Amplified ribosomal DNA restriction analysis and repetitive element polymorphism-polymerase chain reaction ( rep-PCR ) DNA fingerprinting of members of Frankia genus

Slow growing actinobacteria of the genus Frankia are best known for their nitrogen-fixing mutualism with dicotyledonous host plants called actinorhizal plants. Twenty nine (29) strains obtained from diverse host plants and geographic area, have been studied based on amplified ribosomal DNA restriction analysis (ARDRA) and repetitive element polymorphism-polymerase chain reaction (repPCR) DNA fingerprinting using BOX A1R primer. The collection has been classified into 28 ARDRA haplotypes clustered into three genogroups. The first genogroup 1 contains Frankia strains infecting Elaeagnus, genogroup 2 includes strains infecting Casuarina, while Frankia strains infective on Alnus, Comptonia and Myrica were grouped in genogroup 3. The results of BOX-PCR fingerprinting, supported the observation that BOX-PCR seems to be able to discriminate Frankia at strain level but are not useful for assigning strain to their respective genogroups or host infection groups.


INTRODUCTION
Frankia genus contains diazotrophic actinobacteria that are able to establish root nodules with diverse dicotyledonous host plants known as actinorhizal plants that have attracted interest with regard to the input of fixed nitrogen on marginal soil where indigenous legumes are absent (Gtari and Dawson, 2011).Frankia strains grow slowly with doubling times of 15 to 48 h or more leading to difficulty in arranging strains into phenotypically related groups (Benson and Silvester, 1993).However, the host-specific responses of the Frankia strains remain mostly the only useful criteria to group them into four major host infectivity groups (HSGs) (Backer, 1987).HSG 1 is composed of strains that infect Alnus, Comptonia and Myrica; HSG 2 strains infect members of the Allocasuarina, Casuarina and Myrica; HSG3 strains infect members of the Elaeagnaceae, Rhamnaceae, Gymnostoma and Myrica and HSG4 contained strains that nodulate members of the Elaeagnaceae but not Myrica.Phylogenetic analysis based on entire 16S rRNA gene sequences permitted to assign Frankia strains to four clusters (Normand et al., 1996).Cluster 1 includes Frankia strains which form nodules on members of Betulaceae, Myricaceae and Casuarinaceae.In cluster 2 are grouped Frankia strains that only infect members of the Coriariaceae, Datiscaceae, Rosaceae and Ceanothus of the Rhamnaceae.Cluster 3 strains form effective nodules on members of the Myricaceae, Rhamnaceae, Elaeagnaceae and Gymnostoma of the Casuarinaceae.Despite the fact that Frankia strains from cluster 1 and 3 are being routinely isolated and cultivated, those from cluster 2 have not been isolated in pure culture despite many attempts and remain, therefore, considered as an obligate symbionts.Atypical Frankia strains that are unable to infect or fix nitrogen are included in cluster 4.These clustering have been confirmed by other molecular studies such as intertranscribed spacers (ITS) 16S-23S rRNA (Ghodhbane- Gtari et al., 2010), gyrB (Nouioui et al., 2011) and glnII (Gtari et al., 2004;Nouioui et al., 2011) gene sequence analysis.Due to the slow growth of Frankia strains and the limited funding for maintaining several collections especially those containing unidentified and uncharacterized strains, there is need for reducing the costs of shortages without risk of losing biodiversity (Lumini and Bosco, 1999).As a result of this, the present study tests the efficiency of some genetic fingerprinting and low cost based techniques for worldwide and routine characterization of Frankia isolates.

Bacterial strains and culture conditions
Twenty nine (29) Frankia strains were used in the present study (Table 1).Cultures are routinely subcultured at 28°C in DPM medium (Baker and O'Keefe, 1984) modified to contain as carbon source in addition to Na-propionate, Na-pyruvate, Na-succinate, Na-acetate and glucose to accommodate strain specific requirements.

DNA extraction, PCR amplification and amplified ribosomal DNA restriction analysis
DNA extraction was made from one month old liquid culture of the Frankia strains after forcing several bacterial colonies (from a culture volume of 1-5 ml) through a 0.7 × 30 mm sterile needle to homogenize the mycelium.After centrifugation, the resulting cell pellet was washed twice with sterile distilled water, incubated for 30 min in DNA extracting buffer (100 mmol l -1 Tris-HCl, pH 8; 20 mmol l -1 EDTA, pH 8.2; 1.4 M NaCl, and 2% w/v cetyl trimethyl ammonium bromide (CTAB), chloroform extracted and ethanol precipitated.The DNA pellet was dissolved in 50 µl TE (10 mmol l -1 Tris-HCl, pH 8; 20 mmol l -1 EDTA, pH 8.2).PCR reaction of the 16S rRNA gene was carried out by using FGPS56-352 and FGPS1509'-153 following conditions described by Normand et al. (1996).PCR amplification were performed in 100 µl final reaction volume containing 10 ng genomic DNA, 1X Taq polymerase buffer, 1.5 mmol l -1 MgCl 2 , 0.1 µM each dNTP, 0.2 µM each primers and 2 U Taq DNA polymerase.The thermal program consisted of three min at 95°C followed by 35 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 45s.PCR products were digested with restriction enzymes; AluI, HaeIII and RsaI, overnight at the optimal conditions recommended by the manufacturer.Repetitive element polymorphism-polymerase chain reaction (rep-PCR) was performed in 25 µl final volume using 50 ng genomic DNA, 1X Taq polymerase buffer, 2.5 mmol l -1 MgCl 2 , 0.5 µM each dNTP, 0.5 µM BOX-A1R primer (Versalovic et al., 1994), 0.04 U µl -1 Taq DNA polymerase and 5% (v/v) DMSO, and subjected to a thermal program: 95°C for 5 min, 35 cycles consisting of 1 min at 94°C, 1 min at 45°C and 2 min at 72°C.Amplified ribosomal DNA restriction analysis (ARDRA) and BOX-PCR products were electrophoresed in 2.5% agarose in TBE buffer (Sambrook et al., 1989), ethidium bromide stained and photographed under ultraviolet light.
Fingerprints were analyzed using GelCompar II v. 6.5 (Applied Maths NV) and dendrograms were constructed using unweighted pair group method with arithmetic mean (UPGMA) algorithm based on Dice similarity coefficient for ARDRA and Pearson's correlation coefficient for BOX-PCR similarity matrix.

RESULTS AND DISCUSSION
Beside difficulties in isolating Frankia, the nitrogen-fixing actinobacteria and symbionts of actinorhizal plants, identifying and conserving diversity of cultured strains remain problematic in view of scarcity of financial supports in several still devoted laboratories in the field.Moreover, the non useful phenotyping methods especially the time consuming of plant infecting experiments suggest for developing a low cost and world widely used method for routine characterization of Frankia isolates.In the present study, 29 Frankia reference strains and isolates were characterized by ARDRA and repetitive element polymorphism-PCR (BOX-PCR) methods.Individual AluI, HaeIII and RsaI restriction patterns of the 16S rRNA gene amplicon and UPGMA dendrogram were shown in Figure 1.The first genogroup 1 contains Frankia strains NRRL-B16219, NRRL-B16306, NRRL-B16316, NRRL-B16412, EAN1pec, BMG5.3, BMG5.11,BMG5.13 and BCU110501 infecting Elaeagnus, genogroup 2 includes strains; KB5 and CcI3 infecting Casuarina, while Frankia strains infective on Alnus (ArI3, AvcI1, AvsI1, ACN14a and ACN1), Comptonia (CpI1) and Myrica (NRRL-B16386) are grouped in genogroup 3.Moreover, this grouping based on ARDRA is driven by host plant infectivity rather than host plant origin of isolation.Some Casuarina (NRRL-B16306 and NRRL-B16412) and Ceanothus (NRRL-B16316) strains that are non infective on the latter host plant but infective on Elaeagnus (Baker, 1987) grouped accordingly to genogroup 1. Likewise strains isolated from different host plants that cross-infect the same plant sets are included in the same genogroup such as those isolated from Alnus, Comptonia and Myrica that cluster with Alnus infective strains.Such as CpI1 and NRRL-B16386 isolated   (example Casuarina strains) indicate noteworthy genome variability among Frankia strains.While grouping is possible at high cutoff, the generated genogroups are not correlated to host infection groups that are determined in this study based on amplified ribosomal DNA restriction analysis.The high differences observed by BOX-PCR even on closely related Frankia strains was previously reported by Gtari et al. (2004), Murry et al. (1987) and Jeong and Myrold (1999) and seems to be a general feature among Frankia genus that reflect soil effect rather than host effect on Frankia genome variability.This suggests that repetitive element polymorphism-PCR using BOX-A1R primer is a reliable technique for strains discrimination among Frankia genus.Amplified ribosomal DNA restriction analysis has not been often used to characterize Frankia strains.Excepting Gtari et al. (2007) who performed an in silico ARDRA of entire diversity of Frankia 16S rDNA sequences retrieved from GenBank covering all host specificity groups and Huguet et al. (2004) who were interested only on Myricaceae isolated and uncultured strains directly in root nodules.Our ARDRA study may be the first report on assessment of Frankia diversity on a large collection of reference strains and isolates.The study demonstrated the feasibility and utility of ARDRA as fast and low cost based techniques for worldwide and routine characterization of Frankia isolates.

Figure 1 .
Figure 1.UPGMA cluster analysis using Dice coefficient of ARDRA digitized banding patterns generated by restriction digestions with AluI, HaeIII and RsaI enzymes of the 16S rRNA gene amplicon from Frankia reference strains and isolates.

Figure 2 .
Figure 2. Cluster analysis of BOX-PCR fingerprints of Frankia reference strains and isolates using UPGMA algorithm based on Pearson's correlation coefficient.

Table 1 .
Frankia reference strains and isolates used in this study.