OUP user menu

Molecular and functional characterization of the Azorhizobium caulinodans ORS571 hydrogenase gene cluster

Cecilia Baginsky, Jose-Manuel Palacios, Juan Imperial, Tomás Ruiz-Argüeso, Belén Brito
DOI: http://dx.doi.org/10.1111/j.1574-6968.2004.tb09723.x 399-405 First published online: 1 August 2004


In this work, we report the cloning and sequencing of the Azorhizobium caulinodans ORS571 hydrogenase gene cluster. Sequence analysis revealed the presence of 20 open reading frames hupTUVhypFhupSLCDFGHJK hypABhupRhypCDEhupE. The physical and genetic organization of A. caulinodans ORS571 hydrogenase system suggests a close relatedness to that of Rhodobacter capsulatus. In contrast to the latter species, a gene homologous to Rhizobium leguminosarum hupE was identified downstream of the hyp operon. A hupSL mutation drastically reduced the high levels of hydrogenase activity induced by the A. caulinodans ORS571 wild-type strain in symbiosis with Sesbania rostrata plants. However, no significant effects on dry weight and nitrogen content of S. rostrata plants inoculated with the hupSL mutant were observed in plant growth experiments.

  • Azorhizobium
  • hup genes
  • hupE
  • Nitrogen fixation

1 Introduction

During the biological nitrogen fixation process carried out by bacteria from the Rhizobiaceae family in legume nodules, H2 is released as an obligate by-product of the nitrogenase reaction- In the Rhizobium–legume symbiosis, hydrogen evolution is considered a major source of inefficiency [1]. Hydrogen oxidation by symbiotically expressed uptake hydrogenases has been postulated as a way to improve the efficiency of the nitrogen fixation process and enhance legume productivity [2].

Hydrogenase gene clusters have been extensively characterized in the endosymbionts Bradyrhizobium japonicum and Rhizobium leguminosarum bv. viciae [3]. In both species, genetic determinants for hydrogen oxidation are clustered in large DNA regions spanning ∼20 kb. Hydrogenase structural subunits are encoded by the hupSL genes, whereas the 15 downstream genes named hupCDFGHIJK and hypABFCDEX determine accessory proteins involved in hydrogenase synthesis and activity. In addition, specific accessory proteins have been described for each hup system. The hupNOP genes, involved in nickel metabolism in B. japonicum[4], are apparently absent from the R. leguminosarum genome. In contrast, the R. leguminosarum hup system comprises the hupE gene [5], whose predicted product might play a role in nickel uptake. Differences in gene regulation between B. japonicum and R. leguminosarum hup systems are also remarkable. In R. leguminosaum, hydrogenase activity is only expressed in symbiosis. Under such conditions NifA, the key regulator of the nitrogen fixation process, activates hup gene expression [6]. In contrast, B. japonicum oxidizes hydrogen in free-living cultures as well as in soybe-n bacteroids. In this bacterium, as in other hydrogen-oxidizing bacteria like Rho-obacter capsulatus and Ralstonia eutropha, a multicomponent regulatory system formed by HupUV, HoxA (designated HupR in R. capsulatus) and HupT proteins is involved in hup gene activation in vegetative cells [7].

Although symbiotic hydrogenase activity has been detected in several rhizobia groups, no other hup gene clusters have been studied to date. Azorhizobium caulinodans, the endosymbiont of Sesbania rostrata[8], induces hydrogenase activity in bacteroids as well as in microaerobic free-living cells (dissolved oxygen conc-ntration up to 40 μM) [9]. Recently, we have analyzed hydrogenase gene distribution and organization in Azorhizobium sp. and A. caulinodans strains [10]. The presence of hupS, hypB and hupUV genes was only detected in certain strains. The Hup+ strains showed a homogeneous intraspecific hybridization pattern. In addition, phylogenetic analysis with hupS and hupL partial sequences showed that A. caulinodans ORS571 hup genes clustered with those of R. capsulatus and far from other rhizobia. In this work, we have sequenced the A. caulinodans ORS571 hup gene cluster and studied its contribution to the symbiosis with S. rostrata plants.

2 Materials and methods

2.1 Bacterial strains and growth conditions

A. caulinodans ORS571 [8] was routinely grown in YEB [11] or Rhizobium minimal Rm [12] media at 28 °C. E. coli strains HB101 and DH5α were used for genetic manipulations. Antibiotics were used at the following concentrations (micrograms per milliliter): ampicillin 100, kanamycin 50, and spectinomycin 150.

2.2 Plant tests and enzyme assays

S. rostrata seeds were surface sterilized and germinated on 1% water agar plates [13]. Seedlings were grown in Leonard jar-type assemblies under bacteriologic-lly controlled conditions with a nitrogen-free plant nutrient solution [14] supplemented with 20 μM NiCl2 from day 10 after seedling inoculation. Acetylene reduction and hydrogen evolution in intact nodules were determined by gas chromatography. Hydrogenase activity in bacteroid suspensions was measured using an amperometric method with oxygen as electron acceptor [14]. Determinations of protein content of bacteroid suspensions and cell cultures were done by the bicinchoninic acid method [15]. To measure dry weight of S. rostrata plants, the aerial part was excised and dried at 80 °C for 48 h. Nitrogen content was measured in a Kjeldahl apparatus.

2.3 DNA manipulation techniques

Restriction enzyme digestions, PCR amplifications, agarose gel electrophoresis and Southern blot transfers were carried out by standard protocols [16]. Genomic DNA of A. caulinodans ORS571 was extracted as previously described [14]. The R. leguminosarum UPM791 hupS and hypB, and the B. japonicum hupUV gene probes were generated by PCR and labelled with digoxigenin using DIG-11-dUTP (Roche Molecular Biochemica-s,-Mannheim, Germany) at 40 μM final concentration [10]. Hybridizing bands were visualized using a chemiluminescent DIG detection kit as described by the manufacturer (Roche Molecular Biochemicals, Mannheim, Germany).

2.4 Mutant construction

A 1.5-kb DNA fragment containing the hupSL genes from A. caulinodans ORS571 was obtained by PCR amplification using genomic DNA and primers hupSL1 and hupSL2 as previously described [10]. The PCR product was cloned in vector PCR2.1-TOPO (Invitrogen BV, Groningen, The-Netherlands), excised with KpnI–XbaI and ligated to suicide plasmid pSS10 [17] giving construct pSCB1. To generate mutant ORS571.1, pSCB1 was introduced in ORS571 by conjugation. Transconjugant strains were selected with spectinomycin and analyzed for correct plasmid insertion at the hupSL region by Southern blot experiments using the R. leguminosarum UPM791 hupS gene as probe. For mutant ORS571.2, the 1.5-kb DNA fragment was digested with BamH1–XbaI and cloned in the suicide vector pK18mobsac[18]. The resulting plasmid pKCB1 was introduced in ORS571 by conjugation. Insertion of pKCB1 into hup genes in kanamycin-resistant transconjugants was confi-med by Southern hybridization.

2.5 Cloning and sequencing of the A. caulinodans ORS571 hup gene cluster

The DNA region containing the A. caulinodans ORS571 hup genes was isolated from genomic DNA of mutant ORS571.1. DNA samples were independently restricted with BamH1, BglII, NotI, NsiI, SpeI, SphI or XbaI, religated and transformed into E. coli cells. Cells resistant to spectinomycin carried the pSS10 vector and the DNA region adjacent to the hupSL genes. By this procedure, we isolated plasmids pSCB11 from the NsiI digestion and pSCB12 from the BamH1 digestion that harboured the 12- and ∼23-kb DNA region upstr-am and downstre-m the hupSL genes, respectively. For sequencing, these plasmids were mutagenized with the transposon cloned in plasmid pSS20GKT (Kmr) (S. Sidler, unpublished) by an in vitro transposition reaction following the protocol described by the manufacturer (EPICENTRE Technologies). Plasmid DNA was sequenced using two synthetic oligonucleotides 5′-TTTGATTTCACGGGTTGG-3′ and 5′-GCTGGCTT-TT-CTTGTTATCG-3′, c-mplement-ry to each si-e of the transposon. Gaps in the sequence were filled with sequence reactions from specific synthetic oligonucleotides. DNA sequencing was carried out using the BigDye Terminator Cycle-Sequencing Ready Reaction kit and a- ABI377 automatic sequencer (PE Biosystems, Foster City, CA). Sequence analysis was carried out with the Sequencher™ 4.1 and DNA Strider 1.2 programs. Search for homologous proteins was carried out with BLAST software [19]. The 20-kb nucleotide sequence of the hup gene cluster has been deposited in the Genbank with the Accession No. AY581127.

3 Results and discussion

3.1 Sequence analysis of the A. caulinodans ORS571 hup gene cluster

The A. caulinodans ORS571 hup gene cluster was cloned in plasmids pSCB11 and pSCB12 (Section 2.5) that were obtained from genomic DNA of mutant strain ORS571.1 (for mutant generation see Section 2.4). Sequence analysis of a ∼32-kb DNA region cloned in these plasm-ds revealed the presence of 20 open reading frames hupTUVhypFhupSLCDFGHJKhypABhupRhypCDEhupE (Fig. 1). The A. caulinodans hup system started upstream of the hydrogenase structural genes hupSL with three open reading frames whose predicted products showed homology with HupT, HupU and HupV of R. capsulatus[20] and B. japonicum[21,22]. As in the R. capsulatus hup gene cluster, the hypF gene was found between the hupV and hupS genes. The HupS structural subunit showed a leader peptide that contains a double arginine motif typical of proteins exported by the twin-arginine transport (TAT) system [23]. Downstream from the hupS and hupL genes, the physical and genetic organization of the A. caulinodans hup cluster closely resembled that of R. capsulatus, comprising the hupCDFGHJK and hypABhupRhypCDE genes whose deduced proteins shared between 30% and 66% sequence identity with their R. capsulatus counterparts (Table 1).

Figure 1

Physical and genetic maps of the hup gene cluster of A. caulinodans ORS571. Horizontal lines show the restriction map of the DNA regions cloned in plasmids pSCB11 and pSCB12. White and black horizontal arrows represent hup and hyp genes, respectively. Letters below the arrows identify the corresponding genes. Abbreviations: E (EcoRI), B (BamHI), H (HindIII), N (NsiI), Nt (NotI), X (XhoI).

View this table:
Table 1

Characteristics of A. caulinodans hydrogenase gene products

ProteinNumber of residuesPredictedSequence conservationa
  • a Values indicate percentage of identity and similarity (data in brackets) with the corresponding gene products of R.capsulatus (RHOCA). B. japonicum (BRAJA) and R. leguminosarum bv. viciae (RHILV).

  • b Alignment of HupI proteins with the N-terminal part of A. caulino-ans HupJ.

  • c Alignment of HupJ proteins with the C-terminal part of A. caulino-ans HupJ.

Among the Hup proteins, it is noteworthy that A. caulinodans HupJ is a fusion of hupI and hupJ gene products, which are found as two independent proteins in B. japonicum and R. leguminosarum[3]. HupI is homologous to rubredoxins and contains conserved Cx2C motifs possibly involved in Fe binding [24]. These motifs were present at the N-terminal part of A. caulino-ans HupJ, whereas the similarity with HupJ from R. leguminosarum and B. japonicum was restricted to the carboxy-terminus of the fusion protein (Table 1). The function of HupJ in hydrogenase synthesis is still unknown. However, the fact that HupI and HupJ are also fused in the R. capsulatus hup system suggests that they might share a common biological process [20].

Besides the hypABCDE genes, the hyp region contained the hupR regulatory gene located between the hypB and hypC genes (Fig. 1). Hyp proteins exhibit conserved motifs involved in metal binding and ligand assembly to the hydrogenase active center [25]. As in B. japonicum and R. leguminosarum, A. caulinodans HypB displays motifs for a dual function. First, nickel storage through the histidine-rich stretch found in the amino-ter-inus of this protein; and secon-, energization of the nickel insertion process through a sequence showing similarity to the GTP-binding motif. HypF of A. c-ulinodans displayed a sequence (IxHHxAH) similar to the motif involved in the carbamoylation process leading to the formation of the Fe ligands from hydrogenase active center described in E. coli[26,27]. As in other species, the A. caulinodans protein also showed two zinc finger motifs (Cx2Cx18Cx2C) whose function in hydrogenase biosynthesis still remains elusive. The hypC gene encodes a chaperone protein that interacts with the precursor of the large subunit through a cysteine residue found at position 2 [28]. This residue was also conserved in A. caulinodans HypC. Lastly, HypE harboured at its C-terminus the conserved motif PRIC, -hose reactive terminal cysteine is involved in transfer of CN groups to the Fe atom of the active center [26].

Immediately downstream of the hyp operon, an open reading frame showing 43% sequence identity with R. leguminosarum HupE was identified. Presence of hupE at the end of the hup gene cluster was not expected, since this gene, as well as hupR, was not detected in a previous search for hup genes in the A. caulinodans genome by Southern hybridization [10]. An explanation for these negative results is not obvious, but problems related to the reliability of DNA hybridizations for certain strains were already reported in that work. Finding of hupE is also intriguing because this gene has only been found in the hup systems of R. leguminosarum, Methylococcus capsulatus and R. sphaeroides[5,29], Accession No. Y14197]. In these bacteria, hupE is located much closer to the structural genes. The A. caulinodans HupE amino acid sequence predicts a transmembrane protein that contains a canonical signal peptide (Fig. 2) along with six transmembrane domains. HupE shows homology to UreJ, an accesory protein of the nickel-containing urease enzyme, which sug-ests a similar function for these two proteins in hydrogenase and urease biosynthesis, respectively. Although the precise role of HupE and UreJ proteins has not been elucidated, its involvement in nickel uptake has been postulated [30]. In fact, the alignment of HupE proteins revealed a set of conserved residues (Hx5DH, Fig. 2) similar to the motif described in nickel permeases like NixA from Helicobacter pylori, HoxN from R. eutropha, and NikC from E. coli[3133].

Figure 2

Alignment of HupE proteins. The vertical arrow indicates the predicted signal peptide cleavage site. Letters on grey background show the region similar to a conserved motif required for nickel transport in nickel permeases. Asterisks indicate identical amino acid residues. Horizontal lines above the alignment show the location of transmembrane (TM) domains for the A. caulinodans HupE sequence as predicted by the TMHMM2 algorithm. Accesion Nos. for proteins: A. caulinodans (AZOCA), this work; R. leguminosarum (RHILV), P27650; M. capsulatus (METCA), tr: Q8RJ15; and R. sphaeroides (RHOSP), tr: O86469. Amino acid sequences were aligned using CLUSTALW [39].

Inspection of the 7.4-kb region downstream of hup- did not reveal the presence of genes related to hydrogenase synthesis. No homologue to hypX was found in this region, suggesting that this gene is absent in the A. caulinodans hydrogenase system. hypX is found in the hup systems from the aerobic bacteria B. japonicum, R. leguminosarum and R. eutropha[25], and it has been proposed that its protein product might be involved in synthesis and transport of CO and CN groups to the metallic centre of hydrogenase in these bacteria [34]. However, the absence of hypX in the A. caulinodans hup cluster, as well as in that located in the symbiotic island from B. japonicum[35], indicates that there is not a strict correlation between the presence of this gene and aerobic hydrogenases. Finally, sequence analysis of a 3-kb DNA region upstream of h-pT showed the presence of three open reading frames, transcribed in opposite orientation as regards hupT, that encode components of an ABC transport system similar to those involved in nitrate transport whose function is likely not related with hydrogenase biosynthesis.

A. caulinodans hup genes were tightly arranged with short intergenic distances between them. In many cases start and stop codons overlapped, suggesting translational coupling of genes transcribed in the same transcriptional unit. This is the case for the hupTUVhypF genes that apparently defined a single transcriptional unit transcribed from an intergenic region of 1216 bp located upstream of the hupT gene. Intergenic regions of 215 and 263 bp were also identified between the hypF–hupS and hupC–hupD genes, respectively.

The promoter of the hydrogenase structural genes hupSL has been extensively characterized in several hup systems [7]. In R. capsulatus, HupR activates transcription of the σ70-dependent hupSL pr-moter through binding to a TTG-N5-CAA sequence [36]. In our case we have not found such a sequence but, since we have no data on the potential transcription initiation site for A. caulinodans hupSL genes, further comments about the promoter architecture would be too speculative. In any case, the presence of a multicomponent regulatory system consisting of the H2-sensing hydrogenase HupUV, the hist-dine kinase HupT and the transcriptional activator HupR suggests a mechanism for regulation of hupSL genes similar to those of B. japonicum and R. capsulatus. One interesting point will be to evaluate the contribution of this regulatory system to hydrogenase expression in free-living and symbiotic conditions. In-B. japonicum, the HupTUVHoxA signal transduction and regulation system is partially responsible for symbiotic hydrogenase expression [22], and FixK2 has been proposed as the regulator of hup genes in bacteroids [37]. Whether FixK or another regulator is also involved in A. caulinodans hydrogenase expression requires further investigation that will surely shed light on how free-living and symbiotic hup regulation systems are integrated in endosymbiotic bacteria.

3.2 Analysis of the symbiotic hydrogenase activity of A. caulinodans ORS571

A. caulinodans ORS571 and its derivative hupSL mutant strain ORS571.2 were tested for hydrogen metabolism and nitrogenase activity in symbiosis with S. rostrata plants. Hydrogen metabolism was characterized by measuring H2 evolution in root nodules and hydrogenase activity in bacteroid suspensions (Table 2). Bacteroids of A. caulinodans ORS571 showed high levels of hydrogenase activity. This activity completely oxidized hydrogen produced by nitrogenase, since levels of hydrogen evolution from nodules were undetectable. In bacteroids from mutant ORS571.2, hydrogenase activity was drastically reduced and a significant amount of hydrogen evolving from nodules was observed. These data corroborated the inactivation of the hup gene cluster. We also measured nitrogenase activity in S. rostrata nodules (Table 2). In the test conditions used for plant growth, no significant differences in acetylene reduction were observed between ORS571 and the hupSL mutant strain.

View this table:
Table 2

Nitrogen fixation and H2 metabolism of A. caulinodans ORS571 and hupSL mutant strains in symbiosis with S. rostrata plantsa

StrainHydrogenase activitybH2 evolutioncNitrogenase activitydRelative efficiencye
ORS57143,560 ± 5200<0.117.3 ± 3.91
ORS571.2510 ± 1604.5 ± 0.318.5 ± 2.80.76
  • a Values were determined at 30 days of cultivation.

  • b Expressed as nmol H2 h−1 mg prot−1. Values are averages of four determinations ± SE.

  • c Expressed as μmol H2 h−1 g−1 (fresh weight of nodules). Values are averages of four determinations ± SE.

  • d Expressed as μmol C2H2 h−1 g−1 (fresh weight of nodules). Values are averages of 16 determinations ± SE.

  • e Relative efficiency was calculated as 1-[H2 evolution/C2H2 reduction].

3.3 Effect of hydrogenase system inactivation on S. rostrata plant productivity

Contribution of hydrogen oxidation to the A. caulinodansS. rostrata symbiosis was assessed by comparing dry weight and nitrogen content of Sesbania plants inoculated with A. caulinodans wild type and hupSL mutant ORS571.2 (Table 3). Data from three independent experiments did not reveal significant differences either in dry weight or in nitrogen content of 40 days-old plants. Moreover, values were s-ightly higher in the hupSL mutant versus the wild-type strain. The results indicate t-at hydrogenase inactivation did not affect S. rostrata productivity in the plant growth conditions used for these assays. Benefits associated to the presence of the hydrogenase system have been reported in the B. japonicum–soybean symbiosis comparing wild type and isogenic hup mutants [38]. In our experiments, negative results might be affected by the mutant strain tested. The hupSL mutation in strain ORS571.2 was generated by single recombination. By this strategy, hydrogen uptake could not be completely abolished in mutant ORS571.2. The residual activity detected might be sufficient to mask possible benefits of the hydrogenase system on Sesbania plant productivity. An alternative possibility to explain the residual hydrogenase activity detected would be the existence of an additional copy of hup genes in A. caulinodans. Although this was not evident from the Southern analysis performed during the construction of the mutant, it has to be noted that the detection of a second set of hup genes in B. japonicum USDA110 was only possible after the sequencing of the symbiotic island [35]. The construction of a hup mutant strain deleted of the whole hup gene cluster will be required to rule out the possibility of a second set of genes. Plant experiments using a hydrogenase knock-out mutant will help to assess the -eal contribution of hydrogenase activity on Sesbania plant yields.

View this table:
Table 3

Nitrogen content and dry weight of S. rostrata plants inoculated with A. caulinodans ORS571 and hupSL mutant strainsa

StrainExperiment IExperiment IIExperiment III
N contentDry weightN contentDry weightN contentDry weight
  • a Determinations were carried out after 40 days. Values in each experiment are the average of eight determinations. Nitrogen content is expressed in mg plant−1 and dry weight in g plant−1. Different letters in the same column of each experiment indicate significant differences at 1% based on Duncan's test.


This research was supported by grants from the Ministerio de Ciencia y Tecnología (AGL2001-2295) to T.R.A. and from Programa d- Grupos Estratégicos (III PRICYT) of the Comunidad Autónoma de Madrid. C. Baginsky is on leave from the Faculty of Agronomy, Universidad de Chile, Santiago. B. Brito is the recipient of a Contrato Ramón y Cajal from the Ministerio de Ciencia y Tecnología.


  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].
View Abstract