‘Candidatus Liberibacter solanacearum’ haplotypes D and E in carrot plants and seeds in Tunisia

Vegetative disorders similar to those associated with the presence of ‘Candidatus Liberibacter solanacearum’ (CaLsol) were observed in carrot plants in Kairouan, Tunisia from 2014 to 2016. Symptoms including leaf curling, yellowing, bronze and purplish discoloration, stunting of plants and roots, and proliferation of secondary roots, affected 20 to 40% of the carrots in some plots. In order to determine if these symptoms were associated with the presence of CaLsol, and/or ‘Ca. Phytoplasma spp.’ and/or Spiroplasma citri, real-time PCR analyses were conducted using specific primers for these pathogens. CaLsol was detected for the first time in Tunisia and for the first time its haplotypes D and E were detected co-infecting a carrot plant. Furthermore, three samples of carrot seed produced in Kairouan tested positive for the haplotype D, showing a high percentage (35 to 63%) of viable bacterial cells after treatment with propidium monoazide. However, all the tests were negative for ‘Ca. Phytoplasma’ spp. as well as for S. citri. The results highlight that several CaLsol haplotypes are emerging carrot pathogens in new areas.


Introduction
Carrot is one of the most important crops in Tunisia. The average production is about 215,000 tons per year, from an area of 6319 ha (http://www.agriculture.tn, dates from 2015). The main production area of approximately 1730 ha is located in the centre of the country (Governorate of Sidi Bouzid). In 2014, 2015 and 2016, symptoms similar to those associated with the presence of 'Ca. L. solanacearum' (CaLsol) were observed in the Governorates of Kairouan and Sidi Bouzid and, to a lesser extent, in several other carrot-growing areas such as Fahes (Governorate of Zaghouan) and Chott-Mariem (Governorate of Sousse). According to preliminary surveys, there were around 20% symptomatic plants during 2014, that increased to 40% in 2015 and 2016. Symptoms were observed on plants of cultivars grown from local seeds as well as on commercial hybrids such as Maestro, Concerto and Olympus, imported from other continents.
CaLsol belongs to the order Rhizobiales of the α-2subdivision of the Gram-negative Proteobacteria. Species in this bacterial genus alternate hosts in their life cycle between the plant phloem and the psyllid vector hemolymph (Liefting et al. 2009b;Lin et al. 2009) and, with the exception of the new species Liberibacter crescens (Fagen et al. 2014), have not yet been cultivated in axenic medium.
Five haplotypes of CaLsol have been described (A, B, C, D, and E) (Nelson et al. 2011(Nelson et al. , 2012Teresani et al. 2014). Haplotypes A and B are associated with diseases of potatoes and other solanaceous crops (haplotypes A and B in America and only haplotype A in New Zealand), and are transmitted by the psyllid Bactericera cockerelli. In Europe, haplotypes C, D and E are associated with carrot, celery and parsnip disorders (Nelson et al. 2012;Teresani et al. 2014). Haplotype C has been found in Finland and is transmitted by the psyllid Trioza apicalis (Munyaneza et al. 2014;Haapalainen et al. 2016), whereas haplotypes D and E have been found in Spain, France and Morocco (Teresani et al. 2014;Loiseau et al. 2014;Tahzima et al. 2014), and the main vector seems to be B. trigonica, at least in Spain (Teresani et al. 2015). Given that the symptoms observed in Tunisian carrots resembled those associated with CaLsol, a set of analyses on plants as well as on local seeds for CaLsol detection, was carried out. In addition, since 'Ca. Phytoplasma spp.' and Spiroplasma citri have also been reported affecting carrot with similar symptoms , the detection of these two pathogens was also included.

Materials and methods
Plant extract preparation Inspections and samplings were carried out in carrot fields from 2014 to 2016 at Kairouan (Tunisia) (Fig. 1). General stunting, yellowing, purpling ( Fig. 2a) and leaf curling (Fig. 2b), as well as shoot proliferation and hairy secondary roots (Fig. 2c) were frequently observed. A representative selection of symptomatic plants was collected. Mature leaf fragments were preserved in cryotubes with liquid nitrogen, and were sent in dry ice to the Bacteriology Laboratory of IVIA (Spain), where the samples were stored at −20°C before analysis.
Seed extract preparation Seed samples were prepared and analysed according to Bertolini et al. (2014b): 1 g (approximately 450 seeds) per lot was washed by shaking for 30 min in a tube containing 50 ml washing buffer (distilled water with 0.5% Triton X-100). Then, washed seeds were placed into plastic bags with 10 ml extraction buffer plus 0.2% sodium diethyl dithiocarbamate and 2% PVP-10. The seeds were slightly crushed with a hammer, and 1 ml of the extract from each seed lot was stored at −80°C until use. Three seed lots of a local cultivar from the area of Rakada were analysed in 2015 and from Zaafrana in 2016.
Direct sample preparation without DNA purification (spot procedure) Freshly prepared extracts were immobilized on membranes (pieces of Whatman 3 MM filter paper-Healthcare, Europe) on which 5 μl of crude extract were loaded. Then, 100 μl of distilled water were added to each membrane in an Eppendorf tube and vortexed to extract the target DNA (Bertolini et al. 2014a;Teresani et al. 2014). The Eppendorf tubes were placed on ice and 3 μl of the extract was used as a template for real-time PCR assays. For further assays, the membranes were removed before storing the tubes at −20°C.
DNA purification Total DNA was purified from 200 μl of crude plant or seed extract and 100 μl from the spot procedure, using the cetyltrimethyl ammonium bromide (CTAB) protocol (Murray and Thompson 1980). Purified DNA was stored at −20°C until use.
Psyllid extract preparation Psyllids were captured from symptomatic plants using an entomological mouth aspirator. The putative individuals of B. trigonica and B. nigricornis were preserved in 70% ethanol. Psyllids were identified based on the male genitalia, according to Burckhardt and Freuler (2000). They were then squashed on Whatman 3 MM filter paper (HealthCare, Europe) with the rounded end of an Eppendorf tube according to Teresani et al. (2015). The membranes were carefully cut and inserted into Eppendorf tubes, distilled water (100 μl per tube) was added, the tubes vortexed, and the membranes removed before storing the tubes at −20°C, until used for amplification.
Purified DNA and spots of crude extracts prepared from stored symptomatic carrot plants positive for CaLsol were used as positive controls in each PCR assay. DNase-free distilled water and spotted pieces of membrane with a crude healthy carrot extract were used as negative controls. All the samples were analysed at least three times. 'Ca. Phytoplasma spp.' presence was screened by real-time PCR according to Hren et al. (2007) using the universal primers UniRNA. The reaction mix consisted of 1X Path-ID™ qPCR (Ambion, USA), 1 μM of each primer, 0.5 μM of TaqMan MGB probe and 3 μl of template. The amplification protocol consisted of 95°C for 10 min, followed by 45 cycles of 95°C for 15 s, and 64°C for 1 min. Primer and probe sequences were: UniRNA-F, 5'-AAATATAG TGGAGGTTATCAGGGATACAG-3'; UniRNA-R, 5'-AACCTAACATCTCACGACACGAACT-3' and the specific TaqMan MGB probe, 5'FAM-ACGACAACCATGCACCA-3`MGB.
S. citri presence was also screened by real-time PCR, following the protocol of Alfaro-Fernández et al. (2017). The reaction mix consisted of 1X Path-ID™ qPCR (Ambion, USA) 10 μM of each primer, 0.5 μM of TaqMan MGB probe and 3 μl of template. The amplification protocol consisted of 95°C for 10 min followed by 45 cycles of 95°C for 15 s and 58°C for 1 min. Primers and probe sequences were: SCPA-F, 5'-GAAATACTGATAATGGCAAACAGTTTG-3'; SCPA-R, 5'-ACTAGCACTATACCCACGGATATTGT-3'; and the specific TaqMan MGB probe SCPA-P, 5' FA M-ATATGAGCGAAAGCAAAT -3`MGB. The protocols of real-time PCR described for detection of S. citri and 'Ca. Phytoplasma spp.' were used on a StepOne Plus thermal cycler (Applied Biosystems, USA) and the analyses were repeated at least twice.

Determination of bacterial viability
To quantify the live CaLsol bacteria, 0.2 ml aliquots from seed samples were treated with propidium monoazide (PMA) at a final concentration of 100 μM. PMA selectively enters dead cells and, upon photo-activation, it intercalates and binds covalently to DNA strongly inhibiting its PCR amplification. Tubes were incubated at 30°C in the dark for 10 min, with occasional inversion to allow the dye to intercalate with the DNA. After incubation, tubes were vortexed for 5 s and exposed to 100% light intensity for 15 min in a PhAST Blue apparatus (GenIUL, Spain). Subsequently, DNA was extracted from both PMA-treated and untreated samples using the CTAB method and samples were analysed by real-time PCR. Cells number were estimated using cycle threshold (Ct) values from these assays to estimate cells number from a modified standard curve according to Bertolini et al. (2014b), including corrections considering that each bacterial cell contains three copies of the 16S rRNA gene according to Nelson et al. (2015).
Haplotype characterisation and phylogenetic analyses To determine the genetic diversity of Tunisian CaLsol, haplotypes were characterized based on single nucleotide polymorphisms (SNPs) typing of partial regions of the genes 16S rDNA (16S rRNA), 16S-23S rDNA intergenic spacer region (IGS), and rplJ and rplL ribosomal protein genes (50S subunit) by phylogenetic analysis. The amplification of the 16S rRNA was performed using the forward primer designed by Li et al.  (2009), from a region of the 16S rRNA gene that is unique to CaLsol, in combination with the universal Liberibacter reverse primer of Jagoueix et al. (1996). The IGS was amplified using the primers of Ravindran et al. (2011). Finally, the 50S ribosomal protein genes rplJ and rplL were amplified with the new primers: CaLsol 50S-F (5'-CGAGACCATTGAAC ACAACG-3') and CaLsol 50S-R (5'-TGGAGGAG GAGTGAGAAT-3'). These primers were designed in this study from the complete genome of CaLsol strain R1 (GenBank accession No. CP002371) available in GenBank database with SnapGene software (GSL Biotech, USA). The rplJ-rplL locus was amplified by placing the reaction mixture conditions (final concentration of 1X Taq DNA polymerase buffer containing 2 mM MgCl 2 , 400 nM of each primer, 400 μM dNTPs, 1 U Taq DNA polymerase and 2 μl DNA template) in a thermocycler GeneAmp PCR System 9600 (Applied Biosystem/ Perkin Elmer, USA) at 94°C for 5 min, followed by 40 cycles at 94°C for 30 s, 56°C for 30 s and 72°C for 1 min. Final elongation was 10 min at 72°C, giving an amplicon of 946 bp. The results were compared with those obtained following Munyaneza et al. (2009). Purification of the amplified fragment from the PCR products was performed following the instructions provided by the Kit Invisorb Spin DNA extraction (Invisorb, Germany), and then quality tested with a NanoDrop spectrophotometer (Thermo Scientific, USA). The sequencing was performed with two readings (forward and reverse) by the Sequencing Service of the Institute of Molecular and Cell Biology of Plants (IBMCP) (Valencia-Spain) and the quality and detection of secondary peaks in chromatograms were determined with the package SangeranalyseR, tools for Sanger sequencing data (Hill et al. 2014) in R software (R Core Team 2016).
All sequences were subjected to molecular phylogenetic analysis performed with MEGA 6 software (Tamura et al. 2013) by Maximum Likelihood Method performed on 1000 replicates. All analysed carrot plants were negative for 'Ca. Phytoplasma spp.' using universal primers, and also for S. citri using specific primers (Table 1). Table 1) in 2015 were negative for CaLsol, while the three lots of cv. Arbi Zaafrana analysed in 2016 were highly contaminated. Moreover, the viability of CaLsol cells in seeds was very high and ranged from 36% (SE.5) to 65% (SE.4), with an average value of 46% (Table 2, Fig. 3).

Analysis of seeds Three lots of seed samples analysed (SE in
Analysis of psyllids A selection of 49 squashed and immobilized adults of B. trigonica (samples T in Table 1) and eight adults of B. nigricornis (samples N in Table 1) were analysed directly from spots for CaLsol, but all of them were negative.
Haplotype characterisation and phylogenetic analyses The haplotype was determined for 14 samples that were positive in real time PCR for CaLsol, using the three primer sets to ribosomal regions described previously (16S rRNA, IGS, and rplJ and rplL ribosomal protein genes), but in many cases amplification of all three regions was not achieved. Four samples (three carrot seeds SE4, 5 and 6 and the plant sample Z3), where amplification of all three regions was obtained, were selected for further investigation. Interestingly, the new primers designed for amplification and sequencing of rplJ and rplL ribosomal protein genes fragments gave stronger positive amplicons than the primers described by Munyaneza et al. (2009). As an example, the amplification obtained with primers to both rplJ and rplL ribosomal protein genes regions of infected carrot seeds are shown in Fig. 4.

Discussion
The emerging bacterium CaLsol is associated with socioeconomically important diseases in many crops (Munyaneza 2012;Haapalainen 2014). Its presence has been reported in several species of Solanaceae (Liefting et al. 2009b) and in plants of Apiaceae (Munyaneza et al. 2012a, b), being associated with vegetative disorders. In carrots, the bacterium has been detected in European countries such as Spain However, CaLsol was not detected in all the symptomatic plants tested. This may be due to a number of factors: (i) high temperature, because temperatures close to 30°C or higher may affect the prevalence of the high temperature sensitive CaLsol (Munyaneza et al. 2012c). At the time of sampling temperatures were above 30°C, although temperatures during the growing season varied between 24 and 33°C (www. meteo.tn/, dates from 2016); (ii) an uneven CaLsol distribution in the phloem, that could lead to samples testing negative; (iii) variable bacterial titer that may be below the level of detection according to the season and the plant organ infected (data not shown); (iv) time delay between samples collection and their analysis and/or the use of frozen samples, leading to CaLsol DNA degradation. As the disease in carrots has only recently been reported, the influence of these factors on CaLsol levels in plant tissues is not well understood. Furthermore, the presence of inhibitory compounds in plant tissues could also hamper detection, as they are a major impediment to successful DNA amplification. Consequently, 1: 10 and 1:100 extract dilutions were tested to minimize the action of possible polymerase inhibitors but no improvement in amplification was found. In addition, in order to improve the efficiency of detection, different combinations of extraction procedures (spot PCR and modified CTAB extraction) were evaluated to optimize DNA amplification but with no improvement.
Some European laboratories have found difficulties in amplifying some regions used for haplotyping which may be due to the low concentration of the bacterium in the analysed samples. For this reason, a new PCR protocol for amplifying a fragment of the gene rplJ-rpl was designed because our PCR products were weak with protocol described in Munyaneza et al. (2009).  Analysis of CaLsol positives samples showed that only haplotype D was present in carrot seeds but haplotypes D and E were present in one carrot plant (Fig. 6). Due to the limited amount of DNA of this sample (Z3), we could not clone the amplicons to confirm the presence of both fragments of each haplotype. However, co-infection has already been described for haplotypes A and B in potato and potatopsyllid samples (Wen et al. 2013), but this is the first report of co-infection of haplotypes D and E in carrot. Since haplotype E is the most frequent haplotype found in carrot in Southern European countries ) and some of the cultivars found infected, such as cv. Maestro, are the same as those found infected in Europe, and were provided by the same companies, this haplotype may have been introduced through contaminated seeds (Bertolini et al. 2014b).
Using real-time PCR a high level of CaLsol was detected in 2016 in three carrot seed lots of the cv. Arbi Zaafrana, with a high number of quantified viable cells found after PMA treatment, higher than 35% in all cases and reaching 65% in one of the lots. These percentages are much higher than those reported by Bertolini et al. (2014b) and may be due to the lack of seed treatments performed by Tunisian growers on locally produced seed, in contrast to the treated commercially available seeds. The locally produced seeds had not been checked for CaLsol before sowing but, according to information from the growers, they were probably collected from infected mother plants.
Viable CaLsol cells in seeds were shown to be responsible for infection of carrot seedlings germinated from contaminated seeds (Bertolini et al. 2014b). However, transmission of this pathogen by carrot seeds seems to be not a major transmission pathway, in comparison to vector transmission, probably because other unknown parameters are influential in the development of the disease (Loiseau et al. 2017). If the primary inoculum of haplotype D in Tunisia is CaLsol contaminated seed, treatment of carrot seeds would be a key mitigation strategy for preventing the disease (Bertolini et al. 2014b;Ilardi et al. 2016), for example by hot water treatment (Australian Government Department of Agriculture and Water Resources 2017).
In Kairouan, it has only been possible to test a relatively low number of identified B. trigonica and B. nigricornis individuals for CaLsol by direct analysis of squashed Fig. 6 Phylogenetic tree using the Maximum Likelihood method based on the 3-parameter model using MEGA6. Bootstrap test of phylogeny was performed with 1000 replicates. The analysis involved nucleotide sequences from database GenBank of (a) the 50S ribosomal protein genes rplJ and rplL and (b) 16S rRNA partial genes of the haplotypes of CaLsol. Sequence of samples analysed in this study are marked with a large black dot psyllids. As these insects were negative for CaLsol,, more individuals of these and other species should be collected and analysed to determine their possible role in transmission, because different psyllid species have been reported as vectors in several European countries (Munyaneza et al. 2010b;Teresani et al. 2015).
The negative results for the detection of 'Ca. Phytoplasma spp.' and S. citri suggests that CaLsol is the only bacterial agent associated with the symptoms observed in carrots in Kairouan, and with the commercial losses due to the poor quality of these plants. The information from the current study will help carrot growers to develop preventive and pest management strategies, such as analysis of seed lots, elimination of symptomatic plants, and analysis and control of potential vectors to reduce the incidence and spread of this emerging bacterium to other apiaceous or solanaceous hosts, and to other Tunisian regions. The recent discovery of this bacterium in Morocco (Tahzima et al. 2014), together with our results suggests that this emerging pathogen is spreading in North Africa and other Mediterranean countries. Thus, the National Plant Protection Organisations should be alerted to the threat that CaLsol poses to solanaceous and apiaceous cultivated plant species.