OUP user menu

Assessment of the diversity of Paenibacillus species in environmental samples by a novel rpoB-based PCR-DGGE method

Fabio Faria da Mota, Eliane Aparecida Gomes, Edilson Paiva, Lucy Seldin
DOI: http://dx.doi.org/10.1016/j.femsec.2005.01.017 317-328 First published online: 1 July 2005


A specific PCR system based on the gene encoding the RNA polymerase beta subunit, rpoB, was developed for amplification and denaturing gradient gel electrophoresis (DGGE) fingerprinting of Paenibacillus communities in environmental samples. This gene has been previously proven to be a powerful identification tool for the discrimination of species within the genus Paenibacillus and could avoid the limitations of 16S rRNA-based phylogenetic analysis. Initially, the PCR system based on universal rpoB primers were used to amplify DNAs of different Paenibacillus species. A new reverse primer (rpoBPAEN) was further designed based on an insertion of six nucleotides in the Paenibacillus sequences analyzed. This semi-nested PCR system was evaluated for specificity using DNAs isolated from 27 Paenibacillus species belonging to different 16S rRNA-based phylogenetic groups and seven non-Paenibacillus species. The non-Paenibacillus species were not amplified using this PCR approach and one group of Paenibacillus species consisting of strains without the six-base insert also were not amplified; these latter strains were found to be distinct based on 16S rRNA gene phylogeny. In addition, a clone library was generated from the rpoB fragments amplified from two Brazilian soil types (Cerrado and Forest) and all 62 clones sequenced were closely related to one of the 22 sequences from Paenibacillus previously obtained in this study. To assess the diversity of Paenibacillus species in Cerrado and Forest soils and in the rhizosphere of different cultivars of maize, a PCR-DGGE system was used. The Paenibacillus DGGE fingerprints showed a clear distinction between communities of Paenibacillus in Forest and Cerrado soils and rhizosphere samples clustered along Cerrado soil. Profiles of cultivars CMS22 and CMS36 clustered together, with only 53% of similarity to CMS11 and CMS04. The results presented here demonstrate the potential use of the rpoB-based Paenibacillus-specific PCR-DGGE method for studying the diversity of Paenibacillus populations in natural environments.

  • Paenibacillus
  • rpoB
  • Diversity
  • Maize rhizosphere
  • Cerrado soil
  • Forest soil

1 Introduction

Molecular markers, such as the 16S rRNA gene, for investigation of the occurrence and distribution of bacteria in environmental samples, have been extensively applied to explore bacterial diversity and to provide direct information on bacterial community structure [1]. Polymerase chain reaction has been commonly used, in combination with methods that generate fingerprints, such as temperature gradient gel electrophoresis (TGGE), denaturing gradient gel electrophoresis (DGGE), terminal restriction fragment length polymorphism (T-RFLP) and single strand conformation polymorphism (SSCP), for the analysis of bacterial communities [2]. DGGE has gained the preference in many research groups because it can show changes in bacterial community structure that are not detected by methods based on culturing bacteria alone. Furthermore, electrophoretic bands from the 16S-DGGE gels can be selected for sequencing and phylogenetic analysis. However, the presence of different number of copies of the 16S rRNA in different bacterial genera limits 16S rRNA-based phylogenetic analysis. For example, in the genome of different Paenibacillus species, from two to 13 copies of the 16S rRNA gene can be found [3,4]. Several bands per Paenibacillus species will be seen in DGGE, making the profiles confusing and difficult to interpret [3]. Silva et al. [4] clearly demonstrated the heterogeneity in DGGE banding pattern of Paenibacillus durus as a result of the presence of multiple copies of the ribosomal genes. Different markers are therefore proposed as an alternative to the 16S rRNA gene in taxonomic and ecological studies, mainly those that exist in a single copy in all bacterial genomes studied so far and contain conserved as well as variable regions.

The gene encoding the RNA polymerase beta subunit, rpoB, has been introduced in taxonomic studies and community analyses of bacteria because it fulfills the characteristics mentioned above [5]. Gene rpoB has already been used for identification of, for instance, enteric bacteria [6], Staphylococcus spp. [7], Bartonella spp. [8], and Mycobacterium spp. [9]. In Paenibacillus, Mota et al. [10] described the molecular identification of nitrogen-fixing species using a PCR/sequencing approach on the basis of the rpoB gene as an alternative to the 16S rRNA gene. The discriminating power of the rpoB and 16S rRNA sequences has been compared and the partial rpoB sequences used in their study were as informative as the full 16S rRNA gene sequences. However, the value of rpoB as a molecular marker in ecological studies has not been assessed yet for Paenibacillus.

The genus Paenibacillus is formed by 56 species and two sub-species (http://www.ncbi.nlm.nih.gov, July 2004) and harbors strains of industrial and agricultural importance. They are ubiquitous in the environment as different species are usually found in water, plant rhizospheres, inside plant tissues and soil [1114]. Some species were even found in association with arbuscular mycorrhizal hyphae [15]. More recently, three novel species were isolated from blood cultures [16] and from other sources such as food [13], olive-mill wastewater [17], raw and heat-treated milk [18] and cow faeces [19]. Different enzymes (such as crystalline glucanotransferase, chitinase, amylase, cellobiohydrolase, agarase and proteases) and antimicrobial substances (antibiotics, bacteriocins and/or small peptides) are produced by different Paenibacillus spp. [17,2022]. Furthermore, strains of Paenibacillus spp. can degradate polyaromatic hydrocarbons [23], produce phytohormones [24,25], furnish nutrients to plants by nitrogen fixation [13,14,2628], solubilize phosphate [11] and suppress phytopathogens through antagonistic functions [29,30].

Therefore, due to the importance of the genus Paenibacillus, the aim of this study was to develop and apply a specific PCR-DGGE system based on rpoB as a molecular marker for amplification and fingerprinting of Paenibacillus populations in environmental samples. Initially, the PCR system based on the rpoB primers described by Dahllöf et al. [5] was used for amplification of DNAs of different Paenibacillus species. A new reverse primer was then developed and the PCR system was evaluated for specificity using DNA isolated from Paenibacillus and non-Paenibacillus species. After optimization of the method, a PCR-DGGE system was used to assess the diversity of Paenibacillus species in the rhizospheres of four different cultivars of maize, sown in Cerrado soil. Another Brazilian soil, that is Forest soil, obtained from a semi-arid region, was also analyzed by the same PCR-DGGE approach. DGGE profiles obtained in this study were compared to those previously obtained by use of Paenibacillus-specific 16S rRNA gene-based PCR-DGGE method [4]. In addition, a clone library was generated from the rpoB fragments amplified from the two soil types and selected clones were sequenced.

2 Material and methods

2.1 Bacterial strains and culture conditions

The strains used in this study and their growth conditions are listed in Table 1. Strains of Paenibacillus were stored aerobically at room temperature on GB agar slants supplemented with 1% CaCO3 (w/v) while Bacillus strains were stored on LB agar slants. All other strains were stored at −20 °C in their growth media plus 20% glycerol.

View this table:

Bacterial strains used in this study and specificity of PCR system based on rpoB gene

SpeciesStrainsaAccession number of rpoB sequencesbPCR product (240 bp)Growth conditionscOrigin/Referencesd
Paenibacillus maceransLFB126TAY493863+TSB, 30 °CDr. L. Rabinovitch, FIOCRUZ, RJ, Brazil
P. durus= (P. azotofixans)P3L5AY493867+TBN, 30 °CSeldin et al. [26]
P. peoriaeHSCC 353TAY493865+GB, 30 °CDr. O. Shida, Higeta Shoyu Co., Chiba, Japan
P. polymyxaLMD 24.16TAY493864+GB, 30 °CUniversity of Delft, The Netherlands
P. brasilensisPB 172TAY493868+GB, 30 °CDr. I. von der Weid, IMPPG, RJ, Brazil
P. graminisRSA19TAY493861+TSB, 30 °CDr. O. Berge, CNRS-CEA, France
P. borealisKK19TAY493866+TSB, 30 °CDr. S. Elo, University of Helsinki, Finland
P. odoriferTOD45TAY493862+TSB, 30 °CDr. O. Berge
P. pabuliNRRL B-14213AY728291+TSB, 30 °CDr. J.D. van Elsas, Groningen University, The Netherlands
P. glucanolyticusNRRL B-14679TAY728284+TSB, 30 °CDr. L.K. Nakamura, National Center for Agricultural Utilization Research, Peoria, USA
P. campinasensisKCTC 0364BPTAY728283+TSB, 30 °CDr. A.S. Rosado, IMPPG
P. thiaminolyticusDSM7262TAY728285+TSB, 30 °CDSMZ
P. curdlanolyticusDSM10247TAY728290+TSB, 30 °CDSMZ
P. amylolyticusDSM11730TAY728292+TSB, 30 °CDSMZ
P. jamilaeDSM13815TAY728281+TSB, 30 °CDSMZ
P. massiliensis2301065TAY728294+TSB, 30 °CDr. D. Raoult, Hopital de la Timone, Marseille, France
P. sanguinis2301083TAY728287+TSB, 30 °CDr. D. Raoult
P. dendritiformisType CAY728286+TSB, 30 °CDr. D. Gutnick, Tel-Aviv University, Israel
P. timonensis2301032TAY728289+TSB, 30 °CDr. D. Raoult
P. favispousGMP01TAY728288+TSB, 30 °CDr. E. Velasquez, Facultad de Farmacia, Universidad de Salamanca, Spain
P. lactis1871TAY728282+TSB, 30 °CDr. P. Scheldeman, Department of Animal Product Quality, Center for Agricultural Research, Belgium
P. daejeonensisKCTC3745TAY728293+TSB, 30 °CDr. J.S. Lee, Korea Research Institute of Bioscience and Biotechnology, Yusong, Korea
P. naphthalenovoransDSM14203TAY728280TSB, 30 °CDSMZ
P. ehimensisKCTC3748TAY728278TSB, 30 °CDr. J.S. Lee
P. chondroitinusNRS1347TAY728296TSB, 30 °CDr. L.K. Nakamura
P. alginolyticusNRS1351TAY728295TSB, 30 °CDr. L.K. Nakamura
P. validusQ1AY728279TSB, 30 °CDr. L.K. Nakamura
Bacillus licheniformisT6-3LB, 30 °CIMPPG
Bacillus subtilisLF-3LB, 30 °CIMPPG
Bacillus cereus413LB, 30 °CIMPPG
Escherichia coliK12LB, 37 °CIMPPG
Staphylococcus aureusA70TSB, 37 °CM.C. Bastos, IMPPG
Lactobacillus paracaseiATCC 4223TSB, 37 °CIMPPG
  • aType strain (T).

  • bSequences obtained with the primers rpoB1698f and rpoB2041r described by Dahllöf et al. [5].

  • cTSB, Trypticase soy broth (BBL); TBN, Thiamine-biotin medium [26]; GB, Glucose broth [30]; LB, Luria-Bertani medium.

  • dUFRJ, Universidade Federal do Rio de Janeiro, Brazil, IMPPG, Instituto de Microbiologia Prof. Paulo de Góes – UFRJ.

2.2 Maize cultivars and soil samples

One representative soil, which is commonly cultivated with maize in the State of Minas Gerais, Brazil, denoted Cerrado, was planted with four cultivars of maize (Zea mays) in Sete Lagoas, Minas Gerais. Cerrado soil is a dark red latosol (distrophic) of low pH (pH-KCl 4.0), low organic matter content, and clayey texture [4]. The experimental plots consisted of two lines of 5 m length with spaces of 0.9 m between the lines and 0.2 m between plants of the same cultivar. The four maize cultivars used can be briefly described as follows: CMS04 – late-cycle tropical maize, derived from Caribbean, Mexican and Brazilian germplasm, carrying yellow-dented grains; CMS11 – tropical maize of intermediate life cycle, derived from Caribbean, Tuxpeño and Mexican germplasms, with semi-flint yellow grains; CMS22 – subtropical maize of intermediate-late life cycle, derived from Mexican, Caribbean and American germplasm, with yellow-dented grains; CMS36 – late-cycle tropical maize germplasm, created from lines of different origins (Catete, Tuxpeño, ETo amarillo), tolerant to acid soils, and tall. At sampling, 90 days after sowing, three plants of each maize cultivar were harvested and the roots shaken to remove the loosely attached soil. The adhering soil was considered as rhizosphere soil. Forest soil was a red/yellow latosol (distrophic) with medium texture, usually found in semi-arid regions in Brazil. The region sampled contained a complex vegetation cover but with a predominance of arborous Caatinga, characteristic of the Northeast of the country. Samples were kept at −20 °C before DNA extraction.

2.3 DNA extraction

Genomic DNAs were extracted from 3 ml bacterial cultures grown overnight (Table 1) using the Puregene DNA Isolation Kit (Gentra Systems, Minneapolis, USA). DNA concentrations were determined spectrophotometrically using a Gene Quant apparatus (Amersham Pharmacia Biotech, New Jersey, USA). DNA was extracted from bulk and rhizosphere soil samples (0.5 g of each) using the FastDNA Spin Kit for soil (Qbiogene, BIO 101 Systems, Carlsbad, CA, USA) according to the manufacturer.

2.4 Primer design

Sequences of rpoB belonging to members of eight nitrogen-fixing Paenibacillus were previously obtained [10] after amplification of DNAs using a primer set (forward primer rpoB1698f and a reverse primer rpoB2041r) as described by Dahllöf et al. [5]. These rpoB sequences, together with 106 rpoB sequences from different bacterial genomes deposited in the GenBank database (until May 2003), were then aligned using the software Clustal-X [31]. After alignment, one particular region was found as potentially specific for Paenibacillus (Fig. 1). Therefore, a 20-mer reverse primer –rpoBPAEN: 5’-ATG TTG TCI GAI TCC TTG TT-3’ was established. This primer was analyzed by BLAST-N [32] to search for homologous nucleotide sequences in the GenBank database and the only hits related to bacteria were with the eight nitrogen-fixing species of Paenibacillus. A GC clamp [5] was attached to the 5’ end of the rpoB1698f primer so that it could be used for DGGE.


Alignment of rpoB DNA sequences from Paenibacillus nitrogen-fixing species, corresponding to the region chosen for the primer design. The box shows the six nucleotides found only in Paenibacillus. An inosine was incorporated to the primer sequence whenever the variation among the bases was more than two.

2.5 PCR conditions and DGGE

The amplification conditions for pure cultures (Table 1), when using the primers rpoB1698f and rpoB2041r described by Dahllöf et al. [5], were: 1 × (5 min, 95 °C), 6 × (30 s, 94 °C; 1.5 min, 40 °C; 1.5 min, 72 °C), 19 × (30 s, 94 °C; 1.5 min, 50 °C; 1.5 min, 72 °C), with a final 10 min extension period at 72 °C. The reaction mix contained 1 μl of template DNA (50–100 ng), 10 mM Tris–HCl, pH 8.3, 10 mM KCl, 25 pmol of each primer, 2.5 mM of each deoxynucleoside triphosphate, 20 μg of BSA, 2.6 mM MgCl2, and 5 U/50 μl of Taq polymerase. When using primers rpoB1698f and rpoBPAEN, the amplification conditions were the same as described above, changing the annealing temperature of the 25 cycles to 40 °C and the extension time to 30 s. Amplification from soil or rhizosphere DNA extracts was performed in duplicate as a semi-nested PCR, which consisted of a first PCR with the primers of Dahllöf et al. [5]. The PCR product was diluted 1:1,000 and used as the template in the second PCR, performed with primers rpoB1698f (GC clamped or not) and rpoBPAEN. Negative controls (without DNA) were run in all amplifications and PCR products were visualized by 1.4% agarose gel electrophoresis followed by staining with ethidium bromide.

DGGE was performed with the DCode Universal Mutation Detection System (Bio-Rad Laboratories, Hercules, CA, USA). PCR products (15–25 μl) were applied directly onto 6% (wt/vol) polycrylamide gels in 1X TAE buffer (20 mM Tris–acetate (pH 7.4], 10 mM sodium acetate, 0.5 mM disodium EDTA) containing a linear denaturing gradient from 45% to 70%. The gradients were formed with 6% (wt/vol) acrylamide stock solutions [33] that contained no denaturant and 80% denaturant (the 80% denaturant solution contained 7 M urea and 40% (vol/vol] formamide) deionised with AG501-X8 mixed-bed resin (Bio-Rad). The gels were electrophoresed for 19 h at 60 °C and 75 V. After electrophoresis, the gels were stained for 30 min with SYBR Green I nucleic acid gel stain (Molecular Probes, Oregon, USA) and digitalized with the STORM apparatus (Amersham Biosciences, Inc., São Paulo, Brazil).

2.6 Cloning and sequencing of PCR products and DNA sequence analyses

rpoB PCR products of the expected size of 240 bp (without clamp) generated with DNA extracted from soil were cloned using the pGEM T-easy vector according to the instructions of the manufacturer (Promega, Wisconsin, USA). After transformation of Escherichia coli JM109 competent cells, about 50 clones were picked per transformation and the presence of inserts of the expected size was assessed by PCR amplification using rpoB1698f and rpoBPAEN primers. Selected clones were then sequenced using the M13f and M13r primers and an ABI Prism 3100 automatic sequencer (Applied Biosystems, Foster City, CA, USA).

Sequences generated in this study and the sequences recovered from the database were aligned using Clustal X [31]. The phylogenetic trees were calculated by the neighbor-joining method [34], and p-distance method. Trees were constructed with the software MEGA 3 [35]. To identify the clone libraries, a local database (rpoBBANK) was constructed with 2269 rpoB sequences retrieved from GenBank (July 2004) plus 127 rpoB sequences from microbial genomes (Omniome database, http://www.tigr.org) and 27 rpoB sequences of Paenibacillus species (19 sequences obtained in this study and another eight sequences obtained previously by Mota et al. [10]). Sequence identities of the clone libraries were determined by their first hit in BLAST-N analyses [32] using this local database with 2423 rpoB sequences.

2.7 Statistical analysis

Direct 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) to the total number of clones (N) [36]. An Excel worksheet was used to calculate the Shannon–Wiener index, Embedded Image[37] and Evenness, E=H′/ln S[38] 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. DGGE data were collected into a matrix indicating the presence or absence (scored as 1 or 0, respectively) of specific bands in DGGE analysis. A dendrogram was constructed using the DICE coefficient and unweighted pair group method with arithmetic mean (UPGMA). For these analyses, the NTSYS software package (version 2.02, Exeter Software, Setauket, New York) was used.

3 Results

3.1 Primer design and the specificity of primer rpoBPAEN in pure cultures

In a previous study, Mota et al. [10] investigated the usefulness of the rpoB gene for taxonomic studies in nitrogen-fixing Paenibacillus species. Partial rpoB sequences of about 375 bp were generated with the primers described by Dahllöf et al. [5] for the type strains of the eight nitrogen-fixing Paenibacillus species and an insertion of six nucleotides was observed in the sequences analyzed. Fig. 1 shows the localization of these inserted nucleotides in rpoB sequences and the reverse primer designed based on this region. However, some differences were observed among the sequences analyzed and, to increase the universality of the primer designed, two inosines were included in the primer chosen (5′ ATGTTGTCIGAITCCTTGTT 3′) and it was named rpoBPAEN (Fig. 1). One mismatch was found with rpoB sequences of P. macerans and P. odorifer, two with that of P. durus, P. polymyxa and P. brasilensis, while three mismatches were found with P. peoriae. However, they did not affect the specificity of the primer, using the PCR conditions described here.

The specificity of the new primer was tested in combination with the forward primer described by Dahllöf et al. [5], using pure-culture DNAs from 19 other Paenibacillus species and seven non-Paenibacillus strains as templates (Table 1). The Paenibacillus species chosen represent the different phylogenetic groups (I to XI) formed based on sequences of 16S rRNA gene (Fig. 2). At least one strain of each group was tested (seven from group I, five from group XI, four from group V, three from groups VI and IX, two from group III, and one from groups II, IV and X), with the exception of groups VII and VIII. No product was detected with any non-Paenibacillus species and products of the appropriate size (240 bp) were detected with 22 species of Paenibacillus (Table 1). All the five Paenibacillus species that showed negative results in PCR amplification belong to the same 16S rRNA-based phylogenetic group XI (Fig. 2). DNAs obtained from representatives of these five species were amplified with the universal rpoB primers described by Dahlöff et al. [5], PCR products were then sequenced and the lack of the six-nucleotide insertion was observed in those sequences. Therefore, we proposed the division of the 16S rRNA-based tree into two main groups (A and B, Fig. 2) based on the presence (group A) or absence (group B) of the insertion.


Consensus phylogenetic tree showing the relationship between Paenibacillus species based on the full 16S rRNA gene sequences. The tree was constructed based on the neighbor-joining method [34] and p-distance. Bootstrap analyses were performed with 1000 repetitions and only values higher than 50 are shown. The GenBank accession number of each species is enclosed in parentheses. The arrows show the species used in this study. Clusters of different species were assigned Roman numbers and division into groups A and B was based on PCR amplification (solid arrows) or not (open arrows), respectively, using the primer rpoBPAEN.

3.2 Soil-derived rpoB clone libraries and phylogenetic relationship of the sequences

The soil DNA extraction protocol provided sufficient DNA of good quality to be used for PCR amplification. Both Cerrado and Forest soil PCR products were obtained using the rpoBPAEN primer in combination with rpoB1698f and were used for cloning and establishment of rpoB libraries. Two libraries were therefore produced, one with Cerrado soil (29 clones) and the other one with Forest soil (33 clones). The 62 inserts were sequenced and they could be related to one of the 22 sequences from Paenibacillus previously obtained in this study. To assess the diversity of Paenibacillus species in these two Brazilian soils, all clones were checked for their identity by BLAST-N. The results confirmed the great specificity of the system for Paenibacillus, because 100% of the clones were identified as this genus. Considering only the first hit in BLAST-N, clones obtained from Cerrado were closely related to P. amylolyticus (51.7%), P. graminis (44.8%) and P. favisporus (3.4%) with similarity levels varying from 88% to 98%, while clones from Forest soil were closely related to P. favisporus (36.4%), P. macerans (18.2%), P. amylolyticus (12.1%), P. lactis (9.1%), P. graminis (9.1%), P. thiaminolyticus (6%), P. campinasensis, P. sanguinis and P. dentritiformis (3% each) with similarity levels varying from 81% to 98% (Fig. 3). The Shannon–Wiener index values showed a higher diversity of Paenibacillus populations in Forest soil (H′= 1.86) than in Cerrado soil (H′= 0.82). In addition, the evenness value observed was higher in Forest soil (E= 0.84) than in Cerrado soil (E= 0.74).


Histogram showing the species of Paenibacillus from clones obtained from Cerrado and Forest soils. The clone identification was performed by BLAST-N software using a local database with 2423 sequences of rpoB gene. The first hit of BLAST-N for each clone was considered the better identification of species to compare the soils.

To check whether the size of the soil clone libraries was reflecting the real diversity, the coverage of the clones was checked according to Chelius and Triplett [36]. The clone library obtained from Cerrado soil covered 96.5% of the total diversity, while for Forest soil this value was slightly lower, 90.9%, considering the BLAST-N identification.

To compare the clone sequences, phylogenetic analyses were performed (Figs. 4(a) and (b)). The analyses revealed close relationships among a wide range of clones. The sequences obtained from Cerrado soil were grouped in three main branches (Fig. 4(a)), while those obtained from Forest soil in four branches (Fig. 4(b)). Each group could be related to sequences of the 22 species of Paenibacillus studied here. The only exceptions were seven clones (F8, F9, F43, F57, F67, F68 and F70) derived from Forest soil, which formed separate groups without any representative species included in their groups. Some Paenibacillus clones were typical for one soil and not found in the other; for instance, clones related to P. macerans were observed exclusively in Forest soil. Moreover, clones closely related to P. favisporus occurred mainly in Forest soil as only one clone related to this species was observed in Cerrado soil (Figs. 4(a) and (b)). Finally, the species clusters containing the highest number of clones in Cerrado and Forest soils were groups A1 and B1, respectively.


Phylogenetic trees showing the relationship among Paenibacillus species and clones obtained from Cerrado soil (a) and Forest soil (b) based on partial rpoB gene sequences (240 pb). Clusters of related species were assigned by numbers for both trees. The trees were based on the neighbor-joining method [34] and p-distance. Bootstrap analyses were performed with 1000 repetitions and only values higher than 50 are shown.

3.3 Diversity of Paenibacillus-rpoB-specific PCR products in two different soils and in rhizosphere samples of different maize cultivars sown in Cerrado soil

Fingerprints of the most dominant populations were obtained after separation of PCR products in DGGE (Fig. 5(a)). Banding patterns obtained with DGGE from rhizosphere samples (cultivars CMS04, CMS11, CMS22 and CMS36 sown in Cerrado soil) as well as those from soils (Cerrado and Forest) were reproducible among the duplicates; therefore only one lane of each sample was used for further statistical analyses. Twenty-five markers (corresponding to different bands in DGGE profiles) were used for the construction of the dendrogram using DICE and UPGMA. Rhizosphere samples clustered along the soil type from which they have been obtained, and the soil type was responsible for the separation of the two main clusters encompassing all samples, at 45% similarity (Fig. 5(b)). In the “Cerrado” cluster, the rhizospheres were split into three subclusters. Profiles of cultivars CMS22 and CMS36 clustered together, with only 53% of similarity to CMS11 and CMS04. The CMS04 derived profile was about 76% similar to one from Cerrado soil.


(a) DGGE patterns obtained with Paenibacillus-specific PCR-DGGE based on rpoB gene of Cerrado (lanes C soil and C soil*) and Forest (lane F soil and F soil*) soils and maize rhizospheres, 90 days after sown in Cerrado soil (lanes CMS04, CMS11, CMS22 and CMS36). The lane M contains 5 μl of BenchTop 1Kb DNA ladder (Promega), lanes F soil* and C soil* contain 25 μl of PCR product and lanes C and F contain 15 μl of PCR product. Numbers in DGGE correspond to bands that were retrieved from the gel, reamplified and sequenced. (b) Dendrogram showing the levels of similarity of the Paenibacillus communities of Cerrado and Forest soils and maize rhizospheres planted in Cerrado soil. The dendrogram was constructed with unweighted pair group method with mathematical averages and similarity coefficient of DICE.

3.4 Sequences of DGGE bands

DGGE profiles allowed presumptive identification of some of the bands. Thus, 16 bands were retrieved from the gels (marked in Fig. 5(a)), reamplified and sequenced. BLAST-N analyses were performed and bands were identified as bands produced by organisms related to P. graminis (bands 1 and 5, with similarity values of 92% to 90%, respectively), P. amylolyticus (bands 2, 3, 4, 10 and 11, with similarity values varying from 91% to 96%), P. favisporus (bands 6, 7, 8, 9, 15 and 16, with similarity values varying from 89% to 91%), P. lactis (band 14, with similarity value of 89%) and P. durus (bands 12 and 13, both with similarity value of 96%). Although bands migrating to different positions were assigned as belonging to the same species, these results are in agreement with those presented by Mota et al. [10] and they may represent different strains of the same species. Rhizospheres from cultivars CMS11, 22 and 36 showed a strong band (Fig. 5, #13) that could be correlated to P. durus. On the other hand, this band was detected neither in cultivar CMS04 nor in the two soil samples. Data obtained from sequencing of DGGE bands clearly indicated that the DGGE profiles do reflect the sequences that were also found by direct cloning of rpoB amplicons generated on the basis of the soil DNAs.

4 Discussion

Molecular approaches that address the nature of individual bacterial types have been mostly based on sequence determinations of 16S rRNA genes amplified and cloned from environment-derived DNA. Furthermore, the application of PCR-DGGE to generate fingerprints of bacterial communities present in a variety of habitats represents a very important tool in studies of microbial diversity [3941] and this approach is widely used by different researchers. However, many bacterial species have multiple rRNA genes, which might exhibit microheterogeneity. The number of rRNA operons per bacterial genome can vary from 1 to 15 [42] and, therefore the number of bands obtained by PCR-DGGE may be higher than the number of actual species present in a community being studied. The fact that an organism might be represented by more than one band has been previously demonstrated by Silva et al. [4] in Paenibacillus spp. and by Salles et al. [43] in Burkholderia spp. The DGGE profiles obtained in these studies were complex and could not provide an accurate estimate of richness of those populations.

To avoid the limitations of 16S rRNA-based analyses in ecological studies, Mota et al. [10] first investigated the usefulness of the rpoB gene as an alternative to the 16S rRNA gene in nitrogen-fixing Paenibacillus. The data obtained in their study indicated that rpoB was a powerful identification tool, which could be used for discrimination of the nitrogen-fixing species within the genus Paenibacillus. In the present study, the applicability of rpoB for molecular ecological studies was evaluated. First, an rpoB based-primer was designed taking advantage of a region containing a six base-insertion present in the 375 bp rpoB sequence obtained from the eight nitrogen-fixing Paenibacillus species. We intended to develop a specific PCR system for the direct amplification of rpoB sequences belonging to Paenibacillus species from soil suitable for subsequent analysis via DGGE. The great specificity of the primer designed (rpoBPAEN) for Paenibacillus spp. was supported by our data and it is quite possible that the new rpoB-based primer presented here is able to amplify the DNA from the majority of the species of Paenibacillus. However, using this new primer we were not able to amplify five out of the 27 species studied. The failure to amplify these five species (all belonging to the same phylogenetic group based on 16S rRNA gene, Fig. 2) was due to the lack of the insertion used to design the primer. Further studies are necessary to explain why they had a distinct evolution from the other Paenibacillus species analyzed.

The specificity of the primers was also confirmed by sequence analyses of randomly chosen soil-derived clones (see Figs. 3 and 4). All 62 clones sequenced were affiliated with the genus Paenibacillus. However, intraspecific sequence variation was observed among clones of the same species, and different authors working with the rpoB gene [7,10,44] have previously demonstrated this divergence. Even so, clones could be identified by comparison of the rpoB sequence similarity and phylogenetic trees, and all clones were spread among phylogenetic clusters defined in this study (Figs. 4(a) and (b)). The distribution of clones among soil types showed that some groups tend to be more prevalent in one of the soil types. In Cerrado soil, the most abundant clones were those showing sequence similarities to P. amylolyticus and P. graminis, while in Forest soil those similar to P. favisporus, a xylanolytic bacterium isolated from cow faeces [19] were most prevalent. Moreover, clones derived from Forest soil were more diverse and were closely related to species found only in this soil. The observed Paenibacillus diversity in this soil was supported by the high value obtained for library coverage, although the sample size should still be increased to achieve complete coverage. In addition, clones derived from Forest soil that could not be included in any established group (Fig. 4(b)) may represent clones affiliated to species of Paenibacillus not studied here.

The success of the Paenibacillus-specific rpoB-based PCR-DGGE method would thus depend on the presence of DNA targets for amplification. Pure culture or soil DNA extraction methods used in this study have proved to be efficient in producing sufficient amounts of DNA for amplification of Paenibacillus species [4,45]. In addition, a semi-nested PCR procedure, in which the clamped primer was used only in the second PCR, was used to increase the sensitivity of the PCR-DGGE method. The nested approach has been considered to be more convenient for the detection of bacterial communities in soil samples when target organisms are present in low numbers or in environments containing a high organic matter content that can affect the PCR due to the presence of potential inhibitors [43].

The DGGE profiles of the Paenibacillus populations in bulk soils (Cerrado and Forest) and rhizosphere of maize cultivars planted in Cerrado soil indicated that the soil type was responsible for changes observed in the structure of Paenibacillus community. This finding is in agreement with the results obtained by Silva et al. [4], where soil type (Cerrado × Várzea) also influenced the structure of Paenibacillus populations more drastically than the maize cultivar type. Seldin et al. [30] also obtained similar results with P. durus (previously P. azotofixans) isolates from maize planted in Várzea and Cerrado soils. On the other hand, differences among cultivars could be observed more clearly in this study when compared to the results by Silva et al. [4]. Maize cultivars CMS22 and CSM36 formed a cluster in 100% similarity (Fig. 5(b)) indicating very similar predominant Paenibacillus communities within this pair of cultivars. Furthermore, cultivar CMS04 showed a tendency to cluster with bulk soil. Differences in cluster formation might be due to either the different regions of recognition of the two set of primers used (16S rRNA ×rpoB genes) or the high number of markers (bands) derived from 16S rRNA gene sequences used by Silva et al. [4] to construct the dendrogram. In the present study, the number of markers was much lower and fingerprints easier to interpret. One strong band (Fig. 5, #13) was present in three of the cultivars and it was identified as a band produced by organisms related to P. durus. This species harbors nitrogen-fixing strains [26], a very important characteristic in the plant rhizosphere.

In this study, PCR-DGGE, based upon rpoB fragments amplified via semi-nested PCR from community DNA, proved to be a powerful tool for detecting the dominant members of the Paenibacillus community in bulk and rhizosphere soils (maize cultivars planted in Cerrado soil). A clear distinction between soils (Cerrado × Forest) was observed, indicating the selection of specific groups of Paenibacillus in each soil. Moreover, DGGE profiles were much easier to analyze when compared with those obtained from PCR-generated 16S rRNA gene fragments, and problems related to the multiple copies of this gene in Paenibacillus genomes as described by Nübel et al. [3] can now be avoided. We also suggest that the rpoB-based primers presented here are suitable for assessing the prevalence and diversity of many species of Paenibacillus. Studies using this PCR-DGGE system could therefore be very useful to elucidate the role of Paenibacillus species in environmental samples as plant growth-promoting rhizobacteria. However, further studies are warranted regarding the availability of rpoB sequences from the Paenibacillus species that have not been studied yet.


This work was supported by grants from the Brazilian National Research Council (CNPq) and FAPERJ. We thank all researchers who kindly provided us with bacterial strains used in this study and Dr. Irene von der Weid and Dr. Jan Dirk van Elsas for valuable assistance and helpful comments.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
View Abstract