The incidence and type of viral disease patterns in field-grown CA and weeds around CA were analyzed. The study sites were in the main cultivation fields of Jeollabuk-do, South Korea; Gochang-gun (9 sites), Jeongeup-si (9 sites), and Imsil-gun (9 sites). The study period was June-August in 5 years from 2014-2018. The investigation was divided into the early growth and post-harvest phases in compliance with the Application Guidelines for Crop Pest Monitoring and Control (Rural Development Administration, 2014). The samples were collected mainly from leaves (three leaves per plant) suspected of viral infection through visual examination of the symptoms of the disease in CA and weeds around CA. The number of plants during sample collection was not constant due to the differences in viral disease incidence across different fields. The total number of collected samples over the 5 years was 821 for CA and 918 for weeds around CA.

The diagnosis of viral disease was based on reverse transcription polymerase chain reaction (RT-PCR) analysis. First, the collected sample was ground using liquid nitrogen, and the total RNA was extracted using the PowerPrep Viral DNA/RNA Extraction Kit (Kogenebiotech, Seoul, Korea). Each RNA was screened using the primers specific to the 11 viruses (Table 1). For nucleic acid amplification, the TOPscript One-step RT PCR DryMIX kit (Enzynomics, Daejeon, Korea) was used, and the reaction condition was as follows: 50°C, 30 min; 95°C, 10 min; <95°C, 30 s; 50-55°C, 30 s; 72°C, 1 min >35 cycles; 72°C, 5 min. Each PCR product was checked for infection through electrophoresis in 1.2% agarose. Electrophoresis included the 100 bp DNA Ladder (Invitrogen, Carlsbad, CA, USA) to identify the size of each PCR product. The WSE-5300 Printgraph CMOS I (Atto, Tokyo, Japan) was used for measurement.

To analyze the incidence of viral infection in CA and weeds, the infection percent (%) was calculated using the formula in Odum (1971) (Eq. 1). The infection percent was estimated for 821 CA and 918 weed samples collected during 5 years (2014-2018) by checking the state of infection via RT-PCR analysis.

The correlation of viral infection between CA and weeds around CA was analyzed. Based on the viral infection frequency of the study period (5 years), we analyzed six viruses out of eight that simultaneously infected CA and weeds, excluding TMGMV and PMMoV, which are known to be transmitted by seeds. Next, the correlation of annual viral infection was analyzed for CMV and TSWV, whose continuous infection in CA and weeds was confirmed for the study period based on the annual infection frequency of CMV and TSWV in CA and weeds. To analyze the correlation, Spearman’s rank correlation coefficient ρ (rho) was used with the level of significance set at α = 0.05. The IBM SPSS Statistics version 21.0 (IBM Corp., Armonk, NY, USA) was used for statistical analyses.

Data on climate conditions were used to analyze the cause of the large annual variation in the incidence of viral disease in CA. As viral infection is closely associated with the overwintering of insect vectors, the mean, minimum, and maximum temperatures of the overwintering period (from December to February of the following year) during 2014-2018 were analyzed.

Each weed species was identified to analyze the influence of weeds as the primary infection source of the viral disease in CA. The collected weeds were listed and annotated based on the National Biological Species Information System (http://www.nature.go.kr) and the Illustrated Book of Weeds (National Institute of Agricultural Sciences, 2017). The identified weeds were divided based on life cycles into summer annuals, winter annuals, and perennials, and the respective infection percents were analyzed.

To investigate the incidence of viral infection in CA across the fields in Jeollabuk-do, a total of 821 CA samples were collected from Imsil-gun, Jeongeup-si, and Gochang-gun during the 5 years from 2014 to 2018. In the RT-PCR analysis of 11 viruses, it was found 546 CA plants (66.5%) were infected by nine viruses (Table 2).

The highest infection percent among all study sites was observed for BBWV2 and PMMoV at 35.9% and 31.3%, respectively, while CMV also showed a high percent of 26.9%. The infection percents for TSWV and PepMoV were 20.0% and 19.9%, respectively. BWYV had an infection percent of 6.6%, and PVY had 1.9%. TMGMV had 0.6%, and ToCV had 0.4%. No infection by TMV or AMV was detected during the study period.

The results indicated infection by nine viruses, seven of which were previously reported by Kwon et al. (2018): CMV, BBWV2, TSWV, BWYV, PVY, PepMoV, and PMMoV. These viruses were detected during 2015-2016 in Jeollabuk-do (Iksan-si, Imsil-gun, Gochang-gun, and Wanju-si). Additionally, six of the viruses were reported by Choi et al. (2005): PMMoV, TMGMV, PepMoV, BBWV2, TSWV, and CMV. These viruses were detected during 2001-2004 in various regions of South Korea. Furthermore, one virus, ToCV, was reported by Kil et al. (2015), and its first incidence in tomatoes in South Korea was detected in 2013.

The incidence of BBWV2, PMMoV, CMV, and TSWV was detected annually. Among these, BBWV2, PMMoV, and CMV showed a high level of incidence, with 295, 257, and 221 cases, respectively. On the other hand, the incidence of TSWV was relatively low, with n = 164. However, the incidence of PepMoV, BWYV, PVY, ToCV, and TMGMV was not detected annually (Table 2). This finding aligns with the study conducted by Kim et al. (2012) over a period of 5 years from 2007 to 2011, where the incidence of PepMoV, PVY, TMGMV, and other viruses was not detected annually. Therefore, the annual infection frequency of these five viruses, namely PepMoV, BWYV, PVY, ToCV, and TMGMV, was low in CA cultivation fields.

To study the incidence of viral infection in weeds around CA fields, 918 samples were collected from Imsil-gun, Jeongeup-si, and Gochang-gun during the 5 years from 2014-2018. In the RT-PCR analysis of 11 viruses, 190 weeds (20.7%) were infected by eight viruses (Table 3).

The highest viral infection percent across all study sites was shown by PMMoV and CMV at 10.9% and 7.2%, respectively, and the infection percent of BBWV2 was 6.2%. For TSWV, PepMoV, BWYV, ToCV, and TMGMV, the infection percent was 3.9%, 0.7%, 0.2%, 0.2%, and 0.1%, respectively, and no infection by TMV, PVY, or AMV was detected during the study period. The infection frequency was the highest at 100 cases for PMMoV, although no infection was detected in 2017. For CMV with a broad scope of hosts, the incidence during the 5 years was n = 66, with the infection detected annually. For TSWV, the infection frequency was n = 36 for 4 years, excluding of 2014. For PepMoV, BWYV, ToCV, and TMGMV, the infection frequency was ≤6 cases (Table 3).

The incidence pattern in weeds showed that, while the infection percent was low compared to CA, the viruses PMMoV, CMV, BBWV2, and TSWV mainly infected the weeds around CA, as with CA itself. This implied that weeds and wild plants frequently functioned as viral habitats to affect the viral diseases in crops of nearby fields, which agrees with the study by Hasiów-Jaroszewska et al. (2021) reported that weeds and crops can spread viral diseases to one another.

A correlation analysis was performed on the mutual influence of viruses infecting CA and weeds around CA within the agricultural ecosystem during the study period (2014-2018) (Table 4). Except for ToCV and BWYV, which displayed low infection frequency, significant positive correlations were shown by four viruses between CA and weeds. The correlation was the highest between weed-infecting CMV and CA-infecting CMV at ρ = 0.32**, followed by weed-infecting BBWV2 and CA-infecting BBWV2 at ρ = 0.26**; weed-infecting PepMoV and CA-infecting PepMoV at ρ = 0.17**; weed-infecting TSWV and CA-infecting TSWV at ρ = 0.16**. The results indicated that the main CA-infecting viruses could have been influenced by the infection of weeds around CA, thereby implicating the potential spread of viral infection. Thus, it is conjectured that the weeds around CA are the intermediate host and a potent infection source that influence the spread and diversity of the virus (Rist and Lorbeer, 1989; Toyoda et al., 2004).

CMV and TSWV were the viruses with high correlation coefficients and continuous CA and weed infections during the study period (2014-2018), and hence, the potential cause of their spreading was analyzed. First, we determined whether the weeds infected by CMV affected the incidence of CMV in CA in the following year after winter (from December to February of the following year). The CMV-infected weeds in the previous year were positively correlated with the CMV infection frequency in CA in the current year. The correlation was highly significant during the 5 years of the study period (Table 5). A similar trend was found for TSWV, although the level of correlation was lower (Table 6).

The cause of the sudden fall in correlation coefficients in 2018 for both CMV and TSWV was analyzed in connection with the winter climate conditions. In 2018, the maximum temperature was 4.5°C, which was lower by 1.7°C than the average maximum temperature during the 5 years of the study period at 6.2°C. The mean annual temperature was also the lowest at −0.55°C. The average minimum temperature during the 5 years was −3.12°C, and that in 2018 was −4.75°C, which was lower by −1.63°C to indicate a low winter temperature range (data not shown).

In a study analyzing the correlation between winter temperatures and overwintering of Diuraphis noxia in southern parts of Alberta, Canada, the aphid population steadily decreased at temperatures from 0 to −10°C on the soil surface. In another study, the mortality of female imago of Tetranychus urticae in apple orchards was 72-80% on average during overwintering. Considering these findings, the fall of viral infection percent in 2017-2018 was determined to be due to the increased mortality of insect vectors at low winter temperatures (Butts, 1992; Lee et al., 2015b). In contrast, the average temperature during the overwintering period in 2014-2017 was 1.2-2.3°C. This is a higher range than in 2018 at −0.55°C, with a highly significant correlation with the incidence of CA viral infection in the following year. Based on this result, and according to the study by Szostek and Schwartz (2015) which reports that Thrips tabaci, an insect vector of Iris yellow spot tospovirus, ceased to be active at temperatures <0°C then resumed its activity at temperatures ≥0°C to be the infection source of the weeds in the vicinity; Lactuca serriola and Descurainia sophia, in the following crop year, it was determined that the high average temperature in 2014-2017 would have affected the activities of insect vectors and the infections of weeds in the vicinity to affect the incidence of CA viral infection in the following crop year.

Through RT-PCR on the 918 weed samples collected during 5 years, 66 species infected by CMV and 36 species infected by TSWV were detected (Table 3). As shown in Tables 5 and 6, weeds were correlated with the incidence of viral infection as the primary infection source; hence, the pattern of viral spreading was analyzed based on the weed life cycles and infection percents (Table 7). The infection percent for CMV was the highest at 14.0% in summer annual weeds, followed by 11.4% in perennial weeds, and the lowest at 7.8% in winter annual weeds. The high infection percent in summer annual weeds could be attributed to the time of sample collection being the summer season (June and August). Still, considering the overwintering period in combining the infection percents of winter annual and perennial weeds, the percent was 5.2% higher than that of summer annual weeds, which implicated that these weeds were the main habitats of insect vectors during the period without CA. A similar trend was found for TSWV with the infection percent in the following decreasing order: summer annual weeds 10.4%, perennial weeds 6.4%, and winter annual weeds 6.2%. This deviated from a previous study analyzing the life cycles of the TSWV weed hosts that reported winter annual weeds at 42.9%, summer annual weeds at 30.9%, and perennial weeds at 26.2%. However, it may be due to the location and time of sample collection (Kil et al., 2020).

As with CMV, the infection percent for TSWV in winter annual and perennial weeds was higher by 2.2% than in summer annual weeds. Moreover, previous studies have reported that such winter annual and perennial weeds as Capsella bursa-pastoris are the key host plants as the habitat of aphids in the absence of crop plants. The population density of Frankliniella occidentalis female imago is higher in biennial strawberry flowers than in annual flowers due to the overwintering in winter annual and perennial weeds such as Stellaria media, Senecio vulgaris, and Taraxacum officinale. Based on our findings and the aforementioned studies, it is conjectured that winter annual and perennial weeds play a crucial role in the overwintering of CMV and TSWV (Sampson et al., 2021; Satar et al., 2021).

To determine the life cycles of infected weeds as well as the preferred weed species as host plants and the respective scope, each main infected plant was analyzed regarding species (data not shown). As a result, 46 families and 155 species were identified. The Asteraceae family demonstrated the highest population at n = 282, followed by the Fabaceae family at n = 92, the Polygonaceae family at n = 49, and the Gramineae family at n = 35. This agreed with the report by Lee et al. (2015a) that the distribution of weed species in the crop fields of South Korea was Asteraceae 19.5%, Gramineae 11.7%, Polygonaceae 6.7%, Fabaceae 5.3%, and Lamiaceae 4.3%. However, the pattern of infection percent by plant species varied from the distribution of weeds. For summer annual weeds, the highest CMV infection percent of 57.1% was shown by the Convolvulaceae family including Ipomoea triloba and Quamcolit coccinea, followed by 50.0% in Perilla frutescens of the Lamiaceae family and Fallopia dumetorum of the Polygonaceae family, and 22.2% in the Asteraceae family including Bidens frondosa and Xanthium canadense (Table 7). For winter annual weeds, viral infection was detected solely in the Asteraceae family, including Erigeron annuus and Crepidiastrum sonchifolium. For perennial weeds, the infection percent was 100% in Stellaria media of the Caryophyllaceae family and Torilis scabra of the Umbelliferae family and 50% in the Convolvulaceae family, including Calystegia dahurica and Calystegia sepium. The Convolvulaceae family belongs to the order Solanales as with the Solanaceae family. Considering the report by Hobbs et al. (2000) that the weeds of the Solanaceae family are the crucial CMV infection source, it is likely that a correlation exists with the phylogenetic classification of host plants. In addition, the highest TSWV infection percent was shown by Mosla dianthera of the Lamiaceae family among summer annual weeds, followed by 50.0% in the Asteraceae family, including Bidens tripartita, 42.9% in the Fabaceae family including Glycine max, and 16.7% in the Gramineae family including Digitaria ciliaris. D. ciliaris, in particular, is the most dominant species in crop fields; hence, it is likely to be the primary infection source of TSWV. Only the weeds of the Asteraceae family were detected among winter annual weeds. Among perennial weeds, the infection percent was 100% in Trichosanthes kirilowii of the Cucurbitaceae family, followed by 50.0% in Leonurus japonicus of the Lamiaceae family. Meanwhile, Kil et al. (2020) found that the infection percent was high in Eclipta prostrata (95.6%) among summer annual weeds, Stellaria media (55.0%) among winter annual weeds, and Stellaria aquatica (54.5%) among perennial weeds.

The results indicated that 34 species of weeds were infected by CMV and 25 species of weeds were infected by TSWV in the fields close to the CA cultivation fields. The list of natural hosts updated in this study could prove valuable in the control of CMV and TSWV. Various species of the Asteraceae family were found to have been infected. However, the infection percent was low due to the large population of plants, and this should be investigated in further studies.

Our findings highlight the importance of ambient weeds as potent, infectious agents influencing the spread and diversity of viruses that infect CA. In other words, by revealing that surrounding weeds are a significant source of transmission of viruses infecting CA, we highlight the importance of weed management in the surrounding environment of CA fields. It also provides essential information that complements and extends knowledge of CA viral disease control and prevention strategy development. Findings from this study may help advance our understanding of the management and prevention of CA viral disease and help prevent the spread of CA viral disease by: examples include (1) preventing the spread of CA viral diseases such as TSWV and CMV by removing weeds or taking measures to prevent weeds from becoming infected with viruses, and (2) developing biological control methods for CA viral diseases.