Correspondence: Lucy Seldin, Laboratório de Genética Microbiana, Instituto de Microbiologia Prof. Paulo de Góes, Universidade Federal do Rio de Janeiro, Centro de Ciências da Saúde, Bloco I, Ilha do Fundão, Rio de Janeiro, RJ, Brazil. Tel.: +55 21 2562 6741; fax: +55 21 2560 8344; e-mails: email@example.com and firstname.lastname@example.org
The diversity of nitrogen-fixing bacteria was assessed in the rhizospheres of two cultivars of sorghum (IS 5322-C and IPA 1011) sown in Cerrado soil amended with two levels of nitrogen fertilizer (12 and 120 kg ha−1). The nifH gene was amplified directly from DNA extracted from the rhizospheres, and the PCR products cloned and sequenced. Four clone libraries were generated from the nifH fragments and 245 sequences were obtained. Most of the clones (57%) were closely related to nifH genes of uncultured bacteria. NifH clones affiliated with Azohydromonas spp., Ideonella sp., Rhizobium etli and Bradyrhizobium sp. were found in all libraries. Sequences affiliated with Delftia tsuruhatensis were found in the rhizosphere of both cultivars sown with high levels of nitrogen, while clones affiliated with Methylocystis sp. were detected only in plants sown under low levels of nitrogen. Moreover, clones affiliated with Paenibacillus durus could be found in libraries from the cultivar IS 5322-C sown either in high or low amounts of fertilizer. This study showed that the amount of nitrogen used for fertilization is the overriding determinative factor that influenced the nitrogen-fixing community structures in sorghum rhizospheres cultivated in Cerrado soil.
Sorghum bicolor is the fourth most important world cereal grain, following wheat, rice, and corn. Worldwide, sorghum is grown for grain, forage, syrup and sugar, and industrial uses of stems and fibers on more than 40 million hectares, especially in China, India, Brazil and Africa. In Brazil, where tropical and subtropical summer rainfall climate predominates, sorghum has emerged as an excellent alternative in agriculture due to its ability to grow well during the dry season.
Fixed nitrogen is often the limiting factor for crop productivity and affects plant communities and ecosystems in all scales (Tan et al., 2003). In order to increase the yield of agronomical crops, nitrogen fertilization is widely used. However, commercial nitrogen fertilizers are usually an expensive means of supplementing soil for plant growth. Furthermore, improper or excessive use of fertilizer can lead to nitrate pollution of ground or surface water. Therefore, biological nitrogen fixation can effectively contribute to the nitrogen uptake, and the identification of the major taxonomic groups of bacteria that might contribute to the nitrogen input in different economic important plants is of great relevance. Numerous nitrogen-fixing bacterial isolates have been obtained previously from sorghum roots and were identified as members of the Enterobacteriaceae, such as Klebsiella pneumoniae, Enterobacter cloacae and Erwinia herbicola (Pedersen et al., 1978). Other studies have demonstrated the presence of nitrogen-fixing organisms in sugarcane (James et al., 1994), maize (da Silva et al., 2003), winter wheat (Pedersen et al., 1978) and rice (Ladha et al., 1997; Hirano et al., 2001), among other grasses.
The biological nitrogen fixation is carried out by prokaryotic organisms harboring the nitrogenase enzyme encoded by nif, anf or vnf genes (Joerger et al., 1991; Waugh et al., 1995). The nifH gene encodes the iron protein subunit of nitrogenase, and a good correlation between nifH gene and 16S rRNA gene phylogeny has been found for diazotrophic microorganisms (Young, 1992; Zehr et al., 2003). However, using the nifH gene as a molecular marker in natural environments can be seen as an advantage, as it provides evidence for potential nitrogen fixation (Young, 1992). Moreover, the nifH gene is the most thoroughly studied among the genes of the nif operon, with an extensive collection of sequences obtained from both cultured and uncultivated microorganisms isolated from multiple environments (Ueda et al., 1995; Zehr et al., 1998, 2003; Hamelin et al., 2002; Tan et al., 2003; Deslippe & Egger, 2006). Cultivation-independent studies using clone library analyses (Ueda et al., 1995), denaturating gradient gel electrophoresis (DGGE) (Piceno & Lovell, 2000) and terminal or complete restriction fragment length polymorphism (RFLP) methods (Poly et al., 2001) have already shown the great diversity of nifH genes in natural environments. Despite this extent of knowledge, little is known about the assemblage of nitrogen fixing communities in the complex soil environment (Izquierdo & Nusslein, 2006).
In a previous study, the diversity of Paenibacillus species was assessed by PCR-DGGE and the cloning library approach based on rpoB gene in the rhizospheres of four cultivars of sorghum sown in Cerrado soil amended with two levels of nitrogen fertilizer (12 and 120 kg ha−1). Two cultivars considered as ‘inefficient’ (IS 5322-C and IS 6320) demanded the higher amount of nitrogen to grow, while the other two considered as ‘efficient’ (FBS 8701-9 and IPA 1011) did not. The results indicated that the cultivar type was the overriding determinative factor that influenced the community structures of the Paenibacillus in those rhizospheres (Coelho et al., 2007). However, the diazotrophic population, which is supposed to be the most affected by the absence of nitrogen, was not analyzed at that time. Therefore, this study aims to evaluate the diversity of potentially diazotrophic bacteria in the rhizospheres of two of these contrasting sorghum cultivars sown with low and high levels of nitrogen fertilizer. Direct PCR amplification from the rhizosphere soils and subsequent cloning and sequence analysis of partial nifH genes were the experimental strategies used. With the results obtained, one can identify the predominant bacterial diazotrophic population present in the habitat studied and determine whether there is an influence of the cultivar and/or the amount of nitrogen fertilization on the composition of bacterial communities that might contribute to the nitrogen input in the rhizosphere of these sorghum plants grown under the conditions mentioned above.
Materials and methods
Sorghum cultivars and experimental conditions
The field experiment was carried out at EMBRAPA Maize and Sorghum, Sete Lagoas, Minas Gerais, Brazil, located at latitude 19°28′S and longitude 44°15′W, at a height of 732 m. The local climate is the savannah type, with a mean temperature in the coldest month above 18 °C, according to the Köppen classification. The soil planted with two different cultivars of sorghum (Sorghum bicolor) was classified as a typical Distrophic Red Latosol, Cerrado stage, with a very clay texture (coarse sand 6%, fine sand 4%, silt 12% and clay 78%), low organic matter content (3.27 dag dm−3), pH 6.2 and phosphorus, potassium, calcium and magnesium contents of 11, 80, 5.85 and 0.87 cmol dm−3, respectively. The characteristics of the sorghum cultivars used in the present study (chosen based on their agricultural importance) have been presented previously (Coelho et al., 2007). Briefly, they can be described as follows: IPA 1011, white grain, three-dwarf, 84 days from planting to flowering and a grain productivity of 5490 kg ha−1; IS 5322-C, white grain, short three-dwarf, 56 days from planting to flowering and a grain productivity of 930 kg ha−1. IPA 1011 is considered as ‘efficient’ (E) cultivar under nitrogen stress, capable of growing well at low concentrations of nitrogen in the soil and IS 5322-C is considered as ‘inefficient’ (I) cultivar under nitrogen stress, showing characteristic symptoms of lack of this nutrient (reduced growth and yellow leaves). The experimental plots consisted of two lines of 5 m of length spaced by 0.8 m between lines and 0.2 m between plants, made in two replicates. The cultivars were planted randomly in each plot. One plot received 12 kg of nitrogen per hectare (low concentration of nitrogen) and the other 120 kg of nitrogen per hectare (high concentration of nitrogen). Five plants of each sorghum cultivar were harvested 90 days after sowing and the roots shaken to remove the loosely attached soil. The adhering soil of the five plants was pooled and considered as the rhizosphere soil. Samples were kept at −20 °C before DNA extraction.
DNA extraction from rhizosphere soil
DNA was extracted from the rhizosphere soil of sorghum cultivars (0.5 g of each sample) using the commercial kit ‘FastDNA Spin Kit for soil’ (Q.Bio gene, BIO 101 Systems) according to the manufacturer's instructions. DNA concentrations were determined spectrophotometrically using a Gene Quant apparatus (Amersham Pharmacia Biotech). The DNA was visualized on 0.8% (w/v) agarose gels (Sambrook et al., 1989) to assess its purity and molecular size.
PCR amplification and clone library analysis of nifH genes
NifH gene fragments (360 bp) were amplified by PCR with the primers PolF (5′-TGCGAYCCSAARGCBGACTC-3′) and PolR (5′-ATSGCCATCATYTCRCCGGA-3′) following the same conditions as described by Poly et al. (2001). The 50 μL reaction mix contained 1 μL of template DNA (50–100 ng), 10 mM Tris-HCl, pH 8.3, 10 mM KCl, 0.5 mM of each primer, 200 mM of each dNTPs, 2.5 mM MgCl2 and 2 U of Taq polymerase (Promega). PCR conditions consisted of 30 cycles at 94 °C, 55 °C (1 min) for the annealing step and 72 °C (2 min), with a 5 min extension at 72 °C for the last cycle. PCR products were analyzed by electrophoresis on 1.4% agarose gel to confirm the size of the bands. PCR products were then purified with the Wizard nucleic acid clean up kit (Promega) and cloned using the pGEM-T easy vector (Promega), following both suppliers' instructions. After transformation of Escherichia coli JM109 competent cells, clones were picked and the presence of inserts of the correct size was verified by PCR using M13f (5′-GTAAAACGACGGCCAG-3′) and M13r (5′-CAGGAAACAGCTATGAC-3′) primers.
Sequencing and phylogenetic analysis
Selected clones were sequenced using M13f and M13r primers by an ABI Prism 3100 automatic sequencer (Applied Biosystems Inc.). The partial nifH sequences of selected representative clones obtained in this study were compared with the GenBank database using the algorithm blast-n to identify the most similar nifH sequences and then aligned with representative nitrogenase sequences obtained from the same database using the software package clustal x (Thompson et al., 1997). bioedit v. 7.0.0 (http://www.mbio.ncsu.edu/Bioedit/bioedit.html) was used for manual editing of the sequences. Nucleotide sequences were also checked for possible chimeric sequences by performing blast searches with partial sequences. All soil clones that were affiliated with nifH sequences (except chimeras) had their identity checked by phylogenetic analyses. The sequences generated in this study have been deposited in NCBI GenBank under accession numbers EU047946–EU048188.
Statistical and data analyses
Analyses of the clone libraries were followed by calculation of the coverage (C), where C is expressed by 1-n1/N, in which n1/N is the ratio of clones that appeared only once (n1) by the total number of clones (N) (Chelius & Triplett, 2001). An excel worksheet was used to calculate the Shannon–Wiener index, (Margelef, 1958) and Evenness, E=H′/ln S (Pielou, 1969), where S is the number of species observed and pi is the number of clones of a given species divided by the total number of organisms observed. The data from sequence libraries were also submitted to Principal Component Analyzes (PCA) using the software statistica 6.0.
Results and discussion
Amplicons of nifH gene fragments (360 bp) were obtained from rhizospheres of two contrasting cultivars of sorghum, IPA 1011 (efficient=E) and IS 5322-C (inefficient=I), sown with low (12 kg=−N) and high (120 kg=+N) nitrogen fertilizer levels, to evaluate the diversity of potentially diazotrophic bacteria in these systems. From a total of 384 clones analyzed (four libraries with 96 clones each), 139 (36%) were excluded from further analyses as they were considered to represent either chimeric structures or to present uncorrected alignment of the nucleotide sequences obtained from the PCR amplification using the forward and reverse nifH primers. The next step consisted of checking the identity of the remaining 245 clones by blast-n. All clones were identified as nifH sequences, which makes them firmly meaningful for phylogenetic analyses (Tables 1–4).
Table 1. Identification by NCBI blastn (first hit and most similar sequences from known diazotrophic bacteria) of nifH sequences obtained from the rhizosphere of sorghum cultivar IPA 1011 (efficient cultivar) sown with low nitrogen level (E−N)
Table 2. Identification by NCBI blastn (first hit and most similar sequences from known diazotrophic bacteria) of nifH sequences obtained from the rhizosphere of sorghum cultivar IPA 1011 (efficient cultivar) sown with high nitrogen level (E+N)
Table 3. Identification by NCBI blastn (first hit and most similar sequences from known diazotrophic bacteria) of nifH sequences obtained from the rhizosphere of sorghum cultivar IS 5322-C (inefficient cultivar) sown with low nitrogen level (I−N)
Table 4. Identification by NCBI blastn (first hit and most similar sequences from known diazotrophic bacteria) of nifH sequences obtained from the rhizosphere of sorghum cultivar IS 5322-C (inefficient cultivar) sown with high nitrogen level (I+N)
The phylogenetic distribution of clones based on nifH gene sequence in the four samples was quite homogeneous and clustered in various taxonomic groups, including Alphaproteobacteria, Betaproteobacteria and Gammaproteobacteria (Tables 1–4). However, when only the first hit in blast-n was considered, most of the clones (57%) were closely related to nifH genes of uncultured bacteria. Many authors have already described sequences corresponding to diverse unidentified diazotrophs (Ueda et al., 1995; Widmer et al., 1999; Piceno & Lovell, 2000; Poly et al., 2001; Zehr et al., 2003). These noncultivated diazotrophs may be the dominant nitrogen-fixing organisms in soil systems (Widmer et al., 1999; Poly et al., 2001; Tan et al., 2003; Bürgmann et al., 2005) and could represent a new and unexplored group that can have an important role in the nitrogen cycle and nitrogen input in soil. The remaining nifH sequences were affiliated with cultivated organisms and they could be distributed among six orders (Tables 1–4): Rhizobiales (26%), Burkholderiales (12.2%), Bacillales (1.6%), Sphingomonadales (1.6%), Rhodospirillales (1.2%) and Enterobacteriales (<1%).
To minimize the number of clones related to noncultivated organisms, the first hit in blast-n matching to a culturable strain was also considered. Tables 1–4 show the distribution of the clones throughout the four libraries. NifH clones affiliated with Azohydromonas spp., Ideonella sp., Rhizobium etli and Bradyrhizobium sp. were found in all libraries. The high abundance of Alphaproteobacteria (especially Rhizobium and Bradyrhizobium spp.) nifH genes found in the libraries could be explained by crop rotation with leguminous plants (soybean) in the sites where the sorghum was collected. Moreover, Rhizobium sp. is well known to colonize roots of a broad range of nonlegumes as rice, maize and wheat (Yanni et al., 1997; Gutierrez-Zamora & Martinez-Romero, 2001), and Bradyrhizobium has been already observed as an active nitrogen-fixing bacterium in rice (Chaintreuil et al., 2000). Although the genus Azoarcus is often described as being associated with grass roots (Hurek et al., 1994), the authors did not detect any sequence related to this bacterium in the libraries. The same observation was made by Hamelin et al. (2002) in the bacterial community associated with the rhizosphere of Molinia coerulea, a perennial grass.
Analyzing each of the four libraries, the influence of the cultivar and/or the amount of nitrogen fertilization on the composition of bacterial communities that might contribute to the nitrogen input in the rhizosphere of these sorghum plants could be determined. Library IPA 1011 obtained from efficient cultivar under low nitrogen level (E−N) presented a higher percentage of Alphaproteobacteria in the community than library IPA1011 obtained from efficient cultivar under high-nitrogen conditions (E+N) (Tables 1 and 2). Sequences affiliated with Delftia tsuruhatensis were found in the rhizosphere of both cultivars sown with high levels of nitrogen (Tables 2 and 4). This species has been already isolated from the rhizosphere of rice and described as a plant-promoting bacterium for its ability to fix nitrogen and produce antimicrobial substance against various plant pathogens (Han et al., 2005). Libraries obtained from cultivars sown with low levels of nitrogen presented 30% more clones affiliated with Bradyrhizobium sp. than those with high amounts of fertilizer. Clones affiliated with Methylocystis sp. were detected only in plants sown under low levels of nitrogen (Tables 1 and 3). Similarly, Mohanty et al. (2006) have shown that in forest and rice field soils Methylocystis species were affected by fertilization. Sequences affiliated with Azospirillum spp. and Sinorhizobium sp. could not be detected in the libraries obtained from I+N and E−N, respectively. Moreover, Pelomonas saccharophila and Xanthobacter flavus were found in all libraries except for E+N (Tables 1–4). Finally, clones affiliated with Paenibacillus durus could be found in libraries from the inefficient cultivar sown either in high or low amounts of fertilizer (Tables 3 and 4). Previous studies have shown that the rhizosphere of many nonleguminous plants is colonized by one or more genera of nonsymbiotic diazotrophs, including Paenibacillus strains (Bürgmann et al., 2005). Besides fixing nitrogen, Paenibacillus species can influence plant growth and health through the production of phytohormones, chitinases, proteases, antimicrobial compounds and solubilizing phosphate (Mavingui & Heulin, 1994; Seldin et al., 1998).
Statistical analyses of the distribution of clones based on nifH gene sequences showed that the values of the Shannon–Wiener index varied from 2.09 to 1.73 in the libraries obtained from efficient cultivar sown in soil amended with high and low levels of nitrogen (E+N and E−N), respectively. The highest evenness value was observed in IPA 1011 treated with high level of nitrogen, while the lowest value was observed in both E−N and I−N (Table 5). To check whether the size of the soil clone libraries was reflecting the real diversity, the coverage index of the clones was checked according to Chelius & Triplett (2001). The four clone libraries obtained in this study covered from 91.8% (I+N) to 98.3% (E+N) of the total diversity, considering the blast-n identification (Table 1–4).
Table 5. Statistical analyses of the clone libraries
E (cultivar IPA 1011)−N, efficient cultivar under nitrogen stress sown in Cerrado soil with low amount of nitrogen (12 kg ha−1); E+N, efficient cultivar under nitrogen stress sown in Cerrado soil with high amount of nitrogen (120 kg ha−1); I (cultivar IS 5322-C)−N, inefficient cultivar under nitrogen stress sown in Cerrado soil with low amount of nitrogen; I+N, inefficient cultivar under nitrogen stress sown in Cerrado soil with high amount of nitrogen.
Data from clone libraries were also subjected to PCA. Three factors represented 100% of the variance (factor 1=39.1%; factor 2=34.4% and factor 3=26.5%). For this approach, the first hit in blast-n matching a culturable strain was considered in each clone library obtained. Figure 1 corresponds to the three-dimensional plot of PCA, which demonstrates that the cultivar considered as ‘efficient’ (IPA 1011) deviated from the ‘inefficient’ cultivar (IS 5322-C), but the level of nitrogen fertilizer had a major effect on variance. Therefore, the data presented here are not in agreement with those observed in the community structure of Paenibacillus, in which the sorghum cultivar type was the overriding determinative factor that influenced these populations (Coelho et al., 2007). Similar results have been reported by Carelli et al. (2000) for alfafa cultivars and Sinorhizobium populations.
The analysis of the four libraries containing nifH bacterial sequences obtained from the cultivation-independent approach used in this study provided information about the bacterial diazotrophic community present in sorghum rhizospheres sown in Cerrado soil. It also led to the conclusion that the most important factor affecting the nitrogen-fixing bacteria community in these habitats was the amount of nitrogen applied. However, considering that this work was based on PCR amplification using total DNA extracted from rhizosphere soil of the sorghum cultivars, it could not be confirmed whether nif genes were actually expressed. Therefore, in future studies, it is necessary to investigate whether nitrogenases of diazotrophic bacteria are active in sorghum and whether they really contribute to nitrogen uptake.
The authors wish to thank Fabio Faria da Mota for his assistance in statistical analyses. This study was supported by grants from the National Research Council of Brazil (CNPq), FAPERJ and EMBRAPA.