Introduction

The genus Brassica contains many important crop species with a range of adaptations to cultivation under varied agronomic conditions. They include standard Brassicaceae vegetables grown worldwide, such as Brassica oleracea var. botrytis (cauliflower), B. oleracea var. capitata (cabbage), B. oleracea var. gemmifera (brussel sprouts), B. oleracea var. italica (broccoli) and Raphanus sativus (radish) (e.g. Cai et al. 2004; Dixon 2007; Martinez-Villaluenga et al. 2010). They also include Asian vegetables, such as B. rapa var. pekinensis (chinese cabbage), B. chinensis (pak choy), B. parachinensis (choy sum), B. rapa var. nipposinica (mizuna) and B. oleracea gongylodes (kohlrabi) (e.g. Lewis and Fenwick 1988; Chen et al. 2008). The genus Brassica also includes species commonly used as supplements and alternatives to pastures in animal production systems, such as B. rapa var. campestris (turnip), B. napus var. napobrassica (swede) and B. oleracea var. acephala (kale) (Thomas et al. 1990; Vipond et al. 1998; Rowe and Neilsen 2011; Keogh et al. 2012). In addition, other Brassicaceae species are suited to oilseed and biofuel production. These include edible and nonedible oilseeds such as B. juncea (Indian mustard) and Camelina sativa (false flax) (e.g. Si and Walton 2004; Ghamkhar et al. 2010; Moser 2012). Many oilseed Brassica producing countries worldwide have recognized the potential benefits of introducing more high temperature tolerant Brassica species, e.g. B. juncea and B. carinata, than B. napus (Malik 1990; Potts et al. 1999; Burton et al. 2003; Cardone et al. 2003). In southern Australia, B. juncea is being developed for both oilseed and biofuel production (e.g. Burton et al. 2004; Si and Walton 2004; Gunasekera et al. 2006; Odeh et al. 2011).

Turnip mosaic virus (TuMV) is one of the most economically important plant viruses worldwide (Tomlinson 1987). Depending on geographical region and crop type, it causes a damaging disease of Brassica crops (Walsh and Tomlinson 1985; Hardwick et al. 1994; Gorecka and Lehmann 2001; Spence et al. 2007; Shattuck and Stobbs 1987). Quantified damage induced by TuMV epidemics includes a 50 % reduction in number and weight of marketable B. oleracea var. capitata heads in Kenya (Spence et al. 2007), root yield losses of c.40 % in Cochlearia arnoracia (horseradish) in Poland (Gorecka and Lehmann 2001) and 30 % in B. napus var. napobrassica in Canada (Shattuck and Stobbs 1987), and seed yield losses of up to 70 % in B. napus in the UK (Hardwick et al. 1994). It decreases numbers and sizes of seed pods, numbers of seeds, and seed size in B. napus (Walsh and Tomlinson 1985). It also causes quality defects such as internal necrotic disorders in stored B. oleracea var. capitata (Hunter et al. 2002). TuMV infection is particularly widespread in China where it causes major losses in Brassica crops (Liu et al. 1990; Yoon et al. 1993; Jiang et al. 2007).

Integrated Disease Management has proven by far the most successful way to manage virus diseases such as those caused by TuMV globally (e.g. Thresh 1982, 2006; Jones 2001, 2004, 2006, 2009; Coutts et al. 2011). Chemical control is unlikely to be effective with non-persistently aphid-borne viruses like TuMV in Brassicaceae crops, can increase resistance to pesticides in insect vectors, and poses a health hazard to humans and the environment (e.g. Jones 2004, 2006). When available, virus-resistant cultivars, provide an important approach to controlling TuMV epidemics (Hughes et al. 2002; Walsh et al. 2002; Gladysz and Hanus-Fajerska 2009), although resistance breaking TuMV pathotypes that overcome single gene resistance may arise (e.g. Jenner and Walsh 1996). Moreover, temperature sensitive single gene resistance may breakdown at higher temperatures as a result of climate change (Jones and Barbetti 2012). As some Brassicaceae producing regions in different continents are projected to experience increases in temperature, such resistance breakdown may decrease their productivity in future (Barbetti et al. 2012). Cultural and phytosanitary (=hygiene) control measures should therefore always be combined with host resistance to ensure effective TuMV-induced disease management (e.g. Jones 2004, 2006).

Examples of research to identify useful TuMV resistance phenotypes in vegetable Brassica species include studies which evaluated 88 B. oleracea var. capitata lines (Walkey and Neely 1980) and 3000 Chinese cabbage (B. rapa var. pekinensis and var. chinensis) lines (Liu et al. 1996). They also include efforts to identify TuMV resistance phenotypes in oilseed and forage Brassica species, including evaluation of 26 B. rapa (Hughes et al. 2002), 14 B. rapa and 2 B. juncea (Fjellstrom and Williams 1997) lines. In Australia, three recent studies evaluated Brassica species for presence of TuMV resistance phenotypes: (i) Coutts et al. (2007) tested 43 B. napus lines; (ii) Kehoe et al. (2010) assessed a collection of 44 B. juncea, nine B. carinata and five B. nigra (black mustard) lines, and the progenies of six crosses between B. juncea and B. napus; and (iii) Nyalugwe et al. (2014) evaluated 99 B. napus and 32 B. carinata lines from different continents. The studies by Coutts et al. (2007) and Kehoe et al. (2010) both identified additional TuMV resistance phenotypes to those named previously by Jenner and Walsh (1996).

Single dominant resistant genes TURB01-05 in B. napus and TURB01b in B. rapa give pathotype-specific TuMV resistance responses. Recessive TuMV resistance gene (retr01) and dominant TuMV resistance gene (conTR01) provide broad spectrum resistance to all TuMV pathotypes in B. rapa. These eight resistance genes all occur in the A genome originally derived from B. rapa, but TURB02, which is a quantitative trait locus controlling degree of TuMV susceptibility, occurs in the C genome originally derived from B. oleracea (Jenner et al. 2002; 2003; Jenner and Walsh 1996; Walsh et al. 1999, 2002; Walsh and Jenner 2006; Rusholme et al. 2007). Information on TuMV resistance genes in other Brassica species is lacking.

As B. juncea is an important crop in countries such as Australia, China, India, Canada and Russia, identification of new sources of host resistance and knowledge of its genetic basis constitute a key priority. This study therefore defined the range and occurrence of TuMV resistance phenotypes present in a broader selection of B. juncea lines than those evaluated previously by Kehoe et al. (2010). It also identified modes of inheritance for two of the TuMV resistance phenotypes found (+ND and +N), a single incompletely dominant resistance gene being responsible for +ND and +N in the homozygous or heterozygous conditions, respectively. In addition, the range and occurrence of TuMV resistance phenotypes was defined in lines belonging to four other important Brassicaceae crop species: B. oleracea (vars botrytis, capitata, gongylodes, gemmifera and italica), B. rapa, C. sativa and R. sativus.

Materials and methods

Germplasm and general plant growth conditions

Seeds of diverse accessions, breeding lines, and cultivars (=lines) of B. juncea [69 lines: Australia (27), China (12) and India (30)] were obtained from: a collection at The University of Western Australia as previously described (Nyalugwe et al. 2014), or a previous Australian study (Kehoe et al. 2010). Seeds of four other Brassicaceae species: [B. oleracea (11 cultivars), B. rapa (five lines), C. sativa (11 lines) and R. sativus (three cultivars)] were supplied by the Asian Vegetable Research Centre (AVRDC) Tanzania, or a University of Western Australia (UWA) Brassicaceae seed collection. Seeds of all lines were grown in a standard potting mix consisting of 2.5 m3 fine composted pine bark, 1 m3 coco peat, 5 m3 brown river sand, 10 kg slow release fertilizer Osmoform® NXT 22 N + 2.2 P2O5 + 9.1 K2O + 1.2 Mg + trace elements (Everris International B.V.), 10 kg Dolomite (CalMag®), 5 kg gypsum clay breaker, 5 kg extra fine limestone, 4 kg iron hepta sulphate, and 1 kg iron chelate. Potting mix was pasteurized at 63 °C for 30 min. Both culture hosts and experimental plants were grown in air-conditioned glasshouses or controlled environment rooms at UWA, or glasshouses with evaporative cooling at the Department of Agriculture and Food Western Australia (DAFWA). All these plant growing facilities were aphid-proofed.

TuMV cultures and inoculations

The TuMV isolates used were WA-Ap1 belonging to pathotype 8 (Coutts et al. 2007), and NSW-1 and NSW-2 belonging to pathotypes 7 and 1, respectively [pathotyping by J.A Walsh, cited by Kehoe et al. (2010)]. TuMV pathotypes are differentiated according to their reactions when inoculated to four differential lines of B. napus, rape S6 and R4, and swede S1 and 165 (Jenner and Walsh 1996). Pathotype 1 induces a systemic susceptible reaction in S6 and S1, but extreme resistance in R4 and 165. Pathotype 7 responds similarly in R4, 165 and S6, but induces a localised hypersensitive response in S1. Pathotype 8 induces a systemic susceptible response in S6, systemic invasion with necrosis in R4, a localised hypersensitive response in 165 and extreme resistance in S1 (Jenner and Walsh 1996). Kehoe et al. (2010) found pathotype 8 to be milder than pathotypes 1 and 7 in B. juncea and Schwinghamer et al. (2014) also reported differences in severity between TuMV pathotypes in this species, but Nyalugwe et al. (2014) reported no differences in virulence when these three pathotypes were inoculated to B. napus and B. carinata lines. Isolates were maintained by serial subculture using sap inoculation to plants of B. juncea Tendergreen, B. rapa Evigwe 2013 or B. alba / B. hirta (white mustard). Sap inoculation was done by grinding TuMV infected leaves from culture hosts in an extraction buffer containing disodium hydrogen phosphate (Na2HPO4) (11.5 g/l) and monosodium phosphate (NaH2PO4) (3 g/l), and the infective sap mixed with ‘celite’ (diatomaceous earth) before being rubbed onto the leaves of healthy plants. Plants were inoculated when they had 2–3 true leaves (Nyalugwe et al. 2014).

Enzyme-linked immunosorbent assay (ELISA)

Leaf samples from inoculated and/or uninoculated tip leaves of plants were tested for TuMV by double-antibody sandwich enzyme-linked immunosorbent assay (ELISA) as described by Clark and Adams (1977). Leaf samples were extracted (1 g per 20 ml) in phosphate-buffered saline (10 mM potassium phosphate, 150 mM sodium chloride, pH 7.4), Tween 20 at 5 ml/l, and polyvinyl pyrrolidone at 20 g/l using a mixer mill (Retsch, Germany). All samples were tested in duplicate wells in microtitre plates, and sap from TuMV-infected and healthy B. juncea Tendergreen leaf samples was always included in paired wells to provide controls. The substrate used was p-nitrophenyl phosphate at 0.6 mg/ml in diethanolamine, pH 9.8, at 100 ml/l. Absorbance values at 405 nm (A 405) were measured in a Bio-Rad microplate reader (model 680; Bio-Sys, Australia). Absorbance values of positive samples were always more than three times those of the healthy sap control. The polyclonal antiserum specific to TuMV used was obtained from Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Germany.

Responses to inoculation

Evaluation of germplasm

Plants were maintained between May and October 2012 in glasshouses at c.19 °C (+/−2 °C) at UWA, or c.22 °C (+/−3 °C) at DAFWA. The C. sativa lines were grown at UWA whereas the majority of B. juncea lines (62/69) and all other species were grown at DAFWA. Plants (7–14/line) were inoculated with sap extracted from plants infected with isolate WA-Ap1 (pathotype 8). Two plants of each line were kept uninoculated as healthy control comparisons. For 4–6 weeks after inoculation (wai), plants were inspected weekly for symptoms. Samples from inoculated and tip leaves were tested by ELISA at 3–4 wai, or, if systemic symptoms were severe and developed rapidly, as early as 10 to 14 days after inoculation. Uninoculated leaves recorded as uninfected were always re-sampled and tested again at 6 wai. Symptom reactions were classified using the phenotype codes of Jenner and Walsh (1996) and the subsequent code additions of Coutts et al. (2007) and Kehoe et al. (2010). The phenotype codes sought (not all of which were found) were: O, extreme resistance; +, susceptibility; RN, localised hypersensitivity; R, resistance to systemic movement without necrosis; RN/+, systemic infection with necrosis limited to inoculated leaves; +N, systemic infection with some necrosis; +ND, systemic hypersensitivity and plant death; RN/St, severe variant of RN with systemic necrotic stem streaking; RN/St/+, severe variant of RN/+ with systemic necrotic stem streaking, chlorotic mottle and leaf deformation; +N 1, mild variant of +N; and +St, severe stunting variant of +.

Effect of temperature on phenotype

Six lines in which phenotypes were expressed uniformly when inoculated with isolate WA-Ap1 (pathotype 8) [B. juncea Oasis Cl, 3117 and C. sativa 4160 (+ND), B. juncea, JM 06006, B. rapa Evigwe 2013 (+) and B. oleracea var. acephala unknown cultivar (O)] were selected. A group of 18 plants of each of these lines was sub divided into two lots of nine plants, seven were inoculated with infective sap containing isolate WA-Ap1 and two served as uninoculated healthy controls. Immediately after inoculation each lot was placed in controlled environment rooms operating at 16 or 28 °C. The experiment was repeated twice. Procedures for resistance phenotype classification and testing for virus infection using ELISA were as described above under ‘evaluation of germplasm’.

Effect of pathotype on phenotype

Plants of the same six lines in which phenotypes were expressed uniformly when inoculated with isolate WA-Ap1 were inoculated with isolates NSW-1 (pathotype 7) and NSW-2 (pathotype 1). Nine plants of each line were inoculated and two plants of each line were left uninoculated as controls. These plants were grown in a glasshouse at c.22 °C at UWA. Procedures for resistance phenotype classification and detection of the virus in leaf samples by ELISA were as described above under ‘evaluation of germplasm’.

Inheritance of +ND and +N resistance phenotypes in B. juncea

A cultivar which responded when inoculated with isolate WA-Ap1 (pathotype 8) by developing phenotype +ND uniformly (Oasis Cl) was crossed with a line which developed phenotype + uniformly (JM 06006) mostly following the procedure of Hughes et al. (2002). To obtain the F1 generation, JM 06006 was bud–pollinated with pollen from Oasis CI. Four to six buds were pollinated per inflorescence, and the remaining buds were removed. Each pollinated flower bud was labelled with details of the cross, and the inflorescences were enclosed in perforated polythene bags to prevent subsequent pollination by air circulation or bees and allowed to mature. When the seed pods were ripe (c. 8 weeks after the last bud was pollinated), they were dried, harvested, and the seed collected. To obtain the F2 generation, at the unopened bud stage inflorescences, on each individual F1 plant were enclosed within perforated polythene bags to prevent cross pollination. As the buds opened, the plants were shaken to encourage transfer of pollen from stamen to stigma. Seed was collected as described above. Mendelian segregation ratios were determined in four different tests in which F2 generation seeds were sown in standard potting mix in aphid-proofed facilities at UWA maintained at 21 °C constantly with a photoperiod of 12 h (test 1) or at 28/23 °C day/night temperatures (tests 2 to 4). For each test, the numbers of F2 plants and plants of the two parents (Oasis Cl and JM 06006) inoculated with infective sap containing isolate WA-Ap1, and of F2 control plants mock-inoculated with healthy sap, are shown in Table 1. Numbers of plants with characteristic systemic necrotic (hypersensitive) or non-necrotic (susceptible) responses to this inoculation were recorded from 0 to 11 days after infection (dai) (early infection) and 12–21 dai (late infection). Tip leaf samples from infected plants without necrosis (+ phenotype) and mock-inoculated plants were tested individually for TuMV at 14 dai by ELISA. Symptoms were not recorded after 21 dai as after that time older plants began to senesce and die from old age. The early and late infection results from each test were each analysed separately, and the combined early and late infection results collectively. This was done using the Χ 2 test to determine their fit to a 3:1 ratio (systemic necrotic:non necrotic responses) with early infection, and a 1:2:1 ratio (+ND:+N:+) with later infection.

Table 1 Number plants used per test in genetic studies with Turnip mosaic virus and the Brassica juncea cross Oasis CI (resistant) x JM 06006 (susceptible)
Full size table

Results

Evaluation of germplasm

B. juncea

When plants of 69 B. juncea lines were examined after inoculation with isolate WA-Ap1 (pathotype 8), the phenotypes recorded were +, +N (Fig. 1a), RN/+ (Fig. 1b), RN/st/+, +ND (Fig. 1c), +N 1, and +st (Table 2). Twenty two B. juncea lines developed one phenotype uniformly. The phenotypes that occurred alone were (numbers of lines with each phenotype in parentheses): RN/+ (10), +N (8), + (1), +ND (2) and RN/st/+ (1). Eleven of these lines came from Australia [RN/+ (5), +N (3), + (1) and + ND (2)], nine from India [RN/+ (3), +N (5) and RN/st/+ (1)], and two from China [RN/+ (2)]. The remaining 47 lines segregated for two (36) or three (11) different phenotypes, and came from Australia (16), China (10) or India (21). Phenotypes RN/+, +N, +, +ND, RN/st/+, +N 1 and +st were all represented among these segregating lines. The commonest combination of segregating phenotypes was RN/+ and + (16): ten from Australia, four from India and two from China. The second commonest combination was +N and + (6): five from India and one from China. The remaining 25 lines segregated for 12 combinations of 2–3 phenotypes, each combination occurring in 1–5 lines. When the frequencies of occurrence of each segregating phenotype were compared, RN/+, +, +N and +ND occurred most frequently. As with the lines with uniform phenotypic reactions, phenotypes RN/+ and +N were common among Australian and Indian lines, respectively. Phenotype RN/st/+ occurred less often, phenotype +N 1 only occurred segregating in two plants of cv. Kranti and +st in three plants of cv. Durgamini, both of Indian origin. Phenotypes O, RN, R and RN/st were absent.

Fig. 1
figure 1

Symptoms caused by infection of B. juncea plants following sap inoculation with a pathotype 8 isolate of Turnip mosaic virus (TuMV). a ‘JO 009’ showing phenotype +N (systemic necrosis, leaf deformation, severe mosaic and leaf drop) at 15 days after inoculation (dai); b ‘TABP 15’ showing phenotype RN/+ (necrotic spot local lesions in inoculated leaves and mosaic in uninoculated leaves) at 9 dai; and c ‘Oasis Cl’ showing phenotype +ND in a row of seven plants (systemic necrosis and plant death) at 30 dai (front), and two healthy control plants (rear)

Full size image
Table 2 Reactions of 69 Brassica juncea lines from three different continents following sap inoculation with isolate WA-Ap1 (pathotype 8) of Turnip mosaic virus (TuMV)
Full size table

Other Brassicaceae species

Phenotype O developed uniformly in all six types of B. oleracea inoculated with isolate WA-Ap1 (pathotype 8) (numbers of cultivars within each type in parentheses): acephala (1), botytris (2), capitata (3), gongylodes (1), gemmifera (1) and italica (2), no virus being detected in any of them (Table 3). However, B. oleracea var. italica cv. Romanesco was an exception as it segregated for phenotypes R and O. Among the B. rapa lines, one gave a uniform phenotype + response (B. rapa Evigwe 2013), two segregated for phenotypes + and RN/+ [B. rapa var. chinensis (Fig. 2a and b) and var. nipposinica], one for RN and O (B. rapa var. pekinensis) and one for phenotypes +, +N and +st (Fig. 2c) (B. parachinisis). All the 11 C. sativa lines developed uniform phenotype +ND (Fig. 2d, e and f), but there were differences between them in the number of days taken for plant death to occur, C. sativa cv. China and 4160 dying 7 days earlier than the others. R. sativus cv. Daikon Mooli segregated for phenotypes RN and O. The other two R. sativus cultivars developed phenotype O uniformly.

Table 3 Reactions of plants belonging to four different Brassicaeae species following sap inoculation with isolate WA-Ap1 (pathotype 8) of Turnip mosaic virus (TuMV)
Full size table
Fig. 2
figure 2

Symptoms caused by infection of plants of other Brassicaceae species following sap inoculation with a pathotype 8 isolate of Turnip mosaic virus (TuMV). a and b, B. rapa var. chinensis showing (A) phenotype RN/+ (necrotic spot local lesions in inoculated leaves and faint mosaic in uninoculated leaves) at 15 days after inoculation (dai), and (B) phenotype + (severe leaf deformation and systemic chlorotic mottle in uninoculated leaves) at 37 dai; c, B. rapa var. parachinsis showing a plant with phenotype +st (stunting, mosaic and leaf deformation) at 15 dai (right), healthy (left); d, healthy C. sativa China plant in comparison with e a plant showing initial symptoms of +ND (severe mosaic and leaf deformation) at 9 dai; f healthy C. sativa China plant (left) in comparison with an infected plant showing phenotype + ND (plant death) at 30 dai (right)

Full size image

Effect of temperature on phenotype

B. oleracea var acephala

(unknown cultivar) developed phenotype O uniformly at 16° and 28 °C when inoculated with isolate WA-Ap1 (pathotype 8), no virus being detected in inoculated or uninoculated leaves (Table 4). B. juncea Oasis Cl and 3117, and C. sativa 4160 developed phenotype +ND uniformly at 16 °C and 28 °C. They formed necrotic (B. juncea Oasis Cl and 3117) or chlorotic (C. sativa 4160) spots in inoculated leaves at 3–4 dai at 28 °C and 6–7 dai at 16 °C. The virus moved systemically by 5 dai at 28 °C and 8 dai at 16 °C causing mosaic and leaf deformation followed by systemic necrosis leading to plant death. Death occurred by 14–15 dai at 28 °C and 16–17 dai at 16 °C (Fig. 3a, b and c). B. juncea JM 06006 and B. rapa Evigwe 2013 developed phenotype + uniformly at 16 °C and 28 °C. Faint chlorotic spot local lesions formed in inoculated leaves followed by systemic mosaic, local and systemic symptoms appearing 3 (B. juncea JM 06006) and 6 (B. rapa Evigwe 2013) days earlier at 28 °C than at 16 °C, respectively (Fig. 3d). Thus, in B. oleracea, B. juncea, B. rapa and C. sativa, increasing the temperature from 16° to 28 °C decreased the time lag between inoculation and symptom development, but did not alter expression of phenotype.

Table 4 Effects of two temperatures on symptom development following inoculation of a pathotype 8 Turnip mosaic virus (TuMV) isolate to plants of three Brassica juncea lines, and one line each of C. sativa, B. oleracea and B. rapa, all with uniform phenotypic responses
Full size table
Fig. 3
figure 3

Symptoms caused by inoculation with a pathotype 8 isolate of Turnip mosaic virus (TuMV). a, B. juncea Oasis CI plants at 7 days after inoculation (dai) showing numerous necrotic spot local lesions in inoculated leaves (inset) with no systemic infection yet at 16 °C (left), or necrotic spot local lesions in inoculated leaves (inset) and systemic necrosis and leaf deformation at 28 °C (right); b, C. sativa 4160 plants at 15 dai showing phenotype +ND (plant death) at 28 °C (left), or chlorotic spot local lesions in inoculated leaves and systemic necrosis and stunting at 16 °C (right); c, B. juncea Oasis CI plants at 15 dai showing phenotype + ND (plant death) at 28 °C (left), or necrotic spot local lesions in inoculated leaves and systemic necrosis and stunting at 16 °C (right); and d, B. rapa Evigwe 2013 plants at 9 dai showing severe phenotype + (severe mosaic and leaf deformation) at 28 °C (left) and symptomless systemic infection at 16 °C (right)

Full size image

Effect of pathotype

As shown in Table 3, plants of C. sativa 4160 inoculated with isolate WA-Ap1 (pathotype 8) developed phenotype +ND uniformly, local chlorotic spots forming in inoculated leaves followed by systemic mosaic, systemic necrosis and, finally, plant death. In contrast, when plants of this line were inoculated with isolate NSW-2 (pathotype 1) and NSW-1 (pathotype 7), only phenotype + developed (systemic mosaic). In the other five lines inoculated (B. juncea Oasis Cl, 3117 and JM 06006, B. rapa Evigwe 2013 and B. oleracea var. acephala unknown cultivar) isolates NSW-1 and NSW-2 produced the same phenotypic responses and symptom severities as those which occurred previously with isolate WA-Ap1 in these same lines (Tables 2 and 3). B. rapa Evigwe 2013 and B. juncea JM 06006 developed only phenotype +, B. oleracea var. acephala (unknown cultivar) only phenotype O, and B. juncea cv. Oasis CI and 3117 only phenotype +ND. Thus, in the six lines inoculated with all three isolates, the phenotypic responses and symptom severities to inoculation with TuMV pathotypes 1, 7 and 8 remained the same in B. juncea (3 lines), B. rapa (1 line) and B. oleracea (1 line), but infection with pathotype 8 caused a different phenotype in C. sativa 4160.

Inheritance of resistance phenotypes +ND and +N in B. juncea

In the F2 progeny plants inoculated with isolate WA-Ap1 (pathotype 8), a clear differentiation between systemic necrotic (Fig. 4a and b) and non-necrotic (Fig. 4c and d) responses became evident during the early stage of infection (up to 11 dai). With the necrotic response, inoculated leaves developed necrotic spot local lesions which later coalesced to form dead patches before being shed. Such plants developed systemic necrosis at 3–6 dai starting from the leaf margins at the leaf apex (Fig. 4b). With the non-necrotic response, the plants developed local chlorotic spots in inoculated leaves at 3–5 dai followed by mosaic in their uninoculated leaves at 6–8 dai. Plants of parental lines cv. Oasis CI and JM 06006 gave similar initial necrotic (Fig. 4e) and non-necrotic responses (Fig. 4f), respectively. The mock-inoculated F2 progeny plants retained a healthy, vigorous appearance throughout (Fig. 4g). Between 12 and 21 dai, as infection progressed, the F2 progeny plants with a systemic necrotic response gradually separated into two groups with resistance phenotypes +ND or +N. Within the +ND grouping, plant death was staggered, some plants started dying as early as 7 dai (Fig. 5a right), the others all being killed by 21 dai (Fig. 5a left, and b). Symptoms within the phenotype +N group ranged from plants with a bunchy appearance of lower leaves, leaf deformation and leaf drop, stem necrosis, flower abortion, no or poor pod formation and general plant weakness (Fig. 5c), to plants with a stunted bunchy plant structure, flower abortion, leaf deformation and leaf drop (Fig. 5d), or plants which lost their initial leaves developing new leaves showing less severe TuMV symptoms (Fig. 5e). In all four tests, plants of parent Oasis Cl died at 7–11 dai (Fig. 5f right, circled), whereas plants of parent JM 06006 developed systemic mosaic, plant stunting also being recorded in few of them but the majority producing flowers and pods (Fig. 5f left; Table 5). ELISA tests on leaf samples taken at 14 dai detected TuMV in uninoculated leaves of plants within the + grouping (Table 5). Mock-inoculated F2 control plants developed no visible symptoms and TuMV was never detected in their uninoculated leaves (Fig. 5g). It was impossible to study segregation ratios in F3 progeny plants due to unavailability of F3 seed. F2 plants with phenotype +ND were all killed while plants with resistance phenotype +N had aborted flowers or poor pod set, so produced insufficient seed.

Fig. 4
figure 4

Plants of B. juncea inoculated with a pathotype 8 isolate of Turnip mosaic virus (TuMV) at 10 days after inoculation (dai) (test 4). F2 progeny plants from the cross of Oasis Cl (phenotype + ND) × JM 06006 (phenotype +) (A, B, C and D); plants of parents Oasis Cl (E) and JM 06006 (F); and healthy F2 progeny control (G). a, initial stage of necrotic symptom development characterised by coalescing necrotic spot local lesions forming patches in inoculated leaves and tip leaf necrosis starting to appear; b, magnified view of leaf deformation and necrosis symptoms in tip leaves; c, showing phenotype + involving mosaic development in leaves; d, magnified view of phenotype +; e, showing initial systemic necrosis (phenotype +ND parent); f, showing mosaic (phenotype + parent); and g, healthy F2 control plant without symptoms

Full size image
Fig. 5
figure 5

Plants of B. juncea inoculated with a pathotype 8 isolate of Turnip mosaic virus (TuMV). F2 progeny plants from the cross of Oasis Cl (phenotype +ND) × JM 06006 (phenotype +) (A, B, C, D and E); plants of parents Oasis Cl (F, right) and JM 06006 (F, left); and healthy F2 progeny control (G). a and b, plants in the final stages of development of phenotype + ND (plant death) showing leaf death (left) or completely killed (right) at 14 days after inoculation (dai) (from test 3) (A), or completely killed at 21 dai (from test 2) (B); c and d, two plants showing different variants of phenotype +N (necrosis, mosaic, reduction in leaf size, leaf deformation, leaf drop, bunchy plant growth and stunting) at 21 dai (from test 2); e, showing variant of phenotype +N (partially recovered plant showing loss of earlier leaves and regrowth of newer ones) at 21 dai (from test 3); f, showing comparison of parental plants with mosaic (B. juncea JM 06006, left) and systemic plant death (B. juncea Oasis Cl, right in circle) at 10 dai (from test 1); and g, healthy control plant at 21 dai (from test 3)

Full size image
Table 5 Segregation data for F2 progeny plants from the Brassica juncea cross Oasis Cl (phenotype + ND) × JM 06006 (phenotype +) following inoculation with isolate WA-Ap1 (pathotype 8) of Turnip mosaic virus (TuMV)a
Full size table

When the data for all four tests were combined, with early infection 84 F2 plants developed non-necrotic responses, whereas 244 plants developed necrotic responses (Table 5 and Supplementary Fig. 1). With subsequent infection, for all four tests the combined numbers of plants with different resistance phenotypes were 73 with +ND, 174 with +N, and 81 with +. When the individual or pooled initial data sets for F2 progeny plants from tests 1–4, were analysed, they always segregated for the ratio 3 necrotic: 1 non-necrotic responses. However, at the later stages of infection, pooled data for the four tests fitted a 1:2:1 ratio for phenotypes +ND:+N:+ (P = 0.6). The segregation ratios for individual tests 1–4 also all fitted a 1:2:1 ratio (P > 0.3) (Table 5; Supplementary Fig. 1). The symptom reactions, appearance and segregation ratios obtained with plants grown at 21 °C constantly (test 1) were no different from those grown at 28 °C/23 °C day/night temperatures (tests 2–4).

Interpretation of segregation data for inheritance of resistance phenotypes +ND and +N in B. juncea

In F2 progeny plants of the cross between resistant parent B. juncea Oasis Cl (+N phenotype) and susceptible parent B. juncea JM 06006 (+ phenotype), two readily distinguishable resistance phenotypes developed directly after inoculation up to 11 dai with isolate WA-Ap1 (pathotype 8), one with and the other without systemic necrosis. When the numbers of plants with each phenotype were compared, they were in agreement with a 3:1 ratio for necrosis (resistance):lack of necrosis (susceptibility). However, only a third of the plants with systemic necrosis were killed (+ND phenotype), the remainder remaining alive (+N phenotype), giving a final segregation ratio of 1:2:1 for phenotypes +ND:+N:+. These results suggested the presence of a single hypersensitive resistance gene which was fully dominant at the early stage of infection but only exhibited partial dominance later. Thus, when two dominant alleles (homozygous condition) were present, systemic plant death developed (phenotype +ND). However, when only one dominant allele was present (heterozygous condition), the infected plants developed additional systemic symptoms. When two recessive alleles were present (recessive homozygous condition), the plants developed only mild symptoms. We propose the name TuRBJU 01 (TuMV resistance in B rassica ju ncea 01) for this gene, which follows the same nomenclature (TuRB01-05) used previously used to name TuMV resistance genes in B. napus and B. rapa (e.g. Walsh and Jenner 2002).

Discussion

This study of the responses of five economically important Brassicaceae species to inoculation with a TuMV pathotype 8 isolate: (i) identified the first TuMV resistance gene for B. juncea designated TuRBJU 01, an incomplete single dominant gene responsible for systemic resistance phenotypes +ND in the homozygous condition and +N in the heterozygous condition; and (ii) found potentially useful TuMV resistance phenotypes in all of the other four species evaluated. Increasing the temperature from 16° to 28 °C decreased the time lag between inoculation and symptom development in inoculated B. oleracea, B. juncea, B. rapa and C. sativa plants, but did not alter phenotype. Similarly, when plants of some lines of these four species were inoculated with TuMV isolates belonging to pathotypes 1 and 7, except in C. sativa, phenotypic responses and symptom severities remained the same as those obtained with pathotype 8. Resistance gene TuRBJU 01 is potentially valuable in breeding for TuMV resistance in B. juncea. Ten B. juncea lines developed +ND (2) or +N (8) phenotypes uniformly and so are suitable for use as breeding parents. Moreover, once the genes responsible are identified, some of the TuMV resistance phenotypes found among the other four species not only are likely to be valuable to use in breeding TuMV-resistant cultivars, but also may provide sources of additional new resistance genes. Ten B. oleracea and two R. sativus lines developed extreme resistance phenotype O uniformly, while all 11 C. sativa lines developed phenotype +ND uniformly. They will be more useful as parents in TuMV resistance breeding than lines that segregated for different phenotypes, as occurred with, for example, the four different kinds of Asian vegetables (B. rapa) tested, such segregation being due to outcrossing.

Examples of other resistance genes resembling TuRBJU 01 in B. juncea by conferring systemic hypersensitive resistance phenotypes include: ccd1 in Nicotiana clevelandii; cyn1 in Pisum sativum (pea); Nbm-1 in Lupinus angustifolius (narrow-leafed lupin); Nc, Nx, Nv and Nytbr in Solanum tuberosum (potato); Nam-1 in Medicago orbicularis (button medic); and Rsv1-s in soybean (Glycine max) (Cockerham 1970; Chen et al. 1994; Jones 1985, 1990; Ma et al. 1995; Pathipanawat et al. 1996; Király et al. 1999; Jones and Smith 2005; Ravelo et al. 2007). Such genes have been used for many years by plant breeders, e.g. Nc and Nx in breeding for virus resistance in S. tuberosum (Cockerham 1970). They are effective as they kill infected plants quickly thus removing them as internal sources of virus infection for acquisition and further spread by vectors or contact in crops. This reduces the rate of epidemic development and final infection incidence resulting in much decreased yield losses (e.g. Jones et al. 2003; Jones 2005; Ma et al. 1995). Pallett et al. (2002) suggested that the rarity of TuMV infections in wild populations of B. nigra and B. rapa in the UK is probably because they cause death of the infected plants, and Coutts et al. (2007) attributed the low incidence of TuMV infection found in B. napus crops in Australia to widespread occurrence of local cultivars expressing TuMV resistance phenotypes which included +N.

Following inoculation of 69 lines from different continents with TuMV pathotype 8, we found resistance phenotypes +N, +N 1, RN/+ RN/st+ and +ND in B. juncea, but phenotypes O, RN and R were absent. The commoner phenotypes were +N, RN/+ and +ND, and susceptibility phenotype +. Previously, Kehoe et al. (2010) also reported finding additional resistance phenotypes RN and RN/st in this species. In B. napus, dominant TuMV resistance genes TURB01 (+N), TURB03 (+N), and TURB04 with TURB05 (RN) are all on the A genome and no TuMV resistance genes are reported on the B genome (Jenner and Walsh 1996; Walsh and Jenner 2002, 2006). As B. juncea carries both A and B Brassica genomes, occurrence of TuMV resistance phenotypes in this species reflects presence of resistance genes located within one or other of these genomes. Phenotypes +N 1 and +ND are variants of +N (Kehoe et al. (2010). Whether resistance gene TuRBJU 01 controlling the +N and +ND resistance phenotypes in B. juncea also controls phenotype + N 1 is unknown. Most likely, TuRBJU 01 is located on the A genome and in some way related to genes TURB01 or TURB03 controlling the +N phenotype in B. napus, but this is yet to be shown. Alternatively, it may reside on its B genome despite most known pathogen resistance genes being in the A genome. Inoculation of isolates belonging to TuMV pathotypes 1, 7 and 8 to the B. juncea +ND parent of the cross used to study segregation of the +ND phenotype always produced +ND uniformly, but studies with additional pathotypes are required to determine whether any overcome resistance gene TuRBJU 01. As +ND was expressed uniformly by B. juncea Oasis CI and JR 042 in this study, they are likely to be useful parental lines to use in breeding new cultivars of B. juncea with resistance to TuMV. The uniform +ND phenotypic responses of the 11 C. sativa lines tested resembled those in B. juncea Oasis CI and JR 042. However, inoculation of isolates belonging to TuMV pathotypes 1 and 7 to one of these lines (C. sativa 4160) only induced phenotype + (mosaic). Similarly, when inoculated with a Brassica TuMV isolate, the two C. sativa cultivars tested by Nguyen et al. (2013) also developed phenotype + (symptomless infection or vein clearing mottle and stunting). As C. sativa is a significant alternative oilseed crop (Ghamkhar et al. 2010), once the genetic basis of the +ND phenotypic trait is understood, it may prove useful in breeding TuMV resistant cultivars of this species.

When pathotype 8 was inoculated to 11 cultivars belonging to six types of B. oleracea (vars acephala, botrytis, capitata, gonglodes, gemmifera and italica), one of var. italica (cv. Romanesco) segregated for phenotypes R and O, but the other 10 lines all gave a uniform phenotype O response. Inoculation of isolates in pathotypes 1, 7 and 8 to B. oleracea var. acephala (unknown cultivar) also gave a uniform phenotype O response in all instances. These results resemble those of other recent studies in which diverse types of B. oleracea often proved difficult to infect with TuMV (e.g. Korkmaz et al. 2008; Farzadfar et al. 2009; Nguyen et al. 2013). However, they differ from earlier evaluations of diverse accessions and cultivated types of B. oleracea, especially B. oleracea var. capitata in which partial TuMV resistance but no extreme resistance were found (Pink and Walkey 1986; Walkey and Pink 1988; Walkey and Pink 1988). Differences between TuMV pathotypes used are likely to explain why results with B. oleracea often differed between published studies. Phenotype O also developed in the three R. sativus cultivars we tested with our pathotype 8 isolate (which was originally from R. raphanistrum), although it segregated with phenotype RN in one of them. Finding phenotype O in R. sativus resembles results obtained previously by others using different TuMV isolates as it often proved difficult to infect (e.g. Korkmaz et al. 2008; Farzadfar et al. 2009; Nguyen et al. 2013).

When an isolate of TuMV pathotype 8 was inoculated to plants of single lines of different types of Asian vegetables in our study, there was segregation for phenotypes: (i) O and RN in the single line of B. rapa var. pekinensis, (ii) RN+ and + in B. rapa var. chinensis and var. napobrassica, and (iii) +N, +st and + in B. rapa var. parachinensis. In contrast, the oilseed B. rapa Evigwe 2013 plants gave a uniform phenotype + response. The responses of B. rapa var. pekinensis reflect the presence of strain-specific resistances leading to different types phenotypic reactions depending on the isolate used (Provvidenti 1980; Green and Deng 1985; Yoon et al. 1993; Liu et al. 1996; Korkmaz et al. 2008; Farzadfar et al. 2009; Nguyen et al. 2013). The segregation for susceptibility (phenotypes + and/or +st) and resistance (phenotypes +N or RN+) we found in plants of the other four types Asian vegetables indicates that they too carry strain specific resistances. Only one B. rapa line, the oilseed B. rapa Evigwe 2013, gave a uniform reaction (phenotype +), which  may also developed when it was inoculated with isolates belonging to TuMV pathotypes 1 and 7. TuMV resistances present in B. rapa var. pekinensis are already in use in breeding TuMV resistant cultivars of Asian vegetables (Provvidenti 1980; Niu et al. 1983; Green and Deng 1985; Fjellstrom and Williams 1997). The + N resistance found in B. rapa var. parachinensis should also be suitable for this purpose.

Nyalugwe et al. (2014) found that, in most cases, plants of B. napus and B. carinata lines that developed phenotype O uniformly at 16 °C following inoculation with pathotype 8 behaved differently at 28 °C. At the higher temperature, phenotype O was either replaced entirely by one or more other phenotypes or segregated with other phenotypes. In this study, however, phenotype O developed uniformly at both 16° and 28 °C in plants of B. oleracea var. acephala, and, although the higher temperature facilitated earlier appearance of symptoms, there also was no alteration in phenotype in plants of three other species (B. juncea, B. rapa and C. sativus). These studies at 16° and 28 °C included B. juncea Oasis CI which carries resistance gene TuRBJU 01 and was the parental cultivar used in our crosses. Plants of this cultivar and B. juncea 3117 both developed phenotype +ND uniformly at either temperature suggesting that TuRBJU 01 is unlikely to be temperature sensitive. This finding is important, as is its effectiveness against three different TuMV pathotypes (see above), providing strong indications that it is likely to be useful in breeding for TuMV resistance in B. juncea. A future focus on identifying molecular markers linked to this resistance gene would be worthwhile to expedite development of TuMV-resistant B. juncea cultivars in plant breeding programmes. Moreover, as climate change progresses into the future, virus resistance breeding using resistance genes that are not temperature sensitive is likely to become increasingly important (Jones and Barbetti 2012). This applies particularly to crop species like B. juncea because it is better adapted than B. napus to the increasing temperatures and increasing moisture stress expected in coming decades. For example, in Australia B. juncea production is already expanding to extend oilseed Brassica production into cropping areas within lower rainfall or that are becoming warmer, and providing greater yield reliability there. As rainfall becomes even less predictable, B. juncea is expected to replace B. napus as the preferred crop in drier regions (Barbetti et al. 2012). Development of TuMV-resistant B. juncea cultivars therefore is likely to become a key ingredient of Australian Brassica breeding programmes. Furthermore, in addition to B. juncea resistance gene TuRBJU 01, lines of B. oleracea and R. sativus that developed phenotype O uniformly, and of C. sativa that developed phenotype +ND uniformly are potentially valuable for use in breeding TuMV-resistant cultivars of these three species.