OUP user menu

Biotin biosynthesis, transport and utilization in rhizobia

Karina Guillén-Navarro, Sergio Encarnación, Michael F. Dunn
DOI: http://dx.doi.org/10.1016/j.femsle.2005.04.020 159-165 First published online: 1 May 2005


Biotin, a B-group vitamin, performs an essential metabolic function in all organisms. Rhizobia are α-proteobacteria with the remarkable ability to form a nitrogen-fixing symbiosis in combination with a compatible legume host, a process in which the importance of biotin biosynthesis and/or transport has been demonstrated for some rhizobia–legume combinations. Rhizobia have also been used to delimit the biosynthesis, metabolic effects and, more recently, transport of biotin. Molecular genetic analysis shows that an orthodox biotin biosynthesis pathway occurs in some rhizobia while others appear to synthesize the vitamin using alternative pathways. In addition to its well established function as a prosthetic group for biotin-dependent carboxylases, we are beginning to delineate a role for biotin as a metabolic regulator in rhizobia.

  • Biotin biosynthesis
  • Rhizobia
  • Rhizobia–legume symbiosis

1 Introduction

Root nodule bacteria, collectively known as rhizobia, are a crucial component of the global nitrogen cycle because they reduce atmospheric nitrogen to ammonia in symbiotic association with a compatible plant host and thus reduce the need for synthetic nitrogen fertilizers. Before establishing symbiosis, rhizobia must survive in the soil awaiting the presence of a suitable host legume. Infection of the host requires multiplication in the rhizosphere as well as during early phases of the infection. Mature, nitrogen-fixing intracellular rhizobia (bacteroids) require large amounts of energy and reductant derived from the catabolism of plant-supplied organic acids. Consequently, the metabolism and growth factor requirements of rhizobia have long been studied (for reviews, see [[[]).

Biotin (vitamin H) has an essential metabolic function as the CO2-carrying prosthetic group of selected carboxylases, decarboxylases and transcarboxylases [[]. De novo biotin biosynthesis occurs in many prokaryotes while others are partly or totally dependent on external sources. The purpose of this review is to summarize what is known about biotin biosynthesis, transport and utilization in rhizobia. Rhizobia are the only prokaryotes in which novel regulatory roles for biotin have been investigated. Biotin transport is important for the establishment of symbiosis in some rhizobia, and they are the only prokaryotes in which genes encoding biotin transport proteins have been identified. A new aspect of biotin biosynthesis in rhizobia is the probable presence of novel pathways in some species.

2 Biotin requirement of rhizobia

Early studies on biotin used Rhizobium leguminosarum bv. trifolii to demonstrate that “heat-stable Rhizobium growth factor” was identical to “coenzyme R” from Azotobacter and that both were, in fact, identical to biotin [[]. Based on their growth response to biotin in defined media, rhizobia may be grouped with respect to their ability to biosynthesize the vitamin. Biotin auxotrophs are incapable of biotin biosynthesis and require external sources for growth [[[]. An ecologically interesting example is provided by the non-symbiotic Mesorhizobium loti strains isolated from soils [[,[0]. These isolates lack a 500 kb region of their chromosome termed the “symbiosis island” which, in addition to a variety of symbiosis-specific functions, encodes the biosynthesis of thiamine, nicotinic acid and biotin.

Biotin prototrophs synthesize biotin de novo and show neither a growth nor a significant metabolic response to exogenous biotin [[[,[1]. For example, Rhizobium tropici CFN299 grows well in minimal medium subcultures in the absence of biotin and maintains a high level of pyruvate carboylase activity and holo-enzyme protein regardless of biotin supplementation [[2,[3].

Biotin bradytrophs synthesize biotin but either do so inefficiently or only under certain growth conditions [[1,[4]. A controversy exists as to whether Rhizobium etli and Sinorhizobium meliloti fit into this class [[1,[2,[5] or with the biotin auxotrophs [[6]. It is important to note that when S. meliloti Rm1021 was grown in biotin-free medium, a concomitant several-fold increase in biomass and extracellular biotin, detected with an ELISA assay, were found, indicating that this strain can produce the vitamin de novo [[5]. Growth studies show that S. meliloti strain GR4B is a biotin bradytroph whose synthesis of biotin, detected with a standard bioassay, was dependent upon growth conditions [[1].

Wild-type R. etli strain CE3 behaves as a biotin auxotroph when serially subcultured in minimal medium, where very low biotin-dependent carboxylase activities and protein levels confirm the presence of a biotin starved state. Growth is restored not only in the presence of exogenous biotin but also by supplementation with thiamine, pimelic acid (a biotin precursor), fumarate plus malate, cAMP, glutamate, proline, or oxygen ([[2]; unpublished results). S. meliloti Rm1021 behaves similarly to R. etli CE3 with respect to the ability of thiamine to prevent biotin auxotrophy [[2]. Given that both S. meliloti and R. etli lack genes homologous to most or all of the orthodox biotin biosynthesis genes (see Section 6), the challenge of providing an unequivocal demonstration of their ability to synthesize the vitamin remains.

3 Biotin-dependent carboxylases and biotin–protein ligase in rhizobia

Biotin and carbon dioxide are essential for the growth of rhizobia [[,[,[7]. Genome sequence and biochemical analysis show that rhizobia contain the biotin-dependent enzymes pyruvate carboxylase (PYC), acetyl-CoA carboxylase (ACC), and two or more acyl-CoA carboxylases, including propionyl-CoA carboxylase (PCC) [[8]. PYC is required for growth on sugars or pyruvate and, although its inactivation has no effect on nodulation and nitrogen fixation in S. meliloti, R. etli or R. tropici[[3,[9], it would be interesting to determine whether it plays a role in rhizosphere competition, since sugars are excreted to the rhizosphere by legume roots [[0]. The symbiotic phenotype of a rhizobial PCC mutant has not been determined but inactivation of the S. meliloti methylmalonyl-CoA mutase, which catalyzes the step following that of PCC during propionyl-CoA degradation, does not affect symbiotic performance [[1,[2]. ACC has not been characterized but would be expected to be essential for fatty acid synthesis [[8] and thus viability.

Apo-biotin-dependent carboxylases are converted to their active holo-enzymes by biotin–protein ligase (BPL) [[,[3]. The BPLs of Bacillus and enteric bacteria are called BirA's (biotin regulators) because their N-terminal helix-turn-helix motif binds to and represses biotin operon transcription [[3,[4]. In these organisms, BirA catalyzes the conversion of biotin into biotinoyl-AMP, which functions with BirA as a co-repressor: when the intracellular concentration of biotin is elevated (e.g., in biotin-supplemented cultures), more BirA-biotinoyl-AMP is formed and transcription is repressed. The middle and C-terminal portion of BirA contain the catalytic residues for ligating biotin to target proteins [[3].

Genome sequence analysis of M. loti, B. japonicum, S. meliloti (http://www.kazusa.or.jp/rhizobase) R. etli (G. Dávila, V. González, R. Gómez and P. Bustos, unpublished), and R. leguminosarum bv. viciae (http://www.sanger.ac.uk/Projects/R_leguminosarum) shows that their deduced BPL gene products lack the N-terminus found in BirA's and retain only the catalytic motifs required for biotinylating apo-carboxylases. These monofunctional BPLs are present in many prokaryotes and all eukaryotes.

The fact that rhizobia contain multiple biotin-dependent carboxylases raises the question of how biotin is partitioned among them by a single BPL. Western blotting experiments designed to follow the biotinylation of the carboxylases in biotin-starved R. etli cells pulsed with biotin suggest that the relative level of each apo-carboxylase determines the amount of holo-carboxylase formed (M. Dunn, unpublished). It is not known if the BPL has the same affinity for each of the different apo-carboxylases.

4 Effect of biotin on gene expression

Biotin affects gene expression in eukaryotes [[5] but little information exists on biotin as a BirA-independent regulatory molecule in prokaryotes. Proteome analysis shows that biotin markedly alters global protein expression in R. etli[[6] and S. meliloti[[7,[8]. In R. etli, however, most of the changes in protein expression caused by biotin were similar to those observed with thiamine supplementation or growth in complex medium [[6]. Thus most of the changes observed with biotin reflect the general metabolic state of the cells rather than a specific effect of the vitamin, and so without appropriate controls claims of biotin-dependent gene expression must be interpreted with caution.

Gene fusion assays show that S. meliloti bhdA (encoding β-hydroxybutyrate dehydrogenase), bioS (putative biotin-responsive regulatory gene), and copC (possible copper resistance gene) are induced in response to culture biotin supplementation [[7[9]. In contrast, pcm (encoding l-isoaspartyl protein repair methyltransferase), sinI (homoserine lactone autoinducer synthatase) and sinR (homoserine lactone autoinducer transcriptional regulator) are repressed under these conditions [[7]. Proteome analysis revealed that proteins whose levels decreased under biotin limitation included the gene product of the down-regulated copC mentioned above, 50S ribosomal protein L7/L12, RNA polymerase ω subunit, periplasmic transporter substrate binding proteins (two for sugars, one for amino acids) and 2-keto-3-deoxy-6-phosphogluconate aldolase (part of the Entner-Doudoroff pathway). As Heinz and Streit [[7] discuss in detail, there is some correlation between these results and the physiological response of S. meliloti to biotin. For instance, the upregulated BdhA participates in the degradation of the carbon storage polymer poly-β-hydroxybutyrate (PHB), consistent with the finding that biotin supplementation prevents PHB accumulation in S. meliloti[[2,[8]. Regulation of PHB degradation by biotin could prevent its accumulation in bacteroids [[] and promote its accumulation and gradual utilization in oligotrophic environments like soil.

5 Biotin and the rhizobia–legume symbiosis

The effect of biotin on symbiosis depends on the rhizobia–legume combination in question, and many naturally-occuring, symbiotically proficient rhizobia are biotin auxotrophs. A M. loti R7A biotin auxotroph (bioA::Tn5) was indistinguishable from the wild-type in colonizing the Lotus corniculatus rhizosphere. However, the biotin phenotype of this mutant is leaky [[4] perhaps due to the presence of a second copy of bioA, as occurs in M. loti MAFF303099 [[0]. Studies with S. meliloti and R. etli bioN and bioM biotin uptake mutants indicate that high affinity uptake is required for efficient nodulation of their respective legume hosts ([[,[5]; K. Guillén-Navarro, submitted). Interestingly, S. meliloti bioN mutants engineered for biotin overproduction by the introduction of the E. coli bio operon were also found to compete poorly with the wild-type in the alfalfa rhizosphere, perhaps due to the reduced viability observed in the overproducing strains [[9].

Determining the absolute symbiotic requirement for biotin in rhizobia is complicated by the fact that the vitamin is excreted from the roots of host plants [[5,[1]. The very low bacteroid activities of biotin-dependent carboxylases in the bradytroph R. etli CE3 indicate that little biotin is synthesized by, or available to, the microsymbiont. In contrast, bacteroids of the biotin prototroph R. tropici CFN299 from nodules formed on the same host (bean) have high activities, indicating de novo synthesis of the vitamin ([[2]; M. Dunn, unpublished).

6 Biotin biosynthesis

Escherichia coli and Bacillus species are the model organisms to which we owe most of our understanding of biotin biosynthesis (Fig. 1). Bacillus spp. are able to take up (apparently by passive diffusion) and use pimelic acid as a precursor of biotin. Pimelic acid is derived through an unknown pathway which may involve the postulated fatty acid synthase-like activities of BioX and BioI [[3]. Pimelic acid is activated to its CoA derivative by pimeloyl-CoA synthetase, the product of bioW[[4]. E. coli is unable to utilize pimelic acid as a biotin precursor and instead synthesizes pimeloyl-CoA, possibly from acetyl-CoA [[5], using BioH, a probable carboxyl esterase [[6] and the yet uncharacterized BioC [[3]. bioW homologs do not occur in the sequenced genomes of M. loti, B. japonicum, S. meliloti or A. tumefaciens.

Figure 1

The orthodox biotin biosynthetic pathway derived largely from studies utilizing Bacillus spp. and E. coli. Where gene homologs encoding the main biosynthesis pathway enzymes exist in rhizobia, they are designated by the following abbreviations: At, A. tumefaciens; Bj, B. japonicum; Ml, M. loti, NGR, Rhizobium sp. NGR234; Re, R. etli; Rl, R. leguminosarum bv. viciae; Sm, S. meliloti.aHomolog with low sequence identity to other BioFs. bHomolog present in genome but does not complement an E. coli bioF mutant. cData obtained from a partial genome sequence [[2].

The M. loti gene encoding BioZ is part of the bioBFDAZ operon (Fig. 2) and shows similarity to β-ketoacid-acyl synthases. BioZ can functionally complement E. coli bioH, but not bioC, mutants. Based on this, Sullivan et al. [[4] proposed that BioZ catalyzes both the condensation of a thioester with an odd number of carbon atoms to produce pimeoyl-ACP and its subsequent transacetylation to pimeloyl-CoA. This hypothesis is consistent with recent enzymological data on the E. coli BioH [[6].

Figure 2

Biotin biosynthesis gene clusters in selected prokaryotes. Data were obtained from the literature cited in the text or by homology searches of the following genome databases: A. tumefaciens, http://www.ncbi.nlm.hih.gov/genomes/MICROBES/Complete.html; B. japonicum, http://www.kazusa.or.jp/rhizobase; R. etli, G. Dávila, V. González, R. Gómez and P. Bustos, unpublished. Contiguous arrows represent gene clusters and spaces denote genes or clusters in other parts of the genome. The drawings are not to scale.

Four enzymes convert pimeloyl-CoA to biotin, namely BioF (7-keto-8-aminopelargonic acid synthase), BioA (7,8-diaminopelergonic acid aminotransferase), BioD (dethiobiotin synthetase) and BioB (biotin synthetase) (Fig. 1). The physical arrangement of gene clusters encoding these orthodox biotin biosynthetic enzymes are presented in Fig. 2. In M. loti R7A, a functional bioBFDA operon was confirmed by complementation of E. coli mutants inactivated in one of these genes [[4]. Entcheva and co-workers [[] used genome sequence analysis and complementation tests with E. coli biotin mutants to identify putative biotin biosynthesis genes in S. meliloti. bioF and bioA homologs, potentially encoding the first two enzymes for the pimeloyl-CoA to biotin pathway, were found, but only the bioF homolog could complement the corresponding E. coli mutant. Homologs for the last enzymes of the pathway (bioD and bioB) were not found. Genes possibly encoding BioH and BioZ, involved in pimeloyl-CoA synthesis, were also found but could not complement their respective E. coli mutants (a bioI homolog was encountered but complementation was not tested since no homolog occurs in E. coli). The genome sequence of R. etli CE3 contains a bioA homolog on plasmid f, but no bioB, bioD or bioF homologs (unpublished results), while that of R. leguminosarum bv. viciae contains bioA and bioF homologs but lacks homologs for bioB and bioD.

7 Biotin transport

Active biotin uptake occurs in E. coli but the transport system involved is not known [[7]. An S. meliloti mutant inactivated in bioS, the biotin-upregulated gene mentioned in Section 4, has a higher level of biotin uptake than the wild-type [[8]. bioS encodes a LysR type protein and its role in biotin uptake would appear to be regulatory [[8]. Both S. meliloti and R. etli contain operons (bioMNB) encoding products involved in biotin uptake or retention which are identically organized and share high sequence identity ([[], Guillén-Navarro et al., submitted). Very similar operon exists in R. leguminosarum bv. viciae and A. tumefaciens but have not been characterized experimentally. The gene originally designated bioB in S. meliloti[[] does not resemble a biotin synthase (the classical bioB product) but instead has similarity to bioY, first implicated in biotin biosynthesis in Bacillus sphaericus because of its proximity to other genes involved in biotin biosynthesis [[9]. We refer here to the S. meliloti and R. etlibioB” homologs as bioY. Sequence analysis and experimental data [[,[0] suggest that bioM and bioN are ABC-type transporters for biotin and encode the ATPase and permease components, respectively. BioY has six probable transmembrane domains like those of transport permeases but constitutes its own family in the Pfam database [[0]. In S. meliloti, uptake experiments with a high concentration of external biotin (40 nM) showed that a bioM mutant was deficient in biotin retention but not uptake [[]. We used low external biotin concentrations (10–100 pM) to show that a R. etli bioM mutant had significantly reduced uptake of biotin but was not defective in retaining it (Guillén-Navarro, submitted). Overexpression of bioY in wild-type S. meliloti Rm1021 allows better than wild-type growth in medias supplemented with dethiobiotin. It was suggested that BioY might play a role in converting dethiobiotin to biotin by a mechanism distinct from that of a classical biotin synthase [[]. However, because commercially available dethiobiotin contains biotin as a contaminant [[1], extra copies of bioY may promote growth in dethiobiotin-supplemented cultures by allowing more efficient uptake of the contaminating biotin.

BlastN analysis was used to determine the presence of homologs of bioB, bioD, bioF and bioA (the orthodox biotin biosynthesis genes) and bioY (the putative high affinity transport component) in 159 sequenced genomes (including 37 incomplete genomes) in the GenBank and KEGG databases. We found that (i) nearly 16% of the genomes contained only bioY, (ii) 39% lacked bioY and contained all of the orthodox biosynthetic genes, (iii) nearly 18% contained bioY and all of the orthodox biosynthetic genes and (iv) the remainer contained bioY plus one or two of the classical genes. The genomes encoding all of the genes included those of A. tumefaciens and M. loti, which could benefit from possessing both the orthodox biosynthetic route and high affinity uptake capability, since both species colonize plant tissues but also survive as saprophytes in soil.

8 Perspectives

Rhizobia make enlightening subjects for the study of biotin metabolism and utilization owing to characteristics which differ from the standard model organisms including (i) their ability to enter into symbiosis, which has been disected at the molecular level and for which the importance of biotin is dependent on the symbiotic combination; (ii) the presence of multiple biotin-dependent carboxylases; (iii) absence of BirA regulatory functions; (iv) preliminary data indicating a metabolic regulatory function for biotin and (v) the apparent presence of novel biosynthetic pathways. We need to persue the work on possible novel biotin biosynthesis pathways with a rigorous biochemical and physiological characterization, including the use purified precursors to demonstrate actual substrate/product relationships. The application of global methodologies such as proteomics and transcriptomics in rhizobia will allow further identification of genes and gene products regulated by biotin. Our knowledge of biotin uptake and the regulation of its utilization can also be greatly expanded with rhizobia as experimental organisms.


We apologize to the authors of papers which were not cited because of the publishers space limitations. K. G-N. was supported by graduate student fellowships 138526 from CONACyT and 202327 and 202363 from DGAPA-UNAM. We thank G. Dávila, V. González, R. Gómez and P. Bustos for access to the R. etli genome sequence prior to publication.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
View Abstract