Cowpea aphid-borne mosaic virus (CABMV) in Passiora by qPCR analysis reveals asymptomatic plants with viruses and new species with immunity

The passion fruit woodiness disease (Cowpea aphid-borne mosaic virus – CABMV) causes socioeconomic problems for Brazilian passion fruit crop. Understanding the temporal progress of the disease and identifying resistance sources to CABMV are essential steps to develop resistant varieties. The objective of the study was to evaluate temporal progress of passion fruit woodiness disease, identify Passiora genotypes with CABMV resistance and to detect virus infection in asymptomatic plants by qPCR. The experiment was conducted in a greenhouse using 128 genotypes belonging to 12 species and three hybrids (inter and intraspecic) of Passiora evaluated in ve periods after inoculation. The symptoms severity was quantied from the disease index (DI%). The CABMV infection in symptomatic plants was conrmed by RT-PCR and in asymptomatic plants by qPCR. Progress rates and disease severity were lower in the species P. cincinnata, P. gibertii, P. miersii e P. mucronata compared to P. edulis, P. alata, Passiora sp. and hybrids. Of the evaluated genotypes, 20.31% were resistant, with emphasis on the accessions of P. suberosa, P. malacophylla, P. setacea, P. pohlii e P. bahiensis that did not show symptoms of virus. The absence of symptoms does not imply immunity of plants to the virus, as the qPCR analysis conrmed infection by the virus in asymptomatic plants of P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, P. malacophylla e P. suberosa. Even after four inoculations, the virus was not detected by qPCR in the upper leaves in plants of P. pohlii and P. bahiensis indicating that these species are immune to CABMV. bahiensis resistant of wild species of Passiora spp. characterized for CABMV reaction.

Introduction during the entire evaluation of the symptoms caused by the passion fruit woodiness disease (Cowpea aphid-borne mosaic virus -CABMV) ( Table 1).

Biological essay and sampling
Approximately 80 seeds of each genotype were soaked in 2 mL of the growth regulator GA 4+7 + N-(phenylmethyl) -aminopurine (Promalin ® ) at concentration of 400 mg.L -1 for 24 hours [31]. After this period, the seeds were sown in 162 cells rigid polypropylene trays (50 mL of vol.) lled with coconut ber mixture (Gold Mix ® ) and commercial substrate (Vivato ® ) in the ratio of 3:1 (v:v) with the addition of 50 g of slow release fertilizer (Osmocote ® ) for each 10 L of substrate. After emergence (40 days after sowing) the 30 most uniform plants were selected to compose the essay. Subsequently, the plants were transferred to polypropylene tubes (100 cm 3 ) and conditioned in a vegetation house with temperature of 28 ± 2 ºC and relative humidity (RU) of 75 ± 5% during the entire essay.

Inoculations and evaluations of the symptoms induced by CABMV
The inoculations with CABMV were performed arti cially when the plants had at least four expanded leaves, approximately 60 days after the emergence of the seedlings. The CABMV isolate was derived from yellow passion fruit matrices plants (P. edulis Sims) with severe symptoms of the disease maintained in a greenhouse. The mechanical inoculation was performed as described by Gonçalves et al. [24], two inoculations per plant were performed with four days interval between the rst and the second (Fig. 1).
The symptoms characterization was performed based on observations on the variation in the intensity of leaf symptoms in each evaluated leaf [24]. For that purpose, the diagram scale was used that varies from 1 (without symptoms) to 4 (severe symptoms) proposed by Novaes and Rezende [32], according to leaf symptoms demonstrated in Fig. 2. The evaluations started 12 days after the rst inoculation (DAI) in all plants from the rst leaf of the fully developed apex, totaling ve leaves per plants. Subsequent evaluations were performed weekly and ended with 40 DAI.

Evaluation of CABMV severity and incidence
The leaf symptoms evaluation data of each plant (grades 1 to 4) in the ve evaluation times (12, 19, 26, 33 and 40 DAI) were used to obtain the disease index (DI). The symptoms severity was measured from McKinney [33] disease severity index, considering the following formula: DI(%) = (DS x L)/ (TNL x HGS); where: DS = degree of the determined scale for each leaf; L = number of leaves with each degree of symptoms (grades); TNL = total number of evaluated leaves, HGS = maximum infection degree (maximum grade). The incidence of the disease in genotypes was evaluated considering the percentage (%) of plants that presented typical CABMV symptoms.

Reinoculation and evaluation of asymptomatic plants
After nishing the leaf symptoms evaluations at 40 DAI, plants that did not manifest viruses symptoms (asymptomatic) were separated for new inoculations, in order to con rm plant resistance to CABMV (Fig.  3). These asymptomatic plants (n = 7 to 25) of the 34 genotypes (Table 3)  Ampli cation e ciency of the real-time PCR essay (qPCR) For the ampli cation e ciency essay a standard sample was prepared for using in decimal serial dilutions (10 n ). The total RNA sample was used (100 ng.μL -1 ) of P. edulis 40 DAI. Then four cDNA syntheses of the same sample were proceeded (2.6.2). In the sequence, PCR reactions of each of the four repetitions (2.6.2) were performed, and analyzed in agarose gel 1% to verify the formation of unique amplicons. Finally, the four cDNA repetitions were gathered forming a composite sample (pool), quanti ed (3x) (NanoVue ™ Plus Spectrophotometer) and the concentration xed to 200 ng.μL -1 .
The essay was performed obtaining a curve with ve points (10 10 to 10 6 ) of dilution in series of ten times; Standard curve of the real-time PCR essay (qPCR) In the assembly of the standard curve essay a standard sample (de ned concentration) was prepared for using in the quanti cation of CABMV in plants. Initially, a total RNA sample was used (100 ng.μL -1 ) of P. edulis 40 DAI. The cDNA synthesis of the sample was proceeded (4 repetitions) and then four PCR reactions were performed for nal volume of 50 μL, next 5.0 μL of each reaction in agarose gel 2% was applied (verify presence of unique amplicons). The samples were gathered originating two samples with 90 μL each, and the nal products of the PCR were puri ed with KitPureLink ™ PCR Puri cation, following the manufacturer speci cations (Invitrogen ™ ). After puri cation, the samples were again gathered (80 μL).
The standard curve was assembled with ve points (10 9 to 10 5 copies) of serial dilutions in ten times; After end of the qPCR reaction and dissociation curve, the technical triplicates of each sample were gathered in a single sample. Then 4.0 μL of dye was applied (5X Green GoTaq ® -Promega ™ ) to the reaction products and it was applied in agarose gel 2%, submitted to electric eld (80 V) for 1h 30 minutes. Therefore, the gel was photo documented to view the presence or absence of amplicons. The fragments size was determined by means of the molecular weight marker (5.0 µL) of 100 pb (Ludwig ™ ).

Design and Data Analysis
The experimental design used was entirely randomized considering each of the 25 plants inoculated as a repetition. Other ve plants were maintained as control (not inoculated with CABMV). The control plants (of all species) and the asymptomatic plants of P. edulis were not considered in the severity analyses.
The means of the disease index (DI%) in each evaluation period (12,19,26,33 and 40 DAI) were plotted in logarithmic curves to determine the disease evolution in the twelve Passi ora species and in the three hybrids (inter and intraspeci c). To calculate the rate of viruses progress in the leaves, the severity values (DI %) were used due to the ve evaluation time for species with symptoms (P. mucronata, P. miersii, P. cincinnata, P. gibertii, Passi ora sp., P. edulis, P. alata, intraspeci c hybrid and interspeci c hybrid [F 1 and BC 3 ]). The original severity or linearized data (Y= severity/100), were tested by the Gompertz empirical models (Y = -nn [(y)]), Monomolecular (Y = ln [1/(1 -y)]) and Logistic (Y = ln [y / (1 -y)]) and adjusted for simple linear regression models [42]. Determination of R 2 and exponit, logit, monit and gompit coe cients were performed through a regression among real values and evaluation period in days (DAI). They were adjusted according to the original or linearized data, obtaining the determination coe cients (R 2 ) of the regression analysis [42]. Using the best adjustment, the disease progress rate was estimated The analyses were performed in R environment using the 'ExpDes.pt' package [43]. The grouping of genotypes was performed based on the Gower index [44] and hierarchical grouping method Unweighted Pair Group Method with Arithmetic Mean (UPGMA). The dissimilarity matrix was obtained with the Genes program [45] and from the matrix the MEGA7.0 program was used to obtain the dendrogram [46].
After the qPCR reactions for asymptomatic inoculated species, the raw data were exported in Excel worksheet (Microsoft ® ), treated, and the estimated viral load in the species were determined from the average value of the viral cDNA quantity (ng.µL -1 ) having as parameter the values relative to each point of the standard curve and after the calculations the number of viral copies was converted to Log 10 scale.

Results
Temporal progress of CABMV in Passi ora species Considering the evolution of the disease, it was null for the species belonging to group 1 (P. suberosa, P. setacea, P. pohlii, P. malacophylla, P. bahiensis, P. gibertii and P. miersii) and stable in group 2 (P. mucronata and P. cincinnata). The evolution of the disease was more progressive in the species of group 3 (P. alata, P. edulis, Interspeci c hybrid -BC 3 , Interspeci c hybrid -F 1 and Passi ora sp.) and of group 4 (Intraspeci c hybrid) (Fig. 5).
Regarding the disease progress rate based on the values of or among pairs of species (Table 2), it was possible to verify signi cant differences among the disease progress rates being lower in P. cincinnata than in P. alata and Passi ora sp. The species P. gibertii presented lower progress rate compared to the intraspeci c hybrid, Passi ora sp. and P. alata, but with higher rate when compared to P. miersii and P. mucronata species. On the other hand, P. miersii and P. mucronata presented lower rate compared to intraspeci c hybrids; P. alata; P. edulis and Passi ora sp., while the interspeci c hybrid (F 1 ) presented higher rate than P. mucronata. P. miersii and P. mucronata species obtained higher rates compared to P. cincinnata (Table 2). Other comparisons did not present signi cant differences ( Table 2).  (Fig. 6c).

Detection of CABMV by qualitative RT-PCR
The molecular analysis by qualitative RT-PCR and primers CABMV/M1MX3726_F/CABMV/M1MX5039_R con rmed the viral infection in symptomatic inoculated plants, with ampli cation of the genomic fragments related to the cylindrical inclusion region (CI) of CABMV with expected size of 1311 pb (Fig.   7a). In asymptomatic inoculated plants and negative control, the systemic replication of CABMV was not con rmed (Figs. 7b and c). Asymptomatic plants were submitted to real time quantitative PCR analysis (qPCR) to con rm the infection (Fig. 7b).

Standardization of the qPCR technique for CABMV quanti cation
The ampli cations with the primer qCABMV07 were quite uniform (Fig. 8a) and the ampli cation e ciency obtained Slope = -3.53, determination coe cient (R 2 ) = 0.997, E ciency (E) = 91.96% and the standard deviation (SD) of the Cycle threshold (Ct) of the technical triplicates in each dilution (10 10 to 10 6 ) was 0.08 to 0.32 (Fig. 8b). The dissociation curve of ampli cations was uniform, without formation of primers dimers or nonspeci c peaks (Fig. 8c).
Point 10 10 despite presenting ampli cations of replicates considerable uniform, did not present linearity in the spacing of -3.32 between the dilutions 10 10 and 10 9 . Point 10 4 did not present uniformity in the ampli cations of the technical repetitions (Figs. 9a and b). Linear regression ratio between the mean values of the Ct and the concentration of the CABMV genic product was observed with determination coe cient (R 2 = 0.998) and an Slope of -3.33, demonstrated e ciency (E) of 99.37% of the qPCR reaction (Fig. 9a). The dissociation curve of ampli cations also demonstrated that there was high speci city of qPCR for the standard curve, observed through the uniformity of ampli cations, formation of unique peaks and without presence of primers dimers (Fig. 9b). The quanti cation of the number of copies of part of the CABMV coat protein gene in the serial dilutions of the standard curve was equivalent to 3.28 x 10 33 to 3.28 x 10 29 copies.µL -1 with mean values of the Ct of the technical triplicates obtained in a very reproducible way and standard deviation (SD) of the Ct varied from 0.03 to 0.34 (Table 4). These quanti cations were used for absolute quanti cation of the viral load in the evaluated species. Detection and quanti cation of CABMV by real-time PCR (qPCR) The qPCR reactions in asymptomatic inoculated plants (Fig. 3b), demonstrate quality and precision in the ampli cation obtaining curve inclination (Slope) of -3.28, determination coe cient (R 2 ) of 0.998, e ciency (E) of 101.74% and the dissociation curve demonstrated the lack of nonspeci c products and absence of primer dimers (Fig. 10a). The standard deviation of the Ct of the three technical replicates of each of the samples varied from 0.07 to 0.65 (Fig. 10b).
Regarding the quantity of CAMBV in asymptomatic species, it was observed variation of 1 x 10 0 to 1.59 x 10 31 copies of viruses per microliter (Fig. 10b). There was little variation in the number of copies of the virus for the interspeci c hybrid (OTH-137) and the species P. cincinnata, P. gibertii and P. miersii, with 1.59 x 10 31 , 1.41 x 10 31 , 1.33 x 10 31 and 1.11 x 10 31 copies of viruses.µL -1 , respectively (Fig. 10b). Regarding the P. mucronata, P. setacea and P. malacophylla species, it was observed lower quantity of virus than the previously mentioned species, with 1.18 x 10 30 , 1.34 x 10 29 and 4.35 x 10 28 copies.µL -1 , respectively (Fig. 10b). In the P. pohlii, P. suberosa and P. bahiensis species, it was not detected fragment corresponding to CABMV, showing that there was no infection after two inoculation attempts (Figs. 10a  and b). The result in agarose gel con rms the speci city in the ampli cations, with unique amplicons of expected size of 121 pb in the interspeci c hybrid (OTH-137) and the P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea and P. malacophylla species (Fig. 10c).
From the molecular analysis of RT-PCR there was con rmation of the viral infection in plants that presented symptoms after the third and fourth reinoculation (Fig. 11a). However, numerous plants of P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, P. malacophylla, P. suberosa, P. pohlii and P. bahiensis did not manifest viruses symptoms, and, therefore, were submitted to the analysis of qPCR, except for interspeci c hybrid (OTH-137) that had all plants with symptoms.
The parameters of quality de nition and accuracy of the qPCR technique for reinoculated asymptomatic plants (Fig. 3c) demonstrated reproductive and accurate results, obtained Slope = -3.35, determination coe cient (R 2 ) = 0.998 and E ciency (E) = 98.84%, and the dissociation curves also demonstrated high speci city of ampli cation of the CABMV coat protein gene (Fig. 11b). The standard deviation of the Ct of the three technical replicates of each of the samples varied from 0.02 to 0.92 (Fig. 11c). The analysis of qPCR in reinoculated asymptomatic plants (Fig. 3d), detected CABMV in P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, P. malacophylla and P. suberosa, but was not detected in P. pohlii and P.
The variation in the number of copies of the virus among the evaluated species was low, with 1.83x10 31 for P. cincinnata, 1.81 x 10 31 in P. gibertii, 1.78 x 10 31 compared to P. miersii and 1.59 x 10 31 for P. mucronata. P. setacea and P. malacophylla species presented mild decrease in the viral load compared to the species mentioned above, with 3.49 x 10 30 and 2.11 x 10 29 copies of viruses.µL -1 . In P. suberosa, there was the viral detection only after plant reinoculation, with viral load of 7.78 x 10 28 copies of viruses.µL -1 (Fig. 10c). The result demonstrated in Fig. 10d, evidences the formation of unique amplicons of CABMV coat protein of 121 pb in species P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, P. malacophylla and P. suberosa.

Discussion
Temporal progress and identi cation of resistance source in Passi ora spp. species and genotypes The differences in evolution curves and viruses progress rates ( Fig. 5 and However, the incubation time of CABMV for most of the Passi ora spp. it is not known with accuracy, and the initial expression of symptoms is dependent on the plant age, genotype or isolated used and affected directly by environmental conditions and nutrition of plants [16,28]. In the initial evaluation (12 DAI), most species had already manifested the typical symptoms of the disease, however, it is possible that some of the species had already expressed symptoms before this period. P. suberosa, P. setacea, P. pohlii, P. malacophylla and P. bahiensis species did not manifest symptoms over time and Passi ora sp. was at 19 DAI.
Species with null or low severity can be used to intensify the genetic improvement programs of Passi ora via interspeci c crosses. However, interspeci c hybridizations with commercial species, in some cases, may not succeed due to differences in the number of chromosomes (P. pohlii [2n = 12 and 36] and P. suberosa [2n = 12, 24 and 36]) and belong to different subgenera [49,50] and in the same subgenre can occur barriers of crossings or lack of synchronism in the owering [51]. However, in some cases the interspeci c barriers of incompatibility are relatively fragile, and there may be success in hybridization [52].
The understanding of cytogenetic aspects involved in the crossing of P. edulis x P. cincinnata demonstrate the possibility of obtaining hybrids and thus transferring resistance alleles or other characters of the wild species [53]. Studies using P. setacea and P. cincinnata (both 2n = 18), as donors of CABMV resistance alleles for P. edulis (2n = 18) were also successful [18-20, 25, 54]. P. malacophylla and P. bahiensis species (belonging to the genus Passi ora) present potential for using genetic improvement programs of passion fruit, because they are resistant to CABMV (DI: 0.0%) and have chromosomal analogy with P. edulis (2n = 18) [55], opening possibilities of obtaining resistant commercial hybrids. However, complementary studies should be performed in these two species to con rm the real chance of obtaining viable hybrids.
It was observed relation between incidence and severity of CABMV for most of the evaluated genotypes, already in P. mucronata (BGP479) and P. edulis (BGP124) even with incidence of 100% the plants presented only mild mosaic (grade 2), being classi ed as moderately resistant (Sup. 1). The occurrence of this case allows to indicate that the incidence cannot always be related to the severity.
Most of the genotypes of P. edulis presented susceptibility to CABMV (Fig. 6), except for the BGP124 genotype that was moderately resistant. Results in the literature demonstrated different resistance levels among P. edulis genotypes, being classi ed from resistant to highly susceptible [22,24,56,57]. This indicates there is intraspeci c genetic variability in the resistance to CABMV. For this reason, the evaluation of P. edulis genotypes in GABs should be performed to explore low severity levels of the disease, and thus reduce the time for obtaining resistant cultivars [18,24,25].
The susceptibility observed in the six intraspeci c hybrids of P. edulis (H09-111, H09-02, H09-110, H09-09, H09-112 and H09-158) is directly related to the selection of parentals, taking into consideration only agronomic attributes of vigor and production [58,59]. Interspeci c hybrids of the third generation of backcrossing (BC 3 ) despite having contrasting genitors, BGP330 (P. edulis; susceptible) and BGP077 (P. cincinnata; resistant) [24] did not present resistance to CABMV, probably related to the low number of evaluated progenies (n = 21) or fact that there are few genes or resistance losses in BC 3 , since 93.75% of the genome involved in the backcrossing belongs to the susceptible recurrent genitor (P. edulis). Fonseca et al. [60] also did not verify the resistance to CABMV in genotypes of the fourth and fth generation of backcrossing (BC 4 and BC 5 ) involving P. edulis x P. setacea, due to the loss of resistance as new backcrossing are performed, probably due to the polygenic heritage of the character [18][19][20]. Therefore, that there are gains associated with resistance to CABMV it is necessary to evaluate a very large number of progenies as the backcrossing generations advance [19]. However, the genetic heritage for resistance to CABMV of most of the Passi ora species is still unknown, being an open eld for researches in genetic improvement programs.
The variation in the resistance level to CABMV within and among the evaluated genotypes can be associated with the genetic variability of the passi oras, since they are self-incompatible plants [51,61,62]. Moreover, different studies attribute the passion fruit response to the use of different isolates [21,56], latency period and virus incubation [16], individual resistance level of genotypes [24], genetic and environmental factors (such as temperature and relative humidity) [19,25,63], nutritional condition and age difference among plants [64]. Alone or together, these factors can in uence the virulence of the pathogen and in the manifestation of the disease symptoms. In this study many plants were observed without viruses symptoms, especially in wild species. However, some of these factors may not be the cause, since the evaluated genotypes had the same age and the environmental and nutritional conditions were the same.

Conclusions
Progress rates and mean disease severity in symptomatic plants were lower in P. cincinnata, P. gibertii, P. miersii and P. mucronata in relation to the species P. edulis, P. alata, inter and intraspeci c hybrids and Passi ora sp. The accessions belonging to the species P. suberosa, P. malacophylla, P. setacea, P. pohlii e P. bahiensis did not manifest visual symptoms of the disease after arti cial inoculation. Some asymptomatic plants (P. cincinnata, P. gibertii, P. miersii and P. mucronata) after new inoculation cycles (third and fourth inoculation) had lower mean disease severity in relation to the severity of the symptoms of the two initial inoculations and this may indicate some control mechanism like pre-immunization. The absence of visual symptoms of the disease in some plants or accessions does not indicate immunity, because asymptomatic plants may exhibit systemic infection by CABMV as detected by qPCR analysis. The species P. pohlii and P. bahiensis even after four inoculations, showed no disease symptoms and no systemic infection of the virus by qPCR analysis, revealing that they are immune to CABMV.