Introduction

Brazil stands out as the largest global producer of yellow passion fruit (Passiflora edulis Sims) [1, 2]. In 2018, the production was 593,429 metric tons from an area of 41,584 ha. Despite being the largest worldwide producer, the average productivity of 14.3 t ha-1 is low [3]. This low yield is partly due to the severity of passion fruit woodiness disease (PWD), caused by cowpea aphid-borne mosaic virus (CABMV) [4,5,6,7].

CABMV (genus Potyvirus, family Potyviridae) has a genome consisting of a single-stranded, positive-sense RNA, which encodes proteins affecting viral replication and accumulation, defense, viral movement, and symptoms in infected plants [8, 9]. The virus is transmitted by aphid vectors (Hemiptera: Aphididae) in a non-circulative and non-persistent manner during probing [10,11,12]. Plants infected by CABMV show inhibited growth, their leaves have a mosaic appearance, blisters, and/or deformations, and their fruits are deformed and smaller, becoming hardened [10, 13]. Diseases caused by viruses are considered to have the highest socioeconomic impact on passion fruit cultivation in Brazil because they reduce the plant’s longevity, productivity, and fruit quality, and there are no effective measures to control them [14, 15], only damage mitigation measures [16]. The use of resistant cultivars is considered the best strategy, because it does not increase production costs for labor or require chemicals to control the vector insect [17,18,19]. However, so far, there are no yellow passion fruit cultivars with this attribute [18,19,20,21]. On the other hand, studies indicate that wild passion fruit tree species carry CABMV resistance genes [19, 22,23,24,25], making them an alternative for developing resistant cultivars through interspecific crosses with susceptible species [18, 26].

The evaluation of the Passion Fruit Active Germplasm Bank of the Embrapa Cassava and Fruits (Embrapa Mandioca e Fruticultura) research unit, with the aim of identifying resistant wild genotypes is considered an indispensable step for the development of CABMV-resistant cultivars [19, 25]. However, the evaluation and accuracy of the quantification of disease severity are highly dependent on the method used, which has a direct correlation with the quality of the data generated for subsequent manipulation and analyses [27]. In this pathosystem, phytopathometric indices have been used to quantify the CABMV-induced symptoms in Passiflora species [25, 27,28,29]. However, inferring the reaction only based on leaf symptoms has not been sufficient to determine the resistance level, because wild species may not develop leaf symptoms and may therefore be classified as immune despite the presence of virus in the tissues. In this context, RT-qPCR is a sensitive technique that allows the viral titre in plants to be determined [30, 31]. Despite its relevance, there are few studies that have identified the resistance of wild passion fruit species by means of RT-qPCR [14, 32].

Thus, this study had the following objectives: i) to evaluate the progression of passion fruit woodiness disease symptoms in Passiflora spp., ii) to perform screening of Passiflora spp. genotypes for identification of CABMV resistance sources, aiming to select genotypes with high resistance for use in interspecific crosses, and iii) to validate the infection in plants with symptoms through reverse transcription polymerase chain reaction (RT-PCR), and in asymptomatic plants by RT-qPCR in cases where standard RT-PCR detection failed.

Materials and methods

Location and plant material

The study was carried out at the facilities of Embrapa Mandioca e Fruticultura, located in Cruz das Almas, Bahia, Brazil (12°40’39” S, 39°06’23” W, 226 m altitude). The region’s climate is transitional from Am to Aw type (tropical sub-humid to dry) according to the classification of Köppen and Geiger [33], with an annual average air temperature of 23.8 ºC. One hundred twenty-eight genotypes of Passiflora spp. were evaluated, using plants from the Passion Fruit Active Germplasm Bank of Embrapa Mandioca e Fruticultura, belonging to 12 species (Passiflora edulis Sims., P. cincinnata Mast., P. mucronata Lam., P. gibertii N.E Brown., P. alata Curtis., P. setacea DC., P. pohlii Mast., P. miersii Mast., P. bahiensis Klotzsch., P. malacophylla Mast., P. suberosa L., and Passiflora sp.) and three hybrids (a simple interspecific hybrid (F1), a third-generation interspecific hybrid from backcrossing – BC3 [(P. edulis × P. cincinnata) × P. edulis] and an intraspecific hybrid). Two other genotypes were used as controls, one susceptible (P. edulis, cv. BRS Gigante Amarelo) and the other resistant (P. cincinnata, BGP200) [25] (Table 1).

Table 1 List of genotypes of Passiflora spp. used in this study
Full size table

Biological assay and sampling

Approximately 80 seeds of each genotype were soaked in 2 mL of the growth regulator GA4+7 + N-(phenylmethyl)-aminopurine at concentration of 400 mg/L for 24 hours [34]. After this period, the seeds were sown in 162 cells of rigid polypropylene trays (50 mL vol.) filled with a combination of a mixture of coconut fiber (Gold Mix®) and a commercial substrate (Vivato®) in a 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 for the assay. Subsequently, the plants were transferred to polypropylene tubes (100 cm3) and acclimatized in a greenhouse with a temperature of 28 ± 2 ºC and relative humidity (RH) of 75 ± 5%.

Plant inoculation and evaluation of disease symptoms

Leaves with severe symptoms induced by CABMV were collected from yellow passion fruit plants from the Passiflora experimental area and inoculated into plants maintained in a greenhouse as a source of inoculum. The mechanical inoculations with CABMV were performed when the plants had at least four expanded leaves, approximately 60 days after the emergence of the seedlings, as described by Gonçalves et al. [25]. To prevent escape, two inoculations per plant were performed with a four-day interval (Supplementary Fig. S1).

The intensity of symptoms in each leaf was scored using the diagrammatic scale scores proposed by Novaes and Rezende [35], ranging from 1 (without leaf symptoms) to 4 (severe leaf symptoms) (Supplementary Fig. S2). The evaluations started 12 days after the first inoculation (DAI) in all plants, using the first leaf of the fully developed apex, totaling five leaves per plant. Subsequent evaluations were done weekly until 40 DAI.

Evaluation of disease severity and prevalence

Symptom severity was measured by the McKinney disease severity index [36], where disease index (DI%) = (DS × L)/(TNL × HGS); where DS = degree of the determined scale for each leaf; L = number of leaves with each degree of symptoms (score); TNL = total number of evaluated leaves; and HGS = highest grade of the scale (maximum scale score). The prevalence of the disease in various genotypes was determined as the percentage of plants that exhibited typical disease symptoms.

Plants that did not show symptoms at 40 DAI were reinoculated to confirm their resistance to CABMV (Fig. 1). These asymptomatic plants (n = 7 to 25) of the 34 genotypes (Table 2) were pruned to 15 cm (Fig. 1c). Forty days after pruning (DAP), when the plants had at least four leaves, inoculations and leaf symptom evaluations were performed as described above.

Fig. 1
figure 1

Representative scheme of the steps for the detection of cowpea aphid-borne mosaic virus (CABMV) infection in symptomatic and asymptomatic inoculated plants. A) Inoculation (1 and 2) and evaluation of symptoms in passion fruit plants. B) Apical leaf tissue collection from inoculated plants with and without symptoms at 40 DAI. C) Separation of plants that did not show symptoms, pruning and reinoculation (inoculation 3 and 4) of plants at 40 DAP. D) Apical leaf tissue collection of symptomatic and asymptomatic reinoculated plants at 40 days after reinoculation (DARI) for detection of CABMV by reverse transcription polymerase chain reaction (RT-PCR) and reverse transcription quantitative polymerase chain reaction (RT-qPCR). Inoc., inoculation; DAI, days after inoculation.

Full size image
Table 2 Minimum, maximum, and average severity and phenotypic resistance classification of 34 genotypes of Passiflora spp. reinoculated with cowpea aphid-borne mosaic virus
Full size table

Detection of CABMV by RT-PCR

At 40 DAI, apical leaf tissues were collected from symptomatic (S-IP) and asymptomatic (A-IP) inoculated plants, symptomatic reinoculated plants (S-RIP), asymptomatic reinoculated plants (A-RIP), and non-inoculated plants (NIP – negative controls) of the 12 Passiflora spp. and hybrids. RNA extractions were performed from pools containing five apical leaves, representative of five plants of each of the sets (S-IP, A-IP, S-RIP, A-RIP and NIP) (Fig. 1a-d), following the protocol of Ferreira et al. [37]. The treatments were performed with 10 µL of total RNA (2 µg) and 1.5 µL of DNase (2 U/μL) according to the manufacturer’s recommendations (Ambion), with the RNA concentration adjusted to 100 ng/µL.

cDNA was synthesized from 3.0 μL of total RNA, using an with M-MLV Reverse Transcriptase Kit (Invitrogen). PCR reactions were performed with 3.0 μL of cDNA (30 ng/μL) and 10 μM primers for amplification of part of the CABMV cylindrical inclusion gene (CI) to yield a product of 1311 bp [38] (Supplementary Fig. S3). The CABMV CI gene amplification program and procedures for visualization of the PCR products were described previously by Gonçalves et al. [25].

Reverse transcription quantitative polymerase chain reaction

Amplification efficiency was determined from a tenfold serial dilution series from 200 to 0.02 ng/μL. The primer pair qCABMV07_For (5’ CTGGTAGAGTGCTTCTCAATTTGG 3’) and qCABMV07_Rev (5’ CTCTCCCTTGATGGCCTCAA 3’), was used to amplify part of the CABMV coat protein (CP) gene to produce a product of 121 bp [39] (Supplementary Fig. S3). The amplification efficiency was calculated automatically using 7500 Fast v2.0.6 software, using the slope obtained by linear regression according to the following formula: % Efficiency = [10(-1/slope) -1] × 100 [40, 41]. A standard curve of purified CABMV RT-PCR product was generated using a tenfold serial dilution from 1 to 0.0001 ng/μL. The number of molecules of the CABMV (copies/μL) at each dilution point was determined using the following formula: Copy number = (Sample concentration [ng/µL] × 6.022 × 1023)/(Fragment size [bp] × 1 × 109 × 660 g/mol) [42].

The qPCR assays were performed using a 7500 Fast Real-Time PCR System (Applied Biosystems) programmed for analysis of the “Quantitation - Standard Curve” type. The reactions were performed in a MicroAmp™ Fast Optical 96-Well Reaction Plate (0.1 mL), with 10 µL reactions containing 1.0 μL of cDNA (generated from 100 ng/μL of RNA), 0.1 μM each primer, GoTaq® qPCR Master Mix (Promega), and CXR Reference Dye. A cDNA sample derived from a plant that was not inoculated with CABMV and a control without template were included as negative controls. All reactions were conducted in technical triplicate. The cycling conditions for the CABMV coat protein gene were 50 ºC for 2 min, 95 ºC for 10 min, 40 cycles at 95 ºC for 15 s, and 60 ºC for 1 min.

Design and data analysis

The experimental design used was completely randomized, considering each of the 25 plants inoculated one repetition. Another five plants that were not inoculated with CABMV were used as controls. Some plants of P. edulis that were inoculated with CABMV and remained asymptomatic (n = 1 to 4) were not included in the severity analysis because P. edulis is generally susceptible to CABMV [25], and they might have escaped infection.

The mean disease index at each time point (12, 19, 26, 33, and 40 DAI) was plotted on a logarithmic scale to examine the evolution of disease in the 12 Passiflora species and three hybrids (inter- and intraspecific). To calculate the rate of disease progression in the leaves, severity values (DI%) were used at the five evaluation time points only for species with symptoms. The original severity or linearized data were tested for some models and adjusted as described previously [43]. Using the best adjustment, the disease progression rate was estimated (r) and determined by the angular coefficient (\({\theta }_{1}\) and \({\theta }_{2}\)) of the regression equation (R2) [43].

The average DI (%) estimates at 40 DAI were compared using the Scott-Knott test (p ≤ 0.05). The genotypes were classified as [19] resistant (R; DI 0.0 – 15.9%), moderately resistant (MR; DI 16.0–31.9%), susceptible (S; DI 32.0–50.9%) or highly susceptible (HS; DI ≥ 51.0%). The analyses were performed in R, using the ‘ExpDes.pt’ package [44]. Genotypes were grouped based on the Gower index [45] and the unweighted pair group method with arithmetic mean (UPGMA). A dissimilarity matrix was made using the Genes program [46], and from the matrix, MEGA7.0 software was used to generate a dendrogram [47]. The viral titres of asymptomatic inoculated plants were determined by measuring the concentration of the viral cDNA (ng/µL) a standard curve [42].

Results

Temporal progression of PWD caused by CABMV in Passiflora species

The progression of the disease was scored as 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 for those in group 2 (P. mucronata and P. cincinnata). The progression of the disease was faster in the species of group 3 (P. alata, P. edulis, the interspecific hybrid BC3, the interspecific hybrid F1, and Passiflora sp.) and of group 4 (intraspecific hybrid) (Fig. 2).

Fig. 2
figure 2

Logarithmic regression of the disease index in 12 species of Passiflora spp. and three hybrids (inter- and intraspecific) on different days after inoculation with cowpea aphid-borne mosaic virus. Group 1 (resistant – R: P. suberosa, P. setacea, P. pohlii, P. malacophylla, P. bahiensis, P. gibertii, and P. miersii); group 2 (moderately resistant – MR: P. mucronata and P. cincinnata); group 3 (susceptible – S: P. alata, interspecific hybrids BC3 and F1, P. edulis, and Passiflora sp.); group 4 (highly susceptible – HS: intraspecific hybrid). The values in parentheses represent the range of the disease index for the species in each group.

Full size image

Based on the values of \({\theta }_{1}\) or \({\theta }_{2}\) between pairs of species (Supplementary Table S1), it was possible to identify significant differences, with disease progression rates being slower in P. cincinnata than in P. alata and Passiflora sp. The species P. gibertii had a slower progression rate than the intraspecific hybrid, Passiflora sp., and P. alata, but a faster rate than P. miersii and P. mucronata. On the other hand, P. miersii and P. mucronata had slower rates than the intraspecific hybrids, P. alata, P. edulis, and Passiflora sp., while the interspecific hybrid (F1) had a faster rate than P. mucronata, while P. miersii and P. mucronata had faster rates than P. cincinnata (Supplementary Table S1). Other comparisons did not reveal significant differences (Supplementary Table S1).

Classification of Passiflora spp. and genotypes based on disease severity

The disease index (DI) ranged from 0.0 to 70.80%, with 26 genotypes (20.3%) classified as resistant (R; DI 0.0 to 15.75%). Within this same class, the genotypes that should be highlighted are BGP152 (P. suberosa), BGP170 (P. malacophylla), BGP434, BGP244, BRS Pérola do Cerrado (P. setacea), BGP454 (P. pohlii), and BGP477 (P. bahiensis), which did not exhibit symptoms of PWD (DI 0.0%). Another 12 genotypes (9.4%) were moderately resistant (MR; DI 16.3 to 31.1%), 42 (32.8%) were susceptible (S; DI 33.3 to 50.9%), and 48 (37.5%) were highly susceptible (HS; DI 51.2 to 70.8%) (Figs. 3a-b and Supplementary Table S2).

Fig. 3
figure 3

Cluster and severity of passion fruit woodiness disease (PWD) caused by cowpea aphid-borne mosaic virus (CABMV) in 128 genotypes of Passiflora spp. A) Mean values and the number of genotypes in the classes resistant (R), moderately resistant (MR), susceptible (S), and highly susceptible (HS). B) Mean values of disease severity and their distribution in the classes R, MR, S and HS. C) Average severity of PWD in 12 species of Passiflora spp., an intraspecific hybrid, a simple interspecific hybrid (F1), and an interspecific hybrid of the third backcross generation – BC3 [(P. edulis × P. cincinnata) × P. edulis]. D) Prevalence of the disease in the species. 1,2P. cincinnata genotype (BGP200) and yellow passion fruit cultivar (P. edulis, cv. BRS Gigante Amarelo) used as resistant and susceptible controls during the evaluations of leaf symptoms induced by CABMV.

Full size image

Regarding the P. edulis genotypes, only BGP124 was considered moderately resistant, with a DI of 31.11%. The other 53 genotypes (41.4%) showed some degree of susceptibility. Of these, 27 (21.1%) were susceptible and 26 (20.3%) were highly susceptible. All genotypes belonging to the intra-specific hybrids, interspecific hybrids (F1), and interspecific hybrids from the third generation of backcrossing (BC3) were classified as susceptible or highly susceptible (Fig. 3 and Supplementary Table S2). The genotypes used as controls for resistance (BGP200) and susceptibility (cv. BRS Gigante Amarelo) showed severity within the expected level, with mean DI values of 15.30% and 62.20%, respectively.

Of the species evaluated, seven (46.67%) were classified as resistant (P. bahiensis, P. malacophylla, P. pohlii, P. setacea, P. suberosa, P. gibertii, and P. miersii), with a DI of 0.0 to 14.80%; two (13.33%) were classified as moderately resistant (P. cincinnata and P. mucronata), with a DI of 16.10 to 18.70%; three (20%) (Passiflora sp., P. edulis and interspecific hybrids of the third backcross generation [BC3]) were classified as susceptible to CABMV, with DI values ranging from 37.60 to 50.40%; and three (20%) were considered highly susceptible to CABMV, with a mean DI of 51.90 to 64.90% (hybrids [F1], P. alata, and intraspecific hybrids) (Fig. 3c).

The average severity of 34 reinoculated genotypes ranged from 0.0 to 60.74% (Table 2). Some genotypes demonstrated typical disease symptoms, specifically those belonging to the species P. gibertii, P. cincinnata, P. mucronata, P. miersii and the interspecific hybrid (OTH-137), the latter with 100% prevalence in reinoculated plants (Table 2). Among the genotypes evaluated, 55.88% (n = 19) were considered resistant, with DI 0.0 to 15.83%. Within this same group, the genotypes BGP152 (P. suberosa), BGP170 (P. malacophylla), BRS Pérola do Cerrado, BGP434, BGP244 (P. setacea), BGP454 (P. pohlii), and BGP477 (P. bahiensis) remained asymptomatic even after reinoculation, maintaining their classification as resistant. Another 13 genotypes (38.24%) were classified as moderately resistant, with a DI of 17.50 to 30.29%, and two genotypes (BGP275 and OTH-137) were classified as susceptible (DI: 36.74%) and highly susceptible (DI: 60.74%), respectively (Table 2).

Detection of CABMV

Viral infection in symptomatic inoculated plants was confirmed by RT-PCR with amplification of a 1311-bp fragment of the CABMV CI gene (Fig. 4a). In the asymptomatic inoculated and negative control plants, the systemic replication of CABMV was not confirmed (Fig. 4b and c). Asymptomatic plants were subsequently tested by RT-qPCR to confirm the infection (Fig. 4b). The amplifications using the primer qCABMV07 were very uniform (slope = -3.53, determination coefficient [R2] = 0.997, and efficiency of 91.96%). The standard curve for the purified CABMV RT-PCR product, the R2 value was 0.998, the slope was -3.33, and the qPCR efficiency was 99.37% (Supplementary Fig. S4). The amount of the CABMV CP gene in the serial dilutions of the standard curve ranged from 7.54 × 109 to 7.54 × 105 viral copies per microliter (Supplementary Table S3). These measurements were used to determine the viral titre in the evaluated passion fruit species.

Fig. 4
figure 4

Products of amplification of the cylindrical inclusion gene (1311 bp) of cowpea aphid-borne mosaic virus by reverse transcription polymerase chain reaction analyzed by electrophoresis in a 1% agarose gel. A) Pool of leaf samples of symptomatic inoculated plants (S-IP) at 40 days after inoculation (DAI). B) Pool of leaf samples of asymptomatic inoculated plants (A-IP). C) Pool of leaf samples of plants not inoculated with CABMV (NIP, negative controls). PC: positive control (plants of P. edulis with severe PWD symptoms – cDNA 200 ng/µL). M, 1 kb DNA Marker Ladder (Invitrogen). P. ed, Passiflora edulis; Intra H., intraspecific hybrid; Inter H., interspecific hybrid; P. ala, P. alata; P. mie, P. miersii; P. muc, P. mucronata; P. cin, P. cincinnata; P. gib, P. gibertii; P. sp, Passiflora sp.; P. set, P. setacea; P. poh, P. pohlii; P. mala, P. malacophylla; P. sub, P. suberosa; P. bah, P. bahiensis

Full size image

The CABMV titre in asymptomatic species ranged from 1 × 100 to 3.66 × 107 copies/µL (Fig. 5a). There was little variation in viral titre in the interspecific hybrid (OTH-137) and the species P. gibertii, P. cincinnata, and P. miersii, with 3.66 × 107, 3.24 × 107, 3.05 × 107 and 2.55 × 107 copies/µL, respectively. The species P. mucronata, P. setacea, and P. malacophylla had lower virus titres than the above-mentioned species, with 2.70 × 106, 3.07 × 105, and 9.99 × 104 copies/µL, respectively (Fig. 5a). In P. pohlii, P. suberosa, and P. bahiensis, no CABMV CP amplicons were detected, demonstrating there was no infection after two inoculation attempts (Fig. 5b). The specificity of the qPCR amplifications was confirmed by agarose gel electrophoresis, with single amplicons of the expected size of 121 bp obtained for the interspecific hybrid (OTH-137) and the species P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, and P. malacophylla (Fig. 5b).

Fig. 5
figure 5

Detection and quantification of CABMV in inoculated but asymptomatic plants of Passiflora spp. by reverse transcription quantitative polymerase chain reaction (RT-qPCR). A) Number of CABMV molecules (values above the columns – copies/μL) in inoculated asymptomatic plants. B) 2% agarose gel showing qPCR products from the amplification of a 121-bp fragment of the coat protein gene of CABMV. M, 100 bp DNA Marker Ladder (Ludwig). NTC (non-template control), without cDNA as template; NC (negative control), sample of P. edulis not inoculated with CABMV; Inter H, interspecific hybrid (OTH-137); P. gib, P. gibertii; P. cin, P. cincinnata; P. mie, P. miersii; P. muc, P. mucronata; P. set, P. setacea; P. mal, P. malacophylla; P. poh, P. pohlii; P. sub, P. suberosa; P. bah, P. bahiensis

Full size image

Viral infection in plants that showed symptoms after the third and fourth reinoculation was confirmed by RT-PCR (Fig. 6a). However, some plants of P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, P. malacophylla, P. suberosa, P. pohlii, and P. bahiensis did not exhibit PWD symptoms, so they were analyzed further by RT-qPCR. For the interspecific hybrid OTH-137, all plants had symptoms (Table 2). CP amplicons were detected by RT-qPCR in reinoculated asymptomatic plants (Fig. 1d) of P. cincinnata, P. gibertii, P. miersii, P. mucronata, P. setacea, P. malacophylla, and P. suberosa, but not in P. pohlii and P. bahiensis (Fig. 6b-c). The viral titre ranged from 1 × 100 to 4.20 × 107 copies/µL (Fig. 6b). The variation in viral titre among the species was low, with 4.20 × 107 copies/µL in P. cincinnata, 4.16 × 107 in P. gibertii, 4.09 × 107 in P. miersii, and 3.65 × 107 in P. mucronata. In turn, the viral titre in P. setacea, P. malacophylla, and P. suberosa was lower than in the preceding group, with 8.00 × 106, 4.85 × 105 and 1.79 × 105 copies/µL, respectively (Fig. 6b).

Fig. 6
figure 6

Detection and quantification of CABMV in symptomatic and asymptomatic reinoculated plants by reverse transcription polymerase chain reaction (RT-PCR) and reverse transcription quantitative polymerase chain reaction (RT-qPCR) at 40 days after reinoculation (DARI). A) 1% agarose gel showing RT-PCR amplification products of the cylindrical inclusion gene of CABMV of 1311 bp in symptomatic reinoculated plants (S-RIP). M, 1 kb DNA Marker Ladder (Invitrogen); PC, positive control (plants of P. edulis with severe PWD symptoms – cDNA 200 ng/µL). B) Number of the CABMV molecules (values above the columns – copies/μL) in asymptomatic reinoculated plants of Passiflora spp. C) 2% agarose gel showing qPCR products from the amplification of a 121-bp fragment of the coat protein gene of CABMV. M, 100 bp DNA Marker Ladder (Ludwig); NTC (non-template control), without template cDNA; NC (negative control), sample of P. edulis not inoculated with CABMV; Inter H., interspecific hybrid (OTH-137); P. mie, P. miersii; P. muc, P. mucronata; P. cin, P. cincinnata; P. gib, P. gibertii; P. set, P. setacea; P. mal, P. malacophylla; P. sub, P. suberosa; P. poh, P. pohlii; P. bah, P. bahiensis

Full size image

Discussion

The differences in disease progression in the passion fruit species tested demonstrated the high degree of variability in the resistance or susceptibility of Passiflora spp. to CABMV infection (Fig. 2 and Supplementary Table S1). In the initial evaluation, plants of most of the species tested (n = 9) exhibited typical symptoms of PWD by 12 DAI, and plants of Passiflora sp. (BGP482) showed symptoms at 19 DAI (Fig. 2). Plants of P. suberosa, P. setacea, P. pohlii, P. malacophylla, and P. bahiensis did not exhibit symptoms before the end of the experiment (40 DAI) (Fig. 2). The incubation time of CABMV for most of the Passiflora spp. is not known with accuracy, and the initial expression of symptoms is dependent on the age of the plant and the genotype or isolate used, and it is affected directly by environmental conditions and the nutrition of the plants [16, 29].

Species classified as resistant and moderately resistant are uncultivated and naturally carry resistance alleles [48]. This study is a pioneer in reporting the probable immunity to CABMV of P. bahiensis and P. pohlii and the moderate resistance of P. miersii, increasing the number of wild Passiflora species evaluated for CABMV resistance. Studies have reported resistance to CABMV in genotypes belonging to P. setacea [23, 24], P. cincinnata [23, 25], P. gibertii [25], and P. suberosa [22].

Species with no symptoms or low disease severity can be used in breeding programs of Passiflora for interspecific crosses with commercial species. However, interspecific crosses may not succeed if the species differs in their number of chromosomes or belong to a different subgenus [49, 50], as has been reported in the case of P. suberosa [49] and P. pohlii [51], which can result in genetic barriers to crossings or a lack of synchronization in flowering [52]. In some cases, the interspecific compatibility barriers are relatively weak, so successful hybridization can be achieved [53]. The cytogenetic aspects involved in the crossing of P. edulis and P. cincinnata demonstrate the possibility of obtaining hybrids and thus transferring resistance alleles or other traits of the wild species [54]. Studies using P. setacea and P. cincinnata (both 2n = 18) as donors of CABMV resistance alleles for P. edulis (2n = 18) have also been successful [18,19,20,21, 26, 55]. P. malacophylla and P. bahiensis can potentially be used in breeding programs because they are resistant to CABMV and same chromosome number of P. edulis, opening the possibility of obtaining resistant commercial hybrids. However, complementary studies should be performed with these two species to confirm that viable hybrids can be obtained.

There is a relationship between prevalence and severity of disease for most of the evaluated genotypes. However, P. mucronata (BGP479) and P. edulis (BGP124) plants with disease prevalence of 100% showed only mild mosaic symptoms (score 2) and were classified as moderately resistant (Supplementary Table S2). This indicates that the prevalence is not always linked to severity. The genotypes of P. edulis showed susceptibility to CABMV (Fig. 3 and Supplementary Table S2) except for the BGP124 genotype, which was moderately resistant. Results in the literature have demonstrated different resistance levels among P. edulis genotypes from resistant to highly susceptible [23, 25, 56, 57]. This indicates intraspecific genetic variability of the resistance to CABMV. For this reason, P. edulis genotypes in active gene banks should be tested to identify those with low disease severity, thereby reducing the time required to obtain resistant cultivars [18, 25, 26].

The susceptibility observed in plants of the six intraspecific hybrids of P. edulis is directly related to the selection of parents, taking into consideration only agronomic attributes of vigor and production [58, 59]. The interspecific hybrids of the third backcross generation – BC3 [(P. edulis × P. cincinnata) × P. edulis], despite having contrasting genitors – BGP330 (susceptible) and BGP077 (resistant) [25] – did not show resistance to CABMV. This is probably due to the small number of BC3 progeny (n = 21) evaluated. For gains in resistance to CABMV, it is necessary to evaluate a much larger number of progenies, since 93.75% of the genome involved in the backcrossing belongs to the susceptible recurrent genitor (P. edulis), leading to resistance losses in the progeny. Indeed, [60] evaluated a larger number of progenies and identified CABMV resistant plants in BC3. Genotypes of the fourth and fifth backcross generation involving P. edulis × P. setacea are not resistant to CABMV due to the loss of resistance as new backcrosses are performed [61], possibly due to the polygenic heritage of the trait [18,19,20,21]. The genetic heritage for resistance to CABMV of most Passiflora species is still unknown and is therefore an open field for research in breeding programs.

The variation in the resistance to CABMV among the evaluated genotypes is associated with the genetic variability of the Passiflora species, since they are self-incompatible [52, 62, 63]. Moreover, different studies have attributed differences in the passion fruit response to CABMV to the use of different viral isolates [22, 56], the latency period [16], the individual resistance levels of genotypes [25], genetic and environmental factors (such as temperature and relative humidity) [20, 26, 64], and differences in nutritional condition and age among plants [65]. Alone or together, these factors can influence the virulence of the pathogen and the manifestation of the disease symptoms. In this study, many plants without disease symptoms were observed, especially those of wild species. However, some of these factors may not be the cause, since the evaluated genotypes were the same age and the environmental and nutritional conditions were the same.

The observation that some genotypes showed slightly lower severity after the first two inoculations but were still classified as resistant can be attributable to pre-immunization of these plants, since the same viral isolate was used in two reinoculations (third and fourth inoculation) (Fig. 1c). Crotalaria juncea plants infected with two mild strains of passion fruit woodiness virus (PWV; currently recognized as CABMV [7]) were shown previously to be protected against infection by a severe strain and/or expression of symptoms [66]. This may be related to the uniform distribution of viruses in leaf tissues, limiting the availability of infection sites for the severe strain and thus inhibiting replication and establishment of systemic infection by the severe strain [66]. The competition for replication sites between mild and severe strains of papaya ringspot virus, cucumber green mottle mosaic virus, and citrus tristeza virus has also been demonstrated [68,69,69]. The selection of passion fruit plants that allow a higher rate of multiplication of mild strains or the selection of other mild strains with greater invasive power can enable pre-immunization to control viruses causing PWD under field conditions [35]. In the specific case of the genotypes BGP275 (P. cincinnata) and OTH-137 (interspecific hybrid), after reinoculation, the plants showed increased disease severity, leading to their reclassification as susceptible and highly susceptible, respectively (Table 2). This indicates that the initial tolerance of these genotypes was overcome as the plants were challenged with a high dose of virus after the third and fourth inoculation.

Researchers have attributed the occurrence of asymptomatic plants to escape from inoculation [20, 38]. However, in this study, this explanation is unlikely, because four inoculations were performed, similar to the procedure reported by Correa et al. [14]. The occurrence of asymptomatic plants can be related to the individual resistance features of genotypes, since the passion fruit resistance to CABMV has been shown to be genotype-dependent [29] or due to a viral RNA silencing mechanism, as suggested by Correa et al. [14] and verified in another study [70, 71]. On the other hand, the manifestation of symptoms after reinoculation of initially asymptomatic plants may be due to viral suppressors of gene silencing [72], which suppress host resistance and favor viral replication and movement to the apical leaves of the plants. However, this hypothesis needs to be validated in Passiflora spp..

Despite not having sequenced the amplicons, the unique DNA bands and expected size of 1311 bp in all species with PWD symptoms (Fig. 4a) indicated that the virus used in the artificial inoculations belongs to the species Cowpea aphid-borne mosaic virus. The lack of detection of CABMV in inoculated but asymptomatic plants (Fig. 4b) may be related to the low sensitivity of the RT-PCR when the viral titre in the leaf is low [73, 74].

The small variation in viral titre among asymptomatic wild passion fruit species (Fig. 5a and Fig. 6b) indicated high resistance to CABMV is capable of restricting virus replication to basal levels. In fact, asymptomatic plants may not be free from viral infection [75]. In resistant species, the viruses accumulate to some extent without causing significant negative effects in their hosts. Although a viral titre is maintained, plant growth and fruit yield are minimally affected, and the symptoms of the disease are absent or mild [76].

The slight decrease in viral titre when asymptomatic plants were reinoculated with CABMV (Fig. 6b) may have been due to a previously active molecular signaling system conferring faster recognition and response against viral infection, probably by the effect of the systemic acquired resistance caused by the first inoculation or viral RNA silencing [77, 78]. In the specific case of P. suberosa, this species was reported to be immune to CABMV [22]. In our study, viral infection in this species was confirmed after the third and fourth inoculations, indicating that the high inoculum pressure resulted in infection. On the other hand, plants of P. bahiensis and P. pohlii were not infected by CABMV, even under these conditions, indicating probable immunity. However, it will be necessary to test this immunity with inoculation of CABMV in protoplasts to look for viral replication at the cellular level, as observed in citrus protoplasts infected with citrus tristeza virus [79]. Furthermore it is prudent to test the effective immunity of these species in the field, since the environment is more heterogeneous and the natural infection by aphid vectors is more specialized [12, 80, 81].

A study by Carvalho et al. [82] demonstrated that proteins linked to the regulation of proteasomes, heat shock proteins, and ubiquitination are involved in defense and resistance signaling of P. setacea to CABMV. It is possible that these proteins were involved in the signaling and tolerance of CABMV in the wild species evaluated in this study, but many other proteins can be involved, since they are different species and the response patterns to the virus can be distinct. Defense responses can involve numerous signaling pathways, culminating in the limitation of viral replication. These responses are varied, depending on the species, phenological phase, and environmental conditions. Collectively, the mechanisms mentioned above can be involved in the tolerance of wild species to CABMV, and further studies are needed to determine whether this is the case in Passiflora spp..

Conclusions

Mean PWD progression rates and disease severity in symptomatic plants were lower in P. cincinnata, P. gibertii, P. miersii, and P. mucronata than in P. edulis and P. alata, inter- and intraspecific hybrids, and Passiflora sp. The accessions belonging to P. suberosa, P. malacophylla, P. setacea, P. pohlii, and P. bahiensis did not show visual symptoms of the disease after mechanical inoculation. Some asymptomatic plants (P. cincinnata, P. gibertii, P. miersii, and P. mucronata), after additional inoculations, exhibited lower mean disease severity in relation to the symptoms of the two initial inoculations, which may indicate a control mechanism such as pre-immunization. The absence of visible symptoms in some plants or accessions does not indicate immunity, because asymptomatic plants can still be infected by CABMV, as demonstrated by RT-qPCR analysis. However, P. pohlii and P. bahiensis, even after four inoculations, remained asymptomatic and free of the virus, suggesting that they are probably immune to CABMV.