Phytoplasma detection and identification in grapevine by deep amplicon sequencing

Grapevine yellows is one of the major diseases affecting the vineyards in Europe as well as in the majority of grapevine cultivation areas worldwide. Phytoplasma presence seriously affect quality and quantity of production therefore accurate and sensitive pathogen detection and identification are relevant for the disease management. Phytoplasmas within single plants are populations of individuals; routine techniques such as PCR followed by RFLP or sequencing of uncloned or cloned products do not show the diversity of these populations. These techniques only allow identification of the most frequent genotypes in the samples and not mixed phytoplasma or pathogen infection. A number of grapevine samples in which single or mixed phytoplasma infection was detected by routine methods were employed for deep amplicon sequencing on the Roche Genome Sequencer FLX system to compare the two detection systems. Materials and Methods Samples employed were selected from those that are routinely processed during surveys for phytoplasma detection and identification in Northern Italy where both ‘flavescence dorée’ (FD) and ‘bois noir’ (BN) diseases are widespread or epidemic (Botti and Bertaccini, 2007). Total DNA was extracted from 1 g of mid-vein leaf tissue following the procedure of Angelini et al., (2001). Phytoplasma detection was carried out by direct PCR on ribosomal gene and spacer region (Martini et al., 2002) followed by nested amplification with R16(I)F1/R1 (Lee et al., 1994) and 16R758F/V1730 (Martini et al., 1999) primer pairs. RFLP analyses with TruI on the first amplicons and TaqI on the second one allow identification of BN and FD phytoplasmas. Samples showing single and mixed phytoplasma infection were selected for deep amplicon sequencing. Tagged primers were used for generating pyrosequencing samples following described PCR protocol (Nicolaisen et al., 2011). PCR products were pooled in equimolar amounts, run on an agarose gel and a band of the correct size was excised from the gel and purified using QIAquick gel extraction kit from QIAGEN. Two pools of 13 and 11 samples respectively were sequenced on a GS FLX plate at Eurofins MWG. Tag-sorted sequences were quality filtered using CLOTU software at the Bioportal webportal (http://www.bioportal.uio.no/). To minimize sequencing errors, only the first ~200 nucleotides of each sequence were used. Accepted sequences were clustered using CD-HIT with a 99% similarity threshold, and singleton sequences were discarded. To identify sequences, these were aligned together with reference sequences from GenBank using MEGA and phylogenetic trees were constructed. Results and Discussion A total of 35521 sequences were generated from the first pool of 13 field collected samples of phytoplasma infected grapevine in which single infection was detected by routine procedures. Four of these samples were tested twice to verify result consistency (table 1). These data overall confirmed RFLP results. Table 1. Number of sequences determined in each sample with single phytoplasma infection by routine methods. FD-C 46 FD-C 51 FD-D 53 FD-D 64 FD-D 66 FD-D 54 BN 33 BN 43 BN 59 FD-C 56 FD-C 63 FD-C 68 FD-C 71 FD-C 46 FD-D 53 BN 33 FD-C 56 16SrV-C/D 2109 1709 3032 885 1559 42 19 15 10 2040 1876 1520 1356 2126 2240 0 2540 16SrXII-A 0 0 0 0 24 0 2213 2730 4716 3 1 0 0 0 0 2756 0 Table 2. Number of sequences determined in each sample with mixed phytoplasma infection determined by routine methods. FD+BN 63 FD+BN 71 FD+BN 73 FD+BN 60 FD+BN 72 BN 67 FD+BN 76 FD TV1 FD+BN 112 FD+BN 120 FD+BN 80 16SrV-C/D 3009 414 54 612 359 2158 130 201 914 3846 33 16SrXII-A 361 5447 4442 1424 3964 1019 9727 4574 698 0 6823 16SrX-B 189 0 0 0 0 109 0 0 278 0 0 Chimeras 39 17 16 8 26 0 11 0 24 0 0 Total 3598 5878 4512 2044 4349 3286 9868 4775 1914 3846 6856 A total of 50,926 sequences were generated from the pool of 11 field collected samples in which mixed BN and FD phytoplasma infection was ...