Tools for genetic manipulation of the plant growth-promoting bacterium Azospirillum amazonense

Isolation of glnB and glnK genes from A. amazonense

The PII proteins are pivotal regulators of the nitrogen metabolism, controlling the activities of transporters, enzymes and transcriptional factors implicated in this process [9, 10]. These proteins are highly conserved and are widely distributed throughout prokaryotes [11]. In Proteobacteria in particular, there are two main types of PII proteins, GlnB and GlnK. In this work, two PII protein encoding genes from A. amazonense were isolated. Southern blot analysis utilizing a PCR-generated glnB fragment as the probe revealed two distinct signals in the genomic DNA of A. amazonense digested with SalI: the strongest at the ~2 kb DNA fragments and the weakest at the ~3 kb DNA fragments (data not shown). Based on these results, a genomic library enriched with 2-3 kb SalI fragments was constructed. The library was partially sequenced and a PII protein homolog was identified. The deduced amino acid sequence of this gene was found to be highly similar to that of the GlnZ proteins (GlnK-like homologs) from A. brasilense and Azospirillum sp. B510 (75% identity and 86% similarity), and Rhodospirillum. centenum (73% identity and 86% similarity). Arcondéguy et al. (2001) [12] suggested that the glnZ genes should be termed glnK, since their deduced proteins are highly similar to the GlnK proteins. Furthermore, there is a functional correspondence between these proteins, as both regulate the uptake of ammonium through the AmtB transporters [13–15]. Therefore, we adopted the glnK designation for this A. amazonense homolog, mainly because this nomenclature could facilitate comparisons between other bacterial systems.

The glnK gene from A. amazonense is flanked by the aat gene in the downstream region, which codes a putative aspartate aminotransferase and the ubiH gene in the upstream region, which codes an enzyme implicated in ubiquinone biosynthesis (Figure 1). This genetic organization resembles that found in other species from the Rhodospirillales order, namely A. brasilense, Azospirillum sp. B510 and R. centenum.

Since the glnB gene was not found in the genomic library, the Inverse PCR methodology was carried out to isolate this gene. A ~2 kb amplicon that contained the glnB gene was obtained (data not shown). It was found that the protein of this gene displays 92% identity and 98% similarity to the GlnB proteins from Azospirillum sp. B510 and A. brasilense, and 96% identity and 98% similarity to the GlnB protein of R. centenum. The glnB gene is located upstream of the glnA gene (glutamine synthetase), the same genetic context observed in these bacteria (Figure 1).

In A. brasilense, glnB has a key role in nitrogen fixation because its protein product regulates the activity of NifA, the transcriptional factor of nitrogen fixation [16, 17].

Furthermore, both of the GlnZ (GlnK-like homolog) and GlnB proteins are also implicated in the DraT/DraG system, which regulates dinitrogenase reductase activity by covalent modifications [15]. However, Fu et al. [18] verified that A. amazonense does not have the DraT/DraG system. Hence, in the near future, the interaction targets of the PII protein in A. amazonense should be determined to better understand their roles in the nitrogen metabolism of this microorganism.

Antibiotic minimum inhibitory concentration

Most DNA manipulation is dependent on the use of vectors containing resistance markers to antibiotics [19, 20]. In a previous work using antibiotic susceptibility test discs, Magalhães et al. (1983) [5] showed that A. amazonense is sensitive to kanamycin and gentamicin, tolerant to tetracycline, and resistant to penicillin. In this work, we determined the minimum inhibitory concentrations of A. amazonense to antibiotics that are normally used to provide a selective pressure for vectors.

The susceptibility of A. amazonense to kanamycin and gentamicin was confirmed, since no growth was observed in concentrations of these antibiotics of 0.25 μg/mL; therefore, vectors that contain selection markers for these compounds are appropriate for use.

High concentrations of ampicillin (128 μg/mL) were required for complete growth inhibition, showing that A. amazonense is also resistant to this beta-lactam antibiotic.

It is worth noting that the growth of A. amazonense was absent in a relatively high concentration of tetracycline (32 μg/mL), indicating that this species is, in fact, resistant to this antibiotic, instead of tolerant, as pointed out by Magalhães et al. [5]. These findings about the latter two antibiotics are relevant because they could be used in counter-selection procedures in conjugation experiments, as there is a variety of E. coli strains that are susceptible to them.

Conjugation

Conjugation mediated by E. coli is the standard DNA transfer technique of the Azospirillum genus [21]. Therefore, in this work the conjugation ability of A. amazonense was evaluated.

Unlike A. brasilense, A. amazonense cannot grow in LB medium. Furthermore, E. coli cannot grow in M79 medium; therefore, the first concern was to establish a medium that provided appropriate growth conditions for the donor and recipient strains. Hence, different medium compositions, containing distinct ratios of M79 and LB media (varying from 1:1 to 9:1), were prepared. The medium mixture of M79:LB at a proportion of 8:2 was the most suitable for culturing both bacteria and it was designated as MLB medium.

Another requisite for the conjugation procedure is to select vectors that contain proper selection markers that are mobilizable and able to replicate inside the receptor cell [19, 20]. Therefore, the pHRGFPGUS (pBBR1 replication origin) and the pPZPLACEYFP (pVS1 replication origin) plasmids were tested by tri-parental conjugation. These plasmids are mobilizable broad-host vectors harboring kanamycin resistance markers and fluorescent protein coding genes, which could promptly report achievement of the DNA transfer. The transconjugants exhibited kanamycin resistance and fluorescence. The conjugation frequencies were 3.8 × 10^-8 per recipient cell for the pHRGFPGUS vector and 3.8 × 10^-7 for the pPZPLACEYFP vector.

Different ratios of recipient to donor and helper strains (1:1:1, 5:1:1, 10:1:1 and 20:1:1) were also tested. The best efficiencies were obtained with the ratios 10:1:1 and 5:1:1; however, no obvious differences between these latter ratios were observed (data not shown).

In conclusion, conjugation is an appropriate method for DNA transfer to A. amazonense. Although only tri-parental mating was tested in this work, it is important to mention that bi-parental conjugation could be an alternative test, due to the possibility of increasing the conjugation efficiencies.

Electrotransformation

Since suitable vectors for A. amazonense were defined and since conjugation is a time-consuming procedure, the transformation of A. amazonense via electroporation was tested.

The eletrocompetence of the cells is greatly influenced by the growth phase [22]. Therefore, A. amazonense cells were harvested at different growth phases to evaluate their effect on electroporation efficiency. Cells from the late-log phase (OD₆₀₀ 1) and the stationary phase (OD₆₀₀ 2) were not electrocompetent. Electroporation utilizing cells from the early-log growth phase (OD₆₀₀ 0.12) generated a significant number of transformants. Therefore, all subsequent tests were performed utilizing cells cultivated at this growth phase.

In the electrocompetent cell preparation, the cells were harvested and washed continuously until the solution had a low-ionic strength. The MgCl₂ HEPES-sucrose buffer was found to be the most suitable solution for the preparation of A. amazonense electrocompetent cells. Although 10% glycerol solution is commonly used for electrocompetent cell preparation in a diverse number of species (including A. brasilense), it was not appropriate for A. amazonense, as no transformants were obtained when this solution was used.

Different electroporation parameters were tested. The increase in electrical field strength had a positive effect on electroporation efficiency (Figure 2A). The highest electrical field strength tested was 12.5 kV/cm, and this condition was found to be the most efficient, generating about 8000 transformants/μg of pHRGFPGUS (Figure 2A). The effect of pulse length on electroporation efficiency was also investigated (Figure 2B). A pulse length of 4.3 ms (electroporation apparatus set at 200Ω) was the most efficient. The pulse lengths of 7.3 ms (400 Ω) and 10.5 ms (600 Ω) had a dramatic negative effect on transformation efficiency, where only few transformants were obtained (Figure 2B). These conditions are in agreement with the general parameters of bacterial electroporation [22–24].

In conclusion, the transfer of DNA to A. amazonense by means of electroporation was demonstrated. Although the efficiency of electrotransformation was far from desirable, this result is supported by previous works showing that bacteria closely related to A. amazonense, such as A. brasilense [25], R. rubrum [26] and Magnetospirillum gryphiswaldense [27], are recalcitrant to electrotransformation. Nonetheless, this technique is an easy and a rapid method of DNA transfer to the cells of A. amazonense.

Site-directed mutagenesis

Site-directed mutagenesis is a fundamental tool for correlating cellular functions with specific regions of the DNA. Therefore, once DNA transfer techniques were established for A. amazonense, the next step was to determine a site-directed mutagenesis protocol for this species.

Most of the A. brasilense mutants have been generated by the disruptive insertion of an antibiotic resistance cassette into the target gene [14, 28–30]. This approach is not recommended when the target gene composes an operon, since the resistance cassette could introduce a polar effect on the expression of the surrounding genes and, consequently, make it difficult to assign a mutant phenotype to the disrupted gene [31].

Therefore, in this work, a site-directed mutagenesis methodology that generates in-frame mutants without the disruptive insertion of a resistance cassette was evaluated. The glnK gene was selected for this methodology because subsequent studies of our laboratory will aim to determine the role of the PII proteins in A. amazonense metabolism.

The mutagenesis methodology is depicted in Figure 3A. Firstly, an amplicon containing an in-frame deletion of the glnK gene was generated through Crossover PCR, and it was subsequently cloned in the suicide replacement vector pK19MOBSACB, generating the pKΔK plasmid. This vector contains a kanamycin resistance gene (positive selection marker) that allows the selection of bacteria that would have integrated the plasmid into the chromosome. This vector was delivered to A. amazonense by means of conjugation (the carbon source utilized was maltose instead of sucrose) and one colony resistant to kanamycin was obtained, suggesting that the integration of the plasmid was successfully accomplished. The sacB gene (negative marker selection) of the vector is lethal in the presence of sucrose; therefore, the merodiploid strain (containing both wild-type and mutant alleles) was unable to grow in M79 (containing 10 g/L of sucrose). Subsequently, expecting that a recombination event could replace the wild-type allele, the merodiploid strain was cultured for many generations in M79 containing maltose instead of sucrose. Finally, this culture was plated in M79 containing sucrose to eliminate the bacteria that did not accomplish the second recombination event. Seven sucrose-resistant/kanamycin-sensitive colonies were chosen for PCR evaluation of the substitution of the mutant allele for the wild-type gene. Four colonies presented a band of 121 bp, indicating that the wild-type glnK was successfully substituted, whereas three colonies presented the 361 bp band, corresponding to the wild-type allele (Figure 3B). Furthermore, an additional PCR with primers flanking the recombination sites was performed, and it also demonstrated a reduction of the amplicon sizes originated from the glnK mutants in relation to the wild type strain (Figure 3C). This latter result demonstrates that recombination occurred in the target site.

Altogether, these results show that an in-frame glnK gene mutant strain of A. amazonense was successfully generated by this mutagenesis system.

Reporter gene system

The study of promoters is fundamental to elucidation of the genetic regulatory mechanisms of bacterial species. Up until now, there has been neither a report of heterologous gene expression in A. amazonense, nor a reporter system designed for this species. In this work, a reporter system based on expression of the Enhanced Yellow Fluorescent Protein (EYFP) was developed to analyze the regulatory regions of A. amazonense genes in vivo.

In silico analysis using a Sinorhizobium meliloti sigma 70 promoter weight matrix revealed that the genes aat, glnK, and glnB of A. amazonense have putative promoter sequences in their upstream regions (Figure 4). In E. coli, sigma 70 is considered to be the vegetative sigma factor, as it is responsible for the expression of the majority of genes [32, 33]. Therefore, one could expect that these putative A. amazonense sigma 70 promoters could act under standard laboratory growth conditions (aerobic environment, 35°C and M79 medium). Consequently, different vectors were constructed to determine the activity of the upstream regulatory sequences of A. amazonense genes in the expression of EYFP.

The lac promoter was utilized as a positive control since there is a report showing that this promoter has high activity in A. brasilense [34]. Two different vectors were constructed with the lac promoter, one derived from pPZPLACEYFP (pVS1 replicon) and the other derived from pHRGFPGUS (pBBR1 replicon). The upstream regions of the genes glnB, glnK, and aat were cloned into the pHRGFPGUS derivative.

The lac promoter had the best score in the in silico analysis from among the promoters detected, and, as expected, the highest fluorescence levels were observed in the lac constructions (Figure 5). The difference in the fluorescence levels between the pHRLACEYFP and pPZPLACEYFP transformants could be a product of the difference in the copy number between these vectors.

Although the in silico analysis revealed that the glnK promoter had a higher score than the aat and glnB promoters, its in vivo activity under the conditions tested did not differ significantly from the negative controls (without promoter and without plasmid) (Figure 5). One of the possible reasons for this is that this gene was repressed under these conditions. The reporter gene analysis also demonstrated that the aat and glnB promoters were active under the conditions tested, although the aat promoter showed a higher activity than the glnB promoter.

These observations show that a reporter system based on EYFP can be used for in vivo promoter analyses in A. amazonense.

Bacterial strains, plasmids, and growth conditions

Azospirillum amazonense was cultured in M79 medium (10 g/L of sucrose as the carbon source, 0.1 g/L of K₂HPO₄, 0.4 g/L of KH₂PO₄, 0.2 g/L of MgCl₂.7H₂O, 0.1 g/L of NaCl, 0.4 g/L of yeast extract, pH 6.5) [35] at 35°C (unless stated otherwise). The M79 agar plates contained 2.5 mg/L of Bromothymol Blue. Escherichia coli XL1-Blue was cultured in LB medium at 37°C [36].

DNA techniques

Isolation of glnB and glnK genes from A. amazonense

The genomic library enriched with 2-3 kb SalI DNA fragments was constructed as follows: the genomic DNA was digested with SalI and subsequently separated in agarose gel by electrophoresis. The 2-3 kb fragments were excised from the agarose gel and purified. Finally, these fragments were cloned in the pUC18 plasmid. This genomic library was partially sequenced and the glnK gene was identified using BLAST searches.

Inverse PCR for glnB isolation was performed according to Sambrook and Russell (2001) [36]. Azospirillum amazonense genomic DNA was digested with SalI and subsequently circularized. The PCR was performed with the glnB_sfint and revsf_glnBint primers (Table 2) and the circularized SalI DNA as a template. The 5' portion of the glnA gene was isolated by PCR with the revsf_glnBint and glnA_aa_do primers (Table 2) and the genomic DNA as a template.

The DNA sequencing was performed using a MEGABACE automated platform (Centro de Biotecnologia, UFRGS). Sequences were assembled using the Staden software package [39]. Gene annotation was carried out by Artemis software version 12.0 [40] along with BLAST software using the NCBI database http://blast.ncbi.nlm.nih.gov/Blast.cgi. Both sequences were deposited in the NCBI nucleotide database under the following access numbers: glnB region [GenBank:HM161849] and glnK region [GenBank:HM161850].

Antibiotic minimum inhibitory concentration test

The minimum inhibitory concentration of A. amazonense to the antibiotics (gentamicin, kanamycin, tetracycline, and ampicillin) was basically evaluated as described in Andrews (2001) [41]. The antibiotics were serially diluted in 1 mL of M79 medium at concentrations from 256 μg/mL to 0.5 μg/mL. An overnight culture of A. amazonense was diluted to 4 × 10⁴ cells/mL. One milliliter of this dilution was added to one milliliter of M79 medium containing the appropriate antibiotic concentration. The cells were cultivated in a rotary shaker at 150 rpm for 40 h at 35°C.

Conjugation

Conjugation was basically carried out as described by Clerico et al. (2007) [42]. However, some modifications were made as follows: overnight cultures of A. amazonense Y2 (receptor), E. coli XL1-Blue containing the plasmid pRK2013 (helper), and E. coli XL1-Blue containing the appropriate plasmid (donor) were used. Approximately 1 mL of the A. amazonense culture with an OD₆₀₀ = 2 (1.3 × 10⁹ cfu/ml) was mixed with 1 mL of each helper and donor cultures with an OD₆₀₀ = 0.2 (2 × 10⁸ cfu/mL) (ratio 10:1:1), unless stated otherwise. This mixture was harvested by centrifugation at 6000 g for 2 min and then resuspended in 100 μL of MLB medium (LB and M79 mixture at a proportion of 8:2), and this volume was then spotted onto MLB agar and incubated for 20 h at 35°C. Following this, the cell mass was resuspended in 200 μL of M79 medium and plated on M79 medium containing the appropriate antibiotic.

Electroporation

The preparation of cells was based on the protocol described by Schultheiss and Schüler (2003) [27]. A 3 mL overnight culture of A. amazonense was inoculated in 250 mL of M79 and the cells were cultivated to an OD₆₀₀ of ~0.12 (early-log growth phase), unless stated otherwise. From this point, all manipulations were conducted on ice. The cells were incubated in ice for 30 min and then harvested by centrifugation at 5000 g for 20 min at 10°C. The cells were resuspended in 100 mL of electroporation buffer (pH 6.5 HEPES 1 mM, MgCl₂ 1 mM, and sucrose 200 mM) and again harvested by centrifugation (20 min at 5000 g). Subsequently, the cells were resuspended in 40 mL of electroporation buffer and again harvested by centrifugation. At the end, the cells were resuspended in 250 μL of electroporation buffer (final concentration of ~10¹⁰ cfu/mL), distributed in aliquots of 40 μL, and frozen in liquid nitrogen. Cell electroporation was carried out as follows: the 40 μL aliquot was mixed with 50 ng of the pHRGFPGUS vector and electroporated through a Gene Pulser apparatus (Bio-Rad Laboratories Inc.) with 12.5 kV/cm, 25 μF and 200 Ω, unless stated otherwise. After electrical discharge, the cells were resuspended in 500 μL of M79 medium and incubated at 35°C for 3 h in a rotary shaker at 150 rpm. Subsequently, the cells were plated on solid M79 medium containing 20 μg/mL of kanamycin and incubated for 2 days at 35°C.

Gene mutagenesis

Site-directed mutagenesis was based on a protocol described by Eggeling and Reyes (2005) [43]. In summary, the flanking regions of the glnK gene were amplified using the primers KglndelA_EcoRI/KglndelB and KglndelC/KglndelD_BamHI (Table 2). These amplification products were joined by Crossover PCR [31] using the primers KglndelA_EcoRI/KglndelD_BamHI (Table 2) and cloned in pK19MOBSACB digested with EcoRI and BamHI, generating the plasmid pKΔK (Table 1). Subsequently, the vector pKΔK was transferred to A. amazonense by conjugation, as previously described, except that the medium utilized was MLB containing maltose instead of sucrose (10 g/L) and ampicillin (100 μg/mL) for the counter-selection of E. coli. A kanamycin-resistant colony was isolated and cultured overnight in 3 mL of M79 (containing 10 g/L of maltose instead of sucrose). The culture was serially diluted and plated on M79 medium (containing 10 g/L of sucrose). Fifty sucrose-resistant colonies were replica plated onto both kanamycin-containing and pure M79 agar plates. Seven kanamycin-sensitive/sucrose-resistant colonies were submitted to Touchdown-PCR to identify those that had replaced the wild-type glnK gene with the mutant allele. The Touchdown-PCR was performed using the primers glnK_NdeI_up and glnK_BamHI_do (Table 2) under the following conditions: an initial denaturing step of 94°C for 5 min; 15 cycles of 94°C for 30 s, 60°C-56°C for 30 s (for each three cycles one degree was decreased), and 72°C for 30 s; 15 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s. The PCR utilizing the primers Conf_glnK_up and Conf_glnK_do (Table 2), which flank the recombination sites of the glnK region, was carried out in the same way as standard PCR procedures [36].

Gene reporter system

The upstream sequences of the genes utilized in this work were analyzed by Patser (available on the RSAT webserver) [44] with an S. meliloti sigma 70 factor weight matrix [33].

A series of reporter vectors was developed to evaluate the activity of different promoters (Table 1). The upstream regions of the glnB and glnK genes were amplified utilizing the primers listed in Table 2. Subsequently, these amplicons were cloned into the pEYFP vector at the NcoI and BamHI sites, generating pPBEYFP and pPKEYFP plasmids, respectively. After evaluation of the integrity of these amplicons by automated sequencing, the HindIII-EcoRI fragment, containing the promoter-eyfp fusion, was transferred to the HindIII-EcoRI fragment of pHRGFPGUS, which contains the replication origin, the mobilization site, and the kanamycin resistance marker, generating the pHRPBEYFP and pHRPKEYFP plasmids, respectively.

The pHRAATEYFP plasmid was constructed in the following way: the NcoI-BglII fragment of pAAGLNK, containing the upstream region of the aat gene, was transferred to pEYFP, generating the plasmid pAATEYFP. The HindIII-EcoRI fragment from this plasmid was transferred to the HindIII-EcoRI fragment of pHRGFPGUS, generating pHRAATEYFP.

The negative control plasmid, which did not contain a promoter, was constructed as follows: the NcoI-BamHI fragment of pEYFP was transferred to the HindIII-EcoRI fragment of pHRGFPGUS, forming the plasmid pHREYFP.

The positive control plasmid pHRLACEYFP is a fusion of the major EcoRI-EcoRV fragment of pHRGFPGUS with the PvuII-EcoRI fragment of pEYFP.

All of the plasmids were transferred to A. amazonense by tri-parental mating or electroporation. The promoter activity assay was basically performed as described in MacLellan et al. (2006) [33]. Azospirillum amazonense containing the reporter vectors was cultivated in M79 medium overnight in a rotary shaker at 35°C. The cells were washed in sterile saline solution (0.85% NaCl) and resuspended in this same solution to an OD₆₀₀ of between 0.06-0.39. Two hundred microlitres of the cell suspensions were deposited on black microtiter plates and fluorescence was measured with an excitation wavelength of 488 nm and an emission wavelength of 527 nm. The optical densities of the cell suspensions were measured at 600 nm on clear microtiter plates. Specific fluorescence was obtained by dividing the fluorescence by the optical density. Statistical analysis was performed using SAS JMP8 software: the specific fluorescence data was subjected to the natural logarithm to homogenize the variances (tested by Levene's test) and subsequently submitted for ANOVA/Tukey HSD tests (P < 0.01).