Precocious flowering of juvenile citrus induced by a viral vector based on Citrus leaf blotch virus: a new tool for genetics and breeding

Summary The long juvenile period of citrus trees (often more than 6 years) has hindered genetic improvement by traditional breeding methods and genetic studies. In this work, we have developed a biotechnology tool to promote transition from the vegetative to the reproductive phase in juvenile citrus plants by expression of the Arabidopsis thaliana or citrus FLOWERING LOCUS T (FT) genes using a Citrus leaf blotch virus‐based vector (clbvINpr‐AtFT and clbvINpr‐CiFT, respectively). Citrus plants of different genotypes graft inoculated with either of these vectors started flowering within 4–6 months, with no alteration of the plant architecture, leaf, flower or fruit morphology in comparison with noninoculated adult plants. The vector did not integrate in or recombine with the plant genome nor was it pollen or vector transmissible, albeit seed transmission at low rate was detected. The clbvINpr‐AtFT is very stable, and flowering was observed over a period of at least 5 years. Precocious flowering of juvenile citrus plants after vector infection provides a helpful and safe tool to dramatically speed up genetic studies and breeding programmes.


Introduction
Breeding programmes for citrus, the main fruit crop worldwide, are necessary to obtain high-quality competitive varieties and rootstocks. The complex reproductive biology of citrus trees due to apomixis, pollen and/or ovule sterility of several high-quality genotypes and self-and cross-incompatibility that are relatively common among citrus genotypes limits the possibilities to select parents for specific crosses . However, the most important impairment for breeding programmes and genetic studies is the long juvenile period of citrus, often lasting more than 6 years (Krajewski and Rabe, 1995). Plants in the juvenile phase are characterized by their vigorous growth, the presence of thorns and failure to flower even under inductive conditions; they must undergo a transition from the juvenile phase to the reproductive phase to develop flowering buds. Thus, the long juvenile period of citrus plants impairs for years the generation of segregating populations by back crossing, hybrid progeny evaluation to obtain new varieties or studies on inheritance of fruit-related traits.
The genetic factors that control the end of the juvenile phase to start the flowering period in citrus are presently unknown. Recent molecular studies suggested that the FLOWERING LOCUS T (FT) protein in Arabidopsis thaliana and its orthologs in other plant species are the primary component of the flowering signal cascade (Corbesier et al., 2007;Endo et al., 2005;Hecht et al., 2011;Imamura et al., 2011;Jaeger and Wigge, 2007;Kong et al., 2010;Lin et al., 2007;Lv et al., 2012;Tamaki et al., 2007). In the shoot apical meristem, the FT protein forms a complex with the FD protein (with a basic domain/leucine zipper motif transcription factor), which up-regulate the gene APETALA 1 (AP1) in the meristem to induce an inflorescence (Abe et al., 2005;Wigge et al., 2005). In Satsuma (Citrus unshiu (Macf.) Marc.) and Moncada mandarins (C. clementina Hort. ex Tan. 9 (C. unshiu 9 C. nobilis Lour.)), in which floral induction is triggered by low temperatures during the fall and winter, the messenger RNA (mRNA) level of the citrus ortholog of FT (CiFT) increases during the spring just before blooming (Muñoz-Fambuena et al., 2011;Nishikawa et al., 2007Nishikawa et al., , 2009. Similarly, in trifoliate orange (Poncirus trifoliata (L.) Raf.), in which floral induction and flower bud development occur during early summer as in many deciduous trees, the amount of CiFT mRNA increases in early summer (Nishikawa et al., 2009). Ectopic expression of CiFT under the control of the 35S promoter of Cauliflower mosaic virus (CaMV 35S) induces an early flowering phenotype on trifoliate orange (Endo et al., 2005); however, overexpression of FT also affects the growth and architecture of these plants that show a dwarfed aberrant appearance. Precocious flowering was also observed in juvenile transgenic plants expressing the flowering genes LEAFY and APETALA1 (Peña et al., 2001), albeit plants expressing LEAFY also showed an abnormal phenotype. Additionally, recovery of transgenic citrus plants is usually slow and difficult.
Plant virus vectors have been used for both expression of foreign genes (Gleba et al., 2007) and suppression of endogenous target genes by virus-induced gene silencing (VIGS) in the infected plants (Senthil-Kumar and Mysore, 2011). Using viral vectors is a quick system to deliver proteins in plants without any need for the transformation and regeneration techniques required to obtain transgenic plants. Moreover, expressing the FT protein with viral vectors is an emerging technology to induce precocious flowering in plants (McGarry and Ayre, 2012;Yamagishi et al., 2014).
Citrus leaf blotch virus (CLBV), the type member of the genus Citrivirus, family Betaflexiviridae (Adams et al., 2012;Martelli et al., 2007), has filamentous virions about 960 9 14 nm in size composed of a single-stranded, positive-sense genomic RNA (gRNA) of 8747 nucleotides (nt) and a 41-kDa coat protein (CP) Vives et al., 2001). The CLBV gRNA has three open-reading frames (ORFs) and untranslated regions of 73 and 541 nt at the 5 0 and 3 0 ends, respectively. The ORF1, encoding a~227-kDa replicase polyprotein, is translated directly from the gRNA. The ORF 2, encoding a~40-kDa protein with a motif characteristic of cell-to-cell movement proteins (MP) of the 30K superfamily, and the ORF3, encoding the CP, are translated from two 3 0 co-terminal subgenomic RNAs (MP sgRNA and CP sgRNA, respectively). ORFs 2 and 3 are separated by a 64-nt intergenic region (Renovell et al., 2010;Vives et al., 2001Vives et al., , 2002a. Different viral vectors based on a full-genome infectious cDNA clone of CLBV (Vives et al., 2008a) were previously obtained (Ag€ uero et al., 2012(Ag€ uero et al., , 2014. The advantages of CLBV to be used as a viral vector for citrus are as follows: (i) it causes a symptomless infection in most commercial cultivars (Galipienso et al., 2000); therefore, phenotype changes due to endogenous gene silencing or foreign gene expression would not be masked by virus symptoms, (ii) because CLBV is not restricted to the phloem, these vectors are appropriate for gene expression or silencing in nonphloem tissues including meristematic regions (Ag€ uero et al., 2013), and (iii) contrarily to Citrus tristeza virus (CTV), also used as viral vector for citrus (Folimonov et al., 2007), CLBV is not transmitted by vectors and therefore it could be safely used in field experiments.
In this work, we have developed a biotechnology tool to promote an early transition to flowering of juvenile citrus plants by expression of the A. thaliana or citrus FT gene in the clbvINpr vector (clbvINpr-AtFT or clbvINpr-CiFT vectors, respectively). Graft inoculation of juvenile citrus plants with either of these vectors induced flowering in the second flush [about 4 months postinoculation, (mpi)]. The vector clbvINpr-AtFT was very stable, and flowering of the vector-inoculated plants was observed in successive flushes over a period of at least 5 years. Precocious flowering after vector infection provides a helpful tool to dramatically speed up genetic studies and breeding programmes.

Induction of flowering in juvenile citrus plants by means of a CLBV-based vector expressing the FT gene
To investigate whether ectopic expression of the FT gene from a CLBV vector is able to induce flowering in infected juvenile citrus plants, a cDNA of the FT gene from A. thaliana or from Valencia late sweet orange (Citrus. sinensis (L.) Osb.) was cloned into the clbvINpr vector to generate the clbvINpr-AtFT and clbvINpr-CiFT constructs, respectively ( Figure 1) (see 'Experimental procedures'). These constructs were agroinoculated in Nicotiana benthamiana plants, and 1 month later, the resulting recombinant virions were purified and slash inoculated in two juvenile C. excelsa Wester seedlings. At the end of the second flush, about 4 mpi, the citrus plants infected with both constructs flowered ( Figure 2a). No differences were found in either the number of flowers or in the period of flowering between the plants inoculated with the vectors expressing AtFT or CiFT. Later, lateral shoots started growing until the formation of a new floral bud. Some shoots did not produce flowers, may be due to uneven distribution or accumulation of the recombinant virus in host plants (Ag€ uero et al., 2014). As the constructs are very stable, flowering and fruit set occurred in successive flushes during the observation period (at least 5 years). Similar C. excelsa plants noninoculated or inoculated with the wild-type CLBV virus (WT-CLBV) ( Figure 2b) did not produce flowers, but continued their vegetative growth for the same period.
The C. excelsa plants infected with the clbvINpr-AtFT or the clbvINpr-CiFT constructs were used as inoculum sources for other citrus hybrid juvenile plants obtained in the framework of a citrus breeding programme (Navarro et al., 2015), following three different inoculation procedures: (i) seedling plants were graft inoculated with two infected bark patches and then cut back leaving one bud above the inocula (Figure 3a), (ii) buds from the juvenile hybrids were propagated onto a rootstock (e.g. a Rough lemon (C. jambhiri Lush)), simultaneously graft inoculated with two bark patches from the infected C. excelsa plants (Figure 3b), and (iii) the root of small hybrid plantlets was cut-off at the crown and the aerial portion was top-worked onto a rootstock, simultaneously graft inoculated with two bark patches from the infected C. excelsa plants ( Figure 3c). All plants inoculated by either procedure with the clbvINpr-AtFT or the clbvINpr-CiFT constructs became systemically infected, as detected by conventional reverse transcription PCR (RT-PCR) (Galipienso et al., 2004) Figure 6). Fruits of these latter plants were collected when ripe and seeds were germinated. Forty-nine recovered plants were analysed with five Simple Sequence Repeat (SSR) markers (Froelicher et al., 2008; and genetic analysis confirmed that all plants were hybrids of the parental genotypes used (Figure 7).

Genetic stability of the clbvINpr-AtFT vector
Genetic stability of the clbvINpr-AtFT vector was examined by RT-PCR analysis of total RNA (RNAt) from the two originally infected C. excelsa plants and from 10 other citrus genotypes graft inoculated from them, using primers encompassing the insertion site of the AtFT gene in the CLBV genome. This analysis, carried out more than 4 years after inoculation, showed in nine of the 12 plants a unique band of the size expected for the full FT sequence, whereas in the other three plants an additional weak band of smaller size was also observed. To quantify the recombination events, the PCR products were cloned and 422 individual clones were analysed by PCR using the same primers.
Only 36 of these clones (8.5%) had an insert of a size smaller than that expected for the complete AtFT gene. To assess the sequence variability in the population, four small-sized and 156 full-sized clones were sequenced. The four small-sized clones showed an identical deletion between the nucleotides 351 and 462 of the AtFT gene, suggesting that these nucleotides could be a hot spot for recombination. Also, as expected for a RNA virus population, nucleotide changes were observed in 90 of the 156 full-sized clones sequenced (54 clones with a single nt change, 27 with two and 9 with three), with about 90% of these mutations involving an amino acid change. Overall, these results indicate that 4 years after inoculation there were few important changes in the AtFT sequence and the functionality of FT protein remained unaffected.

Expression of flowering-related genes in citrus plants inoculated with the clbvINpr-AtFT vector
To examine how the presence of AtFT in infected citrus plants affected the expression of flowering-related genes, buds of C. excelsa juvenile plants were propagated on ten Rough lemon seedling rootstocks, five of which were simultaneously graft inoculated with the clbvINpr-AtFT vector, and five with the WT-CLBV as control. At the end of the second flush, RNAt from a pool of shoot tips (about 10 mm in length) from each plant was purified and used to quantify the mRNA expression of the AP1 and CiFT citrus genes by RT-qPCR. While significant differences in mRNA accumulation of the AP1 gene were observed (ANOVA, P < 0.05) in the shot tips of plants inoculated with clbvINpr-AtFT in comparison with similar tissues from plants infected with WT-CLBV, no significant difference (P > 0.05) was observed in the mRNA levels of the CiFT gene ( Figure 8). These results indicate that the expression of AtFT in infected plants did not trigger silencing of the endogenous CiFT gene, probably due to the low sequence identity (74%) between AtFT and CiFT and the low CiFT mRNA expression in shot tips of juvenile citrus plants.

Horizontal and vertical transmission of the clbvINpr-AtFT vector
To assess whether the clbvINpr-AtFT vector or a part thereof can become integrated in the plant genome, total DNA was isolated from bark and leaves of infected C. excelsa plants and analysed by PCR using primers designed on different parts of the CLBV genome or on the AtFT gene. No DNA amplification was ever obtained with CLBV-or AtFT-specific primers, indicating that neither of them become integrated into the plant genome. Also, to examine potential recombination between the AtFT gene expressed from the viral vector and the endogenous CiFT gene, the DNA band PCR amplified with primers specific for the CiFT gene was cloned and 10 of these clones were sequenced. All of them had the characteristic sequence of CiFT, and no recombination event was detected.  In a previous experiment, seed transmission of CLBV was observed at a rate of 2.5% . To examine whether the virus could be transmitted by pollen, we analysed RNAt from pollen from clbvINpr-AtFT-infected plants by RT-qPCR with CLBV-specific primers (Ruiz-Ruiz et al., 2009), but no amplification was detected. Moreover, cross-pollination of flowers from noninoculated plants with pollen from clbvINpr-AtFTinfected plants and further analysis of the offspring for the presence of CLBV resulted in none of the progeny plants being infected by the viral vector. Finally, to assess whether clbvINpr-AtFT can be ovule-transmitted, 125 seeds from different infected genotypes cross-pollinated with pollen from healthy plants were grown and individual seedlings were analysed for CLBV by RT-PCR. Only in one of the plants analysed was clbvINpr-AtFT detected, confirming the low seed transmission rate previously detected for CLBV.

Discussion
The juvenile phase characteristic of some woody perennials like citrus has been a limiting factor for traditional breeding and genetic studies. Expression of the FT gene from A. thaliana (AtFT) or citrus (CiFT) by means of a CLBV-based viral vector induced precocious flowering in juvenile citrus plants. When a juvenile citrus plant was graft inoculated with the clbvINpr-AtFT or the clbvINpr-CiFT vectors, the infected plant started flowering within the next 4-6 months depending on the genotype and time of the year, whereas similar citrus plants noninoculated or infected with WT-CLBV did not flower, but remained vegetatively growing under the same conditions. No difference in the juvenile plant response was observed using the clbvINpr-AtFT or the clbvINpr-CiFT vectors; however, to prevent a potential interference of endogenous CiFT silencing induced by the second vector, the clbvINpr-AtFT construct was preferred in most of the experiments. These results suggest that the FT protein delivered by the viral vector could terminate the vegetative growth by inducing the transition from vegetative to reproductive meristematic tissues in the shoot apex. Constitutive expression of CiFT in transgenic P. trifoliata plants also induced early flowering, but it was accompanied by phenotypic aberrations consisting of a dwarfed, highly branched growth, misshapen leaves and altered flowering habit and flower morphology in comparison with the nontransgenic controls (Endo et al., 2005;Nishikawa et al., 2010). Contrastingly, in plants inoculated with either the clbvINpr-AtFT or the clbvINpr-CiFT constructs, the plant architecture was not affected and leaves, flowers and mature fruits were normal. This different behaviour could be due to (i) the slow accumulation and movement of CLBV in the inoculated plants that would allow normal growth during the initial period, before reaching a threshold virus concentration (and FT expression) necessary to induce flowering and (ii) the uneven accumulation and limited virus titre in fully infected plants that would provide reduced amounts of the FT protein in comparison with transgenic plants expressing FT in all their cells. Several observations support the slow accumulation and movement of CLBV: (i) in plants inoculated with a clbvINpr vector expressing the green fluorescent protein (clbvINpr-GFP), CLBV was detected by RT-PCR only at the end of the first flush (about 1.5 mpi) (Ag€ uero et al., 2012), (ii) in plants inoculated with CLBVbased vectors carrying fragments of the endogenous citrus genes phytoene desaturase or sulphur, the silencing phenotype was rarely observed in the first flush and more usually in the second flush, always being more intense in the second flush, and (iii) the symptoms of chlorotic blotching and stem pitting induced by CLBV in Dweet tangor (C. tangerina Hort. ex Tan. 9 C. sinensis) and Etrog citron (C. medica L.), respectively, usually appear in the second flush, not in the first. Finding that while some citrus genotypes infected with clbvINpr-AtFT produced inflorescences, others produced individual flowers, or that some genotypes flowered only in the spring, whereas others repeatedly flowered in successive flushes along the year suggest that, in addition to the FT gene, other factors are likely necessary to induce floral buds.
Genetic instability associated with frequent recombination events and high mutation frequency during RNA replication are limiting factors in the use of plant viral vectors. However, clbvINpr-AtFT was able to induce flowering during the long  , 14, 1976-1985 period studied (almost 5 years). PCR analysis of individual AtFT clones showed that only 8.5% of the clones analysed yielded a DNA band of a size smaller than that expected for the complete gene, confirming that CLBV-based vectors are highly stable (Ag€ uero et al., 2012(Ag€ uero et al., , 2014. Although some of the clones sequenced showed nucleotide substitutions leading to amino acid changes, none of these changes affected the amino acids essential for the FT function (Hanzawa et al., 2005;Ho and Weigel, 2014). On the other hand, genetic variation of CLBV isolates from different geographical origin or citrus host was shown to be very low in comparison with most RNA plant viruses (Vives et al., 2002b).
Although genetic transformation to induce early flowering could be a helpful tool to accelerate breeding, this technique is less practical with citrus trees due to their low transformation efficiency and the long regeneration period required (Peña et al., 2001). While infection of a citrus plant with the viral vector is fast (about 1 month), recovery of a transgenic citrus plant takes more than 1 year. Also, once the viral vector has systemically infected a citrus plant, it can be easily graft inoculated onto hundreds of different  , 14, 1976-1985 genotypes, whereas the transformation protocol has to be adapted and performed for each individual genotype. The juvenile plants inoculated with CLBV-based vectors expressing the FT gene produced flowers at the end of the second flush (about 4 mpi, depending on the inoculated genotype and the season) and these successfully produced fruits with normal seeds. Therefore, this technique allows obtaining seeds for the next-generation progeny within 1 year without any genome modification in the plant.
The viral vector developed in this work is being a helpful tool to speed up genetic studies and breeding programmes. While backcrossing has not been used in citrus breeding because the very long time needed to recover successive hybrid populations, now it has been routinely incorporated in our triploid breeding programme (Navarro et al., 2015). Tetraploid parents were produced by sexual hybridization and screened by molecular markers for some desired traits, for example resistance to the fungus Alternaria alternata (Cuenca et al., 2013) or expression of the Ruby gen for accumulation of anthocyanins in fruits (Butelli et al., 2012). Then, selected hybrids are inoculated with the clbvINpr-AtFT vector to induce precocious flowering and used in crosses with diploid parents to produce triploid progenies. More than 1600 sexual crosses have been made so far following this procedure.
In many countries, the use of genetically modified plants is presently prohibited. The main concerns against the transgenic plants are the use of antibiotic resistance genes for selection of transformants, potential cross-pollination with other sexually compatible plants, plant genome alteration by integration of foreign sequences or by recombination when additional copies of endogenous genes are introduced, and potential transduction of the binary plasmid used for transformation into other bacteria present in the plant or soil. Inoculation of CLBV-based vectors to citrus trees to induce early flowering is performed by grafting, and no antibiotic resistance gene, plasmid or bacteria are introduced in the plants. Neither the viral vector nor the FT gene introduced in the plants become integrated in the plant genome, and the CLBV virus is not transmitted either by insect vectors (Vives et al., 2008b) or by pollen. Although a low seed transmission rate has been detected in a few genotypes, the virus is easily detected in the infected plants by RT-qPCR and it could be easily eliminated by shoot tip grafting in vitro (Navarro et al., 1975) to recover any desired genotype vector free. Overall, our results indicate that this viral vector should not be considered as a risky genetically modified organism and it can be transiently used under field conditions, with essentially no risk, in breeding programmes for early fruit evaluation even under field conditions. This will allow a quick screening of hybrid progeny to eliminate the genotypes with undesirable characters, which are the majority of the progeny in any breeding programme, and concentrate efforts in additional evaluation of selected potentially new varieties. Finally, selected varieties could be recovered vector free for commercial propagation. This approach will save expensive greenhouse space and labour and will accelerate the selection of new varieties, making the breeding programmes much more efficient at lower cost. A field experiment has been recently established in our institute for early evaluation of a triploid hybrid progeny inoculated with the clbvINpr-AtFT vector.
Citrus breeding is also hampered by the scarcity of genetic studies on the inheritance of characters that cannot be phenotyped during the juvenile phase . The use of CLBV-based vectors, together with the new genetic  and genomic (Wu et al., 2014) tools available, will facilitate these studies allowing the development of markers for focused introgression and early selection. These new tools will enable general implementation of efficient breeding programmes based on marker-assisted selection.

Experimental procedures
Construction of the CLBV-based vectors expressing the FT gene The AtFT and CiFT coding regions were amplified by RT-PCR from A. thaliana and Valencia late sweet orange RNAt extracts using appropriate primers (Table 1) and inserted into the PmlI restriction site of the clbvINpr vector (Ag€ uero et al., 2012) to generate the clbvINpr-AtFT and clbvINpr-CiFT constructs, respectively (Figure 1). The clbvINpr vector was built by introducing a unique PmlI restriction site at the intergenic region between the MP and the CP genes and then introducing a duplicate of the CP sgRNA promoter restoring the PmlI restriction site to express foreign sequences by producing an extra sgRNA. In this vector, transcription of the new sgRNA was induced by the native promoter of de CP sgRNA, whereas the CP gene expression was controlled by the duplicated promoter (Ag€ uero et al., 2012(Ag€ uero et al., , 2014. All the constructs were confirmed by sequencing.

Plant growth and inoculations
N. benthamiana plants were grown in small pots with an artificial potting mix (50% vermiculite and 50% peat moss) in a plant growth chamber at 20/24°C (night/day), 60% relative humidity and a 16/8-h light/dark regime. Citrus plants were grown as seedlings, or propagated on a Rough lemon rootstock, in a glasshouse at 18/26°C (night/day), using 2-L plastic containers filled with 50% sand and 50% peat moss and a standard fertilizing procedure (Arregui et al., 1982).
The ClbvINpr-AtFT and clbvINpr-CiFT constructs were agroinoculated into N. benthamiana using Agrobacterium tumefaciens cells, strain COR 308 (kindly provided by Dr. C. M. Hamilton, Cornell Research Foundation). Once confirmed the systemic infection, semipurified virion extracts from N. benthamianainfected plants (Galipienso et al., 2000) were mechanically inoculated onto C. excelsa plants by stem slashing (Garnsey et al., 1977) with scalpel blades dipped in the virion extracts (Galipienso et al., 2000;Vives et al., 2008a). Bark pieces from these plants or from plants infected with the WT-CLBV were used to graft inoculate other citrus species or hybrids.

RNA extraction, RT-PCR detection and real-time RT-qPCR quantification
For RT-PCR amplification and detection, RNAt from 100 mg fresh tissue was prepared using the TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instructions, and after alcohol precipitation, RNA was resuspended in 50 lL of RNasefree water.
The FT sequences inserted in the clbvINpr vector were detected by conventional RT-PCR (Vives et al., 2002b) using the primer pair MpC/MP3U (Ag€ uero et al., 2014), flanking the insertion site. The DNA synthesized was analysed by 2% agarose gel electrophoresis and GelRed staining (Biotium Inc., Hayward, CA). The DNA fragments obtained were inserted into the pGEM-T plasmid vector (Promega Corp. Madison, WI) and cloned in Escherichia coli DH5a cells. A total of 422 individual clones were analysed by PCR using the same primers, and the nucleotide sequence of 160 clones was determined with an ABI PRISM DNA Sequencer 3100 (Applied Biosystems, Paisley, UK). Multiple alignments were performed using the program Clustal W (Thompson et al., 1994).
To assess the endogenous gene expression by RT-qPCR, RNAt from young shoots from noninoculated or vector-infected plants was prepared using a standard protocol with two phenol: chloroform: isoamyl alcohol extractions, followed by RNA precipitation with 6 M lithium chloride (Ancillo et al., 2007). RNAt extracts were treated with RNase-free DNase (Turbo DNA-free; Ambion, Paisley, UK). RNAt concentrations were measured in duplicate in a NanoDrop 1000 UV-VIS spectrophotometer (Thermo Fisher Scientific, Waltham, MA) and then adjusted to the same concentration. DNA-free RNAt from plants infected with the WT-CLBV or the clbvINpr-AtFT constructs was reversetranscribed using oligo(dT) and ThermoScript TM reverse transcriptase (Invitrogen). Apetala (AP1) and CiFT mRNA levels were estimated by RT-qPCR using SYBR GREEN detection and specific primers ( Table 1). The relative mRNA levels were normalized to the RNAt amounts as previously described by Ag€ uero et al. (2012). The expression of each gene in the clbvINpr-AtFTinoculated plants relative to the control WT-CLBV-infected plants was determined by the 2 ÀDDCT method (Livak and Schmittgen, 2001). These values were subjected to multifactorial analysis of variance (ANOVA) using the IBM SPSS Statistics 20.0 program.
Genomic DNA was isolated from the hybrids and their parent genotypes using the Plant DNAeasy kit (Qiagen, Hilden, Germany), following the manufacturer's protocol. PCR amplification of the genomic DNA with the SSR markers was performed using a Mastercycler epgradient S (Eppendorf, Hamburg, Germany). Each reaction contained 0.8 U Taq DNA polymerase (Thermo Fisher Scientific), 2 ng/mL Citrus DNA, 0.2 mM well RED (Sigma-Aldrich, Gillingham, UK) dye-labelled forward primer, 0.2 mM non-dyelabelled reverse primer, 0.2 mM each dNTP, 10 9 PCR buffer and 1.5 mM MgCl 2 in a final volume of 10 mL. The PCR protocol was as follows: an initial denaturation step at 94°C for 5 min followed by 40 cycles of 30 s at 94°C, 1 minute at 50°C or 55°C, 45 s at 72°C and a final elongation step of 4 min at 72°C. Capillary electrophoresis was performed using a CEQ TM 8000 Genetic Analysis System (Beckman Coulter Inc., Fullerton, CA). Data were collected and analysed using GenomeLab â GeXP (Beckman Coulter Inc.) version 10.0 software.

Genome DNA analysis
To assess whether clbvINpr-AtFT integrates within the plant genome, genomic DNA extracts from bark and leaves of clbvINpr-AtFT-infected C. excelsa plants were analysed by PCR using  primers designed on different parts of the CLBV genome and on the CiFT gene (Table 1). The DNA synthesized was analysed by 2% agarose gel electrophoresis and GelRed staining (Biotium Inc., Hayward, CA). The DNA fragments obtained were inserted into the pGEM-T plasmid vector (Promega Corp.) and cloned in Escherichia coli DH5a cells. Positive clones were sequenced.