Development of real-time PCR assay for detection and quantification of Sinorhizobium meliloti in soil and plant tissue


Alessio Mengoni, Department of Evolutionary Biology, University of Firenze, via Romana 17, I-50125 Firenze, Italy. E-mail:


Aims: Sinorhizobium meliloti is a nitrogen-fixing alpha-proteobacterium present in soil and symbiotically associated with root nodules of leguminous plants. To date, estimation of bacterial titres in soil is achieved by most-probable-number assays based on the number of nodules on the roots of test plants. Here, we report the development of two real-time PCR (qPCR) assays to detect the presence of S. meliloti in soil and plant tissues by targeting, in a species-specific fashion, the chromosomal gene rpoE1 and the pSymA gene nodC.

Methods and Results: rpoE1 and nodC primer pairs were tested on DNA extracted from soil samples unspiked and spiked with known titres of S. meliloti and from plant root samples nodulated with S. meliloti. Results obtained were well in agreement with viable titres of S. meliloti cells estimated in the same samples.

Conclusions:  The developed qPCR assays appear to be enough sensitive, precise and species-specific to be used as a complementary tool for S. meliloti titre estimation.

Significance and Impact of the Study:  These two novel markers offer the possibility of quick and reliable estimation of S. meliloti titres in soil and plant roots contributing new tools to explore S. meliloti biology and ecology including viable but nonculturable fraction.


Sinorhizobium meliloti is a soil bacterium, belonging to the alpha subdivision of proteobacterium, able to colonize the roots of leguminous plants where it gives rise to the formation of specialized root structures called nodules where bacterial nitrogen-fixation occurs (Rose 2008), thus increasing plant agronomic yield without the need for nitrogen fertilization. Besides the important role as a model for symbiotic nitrogen-fixation, S. meliloti is a model species also for bacterial population genetics (see, for examples, Carelli et al. 2000; Bailly et al. 2006). The actual population size of S. meliloti in the environment is due to both the fraction of free-living cells in soil and to the symbiotically associated cells in roots. Despite its prominent role among environmental and agricultural bacteria, to date no selective culture media have been developed allowing the direct estimation of S. meliloti titres in soil. Consequently, quantification of S. meliloti in soil samples is currently performed via plant trapping, that is by estimating the nodulation potentiality of the soil (that is the presence of S. meliloti cells) using different Medicago species as host plant in green-house controlled conditions (Thompson and Vincent 1967). This technique, though giving reliable results, is labour intensive and requires facilities and equipments to be performed, hampering large-scale studies about the presence and the abundance of S. meliloti in different soils and habitats. Moreover, plant trapping is ineffective with the viable but nonculturable fraction (Basaglia et al. 2007) and in the last years there has been an increasing interest towards the possibility to analyse rhizobial populations without the need of cultivation (Zézéet al. 2001; De Oliveira et al. 2005).

Quantitative PCR (real-time PCR, qPCR) is a polymerase chain reaction technique which measures PCR amplification kinetics allowing the quantification of template DNA present in a sample. The qPCR assay has been used in many studies for direct quantification of bacteria in different environmental samples by targeting either the 16S ribosomal RNA or other specific gene sequences in the soil (see, for examples, Smits et al. 2004; Duodu et al. 2005; Sauvage et al. 2007; Selleck et al. 2008). The development of a qPCR assay for direct estimation of S. meliloti cells in soil could represent an advantage over the current labour-intensive and time-consuming nodulation-based protocol.

Here, we present the development and the application of two S. meliloti-specific qPCR assays to estimate bacterial titres on soil and plant samples.

Materials and methods

Soil samples, bacterial strains and MPN estimation of Sinorhizobium meliloti titres in soil

Three soil samples were taken from Soliman, Tunisia and Grezzano, Italy. Bacterial strains used in this study were S. meliloti Rm1021 (fully genome-sequenced type strain); S. meliloti BL225C, S. meliloti BO21CC, S. meliloti AK83 and S. meliloti AK58 strains (Giuntini et al. 2005); S. meliloti USDA1002, S. medicae LMG18864, S. medicae WSM419, S. fredii USDA205, S. terangae USDA4101, S. saheli USDA4102, Rhizobium leguminosarum bv. viciae USDA2370, Bradyrhizobium japonicum USDA110, Mesorhizobium huakuii USDA4779, Rh. etli CFN42, Rh. tropici CIAT899 and Azorhizobium caulinodans USDA4892. Moreover, 10 S. meliloti (SA1, SA2, SA3, SA10, SA11, SA12, SA13, SA27, SA40 and SA45) and 10 S. medicae (SS1, SS2, SS21, SS23, SS33, SS34, SS42, SS54, SS55 and SS60) isolates collected from Soliman soil were also included in the study. The most probable number (MPN) estimate of S. meliloti bacteria in soil was carried out by using a plant infection method (Thompson and Vincent 1967) with Medicago truncatula as trapping plant.

Soil spiking, in-vitro nodulation of Medicago truncatula

For soil spiking experiment, 0·5 g aliquots of samples from a clay soil in Grezzano (Italy) sampled with a caterpillar at ∼3-m depth, previously shown to be free from S. meliloti by the application of the present qPCR assay, were spiked with 100 μl of S. meliloti Rm1021 serial dilutions of a logarithmic-phase grown culture in TY (Trypone Yeast) medium (to obtain titres from 3·4 × 106 to 3·4 × 101 CFUs g−1 of soil). Immediately after spiking, soil aliquots were stirred to homogenize soil particles with bacteria and DNA was extracted as described below.

Nodulation of M. truncatula cv. Jemalong plants was performed as reported in Wais et al. (2002), allowing plants to grow up to 20 days after S. meliloti Rm1021 inoculation. S. meliloti titres in plant roots were estimated as CFU g−1 wet weight as in Barzanti et al. (2007) with slight modification (surface sterilization was performed with 0·1% HClO for 3′). Serial dilutions of the samples were plated on TY plates containing 100 mg l−1 streptomycin and incubated at 30°C for 2 days.

DNA extraction, cloning and sequencing

DNA was extracted from soil samples, surface sterilized root tissues and bacterial strains by using standard bead-beating protocols (FastDNA Kit, QBiogene, Illkirch Cedex, France). Quantification of DNA was performed by UV spectrophotometric reading (Biophotometer, Eppendorf).

Two libraries of putative nodC and rpoE1 amplification products from Soliman soil 1 DNA were constructed after cloning in pGEM-T Easy Vector (Promega Italia, Milano, Italy). Fifteen randomly chosen clones per library were sequenced and sequences were compared with those present in GenBank database by using BlastN (; Altschul et al. 1997).

Primer design, qPCR conditions and data analysis

Primer for amplification of rpoE1 and nodC gene fragments were designed by using the clone manager suite ver. 6 software (Sci-Ed, Cary, NC). Primer pairs were tested for similarities with other known sequences present in GenBank database by using BlastN (Altschul et al. 1997). The primer pair specific for S. meliloti rpoE1 gene (rpoE1-fw/rpoE1-rv) was designed to amplify a 100-bp fragment between nt positions 2267138 and 2267237 of S. meliloti Rm1021 chromosome, whereas the primer pair specific for S. meliloti nodC gene (nodC-fw/nodC-rv) was designed to amplify a 149-bp fragment between nt positions 480262 and 480133 of S. meliloti Rm1021 pSymA megaplasmid (Table 1, Fig. 1). Alignments between S. meliloti Rm1021 sequences and the most similar sequences of S. medicae WSM419 were performed with the ClustalW module implemented in bioedit ver. 7·0·9·0 (Ibis Biosciences, Carlsbad, CA) (Hall 1999).

Table 1.   Primers developed in this study*
Primer nameSequence (5′->3′)LengthNucleotide positionsTmGC%
  1. *Name, sequence, length of primer, nucleotide position in S. meliloti Rm1021 genome, Tm and GC% of each primer are reported. rpoE1 primers positions are referred to the S. meliloti Rm1021 chromosome (GenBank accession number NC_003047); nodC primers positions are referred to the S. meliloti Rm1021 pSymA megaplasmid (GenBank accession number NC_003037).

Figure 1.

 (a) Alignment of rpoE1 and nodC primer pairs with the corresponding most homologous regions in S. meliloti Rm1021 and S. medicae WSM419 genomes (see Table 2 for nucleotide positions); (b) Gel electrophoresis stained with ethidium bromide of amplification products on S. meliloti Rm1021 DNA with rpoE1 and nodC primer pairs. Lane 1, size marker; lanes 2 and 4, Rm1021 DNA; lanes 3 and 5, S. medicae WSM419 DNA. Sizes of band markers are indicated.

Real-time PCR was performed in an Applied Biosystems 7300 Real-Time PCR System (Applied Biosystems) programmed with the following temperature profile: 2 min 94°C, followed by 40 cycles composed by 15 s 94°C, 15 s 65°C, 30 s 72°C. Fluorescence data acquisition was done during the extension step at 72°C. A final melting curve was performed to check for product specificity. Reactions were performed in 20 μl final volume containing 10 μl of SYBR Green mix (SYBR Green JumpStart Taq Ready Mix, Sigma-Aldrich), 0·5-μl ROX solution (included in the kit SYBR Green JumpStart Taq Ready Mix) and 10 pmols of primers. Template DNA from spiked and unspiked soil samples was 1 μl out of 100 μl of total DNA solution from 0·5 g of soil. All reactions were done in triplicate. Data were analysed with the software sds version 1·2 (Applied Biosystems). PCR efficiency was calculated as efficiency = −1 + 1 × 10(−1/slope) (Ståhlberg et al. 2003).


Development of Sinorhizobium meliloti-specific markers

For the development of S. meliloti-specific qPCR markers, a list of loci supposed to be conserved in S. meliloti strains based on population genetic (Bailly et al. 2006) and genomic data (Giuntini et al. 2005), was prepared. The loci were selected to be located on both chromosome and on megaplasmids (pSymA and pSymB) of S. meliloti Rm1021 genome to allow the development of a possible multi-marker system based on independent replicons. On the chromosome, the intergenic region named IGSRKP (Bailly et al. 2006) and rpoE1 (Smc01419 encoding for a putative RNA polymerase sigma factor protein) and 16SrRNA (SMc03222) genes were chosen. For megaplasmid loci, intergenic regions IGSNOD, IGSGAB, IGSEXO (Bailly et al. 2006) and genes nodA (SMa0869), nodB (SMa0868), nodC (SMa0866) involved in Nod Factor biosynthesis and required for nodulation were selected.

Primer pairs of all the IGS regions were tested in qPCR with Soliman soil DNA as template. Obtained amplification plots and melting curves indicated poor efficiency of amplification and the presence of spurious amplicons. We consequently focussed on protein-coding genes and 16SrRNA, comparing S. meliloti Rm1021 nodABC, rpoE1 and 16SrRNA sequences with sequences present in GenBank database and in particular with the genome of the sister species S. medicae WSM419. Only the chromosomal gene rpoE1 and the pSymA gene nodC showed sequence divergence between S. meliloti Rm1021 and S. medicae WSM419 (87% and 98% of identity, respectively) and were consequently selected for the design of S. meliloti-specific primer pairs (Fig. 1, Table 1).

To check the selectivity of rpoE1 and nodC primer pairs for PCR amplification of S. meliloti DNA, DNAs from a collection of S. meliloti and other rhizobial strains were tested (see section ‘Materials and methods’). The collection was composed of both laboratory strains and environmental isolates of S. meliloti and S. medicae from Soliman soil (10 isolates each species) and included also S. meliloti strains BL225C, BO21CC, AK83 and AK58 for which comparative genomic data are available (Giuntini et al. 2005). Both rpoE1 and nodC primer pairs gave amplification products only with DNA from S. meliloti strains, confirming the selectivity of the designed pairs on pure culture DNA. To assay the selectivity on DNA extracted from soil and plant, the presence of spurious amplification products was checked in real-time PCR by performing qualitative analysis of melting curves on the amplicons obtained from DNA extracted from the Soliman soil samples and from M. truncatula root nodules. Only single melting peaks with the same Tm (87·5°C for rpoE1, 88·6°C for nodC) were scored from all assays (Fig. 2), suggesting the absence of spurious amplification products from soil microbial communities and plant DNA. Moreover, PCR products obtained after amplification on Soliman soil 1 DNA with rpoE1 and nodC primer pairs were cloned and 15 randomly chosen colonies for rpoE1 and nodC libraries were sequenced. Obtained sequences resulted identical to those of rpoE1 and nodC gene fragments of S. meliloti Rm1021 present in GenBank database, thus confirming the species specificity also on the DNA of a soil bacterial community.

Figure 2.

 Melting curves for (a) rpoE1 and (b) nodC amplicons from soil DNA. The different curves represent different soil DNA samples. The intensity of the peak is proportional to the amount of amplicon present in the PCR tube.

qPCR assays of soil and plant root DNA

External standard curves were obtained using serial dilutions of S. meliloti Rm1021 genomic DNA. Because of rpoE1 and nodC genes present in single copies in the S. meliloti Rm1021 genome, the copy number of the gene can be directly related to the number of copies of the genome (whose size is 6·68 Mbp) and, consequently, to the number of cells present in the sample.

Standard curves showed a linear range from 0·1 to 1 × 10−4 ng of Rm1021 genomic DNA corresponding to 1·39 × 104 to 1·39 gene copies (Fig. 3). The correlation coefficients R2 were very high both for rpoE1 gene (0·984) and for nodC gene (0·992). Slope was −3·618 for rpoE1 and −3·385 for nodC. Calculated PCR efficiencies were 88·8% and 97·4% for rpoE1 and nodC, respectively.

Figure 3.

 Calibration plots on S. meliloti Rm1021 DNA for (a) rpoE1 and (b) nodC. The threshold cycle (Ct) with respect to the amount of S. meliloti 1021 template DNA is reported. Equations and linear correlation coefficients (R2) are reported.

The applicability of the assay to quantify rpoE1 and nodC gene copies in environmental samples was evaluated by spiking known titres of S. meliloti Rm1021 in soil aliquots (soil 3 from Grezzano, Italy) and testing the protocol on two soils (soil 1 and soil 2 from Soliman, Tunisia) in which the number of nodulating S. meliloti cells was previously estimated by MPN. Obtained qPCR results (Table 2) were well in agreement with known bacterial titres spiked in soil 3, down to around 103 cells g−1. Spiking of low number of cells (101 and 102) did not produce positive amplification signals. Moreover, qPCR estimates were also in agreement with the number of CFUs present in nodulated root tissues of M. truncatula. Concerning the Soliman soil 1 and soil 2 in Tunisia, comparing rpoE1 and nodC qPCR estimates, these were similar for soil 1 and one-order magnitude different for soil 2. Interestingly, for both soils, qPCR estimates were higher than MPN estimates.

Table 2.   qPCR estimates of S. meliloti titres in soil and in M. truncatula root tissues*
SamplepH Organic matter (%) Total carbon (%) Total nitrogen (%)Electrical conductivity (EC) (mmho cm−1) Moisture (%)TextureNumber of rpoE1 gene copies†Number of nodC gene copies† Titres‡
  1. *The estimated number of S. meliloti rpoE1 and nodC gene copies (copies g−1 of soil) in three soils with different physico-chemical characteristics; and in M. truncatula nodulated root tissues. Soil 3 from Grezzano was spiked with serial dilutions known quantities of S. meliloti cells (see column ‘Titres’).

  2. Nd, not determined.

  3. †(±)SD from triplicate reactions of two independent experiments (six total reactions).

  4. ‡Titres are: most probable number estimates of S. meliloti cells g−1 in soil 1 and soil 2, spiked Rm1021 cells (CFU g−1) in soil 3, CFU g−1 fresh weight in M. truncatula roots.

Soil 1 (Soliman, Tunisia)8·91·130·71·84·654Clay1·75 ± 0·16 × 1051·26 ± 0·21 × 1053·3 × 104
Soil 2 (Soliman, Tunisia)8·50·90·41·40·732Sandy4·18 ± 0·67 × 1051·50 ± 0·42 × 1043·3 × 103
Soil 3 (Grezzano, Italy) + Rm1021NdNdNdNdNdNdClay1·64 ± 0·87 × 1061·89 ± 0·32 × 1063·4 × 106
Soil 3 (Grezzano, Italy) + Rm1021NdNdNdNdNdNdClay3·58 ± 0·46 × 1052·21 ± 0·34 × 1053·4 × 105
Soil 3 (Grezzano, Italy) + Rm1021NdNdNdNdNdNdClay2·58 ± 0·20 × 1041·89 ± 0·42 × 1043·4 × 104
Soil 3 (Grezzano, Italy) + Rm1021NdNdNdNdNdNdClay5·68 ± 3·21 × 1027·43 ± 0·25 × 1023·4 × 103
M. truncatula root       4·92 ± 0·89 × 1061·44 ± 0·22 × 1063·6 × 106


In this paper, we reported the development of qPCR assays for the detection and quantification of S. meliloti cells in soil and plant samples. Primer pairs designed to amplify portions of the chromosomal gene rpoE1 and of the pSymA gene nodC of S. meliloti gave successful results in terms of specificity on both pure cultures and on DNA extracted from soil and plant tissues. qPCR assays based on both rpoE1 and nodC gave reliable results down to 3·4 × 103 cells g−1 of soil, which is quite lower than for other qPCR protocols in other rhizobial species (Duodu et al. 2005). In plant roots nodulated with S. meliloti Rm1021, both assays were well in agreement with viable titres after surface sterilization. In native soil from Soliman, estimated qPCR titres were one–two orders of magnitude higher than MPN estimates. As reported also for other rhizobial species (Duodu et al. 2005), this higher qPCR estimates could suggest that qPCR approach could be more effective than nodulation-based protocols, allowing the estimation of the whole S. meliloti population which includes both strains nodulating the plants and free-living in soil (Duodu et al. 2005). It is interesting to note here that the titre of Soliman soil 2 estimated with nodC was significantly lower than that obtained with rpoE1, better approximating the MPN-estimated titre. This evidence could suggest the possibility that the chromosomal gene rpoE1 target a wider fraction of S meliloti populations which could not harbour the symbiotic-required gene nodC. Actually, earlier reports on Rh. leguminosarum indicated that only a small fraction of the rhizobial population contains genes required for symbiosis with host plant (Soberòn-Chavez and Nàjera 1989; Segovia et al. 1991). However, further investigations are necessary to elucidate the biological significance of the observed differences.

In conclusion, the qPCR protocols presented here, targeting rpoE1 and nodC genes of S. meliloti can be used as a fast and complementary tool for the estimation of the number of S. meliloti, DNA and cells in soil samples and plant tissues.


We are grateful to Dr Wayne Reeve (Murdoch University, Western Australia) for kindly providing the strain S. medicae WSM419.