Agrobacterium‐mediated inoculation of asymptomatic Apple latent spherical virus as gene silencing vector in pea (Pisum sativum L.)

Apple latent spherical virus (ALSV) has been developed into a virus‐induced gene silencing vector with a broad host range, including legumes. Using Agrobacterium tumefaciens, delivered via stem injection, we introduced ALSV vectors carrying a phytoene desaturase (PDS) sequence from Phaseolus vulgaris, resulting in highly uniform knockdown phenotypes on Pisum sativum L., 23 days postinoculation. The infection rate ranged from 7.6% to 36.3% on five different P. sativum varieties using Agrobacterium stem inoculation. Mechanical inoculation using infected pea sap improved the infection rates to 80% on P. sativum cultivar AAC Lacombe. Reverse transcription polymerase chain reaction and real‐time reverse transcription polymerase chain reaction demonstrated that ALSV virus RNA accumulated in roots, tendrils and leaves, accompanied by decreased PsPDS (Pisum sativum phytoene desaturase) expression level. ALSV virus RNA was also detected from sepals, petals, carpels, pods, and seed coats but not in embryos. This method simplifies the ALSV inoculation and improves the use of ALSV vectors for functional genomics in P. sativum.


| INTRODUCTION
Virus-induced gene silencing (VIGS) has been utilized to analyze and characterize the functions of known and unknown gene products in plants (Baulcombe, 2004;Liu, Schiff, & Dinesh-Kumar, 2002;Lu, Martin-Hernandez, Peart, Malcuit, & Baulcombe, 2003). However, VIGS has not been broadly employed in legume plants because of the limitations associated with virus inoculation efficiency. In the interest of sustainable agriculture, legumes are increasingly being recognized as desirable crops because of their contribution to human dietary protein and their ability to form nitrogen-fixing symbioses with members of the Rhizobiaceae family. With increasing availability of genomic data for a number of legumes, it is necessary to develop effective approaches to facilitate investigation of gene function to improve plant yield, resistance to pests and diseases, abiotic stress tolerance, and food quality. VIGS could be a simple and efficient approach for high-throughput functional genomics in legume plants because it has several advantages compared to other approaches: VIGS can decrease the expression of one or more genes without time-and labor-intensive tissue culture or generating transgenic plants; VIGS has the potential to generate observations associated with novel phenotypes in a single generation; VIGS can silence a series of genes with functional redundancy by including a highly conserved region of a gene family as a target sequence in the viral vector; and finally, VIGS allows targeting of genes that may not produce plants when knocked out or mutated, such as those that result in embryo lethality (Benedito, Visser, Angenent, & Krens, 2004;Burch-Smith, Anderson, Martin, & Dinesh-Kumar, 2004;Godge, Purkayastha, Dasgupta, & Kumar, 2008;Hettenhausen, Baldwin, & Wu, 2014;Robertson, 2004).
Several VIGS vectors have been used for gene silencing in legume plants. Bean pod mottle virus (BPMV) has been used to silence specific genes in soybean [Glycine max (L.) Merr.], common bean (Phaseolus vulgaris L.) and pea (Pisum sativum L.) successfully (Meziadi et al., 2016;Zhang, Bradshaw, Whitham, & Hill, 2010;Zhang & Ghabrial, 2006;Zhang, Yang, Whitham, & Hill, 2009). BPMV has been variously delivered by particle bombardment, direct DNA leaf rubbing, or mechanical rubbing with infected leaf sap. BPMV has proven a valuable tool for gene silencing as well as for expression of gene sequences (Zhang, Whitham, & Hill, 2013). Pea early browning virus (PEBV) infectious clones have been used for gene silencing in P. sativum, introduced by Agrobacterium-mediated infiltration (Constantin et al., 2004). Cucumber mosaic virus (CMV)-based vectors silenced genes involving in flavonoid biosynthesis in soybean (Nagamatsu et al., 2007). Lim et al. (2016) developed a VIGS vector based on the soybean yellow common mosaic virus for use in soybean.
Apple Latent Spherical Virus (ALSV) based vectors hold considerable promise for legume functional studies. ALSV is composed of two single-stranded RNA genomes (RNA1 and RNA2; Li et al., 2000). A number of ALSV infectious clones have been generated, pEALSR1/ pEALSR2 in pE18PGT with an enhanced CaMV 35S promoter, pCALSR1/pCALSR2-XSB in pCAMBIA1300 vector, and pBICAL1/ pBICAL2 in pBICP35 vector under the control of the CaMV 35S promoter (Kawai et al., 2014;Li, Sasaki, Isogai, & Yoshikawa, 2004). Compared with BPMV, pea early browning virus, and CMV gene silencing vectors, ALSV has several advantages. First, the ALSV VIGS vector does not cause obvious symptoms of infection in most plants tested to date, which may otherwise confound interpretation of any phenotype that arises from silencing the target (Igarashi et al., 2009;Li et al., 2004). Because of this, ALSV may be used as a vaccine to prevent subsequent infection from other viruses such as BPMV, Zucchini mosaic virus, CMV, and Tospoviruses (Satoh et al., 2014;Tamura et al., 2013 (Igarashi et al., 2009;Kawai et al., 2014).
In addition, no insect vector of ALSV transmission has been identified, minimizing the risk of cross-contamination between plants. ALSV has been introduced to legume plants by rub-inoculating after two rounds of Chenopodium quinoa infection (Igarashi et al., 2009). Another procedure involves extracting RNA from N. benthamiana leaves for microparticle bombardment of soybean leaves, which requires specialized equipment (Yamagishi & Yoshikawa, 2009). Concentrated ALSVinfected N. benthamiana sap was used by Satoh et al. (2014) to inoculate pea as a vaccine, however this means an additional step instead of direct inoculation of the target species. Recently, agro-infiltration of N. benthamiana leaves was used prior to inoculating soybean (Gedling et al., 2018).
Inoculation with Agrobacterium tumefaciens constitutes a common method to introduce viruses into plants, allowing it to be broadly and quickly used to identify putative gene functions by VIGS (Vaghchhipawala, Rojas, Senthil-Kumar, & Mysore, 2011). Different agroinoculation approaches have been developed such as using a toothpick with an Agrobacterium colony to directly infect a seedling leaf (Lu et al., 2003), syringe infiltration (Liu et al., 2002), and vacuum infiltration (Ekengren, Liu, Schiff, Dinesh-Kumar, & Martin, 2003). Nevertheless, agro-mediated inoculation to deliver viruses to legume plants has not been applied broadly and remains limited by its relative inefficiency.
Here, we tested several inoculation approaches and developed a simple inoculation method using a syringe needle by injecting Agrobacterium carrying ALSV VIGS vectors into the stem of P. sativum seedling. ALSV virus accumulated in roots, tendrils, leaves and pods, accompanying PsPDS (phytoene desaturase) silencing. No other obvious symptoms were induced by ALSV in P. sativum. This is also the first report of direct agro-mediated stem inoculation by ALSV VIGS vector in legumes species. ALSV vectors can be used for highthroughput gene function analysis and for testing the susceptibility of different genotypes by agro-injection.

| Mechanical inoculation
To test the infection of P. sativum, P. vulgaris, and G. max by ALSV using mechanical inoculation, infected pea, N. benthamiana, or C. quinoa upper leaves were ground in liquid nitrogen with extraction buffer (0.1-M Tris-HCl, pH 7.8, 0.1-M NaCl, 5-mM MgCl 2 ) and inoculated to the adaxial surface of two carborundum dusted primary leaves from 15-to 20-day-old plants.

| Agro-infiltration of infectious ALSV clone in N. benthamiana and legume species
Agroinfiltration of VIGS vectors is popular and extremely efficient in some species but not in legumes. Agro-infiltration of ALSV infectious clones has been carried out on N. benthamiana with 100% efficiency

| Agrobacterium stem injection inoculation with infectious ALSV clone in three legume species
As an alternative to leaf infiltration of Agrobacterium containing ALSV, we assessed direct stem injection. Seven-day-old seedlings Pods and seeds were generally visually smaller with lower yields than uninfected plants, presumably because of the decrease in photosynthesis ( Figure 3e).

| Mechanical inoculation from infected pea sap
Mechanical inoculation was carried out using N. benthamiana sap infected with pCALVS1, pCALVS2-PvPDS, and 35S-P19 on N. benthamiana and C. quinoa. Symptoms on C. quinoa appeared at approximately 14 dpi (Figure 4a,b), and virus was detected by RT-PCR ( Figure 4c) on P. sativum (8 out of 10, 9 out of 10, and 9 out of 11, cv. AAC Lacombe) but 0% on P. vulgaris and G. max inoculated with infected pea sap.

| Virus accumulation and PDS expression level in different organs
Virus could be detected from leaves, tendrils, roots, sepals, petals, carpels, pods, and seed coat by RT-PCR (Figures 5a and 6). Virus

| DISCUSSION
Pisum sativum is an economically important legume crop, and an International Consortium has been formed to tackle sequencing of the pea genome (Alves-Carvalho et al., 2015;Gali et al., 2019;Tayeh et al., 2015). VIGS has the potential to be an extremely useful tool, complementing these efforts. The selection of a virus as a VIGS vector depends on its ability to infect and propagate within and between plant cells (Vaghchhipawala et al., 2011). ALSV has a broad host range and could be used to silence genes in many plant species. However, it has been necessary to use infected leaves of N. benthamiana or quinoa as a source of concentrated ALSV, requiring one or two cycles of propagation (Igarashi et al., 2009;Satoh et al., 2014) or RNA extraction for microparticle bombardment (Yamagishi & Yoshikawa, 2009)  This rate includes silent and lethal mutations. A virus isolate results from selection and a majority of "mutants" are less competitive, which overall makes the viral genome stable. Because of this, secondary inoculations have been used widely in a number of studies for many VIGS vectors. In support of these considerations, Kurth et al. (2012) showed that propagation of a virus in a nonnatural host can result in loss of infectivity in the natural host plant.
Using this approach, it appears that no virus was transmitted to the developing seeds. However, VIGS of a soyPDS gene was maintained in the next generation plants by the seed transmission of ALSV-soyPDS (Yamagishi & Yoshikawa, 2009). The ability of seed transmission might be species-specific. Nevertheless, an alternate protocol, inoculating pea plants closer to the timing of flower development might produce different results. Moreover, the presence of the virus in the seed coat and not in the developing embryo may have interesting applications for studying nutrient partitioning. There is doubtless room to improve the efficacy of the ALSV-pea VIGS system, but we see it currently as a promising tool for pea functional geno- mics. An obvious application is gene inactivation as in the present study. Additional potential uses include genome editing, metabolic engineering, particularly of flavor determinants, and targeting of epigenetic modifications at specific genomic loci Pasin, Menzel, & Daròs, 2019).

ACKNOWLEDGMENTS
We thank Deng-Jing Bing from Agriculture and Agri-Food Canada and Tom Warkentin from the University of Saskatchewan for providing seeds of pea cultivars, and K. Peter Pauls for providing seeds of OAC-Rex. We also thank Alex Molnar for help with figures. This project was supported by the Saskatchewan Ministry of Agriculture, the Canada-Saskatchewan Growing Forward 2 bi-lateral agreement, and the Saskatchewan Pulse Growers (Grant 20130262).

CONFLICT OF INTEREST
The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS
RX, FM, and CT contributed to conceptualization; RX and AP to investigation; NY and AW contributed resources; RX contributed to writing-original draft; and FM and CT contributed to writing-review and editing.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.