Genetic-based resistance to viruses is the most acceptable and durable form of crop protection, when it is available. This is particularly important for vegetable crops. This requires genetic resources for resistance (or tolerance), already available in some cultivars of various crops, or in wild relatives, which then can be crossed with the crop species. In a survey of the most important viruses infecting crops in the field, involving 28 countries and regions, turnip mosaic virus (TuMV) ranked as the second most important virus, after cucumber mosaic virus (CMV) (Tomlinson, 1987). The host range of TuMV includes at least 318 species in 156 genera, from 43 eudicotyledon families, including Brassicaceae (formerly Cruciferae), Asteraceae (formerly Compositae), Amaranthaceae (now including the former family Chenopodiaceae), Fabaceae (formerly Leguminosae), and Caryophyllaceae (Edwardson and Christie, 1991). In addition, some isolates can also infect species in at least seven monocotyledon families: Amaryllidaceae, Araceae, Commelinaceae, Iridaceae, Liliaceae, Musaceae, and Orchidaceae. TuMV can be transmitted by at least 89 species of aphid (Edwardson and Christie, 1986). Infection by TuMV is especially damaging to brassicid crops in parts of Europe, Asia and North America, with yield losses of up to 70% (Li et al., 2019a). These crops include oilseed rape (aka canola; Brassica napus), swede (aka rutabaga; B. napus ssp. napobrassica; formerly B. napobrassica; aka B. napus ssp. rapifera), turnip (B. rapa ssp. rapa; formerly B. rapa ssp. campestris; aka B. rapa ssp. rapifera), Chinese cabbage (B. rapa ssp. pekinensis; formerly B. pekinensis; aka B. campestris ssp. pekinensis), mizuna (B. rapa ssp. nipposinica), pak choy (aka bok choy, aka pok choy; B. rapa ssp. chinensis; formerly B. chinensis), choy sum (B. rapa ssp. parachinensis; formerly B. parachinensis), broccoli (B. oleracea ssp. italica), Brussels sprout (B. oleracea ssp. gemmifera), cabbage (B. oleracea ssp. capita), cauliflower (B. oleracea ssp. botrytis), collard (B. oleracea var. acephala), kale (B. oleracea var. acephala), kohlrabi (B. oleracea ssp. gongylodes), Indian mustard (B. juncea; formerly B. japonica), Ethiopian mustard (B. carinata), black mustard (B. nigra), and radish (Raphanus raphanistrum ssp. sativus; formerly R. sativus). For many years, TuMV has been considered the most important virus of cultivated brassicid crops in Asia, causing losses in cash crops such as Chinese cabbage, radish and Indian mustard (Green and Deng, 1985; Sako, 1981), among the other crops mentioned above.

Most of the work on genetic resistance to TuMV has been done in brassicid crops; largely in species of the genus Brassica (brassicas), with only a few studies involving radish. These resistance genes are described below. Many have been mapped, but few have been isolated and characterized. In addition, work on genetic resistance has been done in the brassicid species Arabidopsis thaliana, which is not a crop, but whose genome shows synteny with those of brassicas and radish (Kitashiba et al., 2014; Yu et al., 2017a, 2017b). Thus, with its smaller genome and tools available for gene knockout, as well as other genes identified that inhibit TuMV infection, Arabidopsis serves as a potential source of candidates for homologous genes in other brassicids that can be targeted to obtain resistance to TuMV. Finally, the use of non-conventional resistance will also be considered, since it has been applied in resistance to TuMV.

In describing the systems for differentiating TuMV isolates, based on responses in different brassica species, in particular subspecies of B. napus and B. rapa, as well as characterization of resistance genes from particular host species (see below), generally the English names for these crops will be used here. This is because the specific epithets and subspecies (or variant, or group) names used over the years covered by these activities have changed and the names used in the studies often are not the same names that are used now for those taxa, leading to confusion. Exceptions to this include the analysis of isolates differentiated into their ability to infect local specific cultivars of various brassica species, for which neither English names nor subspecies names were given (see below). There was also a tendency in some publications to use the colloquial term ‘brassicas’ to refer not just to species in the genus Brassica (as used here), but also to include radish, which is not in the genus Brassica, but is a separate genus (Raphanus) in the family Brassicaceae; describing work involving the inclusion of radish with ‘brassica plants’ will be referred to here as ‘brassicids’, meaning members of the family Brassicaceae.

The TuMV genome consists of one single-stranded RNA molecule of (+) polarity and ~9,830 nt (excluding the poly(A) tail); isolates can differ by several nucleotides. The genome organization of TuMV is depicted in Fig. 1. The viral RNA contains a single gene that is translated from an AUG starting 130 nt from the 5ʹ end of the RNA to produce a single polyprotein of ~300 kDa, which ends at nucleotide 9,618, 212 nt before the poly(A) sequence at the 3ʹ end of the RNA (Nicolas and Laliberté, 1992). This polyprotein is processed into 10 mature proteins designated P1, HC-Pro, P3, 6K1, CI, 6K2, VPg, NIa, NIb, and CP (coat protein). In addition, as the result of a transcriptional slippage (in the -1 frame, about ~2% of the time) within in the P3 encoding sequence, during the synthesis of the viral RNA, translation leads to early termination of the concomitant frame-shifted protein, designated P3N-PIPO (Chung et al., 2008; Olspert et al., 2015; Rodamilans et al., 2015). A further truncated protein, produced by +1 transcriptional slippage in the P3 encoding sequence, designated P3N-ALT, has been identified (Hagiwara-Komoda et al., 2016). The functions of these viral proteins have been reviewed previously (Revers and García, 2015; Walsh and Jenner, 2002) and only their names are shown here (Fig. 1), because resistance breakage to infection by TuMV in some crops has been mapped to some of these proteins, as mentioned below. Since these proteins are all derived from one gene, they will be described individually as ‘coding sequences’ rather than as individual genes.

TuMV isolates have been differentiated and classified into strains based on a number of criteria. Initially, this was by symptoms induced in a range of inoculated plants (Pound, 1948; Pound and Walker, 1945). Later, isolates were differentiated by reactions on specific hosts, such as the two groups of strains observed by Yoshii (1963): one (the ‘cabbage strain group’) showing severe necrotic ringspots on cabbage plus severe mosaic on Nicotiana glutinosa, and the other (the ‘ordinary strain group’) producing only mild symptoms on both hosts. Provvidenti (1980) examined various strains using two cultivars and two accessions of Chinese cabbage, and differentiated four groups, designated C1-C4, based on types of symptoms, immunity from infection (i.e., no local or systemic infection), or resistance (i.e., local effects but no systemic infection). A survey of TuMV strains from brassica hosts in Taiwan using various indicator hosts, as well as the same Chinese cabbage indicators, identified the same responses in Chinese cabbage, as well as a fifth response, designated C5 (Green and Deng, 1985). Later, a survey of TuMV isolates from various crops in Ontario, Canada, and their differentiation on Chinese cabbage and rutabaga, identified another Chinese cabbage differential group (Stobbs and Shattuck, 1989), designated S5 by the authors and C6 by others. Two additional groups, later designated C7 and C8 by others, were identified from a survey of brassica vegetables in China (Liu et al., 1990), who identified seven groups, Tu1-Tu7. Using a different approach, involving an oilseed rape line immune to infection by the TuMV UK1 isolate, as well as a susceptible oilseed rape line plus a swede line resistant to TuMV UK1, Walsh (1989) was able to distinguish four groups of TuMV isolates, based on their restricted infection or resistance-breaking properties. Subsequently, this was extended to examine 124 isolates of TuMV from around the world, although principally from Europe, for their responses on four differential lines (two oilseed rape and two swede), resulting in 12 distinct pathotype groups (Jenner and Walsh, 1996). These various studies also examined the effects of location of isolation and the classification as to whether such a correlation existed. There was often a tendency for various countries to contain mostly members of one or two pathotypes, but there were many exceptions to such restrictions. The groupings determined by reactions on B. napus (Jenner and Walsh, 1996) vs. B. rapa (Green and Deng, 1985; Provvidenti, 1980) or various brassica differentials (Liu et al., 1990) cannot readily be compared. As pointed out by Jenner and Walsh (1996), B. napus did not differentiate isolates CHN2-CHN5 (all in pathotype 3), which were differentiated on Chinese cabbage (Green and Deng, 1985) into strains C2-C5, respectively, with isolate CHN1 (strain C5) in pathotype 1, while isolates CHN6-CHN12 corresponding to strains Tu1-Tu7 (Liu et al., 1990), respectively, were only differentiated in B. napas as isolates CHN6-CHN10 plus CHN12 (all in pathotype 3) and CHN11 (in pathotype 1). On the other hand, Walsh et al. (2002) pointed out that strain C1 and pathotype 1 showed genetically identical interactions with the B. rapa and B. napus differential lines, indicating that strain C1 and pathotype 1 are identical.

The first comparison of TuMV CP coding sequences involved 11 isolates and a mutant (UK1M). The phylogenetic analysis showed three distinct groups that generally correlated with geographical and/or the host plant origin of the isolates (Lehmann et al., 1997). With the increase, initially in partial and later in complete, sequence determination of TuMV isolates, other more recent approaches used have involved sequence analysis of the CP coding region (Farzadfar and Pourrahim, 2014; Parmar et al., 2017; Sánchez et al., 2003), the CP and P1 regions (Ohshima et al., 2002), the NIb and CP regions (Korkmaz et al., 2020), or using either most or all of the entire genome (Farzadfar et al., 2009; Gong et al., 2019; Hu et al., 2019; Korkmaz et al., 2008; Tomimura et al., 2003, 2004; Tomitaka and Ohshima, 2006; Wang et al., 2009a; Yasaka et al., 2017), to examine either variation in sequence and infection of particular brassica species or specific changes that affect infection of differential hosts.

In a study involving eight laboratories and 76 isolates of TuMV from around the world, Ohshima et al. (2002) assessed the molecular evolution of TuMV by examining the sequence of the P1 and CP coding regions, and comparing the phylogenetic data for evidence of genetic recombination and correlation with adaptation to specific hosts and geographical distribution. A phylogenetic grouping by host type and geographical origin had been observed previously by Lehmann et al. (1997) using much fewer isolates. The isolates used by Oshima et al. (2002) came mostly from Brassica and Raphanus crops, but also species in other families. Two pathotypes (referred to further on in this section as ‘host types’, since this term is used in most later publications, and is equivalent to ‘strains’) were discerned from the host specificity tests of the various isolates: infecting either brassicas (B) or both brassicas and radish (BR), based on infection in four cultivars of three brassica species and two cultivars of radish. Thus, radish was only susceptible to the BR-host types and not the B-host types; no isolates could infect only radish, while non-brassicid plants were susceptible to both host types. From the sequence analyses, four groups were discerned: (1) a cluster of eight B-host types, which showed the most variation in sequence and were found in both brassica and non-brassica hosts, in Southwest and Central Eurasia, designated the ‘basal-B’ group; (2) a cluster of eight BR-host type isolates from Eurasia, which showed a similar degree of variability and found next to the basal-B group in the phylogeny analyses, designated the ‘basal-BR’ group; (3) a cluster of 22 BRhost type isolates, all found in brassicids, mostly radish plants, from East Asia, largely from Japan, together showing the least variability in sequence, designated the ‘Asian-BR’ group; and (4) a cluster of 36 isolates, mostly of the B-host type (five isolates from Japan were not B-host types) and mostly obtained from brassicas coming from all continents, designated the ‘world-B’ group. The ‘world-B’ group comprised three subgroups: isolates mostly from the ‘New World’ (Brazil, Canada and New Zealand); isolates from all over the world (Africa, Australasia, Europe, and the USA); and isolates from Russia and the Czech Republic. From this work, the authors concluded that like its brassica hosts, TuMV originated in Europe and spread to all parts of the world, while the BR-host type evolved in East Asia. About 10% of the isolates showed evidence of recombination between the P1 and CP coding regions (Ohshima et al., 2002).

In a different study, Sánchez et al. (2003) determined TuMV strain relatedness among 60 isolates by analysis of the viral CP coding sequences, using sequencing and typing based on restriction enzymes. This involved 30 isolates characterized experimentally and 30 isolates from the database characterized in silico. They found four groups, comprising two major clusters and two minor clusters. These consisted of (1) mostly Brassica isolates (MB cluster); (2) mostly radish isolates (MR cluster); (3) intermediates between Brassica and radish clusters (IBR); and (4) isolates outside Brassica or radish clusters (OBR). The MB cluster included isolates from Asia, Europe and the ‘New World’, while the MR cluster contained both European isolates and Asian isolates, although sequences were only obtained from the Asian isolates. The IBR cluster contained isolates from Asia, Europe, and one from Africa, while the OBR cluster contained isolates from Australia, Europe, the ‘New World’, but not Asia. Sánchez et al. (2003) also suggested that the OBR cluster might represent other genetic lineages not considered by Ohshima et al. (2002), and that the basal-B group of Ohshima et al. (2002) might also represent several genetic lineages.

In a subsequent study (Tomimura et al., 2003), the (near) complete sequences of 32 TuMV isolates were determined and compared to six other complete TuMV sequences to further the analysis of relationships, the extent of recombination occurring and to better understand the origin of this virus, the role of recombination in sequence variation and the emergence of new strains. Eight of these isolates showed signs of single [two isolates] or multiple [six isolates] recombination events having occurred in their genomes, between various combinations of world-B, Asian-BR and basal-BR isolates; there were no recombinants found involving basal-B isolates, and all recombinants came from East Asia, predominantly from Japan. The other 30 isolates did not show recombination, but showed variation in sequence such that there was no correlation at any nucleotide or encoded amino acid position with the grouping into B- or BR-host types. [A third host range type was identified as well in this study, designated (B)-host type, which caused a latent infection only in ‘B. rapa’]. The basal-BR and world-BR isolates (probably deriving from Eurasian-B populations), showed the least differences in sequence, while the basal-B isolates (coming mostly from Europe) differed most from each other, up to 18% in sequence, and might represent the oldest population of TuMV isolates. The world-B isolates were also quite diverse, with three subgroups, as indicated previously (Ohshima et al., 2002), but defined as found in East Europe, the New World, and worldwide (Tomimura et al., 2003). Thus, a lineage from the Eurasia ‘B-strain’ population, represented now by the basal-B and world-B groups, was considered to be the origin of TuMV.

A further analysis involving more isolates from both East and West Eurasia (Tomimura et al., 2004), by both sequences analysis (phylogeny and recombination) and host range analysis involving infection of Chinese cabbage, ‘B. rapa’ and radish, confirmed the previously groupings, with West Eurasian isolates in the basal-B, world-B, and basal-BR groups, and East Eurasian isolates in the basal-BR, Asian-BR and the Asian lineage of the world-B group. Further subgroups of basal-B were found in Mediterranean countries, and some Asian world-B group members were either BR- or ‘B(R)’-host types, where ‘B(R)’ represents isolates that infected radish occasionally and latently. From examining the sequences of six coding regions, the West Eurasian isolates were more diverse than the East Eurasian ones, while the diversity estimation between the phylogenetic groups showed that the Asian-BR group was the least diverse and the basal-B was the most diverse. This study also showed that twice as many of the East Eurasian isolates (80%) were recombinants, compared with the West Eurasian isolates (42%), and more of the BR-host type isolates (74%) were recombinants than of the B-host type isolates (54%). Their analysis concluded ‘that TuMV in East Eurasia is a more recently emerged population than that in West Eurasia’ (Tomimura et al., 2004).

Subsequently, Tomitaka and Ohshima (2006) examined 118 isolates of TuMV from East Asia, utilizing similar methodology, as well as additional methods. They found that 26 isolates from Brassica spp. in China and Taiwan were largely of the B-host type [19], with the rest of either B(R)-[4] or BR-host type [3], while 12 isolates from radish were all of the BR-host type. By contrast, 17 isolates from brassica in Japan were largely of the BR-host type [14], with only three isolates of the B-host type, while 55 isolates from radish were all of the BR-host type. In the phylogeny analysis, none of the East Asian isolates were in the basal-B group. The Chinese isolates were in both the Asian-BR group and the world-B group, whereas the Japanese isolates were in the Asian-BR, basal-BR, and world-B groups. The recombination analysis showed that 38% (45 of 118 isolates) of the East Asian isolates were clear recombinants, with 17 recombination patterns, and the [11] Japanese isolates had a greater number of recombination patterns than the [4] Chinese isolates. (Two additional patterns shared by some Chinese and Japanese isolates were considered possible ‘false positives’.) The recombination patterns did not show any additional phylogeographical information. Overall, Chinese and Japanese isolates were part of the same population, but formed discrete lineages (Tomitaka and Ohshima, 2006). Later, two basal-BR isolates were identified in China, both being recombinants, either between basal-BR and Asian-BR parents, or between two basal-BR parents (Wang et al., 2009a).

Other studies have examined similar sequence variations associated with local host types and various groupings from isolates in Asia Minor (or most of Turkey) (Korkmaz et al., 2008, 2020), West Asia (Iran) (Farzadfar and Pourrahim, 2014; Farzadfar et al., 2009), South Asia (India) (from radish; Parmar et al., 2017), Vietnam (Nguyen et al., 2013b), Korea (from radish; Gong et al., 2019), and Australia (Nyalugwe et al., 2015b). The Turkish isolates fell into the basal-B, world-B, and Asian-BR groups; the Iranian isolates were in the basal-B and Asian-BR groups; the Indian (radish) isolates were in the basal-BR group; the Vietnamese isolates were in the world-B group; the Korean (radish) isolates were in the basal-BR group; and the Australian isolates were in the basal-B and world-B groups.

From two studies involving the examination of sequences of TuMV isolates from either the entire world (Nguyen et al., 2013a) or world isolates compared to largely isolates from Europe and Western Asia (Yasaka et al., 2017), coupled with Bayesian coalescence analyses, the authors concluded that TuMV arose about 1000 years ago from a European orchid-infecting TuMV-like potyvirus, to become the brassica-infecting TuMV. The orchidinfecting virus could not infect brassica plants initially and represents a sister lineage of the brassica-infecting TuMV isolates. These brassica-infecting isolates diverged into the four lineages of the current global populations of TuMV (basal-B, world-B, basal-BR, and Asian-BR) by 850 years ago in Western Europe. Inter-continental trade by sea, from 500 years ago, facilitated the spread of TuMV to the rest of the world (Nguyen et al., 2013a). Further work indicated that this change in host from orchids to brassicas may have occurred in Germany, and spread subsequently to Asia Minor via Southern Europe about 700 years ago, coinciding with the expansion of the Ottoman Empire and the spread of new agricultural products in the Old World. The basal-B group in Asia Minor is older than other populations, probably originating from the early European brassica-infecting TuMV isolates. The world-B group seems to have originated soon thereafter, as this was found in Greece, Turkey, and Iran. The Asian-BR group may have arisen in either Turkey or Iran and then moved to the other location, as well as into South Asia, all areas that grow radishes. The basal-BR group may have originated in Italy (Yasaka et al., 2017).

The relationship between pathotypes (largely a measure of the extent of infection and types of symptoms) and the phylogenic groupings was examined only in one early study, in which the CP encoding sequences of 60 isolates of TuMV were compared (Sánchez et al., 2003), where some of the isolates had previously been among those subjected to pathotyping based on the response in four lines of B. napus (Jenner and Walsh, 1996). No correlation was found between the phylogenic group designation and the pathotype (Sánchez et al., 2003). However, as described above, the phylogenetic groups had different names and biological relationships (MB, IBR, MR, and OBR) and thus are not directly comparable with data obtained in other phylogenic studies (grouped into basal-B, basal-BR, Asian-BR, and world-B) involving much larger numbers of isolates, especially from East Asia. The latter studies, some of which examined the entire genomes, assessed the ability of the isolates to infect selected brassica and radish lines, but not the B. napus lines used by Jenner and Walsh (1996) for their pathotyping. An examination of those isolates pathotyped by Jenner and Walsh (1996), which also were subjected to phylogenic analysis by one or more of three studies (Ohshima et al., 2002; Tomimura et al., 2004; Wang et al., 2009a) using the more common ‘basal-B/basal-BR/Asian-BR/world-B’ classification system, is shown in Table 1, along with the results obtained for the same isolates examined by Sánchez et al. (2003). While some isolates with common phylogenic groupings also are in the same pathotypes, especially if they came from the same country (those from the Netherlands and Poland, and for most of those from China or Greece), in other cases, isolates from the same countries can be found in one (or sometimes more) phylogenetic groupings, but in multiple pathotypes. This indicates that there is no general correlation between the pathotype status and the extent of sequence conservation, where the latter allows groupings correlating with infection of brassica and/or radish, in particular in East Asia.

Genetic resistance to TuMV has been used in crops since the early 1970’s (Zink and Duffus, 1970). However, the first resistance gene mapped for resistance to TuMV was the dominant Tu gene in lettuce (Lactuca sativa). The Tu gene was shown to be linked to the Dm5/8 gene for resistance to downy mildew (Bremia lactucae) (Zink and Duffus, 1970), and further mapping by bulk segregant analysis and rapid amplified polymorphic DNA (referred to as RAPD) analysis identified markers close to both the Dm5/8 and Tu loci (Robbins et al., 1994). Subsequent analysis with larger F2 populations identified more markers, including flanking markers 0.8 and 1.5 centiMorgans (cM) from the Tu locus (Montesclaros et al., 1997). However, the Tu gene was not isolated and characterized.

Different Brassica spp. are composed of one or more of three genomes, designated A, B, and C, comprising the diploids B. rapa (AA genome, 2n = 20), B. nigra (BB genome, 2n = 16), B. oleracea (CC genome, 2n = 18), and the amphidiploids, B. juncea (AABB genome, 2n = 36), B. napus (AACC genome, 2n = 38), and B. carinata (BBCC genome, 2n = 34) (Li et al., 2019a). Most TuMV resistance genes found in brassicas have been found in the A genome of B. rapa (Chinese cabbage), but some resistance genes have been found in the A or C genomes of B. napus, and one has been found in a non-specified genome of B. juncea (Table 2). Most TuMV resistance genes characterized have been dominant in nature, although some recessive resistance genes also have been described. For each brassica species below, the dominant resistant genes will be described first (Table 2).

The first identification of resistance to TuMV in radish was actually done in an artificially synthesized Raphanobrassica hybrid (2n = 36, with RRCC genomes), generated from radish (2n = 18, RR genome; susceptible to TuMV) and kale (B. oleracea var. acephala; 2n = 18, CC genome; resistant to TuMV). This hybrid line was backcrossed to susceptible radish to obtain seven monosomic addition lines of radish, each with one kale chromosome (a-g; 2n = 19), as well as revertant radish plants (2n = 18). These various lines, as well as the progeny of further backcrosses of the seven monosomic lines to radish, were each tested for resistance to TuMV. The data showed that only plants of the monosomic line with the f-chromosome from kale (ftype line), were resistant to TuMV (Kaneko et al., 1996). It does not appear as if further work was published on the mapping of this trait.

A survey of brassicid species identified resistance to TuMV of pathotype 8 in three radish cultivars, with extreme resistance in two cultivars (‘Long Scarlet’ and ‘Sparkler’) and extreme resistance segregating with systemic resistance plus local necrosis in the third cultivar (‘Daikon Mooli Minowase’) (Nyalugwe et al., 2015a). Extreme resistance in cultivar ‘Daikon’ was also observed against a pathotype 1 isolate (UK1) and a pathotype 8 isolate (JN1) of TuMV (López-González et al., 2017). In a subsequent study, five isolates of TuMV (one each in pathotypes 1, 7 and 8, and two unpathotyped isolates) were tested against the cultivar ‘Sparkler’, showing extreme resistance to all tested isolates (Guerret et al., 2017). No genetics or mapping work appears to have been done on these resistance responses.

Four of 17 TuMV isolates obtained from a survey of infected radish plants in Korea were not able to infect the radish cv. ‘Iljin’, while all 17 isolates could infect the cv. ‘Chunghwa’ (Gong et al., 2019). The nature of neither the avirulence factor nor the resistance has been identified.

A determinant for infection of radish by TuMV strains was found in the C-terminal region of the P3 protein, with the identities of P3 C-terminal-region amino acids 268, 279 and 280 correlating with the ability to infect radish (Suehiro et al., 2004).

Other examples of genetic resistance to TuMV appear to be limited to reports published as thesis dissertations, as reviewed by Yu et al. (2017b). Using an F2 segregating population of two inbred radish lines (Q07-12P1 × KB07-10P2), which differed in resistance to TuMV, Li (2009) identified two QTLs related to TuMV resistance, one with a positive effect on the resistance and one with a negative effect. Bulk segregant analysis was used to tag the TuMV resistance gene to a marker with a genetic linkage distance of 7.9 cM. Cheng (2013) cloned two radish genes encoding the translation initiation factors eIF4E and eIF(iso)4E and suggested they may be involved in resistance. Cheng et al. (2013) described a method for locating markers by a target region amplification polymorphism (referred to as TRAP) using expressed sequence tags and information concerning resistance gene analogs. They used this method to establish a correlative relationship between marker genotypes and TuMV resistance phenotypes among 30 radish genotypes, producing a cultivar identification diagram.

The first analysis of resistance to TuMV in Aradidopsis was a screen of 106 ecotypes for responses to infection by the TuMV strain UK1, in which four ecotypes (Bay-0, Di-0, Er-0, and Or-0) were found to be resistant to systemic infection of TuMV, while Bay-0 also showed resistance to cell-to-cell movement of the virus (Martin Martin et al., 1999).

The first single dominant resistance gene to TuMV described for Arabidopsis, was a gene responsible for systemic veinal necrosis in the A. thaliana ecotype Landsberg erecta (Ler) (Kaneko et al., 2004). The TuMV necrosis inducer (TuNI) gene was shown to be located on chromosome 1 in a ~105 kb interval between two markers (mXF41 and mRF28), containing 19 putative genes, of which 15 were classified into five gene families. Subsequent chromosomal walking, cloning and sequencing analysis identified a trio of direct repeat variant genes, designated RGX, RG2 and RG3 as being involved in resistance to TuMV (Liu et al., 2015). The three genes were isolated and characterized. They each encoded proteins containing CC-NBSLRR domains and shared nearly identical sequences from the transcription initiation site (TIS) to the NBS domain, while RG2 and RG3 shared 99% sequence identity for 9 kb upstream from the TIS. All three genes shared similar promoter sequences to 200 bp upstream of the TIS. The RG2 gene encoded a truncated protein terminating just before the LRR domain, but might still have an active function. The RG3 gene encoded a rearranged chimeric LRR domain, longer than the one encoded by the RGX gene (Liu et al., 2015). Further work involving either transient expression or transgenic expression of each of three genes indicated that the RGX gene was the primary determinant of the resistance and suggested that the RG2 and RG3 genes played additional roles in regulating the induced necrosis, since RGX alone did not produce systemic veinal necrosis, but only induced necrotic spots and chlorosis after TuMV infection, while RG2 and RG3 themselves induced no necrotic symptoms. Transient expression of RG2 or RG3 together with either TuMV or the TuMV P3 encoding sequence in Arabidopsis protoplasts did not reduce cell viability, while a similar assay involving RGX greatly reduced cell viability (Liu et al., 2015); previous work by this group had identified the P3 as the avirulence factor for the TuNI gene (Kim et al., 2010) and had shown that the TuNI gene induced defense responses associated with a hypersensitive response-like cell death (Kim et al., 2008). Additional work involving promoter analysis showed that the RGX gene was induced by TuMV and suppressed by dark treatment (Liu et al., 2015), as was the case for the TuNI gene response (Kim et al., 2008). Thus, this complex of three genes involved in the TuNI response was the first dominant NB-LRR R-gene cluster isolated for resistance to TuMV.

The first Arabidopsis gene identified, isolated and characterized for recessive resistance to any plant virus was the loss-of-susceptibility to potyviruses 1 (lsp1) gene, which showed resistance to infection by both TuMV and tobacco etch virus (Lellis et al., 2002). The lsp1 locus was mapped to chromosome 5, to an interval of ~200 kb between markers MOK9-39s and MXH1-36116 Taq. The lsp1 gene encoded eIF(iso)4E, which had previously been shown to interact with the TuMV VPg protein (Wittmann et al., 1997). A similar conclusion was reached from an analysis of Arabidopsis plants from a transposon insertion library in which an eIF(iso)4E mutant line was identified and shown to be resistant to TuMV and the potyvirus lettuce mosaic virus (Duprat et al., 2002). These studies led to testing genes encoding eIF4E, eIF(iso)4E, as well as other eukaryotic translation factors, of other plant species, as candidate recessive resistance genes (reviewed by Shopan et al., 2020a, 2020b). Another study also showed that an Arabidopsis mutant line with a mutation in eIF(iso)4E was unable to support infection by TuMV, but supported infection by the potyvirus clover yellow vein virus, while an Arabidopsis mutant line with a mutation in eIF4E was able to support infection by TuMV, but not clover yellow vein virus (Sato et al., 2005). Interestingly, TuMV could infect an Arabidopsis eIF4E knockout mutant line transformed with either eIF4E or eIF(iso)4E from a diploid B. rapa ssp. trilocularis (yellow sarson; aka field mustard), although the mutant lines expressing the B. rapa eIF4E genes were less efficient (Jenner et al., 2010). Other work showed that either Arabidopsis eIF(iso)4G1 or eIF(iso)4G2 could be used by TuMV, but the absence of both, even in the presence of eIF4G, did not allow TuMV infection (Nicaise et al., 2007). However, mutations could occur in the VPg of TuMV to overcome knockout mutants of either eIF(iso)4E or both eIF(iso)4G1 and eIF(iso)4G2, suggesting the presence of an eIF4F-independent pathway (Gallois et al., 2010).

The P3N-PIPO protein of TuMV (Fig. 1) was shown to interact with the Arabidopsis protein PCaP1, a cationbinding protein that attaches to the plasma membrane. In an Arabidopsis PCaP1 knockout mutant (pcap1), TuMV was able to replicate but not able to move efficiently from cell to cell (Vijayapalani et al., 2012). Thus, pcap1 is another potential candidate for a recessive mutant blocking infection of TuMV in other hosts.

Natural mutants of the Arabidopsis gene AtCA1, encoding a salicylic acid (SA) binding protein (SABP3), preventing the binding of SABP3 to the HC-Pro of TuMV (Fig. 1), could provide a candidate gene for enhancing plant defense, since the interaction between HC-Pro and SABP3 led to weakening of the SA-mediated defense response (Poque et al., 2018).

Other candidate resistance genes from Arabidopsis were described from the results of a genome-wide association study involving TuMV infection of 317 accessions of A. thaliana in the field in each of two years (Rubio et al., 2019), including: (1) Restricted TEV Movement 3 (RTM3; AT3G58350) at the end of chromosome 3, involved in blocking long-distance movement (Cosson et al., 2010); (2) a QTL (AT3G01540) at the beginning of chromosome 3, encoding a DEAD Box RNA Helicase 1, shown to be involved in host-virus interactions, including TuMV (Li et al., 2016b); (3) a QTL on chromosome 5 within the gene AT5G27640, encoding a member of the eukaryotic translation initiation factor 3B; (4) a QTL on chromosome 2, located in an intergenic region with one of the adjacent genes, AT2G30880, encoding a protein with a pleckstrin homology domain involved in protein-protein interactions (Naughton et al., 2018); (5) a QTL at the end of chromosome 1 within the gene AT1G79470, encoding a protein with a TIM-barrel domain, such as found in the protein encoded by the Tm-1 gene that binds to the tomato mosaic virus replicase (Ishibashi et al., 2014); and (6) a QTL on chromosome 5, which overlaps with the gene AT5G18170 that encodes glutamate dehydrogenase 1. These genes and others affected viral accumulation, maximum viral accumulation, and/or frequency of infected plants (Rubio et al., 2019).

A number of resistance responses to infection have been identified and the genetics of most of these responses have been determined, mostly in Chinese cabbage and oilseed rape, but also in other Brassica species. From these, a selection of genes has been mapped and markers have been identified to facilitate the use of these genes in breeding for resistance. In a few cases, candidate genes for the resistance have been identified, and in rare instances, the candidate gene has been confirmed as the resistance gene. Still, many sources of resistance identified in various brassicas have neither been mapped nor further characterized, while resistance analysis in radish had largely remained unexplored, with only a few reports. While neither resistance to systemic infection nor extreme resistance has been reported in black mustard (B. nigra; BB genome), such resistance was identified in Ethiopian mustard (B. carinata; BBCC genome), and various subspecies of B. oleracea (CC genome). This suggests that contrary to some earlier speculation, the BB genome may also contain resistance genes, although no such genes have been identified or mapped. Nevertheless, there are numerous resistance responses that await mapping and identification, for either new resistance genes, or potential candidates identified for some of the genes already mapped in the same or other species. Isolating cloned dominant genes and transforming them into the genome of desirable cultivars of the same or related species should make the task of resistance breeding less onerous.

Arabidopsis was shown to contain several genes for resistance to TuMV, some of which have been cloned, while others represent possible candidate genes that could be explored for mapping and evaluation in syntenic crop species, such as brassicas or radish. The identification of genes expressing translation factors [eIF(iso)4E and others] in Arabidopsis, which confer recessive resistance to TuMV and other potyviruses has led to such genes being sought out as candidate resistance genes in various crops (reviewed by Shopan et al., 2020a, 2020b). In addition, various laboratories have developed novel forms of resistance against TuMV, by, intragenic (genes from related and sexually compatible species), cisgenic (genes from related but sexually incompatible species) or transgenic means; the last using expression of either TuMV sequences to induce RNA silencing, or other genes from various sources, usually in a model system, but also in brassica. While intragenics and cisgenics may come to be accepted by more countries, including for the expression of brassicid genes in other genera of the family, it does not appear that transgenic expression of unrelated genes will find their way into our crops in the near future, based on the costs involved in gaining approval; the latter of which moves at glacial speed. Therefore, it may be necessary to identify more sources of resistance as well as better characterize the many resistance responses already detected. Alternatively, various combinations of genes already identified in various Brassica spp. may be crossed into other cultivars of the same species. This will allow pyramiding of resistance genes, preventing infection by numerous pathotypes and obtain truly broad spectrum resistance. Several cultivars already contain more than one resistance gene, including: B. napus, line N-o-1 for TuRB01 and TuRB02; B. napus line S165 for TuBR04 and TuBR05; B. rapa line RLR22 for retr01 and (epistatic to) ConTR01; and B. rapa line AS9 for Rnt1-1 and rnt1-2. In other cases, breeding and selection would have to be done for pyramiding such resistance genes.