Genetic Diversity and Geographic Distribution of Cucurbit-Infecting Begomoviruses in the Philippines

Cucurbits are important economic crops worldwide. However, the cucurbit leaf curl disease (CuLCD), caused by whitefly-transmitted begomoviruses constrains their production. In Southeast Asia, three major begomoviruses, Tomato leaf curl New Delhi virus (ToLCNDV), Squash leaf curl China virus (SLCCNV) and Squash leaf curl Philippines virus (SLCuPV) are associated with CuLCD. SLCuPV and SLCCNV were identified in Luzon, the Philippines. Here, the genetic diversity and geographic distribution of CuLCD-associated begomoviruses in the Philippines were studied based on 103 begomovirus detected out of 249 cucurbit samples collected from 60 locations throughout the country in 2018 and 2019. The presence of SLCCNV and SLCuPV throughout the Philippines were confirmed by begomovirus PCR detection and viral DNA sequence analysis. SLCuPV was determined as a predominant CuLCD-associated begomovirus and grouped into two strains. Interestingly, SLCCNV was detected in pumpkin and bottle gourd without associated viral DNA-B and mixed-infected with SLCuPV. Furthermore, the pathogenicity of selected isolates of SLCCNV and SLCuPV was confirmed. The results provide virus genetic diversity associated with CuLCD for further disease management, especially in developing the disease-resistant cultivars in the Philippines as well as Southeast Asia.


Introduction
Cucurbit is an important economic crop cultivated worldwide [1]. In Southeast Asia, the harvested area of cucurbit crops has reached 240,603 ha with a production of more than 4,479,058 tonnes [2]. Meanwhile, cucurbit crops are also important in the Philippines with 31,454 ha harvested area and a production of 436,849 tonnes [2]. However, plant viral diseases affect the productivity and quality of cucurbit crops, causing significant economic losses [3,4]. More than 70 virus species have caused severe epidemic diseases in cucurbit production areas [5]. Cucurbit leaf curl disease (CuLCD) is one of the most important cucurbit viral diseases. The symptoms are a complex of leaf mosaic, yellowing, curling, enation, vein thickening and plant stunting [6][7][8]. This disease has constrained cucurbit crop production in many tropical and subtropical regions [9][10][11]. The disease incidence often reaches 100%, resulting in significant yield losses (up to 100%) [9][10][11]. CuLCD is caused by whitefly (Bemisia tabaci)-transmitted begomoviruses [7][8][9][10][11][12]. According to the ICTV taxonomic criteria for begomovirus, viruses with nucleotide identity < 91% of fulllength DNA-A are considered as distinct species and those with 91-94% are considered as distinct strains of a virus species [13,14]. More than 17 begomoviruses were associated with CuLCD [14,15], and most of them are bipartite begomoviruses [14].

Sequence Analysis of Cucurbit-Infecting Begomoviruses in the Philippines
Based on the results of begomovirus detection, the crops and locations of the samples collected, 55 begomoviral DNA-As and 50 DNA-Bs were partial sequenced (Table 1). Based on the sequence analysis, partial DNA-A sequences can be grouped into two Clusters (data not shown). Clusters 1 and 2 have the highest nucleotide identity with DNA-As of SLCuPV (>94.2%) and SLCCNV (>93.6%), respectively. The DNA-B partial sequences can also be grouped into two Clusters. Cluster 1 has the highest nucleotide identity (>90.2%) with SLCuPV DNA-Bs. Meanwhile, Cluster 2 has the highest nucleotide identity (>87.0%) with SLCCNV DNA-Bs.
Based on the results of partial sequence analysis, the locations of samples collected and the crops, 39 full-length DNA-A sequences and 38 full-length DNA-B sequences were obtained (Table 1). Conserved begomoviral nonanucleotide sequences (TAATATT/AC) in the stem loop structure of intergenic region (IR) were found in all sequences. Six open reading frames (ORFs AV1, AV2, AC1, AC2, AC3 and AC4) were determined in all 39 full-length DNA-A sequences. Meanwhile, two ORFs (BV1 and BC1) were also determined in all 38 full-length DNA-B sequences (Table S3).
In the phylogenetic analysis, the full-length DNA-A sequences were grouped into five Clusters ( Figure 1 and Table 2). According to the demarcation criteria of DNA-A nucleotide identity < 91.0% for begomovirus species, the newly identified virus isolates in Clusters 1a and 2a were considered as SLCuPV isolates ( Figure 1 and Table 2). Their full-length DNA-A sequences revealed >91.2% identity with the reported SLCuPV isolates. Furthermore, isolates in Clusters 3a, 4a and 5a were SLCCNV and revealed 90.8-99.6% sequence identity ( Figure 1 and Table 2). All newly identified SLCCNV isolates were in Cluster 3a.  cleotide identity <91.0% for begomovirus species, the newly identified virus isolates in Clusters 1a and 2a were considered as SLCuPV isolates ( Figure 1 and Table 2). Their fulllength DNA-A sequences revealed >91.2% identity with the reported SLCuPV isolates. Furthermore, isolates in Clusters 3a, 4a and 5a were SLCCNV and revealed 90.8-99.6% sequence identity ( Figure 1 and Table 2). All newly identified SLCCNV isolates were in Cluster 3a.  Table S2. Both phylogenic trees were  Table S2. Both phylogenic trees were rooted on Squash leaf curl virus (SLCuV). The numbers of each branch indicated the percentage of 1000 bootstraps.  Based on the demarcation criteria of DNA-A nucleotide identity < 94.0% for begomovirus strain in a species, SLCuPV isolates of Clusters 1a and 2a were composed of strains (A and B) ( Figure 1 and Table 2). Virus isolates in Cluster 1a were SLCuPV strain A and shared sequence identity 94.4-99.9% to each other. The isolates of Cluster 2a composed of SLCuPV strain B showing 92.1-99.1% sequence identity to each other. However, isolates among SLCuPV strains A and B revealed sequence identity 91.2-96.1%. For newly identified SLCuPV isolates, isolates from eight bottle gourd, three of chayote and seventeen pumpkin samples were grouped in SLCuPV strain A, whereas SLCuPV strain B contained newly identified isolates from two bottle gourd, two muskmelon, one pumpkin and one wild melon samples. On the other hand, SLCCNV isolates were grouped into Clusters 3a, 4a and 5a and then also considered as distinct virus strains ( Figure 1 and Table 2). The virus isolates in Cluster 3a were SLCCNV strain A which shared sequence identity > 95.9% with each other and 92.0-93.4% with other SLCCNV isolates. The virus isolate in Cluster 4a was considered as SLCCNV strain B showing sequence identity 90.8-93.4% with other SLCCNV isolates. Virus isolates in Cluster 5a were composed of SLCCNV strain C which showed sequence identity 93.2-99.6% to each other and 90.8-93.2% with other SLCCNV isolates. All newly identified SLCCNV isolates from pumpkin samples were grouped in SLCCNV strain A. The isolate SLCCNV-[PH-P54-Cyt-06] (EU487031) was clustered in SLCCNV strain B. The SLCCNV strain C composed of SLCCNV isolates that were reportedly from China, Indonesia, Malaysia, Thailand, Timor-Leste and Vietnam [11,15,17,18,23,24,35].
Based on the phylogenetic analysis, SLCuPV DNA-B sequences can be grouped into Clusters 1b, 2b and 3b ( Figure 1 and Table 2). SLCuPV DNA-Bs in Cluster 1b shared 89.3-99.9% sequence identity and revealed <89.9% sequence identity with others. Virus DNA-B in Cluster 2b shared 85.5-98.6% sequence identity and showed 76.2-89.9% sequence identity with others. Both SLCuPV DNA-Bs in Cluster 3b revealed 96.7% sequence identity and showed <89.4% sequence identity with other DNA-Bs. The newly identified SLCuPV DNA-Bs from four bottle gourd, three chayote and sixteen pumpkin samples were grouped in Cluster 1b. Meanwhile, the newly identified SLCuPV DNA-Bs from six bottle gourd, two muskmelon, six pumpkin and one wild melon samples were in Cluster 2b. Interestingly, the DNA-Bs of SLCuPV-[TW-YL] (EU479711) and SLCuPV-[TW-1-1-Cyt-10] (JF746196) reported from Taiwan were in the distinct Cluster 3b. SLCCNV DNA-B sequences from China, Indonesia, Malaysia, Thailand and Vietnam were in Cluster 4b and revealed >72.8% sequence identity to each other.

Specific Detection of Cucurbit-Infecting Begomoviruses in the Philippines
Based on virus specific detection of virus infectious clones and selected field samples, the designed specific primer pairs could carry out the specific detection effectively:  Figure S1 and Table S1). By the specific detection of all begomovirus-positive samples, the SLCuPV was detected as being the predominant cucurbit-infecting begomovirus in the Philippines (Table 1). The SLCuPV DNA-A and DNA-B were detected in all begomovirus-positive samples including bottle gourd, chayote, muskmelon, pumpkin and wild melon samples collected from Luzon, Visayas and Mindanao (Table 1). However, SLCCNV DNA-A was limited in twelve pumpkin samples collected in Luzon, six bottle gourd and seven pumpkin samples collected in Visayas and one pumpkin sample collected in Mindanao (Table 1). Interestingly, all of the SLCCNV DNA-A was detected as having mixed-infection with SLCuPV (Table 1). However, SLCCNV DNA-B was not present in all samples with SLCCNV DNA-A (Table 1).   Figure 2 and Table 3). Furthermore, tobacco plants infected with SLCuPV-B[PH-BoG216-18] revealed mild symptoms and viral DNAs were also detected ( Figure 2 and Table 3). Meanwhile, symptoms of mosaic and leaf curling were observed on tobacco plants which were inoculated with SLCuPV-A[PH-BoG137-18] (Figure 2). Viral DNA-Bs were also detected in all inoculated plants which SLCuPV DNA-As detected. Since SLCCNV DNA-B was not present in the samples collected from the Philippines, the pathogenicity of SLCCNV isolates were tested by the co-agroinoculation of SLCCNV-[MY-Sq3-5-16] DNA-B with two infectious SLCCNV DNA-As. Both infectious SLCCNV isolates could successfully infect all cucurbit crops ( Figure 2 and Table 3). Virus-infected plants revealed the disease symptoms of mosaic, yellow spot, leaf curling and yellowing. However, the mild symptoms of mosaic, leaf curling and yellowing were observed in the infected tobacco plants (Figure 2). The presence of viral DNAs was also confirmed by PCR detection (Table 3). When the SLCCNV DNA-A was co-inoculated with SLCuPV DNA-B from the same sample, both combinations of infectious clones could not infect cucurbit crops, while only tobacco plants could be infected ( Table 3). The tobacco plants were revealed to be symptomless and only viral DNA-A was detected. In addition, the infectious viral DNA-As of SLCCNV and SLCuPV were also found to infect tobacco plants, but without developing symptoms. The presence of viral DNA-As were further confirmed by PCR detection (data not shown).  [15]. d Viral DNA-A-detected plants/inoculated plants were determined at 28 days after agroinoculation. All plants were detected as having viral DNA-A present, but no DNA-B was detected.
The high CuLCD incidence combined with high infection of begomovirus was detected in bottle gourd, chayote, muskmelon, pumpkin and wild melon samples, however begomovirus was not detected in watermelon and bitter gourd. This may be due to the limited samples surveyed in this study, or bitter gourd and watermelon might not be hosts of begomovirus in the Philippines. The predominant SLCuPV isolates in the Philippines were distinctly separated into two strains (A and B). Meanwhile, SLCuPV distribution was related to geographic area, with strain A distributed throughout the Philippines, whereas strain B was limited to the Luzon area. The SLCuPV-B was also distributed in Taiwan [38], the country north of the Philippines. A geographically related distribution has also been shown in a begomovirus study in Brazil [48]. Here, the geographically related distribution of SLCuPV strains suggests that the virus may originate from Luzon, where CuLCD was observed in 1977 [20]. The virus then evolved strains A and B which were distributed toward the south and north, respectively. The SLCCNV isolates from Southeast Asia and China were diverse with three distinct strains (A to C). Early in the century, SLCuPV was considered as a strain of SLCCNV based on both viruses revealing high sequence identity (90.3%) [21]. Our data also suggest SLCuPV could be evolved from SLCCNV (Figure 1). On the other hand, SLCCNV might have emerged in the Philippines earlier than SLCuPV. In addition, the distribution of SLCCNV was also related to the geographic area, the strains A and B were distributed in the Philippines, and the strain C was in China, Indonesia, Malaysia, Thailand, Timor-Leste and Vietnam. The geographically related distribution of SLCCNV was also confirmed in South Asia [40]. Further analysis of SLCCNV South Asian isolates indicated that they were also distinct to those from Southeast Asia and China (86.7-91.9% sequence identity). In the Philippines, the SLCCNV isolates obtained in this study were the strain A, whereas the previously reported isolate was the strain B. This implies that SLCCNV evolved in this period. The temporal evolution of begomoviruses was also studied in Brazil. The Tomato severe rugose virus (ToSRV) isolates obtained in 2004 and 2008 were in distinct groups [48]. However, further investigation should be conducted for monitoring of the temporal and spatial distribution of cucurbit-infecting begomoviruses in the Philippines as well as Southeast Asia.
Furthermore, the grouping of cucurbit-infecting begomoviral DNA-Bs was related to the viral DNA-As ( Figure 1). However, cucurbit-infecting begomoviral DNA-Bs (71.5-99.9% sequence identity) were more diverse than DNA-As (85.2-99.9% sequence identity) in this study. This may result from the fact that both viral genetic components are facing different evolutionary patterns [48,49]. The accumulation level of DNA-B evolution can be higher than DNA-A and that has contributed to the high variation of the DNA-B population [48,50]. The reduction of the mutation and evolution rate in viral DNA-As may be the result of gene overlapping on the viral genome [49,51]. Thus, DNA-B can conduct greater variation; so far, no gene overlapping was observed in the viral component [48,49,51].
Based on the limited samples and virus specific detection, SLCuPV was determined as the predominant cucurbit-infecting begomovirus in the Philippines and has wider host range including bottle gourd, chayote, muskmelon, pumpkin and wild melon. The pathogenicity of predominant SLCuPV also revealed that all tested SLCuPV isolates can infect bottle gourd, Chinese squash, melon and pumpkin and revealed CuLCD symptoms as those reported previously [21]. In the Philippines, SLCCNV was only found in pumpkin and bottle gourd even when the virus was distributed countrywide. However, SLCCNV can infect more cucurbit crops than SLCuPV in Southeast Asia [8]. This may be due to the fact that SLCCNV was detected without viral DNA-B in the Philippines. The investigation was further confirmed by two SLCCNV infectious viral DNA-As, which could not infect bottle gourd, Chinese squash, melon and pumpkin, even when both viral DNA-As were coinoculated with the SLCuPV DNA-B from same sample. However, when the infectivity of both SLCCNV DNA-As was completed by being co-inoculated with an infectious SLCCNV DNA-B [15], all infected cucurbits also revealed CuLCD symptoms. This implicated that SLCuPV is a more aggressive virus in the Philippines. However, the mixed infection of SLCuPV and SLCCNV provides the possibility for virus recombination, threatening cucurbit crops in the country.
In conclusion, based on the limited cucurbit samples collected in 2018-2019, two cucurbit-infecting begomoviruses, SLCCNV and SLCuPV were found throughout the Philippines. The varied SLCCNV and SLCuPV strains were of a temporal and spatial distribution, respectively. The SLCuPV was determined as predominant with a wide host range. Meanwhile, SLCCNV was detected without DNA-B and limited in bottle gourd and pumpkin. The pathogenicity of both SLCuPV strains was confirmed. The infectivity of SLCCNV DNA-A was also completed with a complementary viral DNA-B. Furthermore, the mixed infection of SLCuPV and SLCCNV implicates the possible recombination of both viruses. The results provided here are important for the development of virus-resistant cucurbit cultivars, and also strengthening the efficiency of CuLCD management in the Philippines as well as Southeast Asia. In addition, the developed virus specific primers can be applied for quarantine purposes. Furthermore, the infectious viral DNAs can also be applied in determining the virulence of viral DNAs, screening virus resistance, studying viral gene function and virus-vector-host interaction, etc. [52][53][54].

CuLCD Survey, Sample Collection and Viral DNA Extraction
Two hundred and forty-nine diseased samples from cucurbits including bitter gourd, bottle gourd, chayote, muskmelon, pumpkin, watermelon and wild melon (Cucumis agristis) with symptoms of mosaic, yellowing, leaf curling, blistering and stunting were collected from 60 locations throughout the Philippines during 2018 and 2019 ( Table 1). The disease incidence in the field was investigated based on the estimated percentage of diseased plants observed in total. One hundred and thirty-one disease samples were collected on 7 cucurbit species from 29 locations in the north Philippines (Luzon). Forty-seven disease samples were collected on four cucurbit species from 11 locations in the central area of the Philippines (Visayas). Seventy-one diseased samples were also collected from four cucurbit species in 20 locations of Mindanao, the Philippines ( Table 1). The samples were dried with silica gel and then preserved under −20 • C. For begomovirus detection, approximately 50 mg of dried leaf were used for viral DNA extraction with the modified Dellaporta method (modified from [55]).

Begomovirus Detection, Cloning and Sequence Analysis of Begomoviral DNAs
Begomoviral DNA-A was detected by PCR with degenerate primer pair-PAL1v1978RYNN/ PAR1c715H with annealing temperature 58 • C [15] (Table S1). Furthermore, the viral DNA-Bs were detected with degenerate primer pair-SLCCNV-BV1/BC1 with annealing temperature 48 • C (Table S1). The virus-associated satellite DNA was also detected by primer pair-Beta01/Beta02 [56]. The PCR reaction mixture contained 5 µL of extracted DNA, 0.5 µL each of 10 mM forward and reverse primers, 12.5 µL of 2X Hieff TM PCR Master mix (Yeasen, China) and adjudged the volume to 25 µL by deionized water. The PCR was conducted using PCR thermal cycle (Applied Biosystems, Thermo Fisher Scientific, USA). The PCR amplification for viral DNA-A detection was conducted for 30 cycles of denaturation at 94 • C for 1 min, annealing at annealing temperature of primers for 1 min and extension at 72 • C for 1.5 min, and followed by a final extension at 72 • C for 10 min. The expected amplicon for DNA-A detection was approximately 1.5 kb in size. The PCR amplification for viral DNA-B detection was the same as for viral DNA-A, with an extension time of 0.5 min, and amplicon was approximately 0.5 kb in size.
Full-length viral DNA was amplified by PCR with abutting primer pair (Table S1). Q5 High-Fidelity DNA Polymerase (New England BioLabs, Ipswich, MA, USA) was used in PCR amplification. The amplified full-length viral DNAs were purified by the Promega Wizard ® SV Gel and PCR Clean-up System (Promega, Madison, WI, USA). The purified viral DNAs were cloned by pGEM ® -T Easy Vector System I (Promega, USA) and sequenced automatically (Genomics, New Taipei City, Taiwan). The related sequences of begomoviral DNAs were also retrieved from GenBank, NCBI for phylogenic analysis (Table S2). Multiple alignment and nucleotide sequence identity of viral full-length sequences were generated using DNASTAR software (DNASTAR, USA). Phylogenic tree was conducted using Molecular Evolutionary Genetics Analysis X (MEGA X) software using the Muscle alignment and the Maximum Likelihood algorithm with 1000 bootstrap replications [57]. Following verification of the begomovirus ORFs, viral DNA sequences were submitted to GenBank, NCBI (National Center for Biotechnology Information).

Specific Detection of Begomoviral DNAs
Based on the sequence analysis results from the Philippines samples and sequences retrieved from GenBank, NCBI (Table S2), primer pairs were designed for the specific detection of SLCuPV and SLCCNV by PCR ( Figure S1 and Table S1). The primer pair-SLCuPV-1-SPAF/-1-SPAC (annealing temperature 47.5 • C) was specific for SLCuPV DNA-A with amplicon in 0.85 kb. The primer pair-SLCCNV-1-SPAF/-1-SPAC (annealing temperature 48.5 • C) was designed for the specific detection of SLCCNV DNA-A with an amplicon approximately in 0.75 kb. The primer pair-SLCCNV-BV1/SLCuPV-2-SPBC (annealing temperature 47 • C) was also used for the specific detection of SLCuPV DNA-B, with the expected amplicon of 0.35 kb in size. For the specific detection of SLCCNV DNA-B, two primer pairs, SLCCNV-BV1/-1-SPBC (annealing temperature 37.5 • C) and SLCCNV-3-SPBV/-3-SPBC (annealing temperature 46 • C) were designed with an amplicon approximately of 1.5 kb and 0.5 kb, respectively.

Construction and Agroinoculation of Begomoviral Infectious DNAs
Based on the sequence analysis results, four SLCuPV isolates and two SLCCNV isolates were selected for development of infectious clones. For the infectious DNA-As of SLCuPV-A[PH-BoG137-18], -A[PH-Pk212- , -B[PH-Pk76-18] and -B[PH-BoG216-18], partial viral DNA-A was released from recombinant plasmids with Sal I and Bam HI digestion (0.23 mer) and inserted into the binary vector pCAMBIA0380 (AF234290) [58]. Consequently, the infectious SLCuPV DNA-A was generated using head-tail ligation of full-length DNA-A, which was released by Bam HI digestion. For the infectious DNA-Bs of SLCuPV-A[PH-BoG137-18] and -B[PH-BoG216-18], partial DNA-B was also released by Sal I and Nco I digestion (0.61 and 0.62 mer, respectively) and then inserted into the pCAMBIA0380. For infectious DNA-Bs of SLCuPV-A[PH-Pk212-  and -B[PH-Pk76-18], partial DNA-B was also released by Eco RI and Nco I digestion (0.58 mer) and inserted into the pCAMBIA0380. The infectious SLCuPV DNA-Bs were constructed by head-tail ligation of full-length DNA-B which was released by Nco I digestion. For infectious DNA-Bs of SLCuPV-A[PH-Pk195-18] and -A[PH-Pk212- , partial DNA-B was released by Bgl II and Hpa I digestion (0.59 mer) and inserted into the pCAMBIA0380. Both infectious DNA-Bs were generated by a head-tail ligation of full-length DNA-B which was released by Hpa I digestion. For infectious DNA-A of SLCCNV-A[PH-Pk195-18], partial viral DNA-A was obtained using Eco RI and Bam HI digestion (0.29 mer) and then cloned into the pCAM-BIA0380. For the infectious DNA-A of SLCCNV-A[PH-Pk212- , partial viral DNA-A was released from recombinant plasmid by Apa I and Bam HI digestion (0.4 mer) and cloned into pCAMBIA0380. The infectious SLCCNV DNA-As were generated by a headtail ligation of full-length DNA-A which was released by Bam HI digestion. All infectious viral DNAs were transformed into Agrobacterium tumefaciens LBA4404. The transformed A. tumefaciens LBA4404 with viral infectious DNA was cultured in YEP broth (10 g peptone, 5 g yeast extract and 5 g sodium chloride per liter) containing 50 µg/mL Streptomycin and