RESEARCHFree Access icon

Development of a DNA-Based Real-Time PCR Assay To Quantify Allorhizobium vitis Over Time in Grapevine (Vitis vinifera L.) Plantlets

    Affiliations
    Authors and Affiliations
    • Trong Nguyen-Huu1
    • Jeanne Doré2
    • Essaïd Aït Barka1
    • Céline Lavire2
    • Christophe Clément1
    • Ludovic Vial2
    • Lisa Sanchez1
    1. 1Unité EA 4707 Résistance Induite et Bioprotection des Plantes, SFR Condorcet FR Centre National de la Recherche Scientifique (CNRS) 3417, Université de Reims Champagne-Ardenne, Reims, France
    2. 2UMR Ecologie Microbienne, CNRS, National Research Institute for Agriculture, Food and Environment, VetAgro Sup, Université Claude-Bernard Lyon, Université de Lyon, F-69622 Villeurbanne, Lyon, France

    Published Online:https://doi.org/10.1094/PDIS-04-20-0732-RE

    Abstract

    Allorhizobium vitis is the primary causal pathogen of grapevine crown gall disease. Because this endophytic bacterium can survive as a systemic latent (symptomless) infection in grapevine, detecting and monitoring its development in planta is of great importance. In plant bacteria studies, plate counting is routinely used as a simple and reliable method to evaluate the bacterial population level in planta. However, isolation techniques are time-consuming and present some disadvantages such as the risk of contamination and the need for fresh samples for research. In this study, we developed a DNA-based real-time PCR assay that can replace the classical method to monitor the development of Allorhizobium vitis in grapevine plantlets. Primers targeting Allorhizobium vitis chromosomic genes and the virulent tumor-inducing plasmid were validated. The proposed quantitative real-time PCR technique is highly reliable and reproducible to assess Allorhizobium vitis numeration at the earliest stage of infection until tumor development in grapevine plantlets. Moreover, this low-cost technique provides rapid and robust in planta quantification of the pathogen and is suitable for fundamental research to monitor bacterial development over time.

    Crown gall disease is one of the most important bacterial diseases of grapevine (Vitis vinifera L.) worldwide, leading to a significant yield decrease and, in the worst case, to plant death (Burr et al. 1998; Schroth et al. 1988). The main pathogenic bacterium responsible for this disease is still frequently known under the name Agrobacterium vitis (Ophel and Kerr 1990) but was reclassified in the genus Allorhizobium (Mousavi et al. 2014, 2015). These bacteria are pathogenic only when they carry a tumor-inducing plasmid that harbors all virulence genes required to establish and develop the tumor. Briefly, after plant wound infection by Allorhizobium vitis, the transfer DNA (T-DNA) region of tumor-inducing plasmids is transferred and incorporated into the plant genome, then expressed in the host plant cells (for review, see Gelvin 2017 and Nester 2015). The T-DNA oncogenes prompt plant cell division leading to gall development. T-DNA also contains genes for the biosynthesis of a diverse group of metabolites called opines. These molecules serve as a specific nutrient source for Allorhizobium vitis that has corresponding catabolic genes present on tumor-inducing plasmids and therefore play a role in host competition with other bacteria (Gelvin 2017; Nester 2015). Opines are key molecules of tumor-inducing plasmid ecology; consequently, tumor-inducing plasmids are often classified according to opine type synthesis by plant cells. For Allorhizobium vitis, the most common types of tumor-inducing plasmids are referenced as nopaline, vitopine, and octopine/cucumopine (Argun et al. 2002; Kuzmanović et al. 2014; Szegedi et al. 1988). Some tumor-inducing plasmids that drive the biosynthesis of both octopine and vitopine were reported, suggesting a possible octopine/vitopine tumor-inducing plasmid (Bini et al. 2008a; Habbadi et al. 2020; Kuzmanović et al. 2018).

    Like many pathogenic bacteria, Allorhizobium vitis thrives primarily within its host plant and can sometimes grow in other biotopes or survive in the soil under favorable conditions (Burr and Katz 1983; Schroth et al. 1971; Spiers 1979). In nature, both biotic (e.g., nematodes; Süle et al. 1995) and abiotic (e.g., freezing; Stover et al. 1997) stresses can generate open wounds, which serve as an entrance for the pathogen. This influences the Allorhizobium vitis-mediated infection process and, consequently, the crown gall disease outbreak. Allorhizobium vitis has the capacity to develop endophytically in grapevine (Lehoczky 1968). The bacteria mainly occupy the xylem vessels, which allow them to migrate even to shoot meristems and leaf surfaces (Burr et al. 1998; Johnson et al. 2016). Once established in vineyards, the endophytic bacterium Allorhizobium vitis may persist permanently (Kuzmanović et al. 2018). To control the disease, detecting the presence of the pathogen in grapevine materials and in the soil prior to planting is of strategic importance. For that purpose, PCR-based methods have been developed and have proven high efficacy (Burr et al. 2017), such as PCR with specific primers (Bini et al. 2008a, b; Schulz et al. 1993; Szegedi and Bottka 2002), quantitative real-time PCR (qPCR) (Bini et al. 2008a; Ferrigo et al. 2017), magnetic capture hybridization qPCR (MCH) (Johnson et al. 2013, 2016; Orel et al. 2017), and digital droplet PCR (ddPCR) (Voegel and Nelson 2018). Among these methods, ddPCR (the latest developed) allows direct absolute quantification of the Allorhizobium vitis population; however, compared with qPCR, ddPCR requires more time to set up, is more expensive, and is not routinely used thus far (Voegel and Nelson 2018).

    In basic research, precisely quantifying pathogen populations in planta during the infection process under controlled conditions is critical to evaluate the response of the plant to a known amount of pathogen, determine the colonization profile (e.g., mutant strains compared with wild-type strains), or screen the efficiency of potential biological control agents. Classically, for fundamental investigations on plant–bacteria interactions, plate counting is used as a simple method for the quantification of bacteria in plant samples. During this process, bacterial cells are extracted from plant tissues and plated on appropriate media in a serial dilution to determine the number of viable bacterial cells as the quantity of CFUs per plant biomass. Nevertheless, this procedure presents some disadvantages (Ross and Somssich 2016). Indeed, a well-defined approach during sampling is needed because of the heterogenic distribution of microorganisms in their hosts and their low quantity in asymptomatic plants. In addition, collected samples cannot be stored and must be processed immediately to ensure the accuracy of the living bacterial populations inside or on the samples. These barriers limit the possibility of measuring very low levels of bacteria or analyzing short-interval time points. Moreover, the plate counting method is labor intensive and requires a sufficient number of replicates because it is vulnerable to repetitive technical mistakes during sample collection, serial dilution pipetting, or colony counting on Petri dishes.

    To characterize the colonization profile of the pathogenic bacterium at different stages of infection more carefully and to further rapidly test the efficacy of some biological control agents, we developed a pathosystem with grapevine plantlets/Allorhizobium vitis strain S4 (AvS4) under controlled conditions in the laboratory. This study aimed to compare two methods of quantification plate counting versus qPCR in this pathosystem. For this purpose, we used a previously known primer pair (virD59 targeting virD2; Bini et al. 2008a) and we defined new primer pairs targeting either virD3 on tumor-inducing plasmids that displayed the largest diversity of conserved vir genes (Vogel and Das 1992) or specific Allorhizobium vitis chromosomic genes (genes present in all strains of this species but absent in other species).

    We here propose an optimized qPCR analysis of Allorhizobium vitis quantification and show that this DNA-based method can replace the conventional plate counting assay to monitor Allorhizobium vitis growth in grapevine plantlets, from the early stage of infection (before symptom emergence) until tumor development.

    Materials and Methods

    Microorganisms.

    All bacterial strains used in this study are listed in Supplementary Table S1. Allorhizobium vitis strains were cultivated overnight in mannitol-glutamate (MG) liquid medium (Keane et al. 1970) at 28°C and 180 revolutions per minute (rpm). Other bacteria species were cultivated overnight at 180 rpm in King’s B liquid medium (King et al. 1954) at 28°C except Escherichia coli K12, which was cultivated at 37°C.

    Plant infection.

    Plantlets of V. vinifera ‘Chardonnay’ were micropropagated as described in Ait Barka et al. (2006). Five-week-old in vitro plantlets were transferred and grown on soil in Magenta boxes under a 16-h/8-h day/night photoperiod at 26°C for 1 additional week before infection with AvS4. For infection experiments, after overnight growth in MG medium, AvS4 cells were collected after centrifugation at 4,500 × g at 4°C for 15 min and washed twice with sterile phosphate-buffered saline (PBS; 10 mM, pH 6.5). The bacteria concentration was determined by spectrophotometry (600 nm) and adjusted in PBS to 109 CFU/ml (OD600 = 1.0 in PBS). The shoots of grapevine plantlets were wounded with a sterile surgical blade and inoculated with the AvS4 suspension.

    DNA extraction.

    Genomic DNA (gDNA) was extracted from 1 ml of bacterial culture suspension or 40 mg of plant powder (ground in liquid nitrogen), using the Wizard Genomic DNA Purification Kit Protocol (Promega, Madison, WI) according to the manufacturer’s instructions. DNA sample concentration and quality was measured with NanoDrop Microvolume spectrophotometers and fluorometers and visualized by migration in 1% agarose gel. Extracted DNA was stored at −20°C until use.

    Primer design and specificity test.

    To develop the qPCR method to quantify Allorhizobium vitis, four original primer pairs were designed based on the genome sequence of AvS4 (GenBank CP000633 through CP000639; Slater et al. 2009) using Primer-BLAST (Ye et al. 2012), except for virD59 designed according to Bini et al. (2008a) (Table 1). The level of specificity of each primer was checked in silico against the currently known sequences using the Nucleotide Collection Database and the Whole-Genome Shotgun Contigs Database with Nucleotide blast (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and with Primer-blast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Next, the primers were tested empirically in qPCR with DNA extracted from different bacterial culture suspensions.

    Table 1. Primers used in this research

    qPCR.

    For qPCR, each well contained 50 ng of template DNA, 0.14 µM of each specific forward and reverse primer, and 7.5 µl of ABsolute Blue SYBR Green ROX (Thermo Scientific) in a total volume of 15 µl. Reactions were performed with the CFX96 Touch Real-Time PCR Detection System (Bio-Rad) under the following thermal cycle protocol. Samples were preheated to 95°C for 15 min. Then, 35 amplification cycles were run: 95°C for 10 s and 60°C for 45 s. Fluorescence (521 nm) was measured at the end of each cycle. After the last amplification cycle, melting curves were acquired by slowly heating from 65 to 95°C at 0.1°C per second with continuous measurement of fluorescence at 521 nm. The threshold to determine quantification cycle (Cq) values was set at 60 relative fluorescence units for all primers. Two technical replicates were performed for each sample from all experiments.

    Phylogenetic analysis.

    The phylogeny of all bacteria used in this study was inferred from the recA gene retrieved from the NCBI database. For Allorhizobium vitis strains, one recA sequence (named avi1 to avi11) was selected for each recA allele as defined by Habbadi et al. (2020) and was complemented with the recA sequence from all Allorhizobium vitis sequenced genomes available in the database. Sequences were aligned using the Clustal ω2 program within the SeaView 4 package (Gouy et al. 2010). Neighbor-joining and maximum likelihood methods were performed by using the SeaView graphical user interface (http://doua.prabi.fr/software/seaview) with parameters as described by Habbadi et al. (2020).

    Primer limits of bacterial detection in vitro.

    To primarily determine their limit of AvS4 detection, primers were tested on a 10-fold serial dilution of pure AvS4 gDNA, from 50 ng to 0.05 pg, either spiked or not with a background of 50 ng of pure gDNA of grapevine. Next, the limit of detection of primers was further tested in terms of AvS4 cells per weight of plant tissue. For this purpose, sterile fresh shoot tissue was collected from 6-week-old in vitro plantlets and ground carefully in liquid nitrogen with sterile pestles and mortars into a fine powder. The suspension of 109 AvS4 CFU/ml (OD600 = 1.0 in PBS) was diluted 10-fold in sterile PBS to obtain a serial dilution from 108 to 102 CFU/ml. Then, every 30 µl of each concentration was added to 30 mg of sterile fine powder of shoot tissue, resulting in a series of concentrations ranging from 105 to 10−1 CFU per milligram of plant fresh weight (AvS4 CFU/mg FW). Sterile plant shoot powder with no AvS4 was used as the control. The gDNA of all mixtures was then extracted as mentioned previously. This experiment was independently repeated three times with two replicates for each repetition.

    In planta bacterial quantification.

    To monitor the in planta growth of AvS4, the shoots of grapevine plantlets were wounded with a sterile surgical blade and inoculated with 3 µl of the AvS4 suspension at 109, 107, and 105 CFU/ml corresponding to the final quantities of 3.106, 3.104, and 3.102 CFU, respectively. At 7 days postinfection (dpi), 1-cm-long shoot segments around the infection point were collected for the plate counting and qPCR assays. For the plate counting method, three replicates of three shoot segments (n = 9) were weighed and ground in 1 ml of sterile PBS with sterile pestles and mortars. The homogenates were diluted 10-fold and cultured on MG medium. The number of AvS4 colonies was counted after 2 days of incubation at 28°C. For qPCR, DNA extraction of the other two replicates of three shoot segments (n = 6) was realized with the Wizard Genomic DNA Purification Kit as mentioned previously. Two independent biological repetitions were done.

    Next, to monitor bacteria growth over time in planta, shoots of grapevine plantlets were infected with 3 µl of the AvS4 suspension at 105 CFU/ml (corresponding to 3.102 CFU inoculated). For each time point (1, 3, 7, 14, and 21 dpi), three replicates of three shoot segments (n = 9) were collected to quantify the bacterial populations by plate counting and three replicates of three shoot segments (n = 9) were collected for qPCR. Three independent biological repetitions were done.

    Statistical analysis.

    Results were statistically analyzed with SPSS Statistics 20 software. A linear regression analysis was performed to assess a possible relationship between AvS4 DNA quantity and Cq values, the number of AvS4 CFU infected, and the log (AvS4 CFU/mg FW). Significant differences in the log (AvS4 CFU/mg FW) and Cq values were analyzed using multiway analysis of variance and the post hoc Tukey test with a 95% confidence level (P < 0.05).

    Results

    Primer specificity.

    To develop a qPCR-based method for Allorhizobium vitis quantification, four original primer pairs were designed and a primer pair (virD59F and virD59R) already used to detect Allorhizobium vitis was also included (Table 1). A recA phylogeny was determined in order to have a general view of Allorhizobium vitis diversity and the relatedness of Allorhizobium vitis strains with other species used in this study (Supplementary Fig. S1). Primers were initially tested in silico before empirical testing with DNA extracted from bacteria that belonged to the genus Allorhizobium (Allorhizobium undicola LMG11875 and Allorhizobium taibaishanense LMG27055), were phylogenetically close to Allorhizobium vitis (Agrobacterium tumefaciens C58, Rhizobium rhizogenes K84, Ensifer meliloti 1021), or were phylogenetically more distant (Burkholderia thailandensis E264, Pseudomonas brassicacearum NFM421, Escherichia coli K12) (Supplementary Fig. S1). In qPCR, primers did not cross with the grapevine DNA but gave signals when tested with samples of grapevine shoot infected with AvS4 (Cq ≈ 19) (Supplementary Fig. S2).

    For all primers, an early amplification (Cq ≈ 12) could only be observed with samples containing Allorhizobium vitis DNA (Supplementary Fig. S2). Moreover, primers targeting chromosomal genes, Avi_1889 and Avi_5072, allowed detection experimentally of all tested Allorhizobium vitis strains (Fig. 1). Indeed, avi_1889 and avi_5072 were conserved among Allorhizobium vitis strains, but these genes were not present in the other bacteria even in the other Allorhizobium species such as Allorhizobium undicola and Allorhizobium taibaishanense (Supplementary Figs. S2 and S3). Primers targeting chromosomal genes were also tested in silico against all Allorhizobium vitis whole-genome sequences. In silico analysis suggested that these primers detected the majority (but not all) of Allorhizobium vitis strains (Supplementary Fig. S3). For CT1-2 and KT1-1 strains with primer targeting avi_5072, we detected amplification (Cq ≈ 12) despite the presence of one primer-template mismatch (Supplementary Figs. S2 and S3).

    Fig. 1.

    Fig. 1. Primer specificity among seven Allorhizobium vitis strains. Primers targeting chromosomic genes (Avi_1889 and Avi_5072) and the tumor-inducing plasmids (virD59, virD3, and virD3cons) were tested on genomic DNA extracted from different Allorhizobium vitis strain culture suspensions (S4, BT2-2, BT3-1, CT1-2, ET2-10, KFB264, and KT1-1) and a no-template (water) control. The experiment was done with two technical replicates. Error bars represent the SD. RFU = relative fluorescence unit.

    Download as PowerPoint

    Primers targeting the virulence genes on tumor-inducing plasmids were also developed. The primer sets virD59 and virD3, targeting virD2 and virD3, respectively, detected mainly vitopine strains (e.g., Allorhizobium vitis S4 strain) (Cq ≈ 12; Fig. 1) but also other strains such as KT1-1, ET2-10, BT2-2, and BT3-1 to a lesser extent (Cq ≈ 24). By using virD3, we developed primers (named virD3cons) to detect both vitopine strains (e.g., Allorhizobium vitis S4, BT3-1), octopine/cucumopine strains (e.g., Allorhizobium vitis KFB264), and octopine/vitopine strains (e.g., Allorhizobium vitis BT2-2, ET2-10, KT1-1, CT1-2) (Fig. 1). However, for virD59, virD3, and virD3cons, a latter amplification (Cq ≈ 23 to 25) was observed for Agrobacterium tumefaciens C58, B. thailandensis E264, and Ensifer meliloti 1021 (Supplementary Fig. S2).

    Primer limit of detection.

    All primers tested using a 10-fold dilution series of pure AvS4 DNA could amplify until 0.5 pg of pure DNA of AvS4 (Fig. 2). For lower template concentrations or the water control, no signal or only unspecific products were noticed according to the melting curves obtained for each primer pair (Table 1). In addition, when mixed with background grapevine DNA, primers were able to detect until a 1:100,000 AvS4:grapevine DNA ratio. The linear trendlines of all primers obtained correlated with those of pure AvS4 template DNA, proving no inhibition effect of the plant DNA matrix on primer activity (Fig. 2).

    Fig. 2.

    Fig. 2. Primer validation for quantification of genomic DNA (gDNA) of Allorhizobium vitis strain S4 (AvS4). Black squares indicate DNA extracted from the AvS4 culture suspension (S4), and orange circles indicate a mixture of DNA extracted from the AvS4 culture suspension and from the grapevine Vitis vinifera plant (Vv+S4). Correlation coefficients (R2) are indicated with the corresponding color of each dataset. Linear regressions were obtained with two technical replicates for each primer pair. The lower limit of the linear range determined the limit of quantification. Cq = quantification cycle.

    Download as PowerPoint

    Furthermore, to determine the detection threshold in planta, the qPCR results showed a highly linear correlation (R2 > 0.98) between the Cq values and the log (AvS4 CFU/mg FW) for all primers tested (Fig. 3). The lowest limit of detection of all primers was 1.102 AvS4 CFU/mg FW. For lower-quantity and no-template control wells, no signal or unspecific products were observed.

    Fig. 3.

    Fig. 3. Limit of detection of primers for quantification of Allorhizobium vitis strain S4 (AvS4) cells in planta. Correlation coefficients (R2) are indicated. Linear equations demonstrate the relation between the quantification cycle (Cq) value and the bacterial quantity for each primer pair. Data represent means ± SD of three biological repetitions realized in duplicate (n = 6) for each primer pair. The lower limit of the linear range determined the limit of quantification. FW = fresh weight.

    Download as PowerPoint

    In planta bacterial quantification.

    To monitor AvS4 growth on grapevine (Fig. 4A), we compared the conventional plate counting method (Fig. 4B) with qPCR analysis (Fig. 4C). Results at 7 dpi showed that the qPCR method gave similar quantification profiles compared with the plate counting method (Fig. 4D; Supplementary Fig. S4) (R2 > 0.98), meaning that the hereby optimized qPCR method can monitor the AvS4 population in grapevine.

    Fig. 4.

    Fig. 4. Comparative analysis of two quantification methods for Allorhizobium vitis strain S4 (AvS4) growth levels at 7 days postinfection (dpi) from different initial numbers of CFU inoculated. A, A 1-cm stem segment collected for plate counting or quantitative real-time PCR (qPCR). Bacterial levels in planta were quantified by B, plate counting and C, qPCR methods with primer Avi_1889. Data represent means ± SDs of three replicates of three stem segments (n = 9) for plate counting and two replicates of three stem segments (n = 6) for qPCR. D, Results of the two assays are plotted against each other. Correlation coefficients (R2) are indicated. These results represent one of two independent biological repetitions. FW = fresh weight and Cq = quantification cycle.

    Download as PowerPoint

    Bacterial quantification over time.

    To monitor bacterial growth in planta over time, the bacterial populations were measured by the two methods at 1, 3, 7, 14, and 21 dpi. Morphologically, until 7 dpi, only slight swelling at the wounding sites on grapevine shoots was observed. Tumors started to appear later and steadily developed until the end of the experiment (Fig. 5A). The bacterial populations increased significantly and reached a peak at 3 dpi (106 CFU/mg FW). They remained stable until 7 dpi before decreasing steadily at 14 and 21 dpi (around 105 CFU/mg FW) (Fig. 5B). The results from qPCR with the five primer pairs also showed the same profile as the plate counting method (Fig. 5C; Supplementary Fig. S5). Plotting the data of both methods in the same graphs resulted in a highly linear correlation with R2 > 0.98, meaning that these two methods are highly comparable (Fig. 5D; Supplementary Fig. S5).

    Fig. 5.

    Fig. 5. Comparative analysis of two quantification methods for Allorhizobium vitis strain S4 (AvS4) growth levels over the time course of infection. A, Tumor growth over 21 days of the experiment. Bars = 5 mm. Bacterial growth was determined by B, the plate counting method and C, quantitative real-time PCR with primer virD59. Data represent means ± SD of three replicates of three stem segments (n = 9) for both methods. D, Results of the two assays are plotted against each other. Correlation coefficients (R2) are indicated. Data with the same letters have no significant differences according to the Tukey test at P < 0.05. These results represent one of three independent biological repetitions. dpi = days postinfection and Cq = quantification cycle.

    Download as PowerPoint

    Discussion

    Allorhizobium vitis is an endophytic pathogenic bacterium that can systemically infect grapevine. Quantification of bacterial populations is essential to study their interactions with plant hosts, especially in the case of endophytic bacteria. Several studies used an “artificial” inoculation of the pathogenic bacterium under controlled conditions to monitor population dynamics in planta, deeper investigate the impact of pathogenic strains on plants, or test potential molecules or biological control agents (Hao et al. 2018; Jung et al. 2016; Kaewnum et al. 2013; Kawaguchi 2014). Most of these studies used the fastidious classical plate counting method to determine the level of pathogenic bacteria in different grapevine tissues. Although many studies have developed primers to identify or detect the pathogenic bacteria in infected or asymptomatic tissues (Bini et al. 2008a, b; Schulz et al. 1993; Szegedi and Bottka 2002), few have compared culture-dependent enumeration and qPCR (Bini et al. 2008a). qPCR is a widespread method used in numerous pathosystems to detect and quantify microorganisms in the interaction with their hosts (Abdullah et al. 2018; Brouwer et al. 2003; Gachon and Saindrenan 2004; Martin et al. 2000; Weßling and Panstruga 2012). This technique provides less variability compared with culturing-dependent methods (Li et al. 2008; Llorente et al. 2010).

    Here, we compared this alternative DNA-based approach and the classical plate counting method to quantify in planta Allorhizobium vitis S4 populations after artificial inoculation on grapevine plantlets shoots in the laboratory. We used one primer pair for virD59 described previously (Bini et al. 2008a), and we also developed primers based on the genome sequence of AvS4 (Fig. 1) targeting both chromosomic genes and virulence genes on tumor-inducing plasmids. The two primer pairs targeting chromosomal genes (avi_1889 and avi_5072) allowed us to exclusively detect Allorhizobium vitis strains tested. Allorhizobium vitis strains can be classified according to the types of opines that their tumor-inducing plasmids produce (Habbadi et al. 2020; Kuzmanović et al. 2018). In this study, we chose seven Allorhizobium vitis strains representing vitopine-, octopine/cucumopine-, and octopine/vitopine-type tumor-inducing plasmids. We worked on three primer pairs targeting the vir region of these three types of tumor-inducing plasmids. The virD3cons primers, targeting a consensus sequence among all three types of tumor-inducing plasmids, proved their ability to amplify all Allorhizobium vitis strains tested. However, one should keep in mind that these results gave just a very primary overview of the specificity of these primers toward Allorhizobium vitis strains and other bacterial species. Because of the low numbers of strains involved and primer-template mismatch, these results are only applicable within Allorhizobium vitis strains tested. For other purposes such as indexing samples from vineyards with the presence of more diverse microorganisms, further investigation must be realized. In the same way, the primers developed here cannot be used to diagnose the presence of Allorhizobium vitis in grapevine crown gall or for epidemiological studies so far.

    In qPCR, primer sensitivity is expressed as the limit of detection, which is the concentration of the template that can be detected with 95% certainty (Bustin et al. 2009). Our results showed that five primer pairs could quantify different levels of AvS4 gDNA until 0.5 pg. These results are comparable to those obtained by Abdullah et al. (2018), who were able to detect until 0.5 pg of fungal DNA in wheat by qPCR. Even in the presence of grapevine DNA matrix, their activity was not influenced (Fig. 2), indicating that these primer pairs can be used for in planta quantification. By testing a series of Allorhizobium vitis S4 concentrations ranging from 105 to 10−1 CFU/mg FW, we showed that the lowest limit of detection of all primers in planta was 102 CFU/mg FW. We also tested different initial bacterial titers (3.102, 3.104, and 3.106 CFU) for shoot inoculation. Our results confirmed the ability of qPCR to monitor the bacterial population level in planta from a low to high level. Moreover, the comparison of the two methods showed a good correlation (R2 > 0.97) at different levels of infection, from the early stage (at 1 dpi) until tumor development (at 21 dpi). However, we observed an overestimation of the AvS4 population by qPCR, owing to amplification of nondegraded DNA material from dead bacterial cells as described previously (Ross and Somssich 2016). Our DNA-based method is not as sensitive as MCH (Johnson et al. 2013) or ddPCR (Voegel and Nelson 2018) used for in planta detection. Nevertheless, it is less complicated than MCH, which needs a step to enrich the pathogen DNA, and it is more popular than ddPCR so far. Moreover, since our method gave robust, repetitive, and high similarity with plate counting assays, it can replace the latter method in basic research (e.g., colonization profile). The qPCR method allows temporary storage of samples and, hence, analysis at various time points. A large number of samples can be processed simultaneously and results can be achieved within only a few hours.

    Conclusion

    In this study, we validated new primer pairs for the use of qPCR as an alternative and highly comparable method to the plate counting assay. The proposed method allows monitoring of Allorhizobium vitis growth in V. vinifera shoots before emergence of the first symptoms until full tumor development. Its accuracy, rapidity, flexibility, and specificity over the plate counting assay are undeniable. To go further, the qPCR assay can be applied in fundamental research to study derivative mutants of AvS4 or to monitor pathogen bacterial populations in the presence of a biological control agent.

    The author(s) declare no conflict of interest.

    Literature Cited

    L. Vial and L. Sanchez are co-last authors.

    The author(s) declare no conflict of interest.

    Funding: This work was co‐funded by Grand Reims, the European Union (“Europe Invests in Champagne Ardennes”) with the European Regional Development Fund (PhD grant, AGROVITIFREE project), and the Centre National de la Recherche Scientifique (CNRS) (EC2CO “Ecosphère Continentale et Côtière” CNRS IntEnd).