OUP user menu

Physical organization of phytobeneficial genes nifH and ipdC in the plant growth-promoting rhizobacterium Azospirillum lipoferum 4VI

Didier Blaha, Hervé Sanguin, Patrick Robe, Renaud Nalin, René Bally, Yvan Moënne-Loccoz
DOI: http://dx.doi.org/10.1016/j.femsle.2005.01.034 157-163 First published online: 1 March 2005


The physical organization of phytobeneficial genes was investigated in the plant growth-promoting rhizobacterium Azospirillum lipoferum 4VI by hybridization screening of a bacterial artificial chromosome (BAC) library. Pulsed-field gel electrophoresis gave an estimated 5.7-Mb genome size for strain 4VI and a coverage level of 9 for the BAC library. The phytobeneficial genes nifH (associative nitrogen fixation) and ipdC (synthesis of the phytohormone indoleacetic acid) are chromosomal, but no BAC clone containing both genes was found, pointing to the absence of any genetic island containing nifH and ipdC. A 11.8-kb fragment containing nifH was analyzed. Neighboring genes implicated in nitrogen fixation (nifH draT draG) or not (arsC yafJ and acpD) were organized as in A. brasilense. In contrast, the region located downstream of acpD contained four housekeeping genes (i.e. genes encoding DapF-, MiaB- and FtsY-like proteins, as well as gene amn) and differed totally from the one found in A. brasilense.

  • PGPR
  • Associative symbiosis
  • Evolution
  • Synteny

1 Introduction

Plant growth-promoting rhizobacteria (PGPR) of the genus Azospirillum are involved in associative symbiosis (cooperation) with the plant, and they have been used as inoculum for phytostimulation of crops, especially in the case of A. brasilense and A. lipoferum strains [13]. In addition to free-living nitrogen fixation, which is catalyzed by an enzymatic complex comprised of a nitrogenase reductase (NifH) and a dinitrogenase protein (NifDK) [4,5], phytostimulation entails also the production of phytohormones, noticeably the auxin indoleacetic acid (IAA) [57]. IAA is mainly produced via the tryptophan-dependent pathway, which involves an indole-3-pyruvate decarboxylase encoded by ipdC[5]. Phytohormone production leads to changes in root morphology and development, which is important for soil nutrient uptake [8,9].

Most information on the genetics of phytostimulation by Azospirillum has been derived from A. brasilense. Indeed, much less is available on A. lipoferum despite its ecological importance in the field [10,11] and the particularity of its interaction with plants [1214]. Therefore, it is important to know more about the genetic make-up of A. lipoferum strains, especially with regards to phytobeneficial genes. In A. brasilense Sp7, three nif operons located on a 40-kb fragment have been identified [15,16]. In bacteria associated with eukaryotic species, clustering of genes important for the interaction with the host is common, and the concept of genetic islands (i.e. pathogenicity and symbiotic islands) has been proposed for clusters of genes involved in distinct but related ecological functions [1719].

Therefore, in light of the recent recognition of the importance of genetic islands in the ecology of plant-associated bacteria, the objective of the current work was to determine the location of key phytobeneficial genes in the PGPR A. lipoferum 4VI and determine whether their clustering level is compatible with the concept of genetic island. To achieve this goal, nifH and ipdC were chosen as key phytobeneficial genes involved in distinct plant-beneficial modes of action. In order to assess whether the two genes are located in the vicinity of one another, a bacterial artificial chromosome (BAC) library of A. lipoferum 4VI was constructed using pulsed-field gel electrophoresis (PFGE) and validated. BAC clones carrying nifH were identified and analysed for the presence of other diazotrophy-relevant genes such as rpoN (encoding a sigma factor regulating nitrogen metabolism [20]) and draT (encoding a nitrogenase regulator [5,21]), as well as ipdC. Part of a BAC clone carrying genes related to nitrogen fixation was sequenced and the sequences compared with those available for A. brasilense.

2 Materials and methods

2.1 Hybridization procedure

A. lipoferum 4VI contains five indigenous plasmids ranging from 8 to 300 MDa [22]. To determine whether nifH and ipdC were chromosomal or plasmid-borne, hybridizations were carried out after transfer of DNA from Eckhardt electrophoretic gel [22] onto Gene Screen Plus nylon membrane (NEN Life Science Products, Le Blanc Mesnil, France), using a nifH or ipdC PCR fragment (see below for primers) as probe. The PCR fragments were extracted using the QIAquick Gel Extraction Kit (Qiagen, Courtaboeuf, France). The membrane was labelled by random priming (Roche, Meylan, France) and washed at 65 °C with SSC 2× for 5 min (twice), then with SSC 2× containing 1% SDS for 30 min (twice), according to supplier's instructions (NEN Life Science Products). Autoradiography was performed according to supplier's instructions (Amersham-Pharmacia, Orsay, France).

2.2 Estimation of genome size by pulsed-field gel electrophoresis

A. lipoferum 4VI[23] was grown at 28 °C with shaking in Tryptone Yeast extract (TY; Serlabo, Bonneuil sur Marne, France) until mid-log phase (OD600 of 0.6) and harvested by centrifugation at 4000 g. The pellet was washed once in TNEE buffer (Tris-base 10 mM, EDTA 10 mM, EGTA 10 mM, NaCl 1 mM) and resuspended in TEC buffer (Tris–HCl 6 mM, NaCl 1 M, EDTA 100 mM, BRIJ 58 0.5%, sodium desoxycholate 0.2%, N-lauryl Sarcosine 0.5%) to obtain an OD600 of 2. An identical volume of melted agarose InCert (1.8%) containing lysozyme (1 mg ml−1) was added to the cell suspension. The mixture was poured into 1-cm3 molds (Bio-Rad Laboratories, Hercules, California) and placed at 4 °C for 30 min. Once solid, the agarose was expelled from the molds and sliced in two with a razor blade. The two pieces were incubated together 1.5 h in 1 ml of TEC buffer containing lysozyme (1 mg ml−1) at 37 °C and then 1 h in 1 ml of TE1 buffer (Tris–HCl 10 mM, EDTA 1 mM) containing proteinase K (0.1 mg ml−1) at 55 °C. The plugs were washed in TE1 buffer and stored at 4 °C.

Restriction enzymes SwaI, SpeI and I-CeuI (New England Biolabs, Hitchin, England) were used to digest genomic DNA within agarose plugs. Individual plugs were placed in the appropriate restriction buffer (Biolabs) and kept 1 h on ice prior to digestion for 17 h with 40 units of enzyme in a liquid volume of 300 μl.

PFGE was carried out using a Chef-DRII System (Bio-Rad Laboratories). Saccharomyces cerevisiae chromosomes (Bio-Rad Laboratories) and λ ladder PFG (Biolabs) were used as molecular weight markers. Each run was performed with 1% agarose gel in 0.5 TBE buffer (Euromedex, Murdolsheim, France) at 6 V cm−1 for 24 h with a switch time of 60–120 s at 15 °C. Genome size was estimated by summing the size of all PFGE bands.

2.3 Construction of a BAC library of A. lipoferum 4VI

Genomic DNA from A. lipoferum 4VI was cloned in the BAC vector pIndigoBAC-5 (Tebu-Bio, Le Perray en Yvelines, France), as follows. Genomic DNA within agarose plugs was subjected to partial digestion for 17 h (one unit of enzyme in 300 μl; 37 °C) using HindIII (Boehringer Mannheim, Germany), an enzyme compatible with the cloning site of pIndigoBAC-5. PFGE was carried out with 0.8% low melting point agarose gel (Sigma, Saint Quentin Fallavier, France) in 0.5× TBE buffer at 4.5 V cm−1 for 24 h with a switch time of 20–40 s at 13 °C. Digested DNA between 100 and 300 kb was extracted from the gel by electroelution (Electro-Eluter 422, Bio-Rad), using 150 V for 5 min followed with 60 V overnight. DNA was eluted in elution buffer (Tris 2 M, EDTA 0.05 M, SDS 0.02 M and glacial acetic acid 5% v/v) after tension inversion at 200 V for 90 s. Ligation with pIndigoBAC-5 was performed in a final volume of 150 μl, using 9 U of T4-ligase (Roche Applied Science, Mannheim, Germany), 300 ng of eluted DNA and 50 ng of dephosphorylated, linearized pIndigoBAC-5. The mixture was incubated overnight at 16 °C. After heat inactivation of the ligase, 2 μl of the ligation mixture was used to transform 20 μl of competent E. coli DH10B cells (Invitrogen, Cergy Pontoise, France) by electroporation at 1.8 kV using an E. coli Pulser transformation apparatus (Gene pulser II, Bio-Rad). Transformed cells were incubated in SOC medium [24] at 37 °C for 1 h and plated on Luria Bertani (LB; [24]) agar with X-gal (50 μg ml−1), IPTG (25 μg ml−1) and chloramphenicol (12.5 μg ml−1). White colonies were picked up and stored in 96-well microtiter plates containing LB and chloramphenicol (12.5 μg ml−1).

The coverage level of the BAC library was assessed based on eight randomly-chosen BAC clones and the three clones described below. Each was digested using 10 units of NotI (Euromedex) in 30 μl volumes containing 500 ng of BAC clone, and the size of the inserts was determined by PFGE. The coverage level was computed as (mean size of the BAC inserts × number of BAC clones)/size of the genome.

2.4 Screening of the BAC library

The BAC library of A. lipoferum 4VI was screened for nifH by PCR. PCR amplification was performed according to the Taq polymerase manufacturer's instructions (Invitrogen), using 100 ng of DNA as template. PCR was done at 95 °C for 5 min, followed with 35 elongation cycles at 95 °C for 30 s, 55 °C for 30 s, 72 °C for 30 s, and an elongation at 72 °C for 5 min, as described by Poly et al. [25]. Primers PolF and PolR were used (Table 1), and they amplify a 360-bp region between positions 115 and 476 of nifH in Azotobacter vinelandii (Accession No. M20568). BAC clones selected based on nifH amplification were assessed for the presence of draT by PCR. PCR amplification was done at 95 °C for 5 min, followed with 80 elongation cycles at 95 °C for 30 s, 66 °C (for the first 35 cycles) or 65 °C (for the next 45 cycles) for 30 s, 72 °C for 30 s, and finally an elongation at 72 °C for 5 min. Primers F1903 and F1904 (Table 1) were designed based on sequences in A. lipoferum FS (Accession No. D55631) and A. brasilense Yu62 and Sp7 (Accession No. AF216815 and M87319, respectively). The expected amplicon was 635-pb long (between positions 91 and 725 of draT) in A. lipoferum FS.

View this table:
Table 1

Primers used in this study

Target genes and primersSequenceReference

2.5 Analysis of selected BAC clones

Selected clones were grown overnight at 37 °C with shaking in 25 ml of LB supplemented with chloramphenicol (12.5 μg ml−1). The BAC clones were extracted, as follows. The cells were harvested by centrifugation (13,000 rpm) for 15 min and the pellet was resuspended in 2 ml of solution 1 (glucose 50 mM, Tris–HCl 25 mM, EDTA 10 mM) containing 50 μl of RNase A (10 mg ml−1). Solution 2 (NaOH 0.2 M, SDS 1%) was added (2 ml). The suspension was mixed by inverting and incubated 5 min at room temperature, prior to adding 2 ml of solution 3 (KCH3 COO 5 M, glacial acetic acid 60% v/v). The suspension was incubated 7 min at room temperature and centrifuged 10 min at 13,000 rpm. The BAC clones were purified from the supernatant by ethanol precipitation and subjected to restriction analysis using either (i) NotI, (ii) EcoRI, (iii) NotI and EcoRI, (iv) HindIII, or (v) NotI and HindIII (Euromedex) in 30 μl volumes containing 10 units of each enzyme and 500 ng of BAC clone.

The DNA bands were transferred from agarose electrophoresis gel onto Gene Screen Plus nylon membranes and hybridized to a nifH amplicon, as described above (see Section 2.1). Hybridization was also carried out with rpoN and ipdC amplicons, which were obtained using primer pairs F1909/F1910 and F2465/F2466, respectively (Table 1).

Positive DNA bands were extracted from the gel using the QIAquick Gel Extraction Kit (Qiagen) and cloned into pBluescript (KS-) (OZYME, Schwalbach, Germany). Presence of nifH was checked by PCR. One positive clone (about 17 kb) was selected and the insert digested with restriction enzyme SalI prior to subcloning into pBluescript (KS-). DNA sequencing (Genome Express, Grenoble, France) was performed on both strands of one clone of each restriction fragment using the universal primers T3 and T7. Sequence assembling was performed by compiling BLAST results and restriction map of the 17-kb DNA fragment.

Sequences obtained in this work are available at accession number AY786992.

3 Results and discussion

3.1 Location of phytobeneficial genes among cell replicons of A. lipoferum 4VI

To assess whether nifH and ipdC were chromosomal or present on one of the five indigenous plasmids in A. lipoferum 4VI, DNA hybridization of the replicons of A. lipoferum 4VI was performed after transfer from Eckhardt electrophoretic gel. Results indicated that both genes were chromosomal (data not shown). This result came as no surprise for nifH, as this gene was shown to be harbored by the chromosome in A. brasilense[16,26] and in A. lipoferum SP59 and UQ1618 [26]. To our knowledge, the chromosomal location of ipdC has never been reported in Azospirillum spp. Other genes (homologous to e.g. nod and genes encoding flagella or polysaccharide synthesis) that may be important for the interaction with the plant are located on the 90-MDA plasmid pRhico in A. brasilense Sp7 [27,28].

3.2 Construction of a BAC library of A. lipoferum 4VI and determination of the size of BAC clones

To assess whether nifH and ipdC were located in the vicinity of one another, a BAC library of A. lipoferum 4VI was constructed. To reach this objective, genomic DNA of A. lipoferum 4VI was digested using HindIII and cloned on 20 separate occasions. A total of 390 BAC clones were obtained and stored in 96-well microtiter plates. The size of the insert was determined by PFGE for eleven clones from different cloning experiments and gave 136 ± 118 kb (from 32 to 400 kb).

3.3 Genome size of A. lipoferum 4VI and coverage level of the BAC library

Knowledge of the genome size of A. lipoferum 4VI was needed to estimate the coverage level of the BAC library. To this end, the genome of A. lipoferum 4VI was cleaved into large DNA fragments with restriction enzymes, and the fragments separated by PFGE. Since the genus Azospirillum has a GC content between 64% and 71% and A. lipoferum 68%[29], restriction enzymes like SwaI, SpeI and I-CeuI were chosen because they recognize AT-rich targets that are respectively 8, 6 and 10 bp long. Digestion was done with one or two enzymes and generated nine or ten fragments (Fig. 1). Summing the size of the DNA fragments in each of the four digestions gave 5330, 5680, 5800, and 6080 kb, i.e. a genome size of 5720 ± 310 kb. This is less than values reported for A. lipoferum strains Sp59b (9.7 Mb) and JA25 (7.9 Mb) [30], but the latter display a very large extrachromosomal genome, exceeding that of strain 4VI by more than 2 Mb [22,30]. Contrasted genome sizes (4.8–9.7 Mb) were also reported when comparing different Azospirillum species [30].

Figure 1

PFGE analysis of the genome of A. lipoferum 4VI (a) and estimation of genome size based on summing the size of the restriction fragments in each of four digestions (b). Sc, size standard consisting in the chromosomes of S. cerevisiae.

Based on a genome size of 5720 kb, the coverage level of the BAC library was about 9. This means that the library was appropriate to assess whether or not nifH and ipdC were located in the vicinity of one another by searching for clones harboring both genes.

3.4 Selection of BAC clones containing nifH and search for ipdC

Three clones were identified following screening of the BAC library of A. lipoferum 4VI for nifH by PCR. The presence of nifH was confirmed by sequencing, and the three sequences were identical to one another. Each of the three BAC clones contained also draT, as indicated by PCR amplification and sequencing. Once extracted, the three BAC clones generated each 7–10 restriction fragments in each digestion (data not shown). Southern hybridization, using PCR-amplified nifH as a probe, gave a positive response for one restriction fragment in each digestion for each of the three BAC clones. The restriction fragments that hybridized (i.e. EcoRI fragments for two clones and HindIII fragment for the third one) were 15–18 kb long based on PFGE results (data not shown). The membranes were also hybridized using a rpoN amplicon as probe, and no hybridization signal was obtained.

The three BAC clones containing nifH were studied by hybridization to ipdC, and no signal was found, meaning that nifH and ipdC are not located on a same BAC clone. Thus, we did not find evidence for a close location of nifH and ipdC on the chromosome of A. lipoferum 4VI, pointing to the absence of any genetic island containing these two genes. It must be kept in mind that the approach followed in this work has two main limitations. First, screening relied on a small number of phytobeneficial genes. For instance, the genes involved in synthesis of phytohormones other than IAA are not documented. Similarly, certain Azospirillum strains may be involved in biocontrol interactions, by protecting plants from the phytopathogenic bacterium Pseudomonas syringae pv. tomato [31] or the parasitic plant Striga[13,32], but the biocontrol genes implicated are unknown. Second, the scope of the assessment was limited by the size of the BAC clones, considering that genetic islands are usually 10–200 kb long [1719]. It will be useful to reassess the location of phytobeneficial genes once whole genome sequencing data are available. For instance, sequence analysis of plasmid pRhico in A. brasilense Sp7 evidenced several genes (but not nifH or ipdC) of potential importance for colonization of plant roots [28], and the authors proposed a comparison with the pSym plasmid(s) of rhizobia.

3.5 Identification of genes located near nifH

A restriction fragment about 17 kb in size and carrying nifH was chosen for further analysis. As expected from the results of nifH hybridization, this BAC insert was chromosomal based on hybridization (done as described in 2.1) performed after separation of plasmids by Eckhardt electrophoresis (data not shown). Sequencing of restriction fragments derived from the 17-kb BAC insert gave a GC content of 68%, which is consistent with the overall GC content of A. lipoferum. It suggests that this region was not acquired by recent horizontal transfer from a bacterium of contrasted GC content. Nine ORFs spanning a 11.4-kb region located next to nifH were evidenced (Fig. 2).

Figure 2

Organization of a 11.4-kb region located upstream of nifH in A. lipoferum 4VI. The name of the ORFs is indicated when available, and the size of DNA fragments is shown (bp). The sequence of the 2320-bp non-coding fragment displays (i) three GG-N12-GC consensus sequences of RpoN-dependant nif promoters (DPE, DPE1 and DPE2; in bold and underlined), (ii) a NifA binding site located within the nifH promoter (UAS; in bold), (iii) three AT-rich regions (highlighted in grey), (iv) two putative ribosome binding sites (AGGA and TCCT) located upstream from respectively nifH and draT (in bold and boxed), and (v) the initiation codons ATG and GAC (in bold and boxed). The detail of a 1808-bp fragment located between two AT-rich regions (II and III) is not displayed because no particular sequence motif was identified.

Part of the region next to nifH was comparable to the one in A. brasilense strains Yu62 (for ORFs 1–5; 6794 bp in total; [21]), Sp7 (for ORFs 1 and 2; [16]) and FS (for ORFs 1–3; [33]) in terms of gene organisation and sequence, whereas the rest differed from the situation found in A. brasilense. The region comparable in A. lipoferum 4VI and A. brasilense is as follows. From the 3′ end, a 2320-bp non-coding sequence comprised of potential regulatory elements was evidenced. These potential regulatory elements include (i) the consensus sequence GC-N12-CC (or GG-N12-GC), located close to nifH, which corresponds to a downstream promoter element (DPE) and is typically present in RpoN-dependant nif promoters [34], (ii) two other DPEs (named DPE1 and DPE2 [34]) found at a further distance from nifH (one of them close to the initiation codon of ORF 1), (iii) the consensus sequence TGT-N10-ACA, which is the NifA binding site located within the nifH promoter and corresponds to an upstream activator sequence (UAS) [34], (iv) three AT-rich regions (I, II and III), with a GC content of respectively 19%, 33% and 28%, thought to be a critical promoter element influencing the frequency of transcription initiation [34], and (v) two putative ribosome binding sites located 9 and 7 bp away from the initiation codons of respectively nifH and ORF 1.

ORFs 1 and 2 are homologous to draT and draG of A. brasilense Sp7 (Accession No. M87319; Table 2). The sequence between draT and draG is 76 bp long and does not contain a ribosome binding site. The next three ORFs are homologous to genes of A. brasilense Yu62 not implicated in nitrogen fixation i.e. arsC (encoding a putative arsenate reductase) for ORF 3, yafJ (encoding a glutamine amidotransferase) for ORF 4 and acpD (encoding a putative acyl carrier protein (ACP) phosphodiesterase) for ORF 5 (Table 2). Therefore, they correspond to genes with seemingly-unrelated functions.

View this table:
Table 2

Percentages of homology (%) between ORFs 1-9 identified in A. lipoferum 4VI and Genbank data

ORFa in A. lipoferum 4VIStrain with highest protein homologyIdentity DNA sequenceIdentity deduced protein sequenceSimilarity deduced protein sequence
ORF 1 (DraT)A. brasilense Sp7837483
ORF 2 (DraG)A. brasilense Sp7868091
ORF 3 (ArsC)A. brasilense FSb837583
ORF 4 (YafJ)A. brasilense Yu62878994
ORF 5 (AcpD)A. brasilense Yu62897384
ORF 6 (Amn)B. bronchiseptica RB50c877987
ORF 7 (FtsY)M. loti MAFF3030994355
ORF 8 (MiaB)M. magnetotacticum MS-16985
ORF 9 (DapF)M. magnetotacticum MS-17588
  • a The corresponding deduced protein is indicated in parenthesis.

  • b A. brasilense Yu62 when considering the DNA sequence.

  • c B. cepacia R1808 when considering the DNA sequence.

In contrast, ORFs 6–9 differ from those found in the corresponding region in A. brasilense Yu62. ORF 6 is homologous to amn of Bordetella bronchiseptica RB50 (Accession No. BX640446) and the deduced protein sequence to the nucleoside phosphorylase protein Amn of Burkholderia cepacia R1808 (Accession No. ZP00223851) (Table 2). The last three ORFs are not homologous to Genbank entries based on BLASTN analysis. However, the deduced protein sequences are homologous to FtsY (involved in cell division; 55% similarity only) of Mesorhizobium loti (for ORF 7), the 2-methylthioadenine synthetase MiaB (involved in translation and ribosomal structure) of Magnetospirillum magnetotacticum (for ORF 8) and the diaminopimelate epimerase DapF (implicated in peptidoglycane synthesis) of M. magnetotacticum (for ORF 9) (Table 2). The presence of house-keeping genes is consistent with a chromosomal location for this DNA fragment.


D.B. and H.S. contributed equally to this work. We are grateful to V. Segealet, D. Mourier des Gayets and M. Boyer for technical assistance. We are indebted to J. Haurat (UMR CNRS 5557 Ecologie Microbienne, Lyon1) and M.S. Mirza (National Institute for Biotechnology and Genetic Engineering, Faisalabad, Pakistan) for help with hybridization and design of ipdC primers, respectively. We thank Jean-Pierre Wisniewski (ENS, Lyon, France), Pierre Pujic (LibraGen S.A.), Patrick Mavingui, Claire Prigent-Combaret and Philippe Normand (UMR CNRS 5557 Ecologie Microbienne, Lyon1) for useful discussion.


  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].
View Abstract