right) Growth behavior of wild-type (top) and bioF deletion strain (bottom) on solid MSM with 1% (wt/vol) gluconate and with (lower plate) or without (upper plate) supplementation of 200 ng/ml biotin. (A) Growth after inoculation with washed cells of R. eutropha ΔbioF grown in MSM with 1% (wt/vol) gluconate and with 200 ng/ml biotin. (B) Subsequent cultivation of the mutant after inoculation from panel A. Growth in the presence of biotin from the beginning, without biotin in two consecutive cultivations, or with biotin after one cultivation without biotin. (C) Subsequent cultivation of the mutant after inoculation from panel B. Growth in the presence of biotin from the beginning, without biotin in three consecutive cultivations, or with biotin after two cultivations without biotin. (D) Subsequent cultivation of the mutant after inoculation from panel C. Growth in the presence of biotin from the beginning, without biotin in four consecutive cultivations, or with biotin after three cultivations without biotin. Growth was performed in duplicates (except for panel D with and without biotin from the beginning), and standard deviations are indicated as error bars.

In order to analyze how long a biotin-auxotrophic R. eutropha strain is able to grow without an external biotin source when transferred from complex medium to MSM medium, depletion of biotin was determined by growth analysis of R. eutropha ΔbioF in liquid MSM medium. A preculture of the mutant was set up in MSM medium with 1% (wt/vol) gluconate and 200 ng/ml biotin. After 24 h, cells were washed twice to remove biotin and utilized to inoculate fresh MSM medium with 1% (wt/vol) gluconate and with and without addition of biotin. Growth was monitored until cells entered the early stationary phase. Second and third consecutive cultures were inoculated with cells from the previous cultivation. In addition, cells that were not provided with external biotin in the previous cultivation were further cultivated in the presence of biotin in the following cultivation to see whether growth of the mutant can be reactivated.

During the first cultivation, the biotin auxotroph mutant did not show any difference in growth behavior whether biotin was supplemented or not (Fig. 3A). In the second cultivation, mutant cells showed reduced growth if still no biotin was added to the medium. The early stationary phase was reached after 15 h of cultivation, and the final density was 260 Klett units (KU) instead of 600 KU after 22 h like in the first cultivation. If biotin is provided in the second cultivation, again no difference in the growth behavior is visible between the mutant that was provided with biotin from the first cultivation and after one cultivation without an external biotin source (Fig. 3B). During the third consecutive cultivation, no cell growth was detected within 24 h in the absence of biotin. Therefore, three cultivations without biotin are necessary to deplete the internal biotin storage of a biotin-auxotrophic R. eutropha strain. If biotin is added in the third cultivation, mutants with no biotin source during the two previous cultivations have an elongated lag phase but reach the same final optical density after 23 h of cultivation (Fig. 3C). Even after three cultivations without biotin and without any detectable cell growth, the biotin auxotroph mutant was able to grow during the fourth cultivation when biotin was provided again (Fig. 3D).

The minimal demand for biotin of R. eutropha has not been documented yet. To determine this demand, the biotin-auxotrophic mutant R. eutropha ΔbioF was cultivated in liquid MSM medium without biotin until the internal storage of biotin was depleted, and a subsequent culture in fresh medium was provided with 0, 40, 80, 120, or 200 ng/ml biotin. In comparison to the biotin-prototrophic wild type, the mutant showed almost no growth without supplementation of biotin (Fig. 4). As observed in the experiments exploring biotin depletion, the mutant was unable to grow after three consecutive cultivations without biotin. If biotin was supplemented in concentrations of 40 to 200 ng/ml, the mutant showed an elongated lag phase in comparison to the wild type. In the presence of 200 ng/ml biotin, the mutant reached a final optical density comparable to that of the wild type in the stationary phase after 26 h. If only 80 or 120 ng/ml biotin was provided, the final optical density at 26 h was slightly decreased and was even lower if 40 ng/ml biotin was supplemented in the medium (Fig. 4). Hence, a concentration of at least 200 ng/ml is necessary to ensure maximum cell growth of a biotin-auxotrophic R. eutropha strain. Nevertheless, lower concentrations of biotin down to 40 ng/ml can also efficiently promote cell growth with an only slightly decreased final optical density. Biotin concentrations between 0 and 200 ng/ml did not significantly influence the growth behavior of the wild-type strain in a positive or negative manner (data not shown).

FIG 4.

Biotin demand of auxotroph mutant R. eutropha ΔbioF. Growth behavior of the mutant in liquid mineral salts (MSM) medium with 1% (wt/vol) gluconate and with 0, 40, 80, 120, or 200 ng/ml biotin after inoculation with washed cells of R. eutropha ΔbioF grown in two consecutive cultivations without biotin. The wild-type strain was cultivated accordingly without biotin. Growth was performed in duplicates, and standard deviations are indicated as error bars.

Complementation of R. eutropha ΔbioF with bioF genes of E. coli and Bacillus subtilis.

In contrast to E. coli, no bioH gene is present in the genome of R. eutropha. Furthermore, the annotated bioC gene of R. eutropha shows only weak similarity to the bioC gene of E. coli, and deletion of bioC did not lead to a biotin-auxotrophic R. eutropha strain (Fig. 5). The strain R. eutropha ΔbioC ΔphaCAB provided by our partners from INVISTA Textiles (UK) showed no differences in growth on minimal medium plates supplemented with 200 ng/ml biotin or without biotin. Differences in colony opacity can be traced back to the deletion of the phaCAB operon. As the early steps of biotin synthesis differ in many microorganisms, R. eutropha could possibly catalyze the formation of the pimelate thioester intermediate by enzymes other than BioC and BioH. As an example, Bacillus subtilis synthesizes biotin from pimelic acid as a precursor. The pimelic acid is derived either from the medium or from the fatty acid biosynthesis pathway, but the responsible enzyme that cleaves pimeloyl-ACP to provide pimelic acid for biotin synthesis is unknown so far (10). Free pimelic acid is activated in B. subtilis by BioW to pimeloyl-CoA instead of pimeloyl-ACP. BioF only accepts pimeloyl-CoA for the synthesis of 8-amino-7-oxononanoate in the following synthesis step (11). The subsequent synthesis steps for the assembly of the fused heterocyclic rings of biotin are the same for B. subtilis and E. coli.

FIG 5.

Growth behavior of R. eutropha ΔbioC ΔphaCAB on solid medium. (A and B) Cultivation of R. eutropha H16 (top) and R. eutropha ΔbioC ΔphaCAB (bottom) on solid mineral salts medium (MSM) without supplementation of biotin (A) or with supplementation of 200 ng/ml biotin (B) for 3 days. Cultivation after growth on solid NB medium (1). Restreaking from the respective previous plate without supplementation of biotin (2 and 3). Differences in colony opacity of the mutant lacking bioC can be traced back to the additional deletion of the phaCAB-operon.

Complementation experiments were performed to identify whether one of the bioF genes from E. coli and B. subtilis or both could enable R. eutropha ΔbioF to grow in the absence of biotin. The bioF genes of R. eutropha, E. coli, and B. subtilis were amplified and cloned into the broad-host-range vector pBBR1MCS-3. The resulting plasmids were transferred to R. eutropha by conjugation, and the recombinant strains were streaked on solid MSM with 1% (wt/vol) gluconate and without supplementation of biotin. The deletion mutant harboring the empty vector showed no growth on MSM (Fig. 6B). R. eutropha ΔbioF homologously expressing bioF as well as the strain that expresses bioF of B. subtilis showed normal growth on solid MSM without supplementation of biotin (Fig. 6B). Comparable to the biotin auxotrophic mutant harboring the empty vector, expression of bioF of E. coli did not promote cell growth of the mutant on medium without addition of biotin. All strains grew on solid MSM with 1% (wt/vol) gluconate and with supplementation of biotin (Fig. 6A).

FIG 6.

Complementation of biotin auxotroph R. eutropha ΔbioF and growth on solid medium. (A and B) Growth behavior of R. eutropha H16 ΔbioF(pBBR1MCS-3) (upper plate, top), R. eutropha ΔbioF(pBBR1MCS-3::bioF_Re) (upper plate, bottom), R. eutropha ΔbioF(pBBR1MCS-3::bioF_Ec) (lower plate, top), and R. eutropha ΔbioF(pBBR1MCS-3::bioF_Bs) (lower plate, bottom) on solid mineral salts medium (MSM) with 1% (wt/vol) gluconate and with supplementation of 200 ng/ml of biotin (A) or without biotin (B) after three restreaks on fresh plates and after 48 h of incubation.

The complementation experiment was repeated in liquid medium to determine if duration of biotin depletion is comparable between R. eutropha ΔbioF harboring the empty plasmid and the mutant expressing bioF of E. coli. After washing the recombinant R. eutropha strains cultivated in liquid MSM with biotin and transferring them to fresh liquid MSM without biotin, a significantly decreased final optical density was only monitored for the biotin-auxotrophic strain harboring the empty vector (Fig. 7A). The mutant expressing bioF of B. subtilis showed a longer lag phase and reduced cell growth during the early exponential phase, but the final optical density was comparable to the strains expressing bioF of R. eutropha or of E. coli. After 44 h of cultivation, the cells were transferred for a second time to fresh MSM without biotin, and cell growth was monitored. No growth was detected in this second cultivation for the mutant harboring empty vector (Fig. 7B). Significantly reduced growth was also observed for the mutant expressing bioF of E. coli. The strain reached a final optical density at 600 nm (OD₆₀₀) of 0.42 after 45 h of incubation starting from 0.08 OD₆₀₀ after inoculation. The mutants expressing bioF of R. eutropha or of B. subtilis showed comparable growth behavior and reached a final optical density of about 1.4 (Fig. 7B). A third consecutive cultivation without biotin was performed which confirmed the successful complementation of R. eutropha ΔbioF with bioF of R. eutropha as well as with bioF of B. subtilis. No growth could be monitored for the mutant harboring the empty vector or the mutant that expresses bioF of E. coli (Fig. 7C).

FIG 7.

Complementation of biotin auxotroph R. eutropha ΔbioF and growth in liquid medium. Growth behavior in liquid mineral salts medium (MSM) with 1% (wt/vol) gluconate and without supplementation of biotin of the recombinant strains R. eutropha H16 ΔbioF(pBBR1MCS-3), R. eutropha H16 ΔbioF(pBBR1MCS-3::bioF_Re), R. eutropha ΔbioF(pBBR1MCS-3::bioF_Ec), and R. eutropha ΔbioF(pBBR1MCS-3::bioF_Bs). (A) Cultures were inoculated with washed cells from previous cultivation in MSM with 1% (wt/vol) gluconate and in the presence of 200 ng/ml biotin. (B) Subsequent cultivation of the recombinant strains in the absence of biotin after inoculation from panel A. (C) Subsequent cultivation of the recombinant strains in the absence of biotin after inoculation from panel B. Growth was performed in triplicates, and standard deviations are indicated as error bars.

The failed complementation of R. eutropha ΔbioF by plasmid-encoded bioF of E. coli could also result from poor heterologous expression of bioF in R. eutropha. However, the presence of a bioF transcript was confirmed by isolation of total RNA of R. eutropha ΔbioF(pBBR1MCS-3::bioF_Ec) and reverse transcription PCR (RT-PCR) with specific primers for bioF of E. coli. A transcript of the bioF gene of E. coli could be demonstrated with this method in the mutant during the exponential growth phase (Fig. 8).

FIG 8.

Detection of transcript of bioF of E. coli in recombinant R. eutropha. RNA was isolated from R. eutropha H16 ΔbioF(pBBR1MCS-3::bioF_Ec) cultivated in liquid MSM with 1% (wt/vol) gluconate, and cDNA was synthesized with unspecific poly-dT primers. Specific bioF primers were used to detect bioF of E. coli (bioF_Ec), and specific phaP1 primers were applied as positive controls (phaP1). RT-PCR was performed with cDNA (+) or total RNA (–) to verify the lack of gDNA in RNA samples.

Furthermore, a bioF deletion strain of E. coli was transformed with either the pBBR1MCS-3 empty plasmid or the plasmid harboring the bioF gene of R. eutropha, E. coli, or B. subtilis. The resulting recombinant strains were cultivated in liquid and on solid M9 medium without supplementation of biotin. After three consecutive cultivations in liquid medium, only the strain harboring the E. coli bioF gene grew without supplementation of biotin (Fig. 9 C). On solid medium, the bioF deletion strain showed growth after a second transfer to fresh medium without biotin supplementation when bioF of E. coli and of R. eutropha was provided on the plasmid (Fig. 9B).

FIG 9.

Complementation of the biotin auxotroph E. coli bioF748::kan and growth in liquid and on solid M9 medium. (Left) Growth behavior in liquid M9 medium with 0.4% (wt/vol) glucose and without supplementation of biotin of the recombinant strains E. coli bioF748::kan(pBBR1MCS-3), E. coli bioF748::kan(pBBR1MCS-3::bioF_Re), E. coli bioF748::kan(pBBR1MCS-3::bioF_Ec), and E. coli bioF748::kan( pBBR1MCS-3::bioF_Bs). (A) Cultures were inoculated with washed cells from previous cultivation in liquid M9 with 0.4% (wt/vol) glucose and in the presence of 200 ng/ml biotin. (B) Subsequent cultivation of the recombinant strains in the absence of biotin after inoculation from panel A. (C) Subsequent cultivation of the recombinant strains in the absence of biotin after inoculation from panel B. Growth was performed in triplicates, and standard deviations are indicated as error bars. (Right) Growth behavior of E. coli bioF748::kan(pBBR1MCS-3) (top), E. coli bioF748::kan(pBBR1MCS-3::bioF_Re) (right), E. coli bioF748::kan(pBBR1MCS-3::bioF_Ec) (bottom), and E. coli bioF748::kan(pBBR1MCS-3::bioF_Bs) (left) on solid M9 medium with 0.4% (wt/vol) glucose and with (upper plate) or without (lower plate) supplementation of 200 ng/ml biotin. (A) Growth after inoculation with washed cells of recombinant E. coli bioF748::kan grown in liquid M9 medium with 0.4% (wt/vol) glucose and with 200 ng/ml biotin. (B) Subsequent cultivation of the mutant after inoculation from panel A.

Bioinformatic analysis of BioF of R. eutropha, E. coli, and B. subtilis.

Multiple sequence alignment was generated of protein sequences of BioF of E. coli, B. subtilis, and R. eutropha (see Fig. S1 in the supplemental material). BioF of R. eutropha showed 44 and 39% identity and 61 and 56% similarity to BioF of E. coli and B. subtilis, respectively. The catalytically important lysine that forms a Schiff base with the pyridoxal 5′-phosphate cofactor is located at position 246 in BioF of R. eutropha. Crystal structures of ACP-dependent enzymes revealed a predominantly hydrophilic interaction by salt bridges between basic residues of the enzyme and acidic residues of the helix II of ACP protein (28). No crystal structure of E. coli BioF bound to pimeloyl-ACP is available so far, but in silico docking studies suggest that five basic residues at the luminal surface of the half-open hand structure of BioF could be responsible for the ACP binding (11). These basic residues are replaced by hydrophobic residues or glutamine in BioF of B. subtilis, which could explain its inability to utilize pimeloyl-ACP instead of pimeloyl-CoA (11). R. eutropha possesses two of these five basic residues at the same position at which the other three are replaced by neutral residues. Furthermore, one of four catalytically important residues in BioF of E. coli is replaced by aspartic acid in BioF of B. subtilis and of R. eutropha (Fig. S1).

Regulation of biotin synthesis in R. eutropha.

To determine whether biotin synthesis in R. eutropha is regulated by the amount of acceptor protein AccB like in E. coli, accB was homologously overexpressed in R. eutropha. It was expected that more biotin would be synthesized when the concentration of AccB was enhanced by homologous expression. In addition, the impact of homologous expression of both accB and accC was analyzed. The R. eutropha strains expressing the accB gene (locus tag A3171), the accB to accC2 gene cluster (locus tags A3171 and A3172), or the control harboring the empty vector showed similar growth in liquid mineral salts medium (MSM) with 1% (wt/vol) gluconate during 35 h of cultivation (data not shown). The internal amount of total biotin (protein-bound and free biotin) of the recombinant strains in the late stationary phase was determined by the Quant*Tag biotin kit (Vector Laboratories, Inc., Burlingame, CA). The biotin content was increased by homologous expression of the accB gene and decreased by expression of the accB to accC2 gene cluster in R. eutropha. The control strain exhibited a biotin content of 18.7 ± 1.1 nmol biotin/mg total protein, whereas expression of accB resulted in a biotin content of 40 ± 4.3 nmol biotin/mg total protein, which is an increase of 115% in comparison to the control strain. The expression of the accB to accC2 gene cluster resulted in a biotin content of approximately 5 ± 5 nmol biotin/mg total protein, which is a decrease of 72% in comparison to the control strain (Fig. 10).

FIG 10.

Biotin content of R. eutropha strains expressing accB or accB and accC2. The strains R. eutropha pBBR1MCS-3 (control), R. eutropha(pBBR1MCS-3::accB) (accB), and R. eutropha(pBBR1MCS-3::accB::accC2) (accB-accC2) were cultivated in liquid MSM with 1% (wt/vol) gluconate and 25 μg/ml tetracycline and harvested in the stationary phase. Biotin content was determined and is represented as the total biotin amount in the soluble fractions per 1 mg of total protein in the respective soluble fraction.

DISCUSSION

The biotin metabolism of several bacteria has been examined in the past years (6, 29, 30). The two best-studied bacteria regarding biotin synthesis and its regulation are E. coli and B. subtilis (for a review, see reference 3). The biotin synthesis pathways in these two bacteria differ in the early steps leading to the intermediate pimeloyl-ACP or pimeloyl-CoA, whereas the late steps for the formation of the two ring systems to yield biotin are identical. In the biotechnological production strain R. eutropha, biotin metabolism has not been studied yet. As this bacterium has been established as a production platform for diverse polymers, fine chemicals, or industrial intermediates, elucidation of its general metabolic pathways has become important. Due to the complete annotation of the genome of R. eutropha (31), genes putatively involved in biotin metabolism were also identified. From these data, it is not clear if R. eutropha possesses a bioC/bioH pathway like E. coli or a bioI/bioW pathway like B. subtilis for the early steps of biotin synthesis. Although a bioC gene has been annotated, its similarity to the bioC gene of E. coli is low, and a bioH homologue is absent in the genome of R. eutropha. On the other hand, no bioI gene is annotated in the genome of R. eutropha, but four genes annotated as (putative) cytochrome P450 proteins are present. Whether one or all four genes could code for a cytochrome P450 that cleaves intermediates of the fatty acid synthesis to form the intermediate pimeloyl-ACP and participate in biotin synthesis pathway is unclear.

The Tn5 mutagenesis method to identify essential genes for biotin synthesis in R. eutropha was chosen because it provides a simple and fast technique to screen for genes possibly involved in metabolic pathways. In this study, a clear biotin-negative phenotype was only observed if the bioB gene was disrupted. Three independent mutants all with disrupted bioB were identified with this method. No other biotin-negative mutants were obtained. The random integration of the Tn5 element should theoretically ensure that gene disruption mutants for every gene of the genome are obtained as long as the number of generated and screened mutants is high enough. Two circumstances can prevent genes involved in the metabolism of interest from being identified with this method. First, this occurs if homologous genes are present in the genome and if the gene product of one gene can complement the disruption of the other gene. Second, although it is known that Tn5 integrates randomly, so-called hotspots were identified where Tn5 integrates more often (32). This could lead to an underestimated number of theoretically required mutants to disrupt every gene in the genome once. However, the number of Tn5 mutants generated and screened in this study was very high and in theory should have been sufficient to disrupt every gene of the genome of R. eutropha at least one to two times. It could be possible that the generation of more mutants could lead to the identification of more biotin-negative mutants possessing the Tn5 element in genes other than bioB. At least the disruption of bioF should result in a biotin-negative mutant, as the generated ΔbioF mutant was biotin auxotrophic. Furthermore, our subsequent experiments showed that thorough depletion of biotin is necessary to identify a biotin-negative phenotype, and every mutant candidate was therefore transferred two times to fresh medium without biotin. The necessity of multiple restreaking to identify biotin auxotrophs is already known from other studies with E. coli (24). In addition, experiments with avidin added to the medium were done to bind traces of biotin and to enhance the phenotype of biotin-negative mutants (33).

The complementation experiments with the biotin auxotroph mutant R. eutropha ΔbioF performed in this study clearly showed that heterologously expressed bioF of B. subtilis complements the mutant, whereas bioF of E. coli does not. BioF of B. subtilis utilizes pimeloyl-CoA as a substrate instead of pimeloyl-ACP (11). Furthermore, bioinformatic analyses revealed that BioF of R. eutropha did not share all of the five relevant basic residues for ACP binding with BioF of E. coli but has at three positions a substitution to a neutral amino acid like in BioF of B. subtilis (Fig. S1). These results suggest that in R. eutropha also, the intermediate of the biotin synthesis is pimeloyl-CoA instead of pimeloyl-ACP. As mentioned before, cytochrome P450 proteins are present in R. eutropha that could possibly cleave intermediates of the fatty acid synthesis pathway, but in B. subtilis, the cytochrome P450 protein BioI plays a minor role in biotin synthesis compared to BioW. B. subtilis strains that lack bioI can grow in the absence of biotin, whereas deletion of bioW leads to a biotin-negative strain (10). It is concluded that in B. subtilis, pimelic acid originating from the fatty acid synthesis pathway is the precursor of biotin synthesis and that BioW activates pimelic acid to pimeloyl-CoA. Cleavage of pimeloyl-ACP is presumably catalyzed by a specific pimeloyl-ACP thioesterase (10). In R. eutropha, the genes with the locus tags B0928 and B1438 are annotated as 6-carboxyhexanoate-CoA ligases, but they exhibit no sequence similarity to BioW of B. subtilis. Whether the corresponding proteins could activate pimelic acid with CoA for biotin synthesis is unclear.

On the other hand, studies demonstrated that BioF of E. coli can utilize either pimeloyl-ACP or pimeloyl-CoA as the substrates (11, 34). If pimeloyl-CoA is formed as an intermediate in R. eutropha, bioF of E. coli should have been able to complement the mutant. In both studies, the substrate specificity of BioF of E. coli was tested in in vitro assays. Hence, the conversion rate of pimeloyl-CoA catalyzed by BioF of E. coli could possibly be too low to complement the bioF-negative mutant of R. eutropha. We clearly demonstrated that bioF of E. coli is transcribed in the mutant of R. eutropha, but it could not be excluded that the protein is not correctly synthesized or is inactive. Another explanation could be that BioF of E. coli is not able to recognize or convert ACP of R. eutropha.

The regulation of biotin synthesis can be influenced by the amount of acceptor protein (AccB) present in R. eutropha. The current study showed that homologous expression of accB leads to a higher total content of intracellular biotin as was already shown for E. coli (13). Expression of accB and accC2 in combination decreased the biotin content in R. eutropha. Further studies have to be done to elucidate the regulation of biotin synthesis at the transcriptional level. Although R. eutropha possesses a birA gene, the corresponding protein lacks the regulator domain with the typical helix-turn-helix motif like in some alphaproteobacteria and actinobacteria (35–37). Furthermore, the genome of R. eutropha does not contain bioR or bioQ orthologues, which are involved in biotin regulation in other microorganisms.

Biotin depletion has already been monitored in the Gram-positive bacterium Corynebacterium glutamicum, which is used for the biotechnological production of amino acids in industry (38). In that study, it was shown that biotin is completely depleted in biotin-auxotrophic C. glutamicum strains after three transfers to fresh medium without biotin supplementation. After the first transfer, the biotin-auxotrophic strain already did not reach the final optical density of the strain that was grown in the presence of biotin, but it was still able to grow. After the second transfer to medium without biotin, the cells grew poorly, but only after the third transfer to medium without supplemented biotin was growth completely prevented (38). The same effect of biotin depletion was published in 1962 for Arthrobacter globiformis, which needed three to four consecutive reinoculations to prevent growth of the biotin-auxotrophic strain (39). In our present study, we showed a comparable effect for a generated biotin-auxotrophic R. eutropha strain. After the first transfer from medium with biotin to medium without biotin, the biotin-auxotrophic mutant did not differ in its growth behavior in comparison to the biotin-prototrophic wild type. During the second cultivation in the absence of biotin, a significant decrease of cell growth was observed, whereas after the third transfer to medium without biotin, almost no growth was detectable. This is consistent with previous studies that found that biotin is only necessary in trace amounts in bacterial cells (3). Apparently, biotin that is taken up from the environment can promote cell growth during several generations when biotin is exhausted.

In addition, the minimal demand of biotin to promote cell growth of auxotrophic bacteria has been studied occasionally in the past but not for R. eutropha cells (12, 39–42). We showed that a biotin-auxotrophic mutant of R. eutropha requires about 200 ng/ml biotin in the medium to show cell growth comparable to the prototrophic wild type. Nevertheless, a concentration of about 40 ng/ml biotin is sufficient to promote general cell growth of the biotin-auxotrophic mutant. Former studies revealed that other bacteria, such as C. glutamicum or auxotrophic mutants of E. coli, can grow efficiently with concentrations as low as 10 ng/ml or 3.6 ng/ml, respectively (40, 42).

In this study, the duration of biotin depletion in a biotin-auxotrophic R. eutropha strain was monitored, as was the minimal demand of biotin for optimal growth of this strain. Both experiments confirm that biotin is only necessary in small amounts in the cells. Furthermore, it was shown that regulation of biotin synthesis can be influenced by the amount of AccB present, which is similar to the regulation in E. coli, although the synthesis pathway seemed to be different. No evidence was found for a comparable bioC/bioH pathway for the early synthesis steps. Instead, complementation experiments and bioinformatic analyses led to the assumption that the early steps of biotin synthesis could be similar to a bioI/bioW pathway for production of pimeloyl-CoA as an intermediate of the biotin synthesis pathway.

MATERIALS AND METHODS

Bacteria and growth conditions.

The bacterial strains used in this study are listed in Table 1. R. eutropha H16 strains were grown in nutrient broth (NB; BD Difco, Heidelberg, Germany) or in mineral salts medium (MSM) with 1% (wt/vol) sodium gluconate (43) at 30°C for 30 h in flasks with baffles on a rotary shaker at 130 rpm or in 96-well C-bottom plates (Brand, Wertheim, Germany) in a microplate spectrophotometer (Epoch 2; BioTek Instruments, Winooski, VT, USA) with double orbital shaking at 282 cycles per minute (cpm) and 30°C. For complementation experiments, R. eutropha was grown on solid MSM containing 1.7% (wt/vol) Bacto agar. Liquid and solid MSM were supplemented with 200 ng/ml biotin if necessary. For plasmid stabilization, a concentration of 25 μg/ml tetracycline was applied. Cell growth was determined spectrophotometrically at 600 nm (UviLine 8100; Xylem Analytics GmbH, Weilheim, Germany) or in a Klett-Summerson photometer (Manostat Corporation, NY, USA) using filter no. 54 (520 to 580 nm). E. coli strains were grown in Luria-Bertani (LB) medium (44) or in M9 medium (modified from reference 45; 33.7 mM Na₂HPO₄, 22.0 mM KH₂PO₄, 8.55 mM NaCl, 9.35 mM NH₄Cl, 0.4% [wt/vol] glucose, 1 mM MgSO₄, 0.3 mM CaCl₂, 1 μg/liter thiamine, 134 μM EDTA, 31 μM FeCl₃, 6.2 μM ZnCl₂, 760 nM CuCl₂, 420 nM CoCl₂, 1.62 μM H₃BO₃, and 81 nM MnCl₂; if necessary, 200 ng/ml biotin was added) with 0.4% (wt/vol) glucose at 37°C in flasks with baffles on a rotary shaker at 130 rpm. For complementation experiments, E. coli strains were cultivated on solid M9 with 0.4% (wt/vol) glucose or in 96-well plates (Brand, Wertheim, Germany) in a microplate spectrophotometer (Epoch 2; BioTek Instruments) with double orbital shaking at 282 cycles per minute (cpm) and 37°C. For cultivation of recombinant strains, 100 μg/ml ampicillin, 50 μg/ml kanamycin, or 12.5 μg/ml tetracycline was added.

TABLE 1.

Bacterial strains used in this study

Strain	Descriptiona	Source or reference
R. eutropha H16	WT, Gmr	DSM 428
R. eutropha HF39	Strr mutant of R. eutropha H16	DSM 15444
R. eutropha bioB::Tn5::mob	Tn5 mutant of R. eutropha HF39 with Tn5 element in bioB	This study
R. eutropha ΔbioF	Marker-free deletion of bioF, biotin auxotroph	This study
R. eutropha ΔphaCAB ΔbioC	Marker-free deletion of phaCAB and bioC	This study
E. coli TOP-10	Cloning strain, F− mcrA Δ(mrr-hsdRMS-mcrBC) galK rpsL Φ80lacZΔM15 Δlac*X74 recA1 ara*D139 Δ(ara-leu)7697 galU endA1 nupG	Life Technologies
E. coli S17-1	thi-1 proA hsdR17 (rK− mK+) recA1; tra genes of plasmid RP4 integrated into the genome	47
E. coli bioF748::kan (JW0759-1)	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− ΔbioF748::kan rph-1 Δ(rhaD-rhaB)568 hsdR514	Keio Collection
E. coli K-12 MG 1655	K-12 laboratory strain	DSM 18039
B. subtilis subsp. subtilis 168	WT	DSM 402

^aWT, wild type.

Tn5 mutagenesis.

Transposon mutagenesis of the streptomycin-resistant mutant strain R. eutropha HF39 using the suicide plasmid technique was performed as described previously (46, 47). In brief, plasmid pSUP5011 was transferred from E. coli S17-1 to R. eutropha HF39 by conjugation (48). Tn5-induced mutants were selected on NB plates containing 500 μg/ml streptomycin and 150 μg/ml kanamycin. Mutants were transferred to fresh MSM agar plates containing 150 μg/ml kanamycin (master plates) and twice to plates without 200 ng/ml biotin. Putative biotin leaky or negative mutants were analyzed by construction of a gene library (46, 49). In brief, the library was constructed by isolation of genomic DNAs (gDNA) of Tn5 mutants, digestion of gDNA with BamHI or SalI, and ligation of the resulting fragments with pBluescript SK(−). E. coli TOP-10 was transformed with these plasmids and selected for kanamycin resistance conferred by the Tn5 fragment. Sequencing of the hybrid plasmids allowed identification of the DNA region in which the Tn5 was inserted.

Construction of recombinant and deletion mutant strains.

The plasmids and oligonucleotides used in this study are listed in Table 2. Genomic DNA of R. eutropha H16, E. coli K-12 MG1655, or B. subtilis subsp. subtilis 168 was isolated using the method of Marmur (50), and plasmids were isolated using the method of Birnboim and Doly (51). Genes were amplified using Phusion high-fidelity DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA) and purified by gel extraction employing the Monarch DNA gel extraction kit (New England Biolabs, Ipswich, MA, USA). DNA fragments were subcloned into pJET1.2/blunt and subsequently digested with FastDigest restriction enzymes (Thermo Fisher Scientific, Waltham, MA, USA). For homologous expression and complementation experiments, genes were cloned into the broad-host-range vector pBBR1MCS-3 and transferred to R. eutropha by conjugation via E. coli S17-1 (48).

TABLE 2.

Plasmids and oligonucleotides (shown in 5′-3´direction) used in this study

Plasmid or oligonucleotide	Description	Source, reference, or locationa
Plasmids
pJET1.2/blunt	Cloning vector, Apr Kmr	Thermo Fisher Scientific
pSUP5011	Apr Cmr Kmr Tn5::mob	57
pBluescript SK(−)	Cloning vector, Apr, lacPOZ′	Stratagene, California
pJQ200mp18Tc	sacB oriV oriT traJ, Tcr	52
pJQ200mp18Tc::bioFupdn	pJQ200mp18Tc carrying up/down fragment of bioF, Tcr	This study
pBBR1MCS-3	Broad host range vector, Tcr, lacPOZ′mobRP4	Life Technologies
pBBR1MCS-3::bioFRe	pBBR1MCS-3 carrying bioF of R. eutropha, Tcr	This study
pBBR1MCS-3::bioFEc	pBBR1MCS-3 carrying bioF of E. coli, Tcr	This study
pBBR1MCS-3::bioFBs	pBBR1MCS-3 carrying bioF of B. subtilis, Tcr	This study
pBBR1MCS-3::accB	pBBR1MCS-3 carrying accB of R. eutropha, Tcr	This study
pBBR1MCS-3::accB::accC2	pBBR1MCS-3 carrying accB and accC2 of R. eutropha, Tcr	This study
Oligonucleotides
bioF Re up	AAAAGGATCCATGCTGCTTGAACAACTGAGGCGG	5′ region of bioFRe
bioF Re down	AAAAGTCGACTCATGCGACCTCCCGCTGCG	3′ region of bioFRe
bioF Ec up	AAAAGGTACCGGAGGGGCATTATGAGCTGGCAG	5′ region of bioFEc
bioF Ec down	AAAAGAGCTCTTAACCGTTGCCATGCAGCACC	3′ region of bioFEc
bioF Bs up	AAAAGGTACCGGAGGGGAGCCATGAAGATTGATTC	5′ region of bioFBS
bioF Bs_dn	AAAAGAGCTCTCAAATGATGTGCAGCTCCTTTCCG	3′ region of bioFBS
bioF up fw	AAAAGAGCTCTTCTTCGCGTGGGAACAGTC	5′ region of bioF up flank
bioF up rv	AAAAGGTACCTCAAGCAGCATGGCGCAC	3′ region of bioF up flank
bioF down fw	AAAAGGTACCTGCGCATCACGCTGTCG	5′ region of bioF dn flank
bioF down rv	AAAACTGCAGGTGGCGATGTTCTCGTCG	3′ region of bioF dn flank
bioF intern fw	GCTGACCTTCTGCAGCAACGACTAC	5′ internal region of bioF
bioF intern rv	CTCCTCTCCTTCGATGATCGCC	3′ internal region of bioF
bioF extern fw	CGGCATGGCGATGTACGACC	5′ external region of bioF
bioF extern rv	CGGCCTCGGCTTGGTATTGG	3′ external region of bioF
phaP1 up	GGAGACCAGCAATGATATCCTCACCCCG	5′ region of phaP1
phaP1 down	CAACGCAGGCAGGAATTCTTATCAGGCAGCCGTCG	3′ region of phaP1
accB up	AAAAGGTACCGGAGGGAAAAGATGGACCTGC	5′ region of accB
accB down	AAAATCTAGATCAGCCGATCACGAACAGCG	3′ region of accB
accC2 down	AAAATCTAGACGTTGCGTTGCGTCAGGC	3′ region of accC2

^aup flank, upstream flanking region; dn flank, downstream flanking region.

To achieve gene deletion of bioF, flanking regions of about 500 bp up- and downstream of bioF were amplified, subcloned, and ligated. The resulting fragment was digested with SacI and PstI and integrated into the suicide plasmid pJQ200mp18Tc (52). The resulting plasmid was used to delete the bioF gene marker free by two events of homologous recombination as described elsewhere (52). Precise deletion was confirmed with control PCR employing primer pairs that bind in the bioF gene and up- and downstream of bioF.

Reverse transcriptase PCR (RT-PCR).

To detect transcript of bioF of E. coli in R. eutropha, RT-PCR was performed. R. eutropha H16 ΔbioF(pBBR1MCS-3::bioF_Ec) was cultivated in liquid MSM medium with 1% (wt/vol) gluconate, and cells were harvested in the exponential growth phase. RNA was isolated from the cells using an RNeasy minikit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RT-PCR was performed utilizing a OneTaq RT-PCR kit (New England Biolabs, Ipswich, MA, USA) with unspecific poly-dT primers for cDNA synthesis. The primer pair bioF Ec up and bioF Ec down was used for the amplification of bioF of E. coli to identify a bioF_Ec transcript, and the primer pair phaP1 up and phaP1 down was used for the amplification of phaP1 as a positive control.

DNA sequencing.

The correct construction of all plasmids for heterologous or homologous gene expression as well as gene deletion and hybrid plasmids of a gene library for analysis of Tn5 mutants were verified by DNA sequencing. Sequencing reactions were carried out according to a standard procedure by Eurofins Genomics (Ebersberg, Germany) and analyzed with Chromas software ver. 1.45 (School of Health Science, Griffith University, Queensland, Australia).

Assay for biotin depletion and requirement of R. eutropha ΔbioF.

The generated biotin-autotrophic mutant R. eutropha ΔbioF was utilized to determine the duration of biotin depletion in and biotin demand of R. eutropha cells.

To determine biotin depletion, a preculture of R. eutropha ΔbioF was set up in liquid MSM with 1% (wt/vol) sodium gluconate and 200 ng × ml⁻¹ biotin and incubated for 24 h at 30°C and 130 rpm on a rotary shaker. Cells were harvested by centrifugation for 10 min at 4,000 rpm in a Universal 320R centrifuge (Hettich Lab Technology, Tuttlingen, Germany), washed twice with 0.9% (wt/vol) sodium chloride solution, and resuspended in MSM medium. A volume of 2.5 ml was employed to inoculate 50 ml of MSM with 1% (wt/vol) sodium gluconate with and without supplementation of 200 ng × ml⁻¹ biotin in 250-ml Klett flasks. After cultivation at 30°C and 130 rpm, a sample of 2.5 ml was taken at the early stationary phase from the culture without biotin supplementation to inoculate in 50 ml fresh MSM medium with or without biotin supplementation. The described steps were repeated three times and led to four consecutive cultivations of R. eutropha ΔbioF with or without biotin. In addition, one culture was set up with the mutant that exhibited biotin supplementation over all four consecutive cultivations.

The requirement of biotin of R. eutropha cells was determined by growing the biotin auxotrophic mutant R. eutropha ΔbioF in two consecutive cultivations in flasks with liquid MSM with 1% (wt/vol) sodium gluconate without supplementation of biotin to deplete the internal biotin content. Cells were harvested and applied for inoculation of 200 μl MSM with 1% (wt/vol) sodium gluconate and 0, 40, 80, 120, or 200 ng/ml biotin in a microplate. R. eutropha H16 was used as the control strain and was grown without biotin supplementation. Growth was monitored at 600 nm with a microplate spectrophotometer (Epoch 2; BioTek Instruments, Winooski, VT, USA) at 30°C for 26 h.

Determination of total biotin content of soluble cell fraction.

The amount of total biotin (protein-bound and free biotin) was determined with a QuantTag biotin kit according to the manufacturer’s instructions (Vector Laboratories, Burlingame, CA, USA). Cells were grown in liquid MSM medium with 1% (wt/vol) sodium gluconate at 30°C, harvested after 30 h, and washed with 0.9% (wt/vol) sodium chloride solution. Cell extracts were prepared by resuspending cell pellets in 4 ml 0.5 M Tris-HCl (pH 7), and cell disruption was achieved with a Sonopuls instrument equipped with an M72 probe for 4 min at 50% output control (Bandelin Electronic, Berlin, Germany). Cell debris was removed by centrifugation, and the resulting soluble cell fraction was utilized for determination of protein concentration by the method of Bradford (53) and for determination of biotin content with a QuantTag biotin kit.

Bioinformatic analyses.

Complete genome sequences of R. eutropha H16, E. coli K-12 MG1655, and B. subtilis subsp. subtilis 168 were downloaded from GenBank (54), and protein sequences of BioF were downloaded from UniProt Knowledgebase (55). A multiple sequence alignment of proteins was obtained with MegAlign Pro ver. 15.2 (DNAStar, Inc., Madison, WI, USA) by applying Clustal_X (56).