Broad-based root-knot nematode resistance identified in cowpea gene-pool two

The test materials were a subset of 48 cowpea genotypes previously selected from a drought tolerance study from a diverse pool of 350 genotypes, which include accessions and landraces from the Mozambique Institute of Agricultural Research (IIAM) and others collected across Mozambique representing the cowpea gene-pool two (Huynh et al., 2013). These cowpea genotypes display very distinct agronomic and morphological traits including seed size, shape and color, stem pigmentation, stem diameter, leaf shape and size, plant architecture, growth habit, biological cycle, yield ability, and drought tolerance. The cowpea genotypes used in all experiments as controls and their responses to M. incognita and M. javanica are given in Table 1.

The control genotypes CB46-Null, NIL-2 genes and NIL-3 genes are near-isogenic lines (NIL) developed in a California blackeye cultivar CB46 background through multiple backcrosses (Roberts et al., 2014; Huynh et al., 2016). The breeding line CB46-Null is susceptible whereas NIL-2 genes and NIL-3 genes are resistant lines. The genotype CB3-gg is a NIL developed in a CB46 background through successive backcrosses, and it carries a root-galling resistance gene derived from California blackeye cv. CB3 (Roberts et al., 2014). The cultivar CB50 also has CB46 background in its pedigree (Ehlers et al., 2009). California blackeye cultivar CB27 is resistant; in addition to the gene Rk, it carries a recessive gene with additive effect (Ehlers et al., 2000a, 2000b). The genotype UCR779 is a cowpea accession originally from Botswana; it lacks resistance to all tested RKN.

Three Meloidogyne incognita and one M. javanica (same isolate used in all experimental conditions) isolates were used in this study. Two of the M. incognita isolates (“Project 77” and “Beltran”) were avirulent to cowpea genotypes carrying resistance gene Rk (Roberts et al., 1995), and an incompatible interaction between these isolates and a genotype carrying Rk, or any genetic resistance factor equivalent to this gene would be expected. The term avirulent is used to indicate nematode populations that reproduce poorly on plants on which virulent nematode populations of the same species reproduce significantly and induce substantial root-galling (Roberts et al., 1995). The third M. incognita isolate, “Muller,” is highly virulent to cowpea genotypes carrying gene Rk, inducing excessive root-galling and reproducing successfully (Roberts et al., 1995), but resistance based on Rk plus other genes in combination (Table 1) provides effective resistance against this isolate (Roberts et al., 2014). The “Muller” isolate was obtained from a cowpea field in which selection for virulence to gene Rk had occurred following repeated planting of cowpeas with Rk-based resistance (Petrillo and Roberts, 2006). The M. javanica isolate “Project 811” is aggressive and able to induce root-galling and to reproduce significantly on plants carrying gene Rk at a level of 50% or more of that observed on susceptible plants, although this ability is not based on selection for Rk-virulence (Roberts et al., 1995; Ehlers et al., 2002). The term “aggressive” refers to the enhanced ability of this M. javanica isolate to cause damage to cowpea plants carrying the Rk gene (Roberts et al., 1995).

Experiments to determine the root-galling response of the test genotypes were conducted in infested field sites and field data were validated in greenhouse pot experiments. Four field experiments were conducted during June – October of 2012, 2014, 2015, and 2016 to determine the response of the test genotypes to root-galling by M. javanica, avirulent, and virulent M. incognita isolates in separate field sites infested with each nematode isolate. These sites were established several years ago by injecting nematode eggs extracted from greenhouse-grown tomato plants into the root-zone of susceptible tomato plants, and are maintained by planting susceptible tomato plants to provide high and uniform nematode infestation levels (Huynh et al., 2016). In the 2012 field experiment (conducted at South Coast Research and Extension Center – SCREC), 24 of the 48 genotypes were tested for root-galling by M. javanica, avirulent, and virulent M. incognita separately in a completely randomized design on each nematode site. In 2014 (at SCREC), all test genotypes were screened in four replicate blocks in a randomized complete block design on each nematode site. In the 2015 and 2016 experiments (at Kearney Agriculture Research and Extension Center – KARE, and SCREC, respectively), only the highly resistant genotypes identified in the 2012 and 2014 experiments plus all controls were tested in each nematode site using the same experimental design as in 2014, to validate the response of these potential RKN resistance donors. The genotype Gile-K-Local was not included in this experiment due to seed shortage. In all field experiments, each replicate plot consisted of 20 to 25 seeds per genotype planted in a 1.5 m-long single-row, and water and fertilizer were supplied as needed through drip-irrigation tape. Sixty-days after emergence, plant tops were cut 2 to 3 cm above the soil line and all root systems (except for the 2012 experiment where plant stand per genotype ranged from 10 to 24) dug and evaluated for root-galling response. The genotype response was assessed as the average of all plants in each plot. The levels of nematode infestation in the field were indicated by root-galling index of susceptible controls CB46-Null and UCR779; these controls in all field experiments were planted at every fifth plot planted to test cowpea genotypes.

Under greenhouse conditions (28°C day and 22°C night temperatures), the entire cowpea collection including all controls was tested for root-galling induced by M. javanica in two separate experiments each with four replicates arranged in a randomized complete block design. Two seeds of each genotype were planted in fiber pots (15-cm-diameter), containing a soil mixture of 80% sand and 20% peat, and thinned to one plant per pot 7 d after emergence. Water and fertilizer were provided as needed using a drip-irrigation system. Fifteen-days after emergence, a 10 ml egg suspension in water containing about 1,000 egg/ml were pipetted per pot in four 3 cm-deep holes around the plant stem. The eggs for these experiments were extracted from roots of susceptible tomato plants maintained in a greenhouse at UC-Riverside. Sixty-days after inoculation, plant tops were cut 2 to 3 cm above the soil line and the roots washed and indexed for root-galling response under 10X magnification following the protocol described next. The greenhouse experiments focused mainly on evaluating the response of the cowpea collection to M. javanica because there was a need to identify a highly effective source of resistance to this nematode isolate for breeding purposes and to validate the results from field experiments with M. javanica. However, a subset of six M. javanica resistant genotypes (including all controls) identified in the 2012 and 2014 field experiments plus a genotype identified with resistance to virulent M. incognita in the 2014 experiment were also evaluated for egg production per root system and root-galling by virulent M. incognita in a separate greenhouse experiment. This was a single experiment with four replications; the genotypes were inoculated with virulent M. incognita isolate “Muller” (following the protocol described above) obtained from greenhouse cultures mantained on tomato plants.

Root-galling assessment followed a 0 to 9 index (GI) modified from Bridge and Page (1980), where 0 = no galls on root system; 1 = very few, small galls and difficult to see; 2 = very few and small galls can be seen; 3 = galls can be easily seen on most roots except the main root, the size varies from very small to small; 4 = root system is obviously galled, some large galls can be seen on secondary roots and very few bumps can be seen on the main root; 5 = generally large galls can be seen on the root system and the main root is slightly galled with galls of different sizes; 6 = large galls, main root heavily galled; 7 = large galls and large coalesced galls on the main and secondary roots, respectively; 8 = generally huge galls and huge coalesced galls on secondary and main roots, respectively, very few secondary roots can be seen; 9 = huge galls and coalesced galls, generally no secondary roots visible. A cut-off between resistant-susceptible genotypes for root-galling response was set at GI = 3; this threshold was established based on the average root-galling on resistant and susceptible controls, and the average root-galling of all tested genotypes; thus, genotypes with GI ≤ 3 were considered resistant, and those with GI > 3 were considered susceptible (modified from Fery et al., 1994; Roberts et al., 2008). In cowpea, the size, location, and the relative number of galls allows identification of resistant from susceptible reactions to RKN. Also, cowpea resistant genotypes do not show galls on main roots, so this phenotypic response has been used to easily distinguish resistant from susceptible plants.

Egg-mass production (EM) by M. javanica “Project 811” and avirulent M. incognita “Project 77” was assessed in the test cowpea collection using seedling growth-pouch tests (Ehlers et al., 2000a; Atamian et al., 2012) in a growth chamber with day and night temperatures set at 28 and 22°C, respectively, under 16 hr day-length. A single seed of each genotype was planted in a plastic pouch, and pouches were minimally watered using a wash bottle to allow seed germination and seedling emergence. After seedling emergence pouches were watered as needed. Second-stage juveniles (J₂) were hatched from nematode eggs placed in an incubator at 26 to 27°C for 7 d. Every 2 to 3 d emerged J₂ were collected, counted and concentrated to the desired inoculum density. Approximately, 12 to 14 d days after emergence, the plants were inoculated with freshly hatched second-stage juveniles (J₂) at a density of 1,500 J₂/plant and laid on a table horizontally for 24 hr in the dark. After inoculation, the plants were fertilized for 3 to 5 d with half-strength Hoagland’s solution (Hoagland and Arnon, 1950) and additional fertilizer was applied as needed for the remainder of the experiment. At 30 to 35 d after inoculation the roots were infused with erioglaucine solution (1 g/l) (Sigma Chemical Co., St. Louis, MO, USA), and 24 hr later the solution was poured off and the stained egg-masses counted under 10X magnification. The distinction between resistant and susceptible reactions to nematode reproduction measured by egg-masses per root system (EM), was determined using EM = 30, which was based on the average EM production on resistant and susceptible controls, and the average EM production in the entire collection (modified from Roberts et al., 1996, 2008). Genotypes showing ≤30 EM per root system were classified as resistant, and genotypes with EM count >30 were classified as susceptible. This assay was conducted to determine the response of the test cowpea genotypes to nematode reproduction, and to determine the relationship between root-galling and nematode reproduction responses (to infer on the relationship between the genetic determinants controlling both phenotypes). Five separate experiments each with four replicates were conducted. In the first three experiments, the entire cowpea collection including controls was screened for nematode reproduction by each nematode isolate, whereas in the additional two experiments only resistant genotypes plus controls were tested to validate their response. These experiments focused mainly on M. javanica and avirulent M. incognita due to the reason stated previously.

Data for root-galling and egg-mass production by M. javanica from greenhouse and seedling growth-pouch experiments were correlated to infer whether both traits are under control by the same or distinct genetic factors.

Nematode virulence herein is defined as the ability of a nematode to successfully establish a feeding site, induce root-galling and reproduce on roots of resistant plants (Roberts et al., 1997; Petrillo et al., 2006). Nematode virulence was determined using virulence index (VI) estimates for each nematode isolate, calculated as the proportion between galling or reproduction on the root systems of resistant genotypes and susceptible genotypes (Petrillo et al., 2006). The spectrum of resistance of tested cowpea genotypes was defined as the relative response of tested genotypes to each RKN isolate. In addition, the relationships between root-galling data for M. javanica, avirulent M. incognita, and virulent M. incognita were analyzed for correlation to infer on the relationship among the resistances underlying response to root-galling by these nematodes using root-galling data from the 2014 field experiment. Virulence indices were computed using root-galling data from the 2014 field experiment and egg-mass data from growth chamber experiments.

Raw data for root-galling and nematode egg-mass reproduction from all experiments were analyzed by ANOVA. Data analysis comprising separate testing within a season to more than one nematode isolate was performed following the procedure for analysis of series of experiments described by Gomez and Gomez (1984), where the nematode isolates were considered as environments. In the first step the data were analyzed separately for each nematode isolate, and in the second step combined ANOVA was performed to test the genotype × environment effect. For separate greenhouse experiments with M. javanica and virulent M. incognita, one-way ANOVA was performed using the average genotypic responses of two separate experiments within each replication and genotypic responses from a single experiment in four replications, respectively. For all experiments (field, greenhouse and growth chamber), the data analysis was performed using SAS University Edition 3.2.2 following the mixed procedure (Proc Mixed) where the blocks were considered as the random factor while nematode isolates and cowpea genotypes were considered as fixed factors. The 2012 field experiment followed the same procedure as the other experiments; 10 to 24 plants of each genotype were evaluated for root-galling under each nematode isolate infestation. This experiment was arranged in a complete randomized design (CRD). The 2014, 2015, 2016, and greenhouse experiments were arranged in a randomized complete block design (RCBD) with four replicate blocks. The growth chamber experiments were conducted five times (three with all test genotypes and two with resistant genotypes), and each was arranged in a RCBD with four replicate blocks. Mean separation in all experiments was performed using Tukey’s multiple comparison test at p < 0.05. The relationship between M. javanica root-galling (greenhouse data) and egg-mass production responses was determined through Pearson correlation following procedure Corr, and procedure Reg was used to fit a linear model using SAS University Edition 3.2.2. Also, the relationships between root-galling by M. javanica, avirulent M. incognita, and virulent M. incognita were analyzed for correlation following the same procedures. However, the relationship between root-galling by M. javanica and avirulent M. incognita was best fit using a logarithmic model.

In a preliminary study conducted at the UC-SCREC in 2012, a subset of 24 of the 48 cowpea genotypes plus controls were screened for root-galling response on sites infested with avirulent M. incognita, virulent M. incognita or aggressive M. javanica (Fig. 1). A total of 16 of the 24 test genotypes and controls CB46, CB27, NIL-2 and NIL-3 genes, and CB3-gg were resistant (GI ≤ 3) to avirulent M. incognita compared to susceptible control CB46-Null (Fig. 1). Three test genotypes (VAR-3A, FN-2-9-04, and Namuesse-D) were resistant (G ≤ 3) to virulent M. incognita, as were controls NIL-3 genes, CB3-gg, and CB27 (Fig. 1), while control genotypes CB46, CB50, and CB46-Null were susceptible.

Figure 1

The control NIL-2 genes showed moderate resistance (GI = 3.8) response to virulent M. incognita (Fig. 1). Of the 24 test genotypes, four (VAR-3A, FN-2-9-04, Namuesse-D, and FAEF-14-INE) were resistant (G ≤ 3) to aggressive M. javanica, while controls CB27, NIL-2, and NIL-3 genes were moderately resistant (GI = 3.3, 3.7, and 4.0, respectively), and CB46, CB50, CB3-gg, and CB46-Null were susceptible (GI = 5.0, 5.0, 6.6, and 6.4, respectively). The average root-galling scores of the genotypes ranged from 1.3 to 6.1, 1.9 to 5.5, and 1.3 to 6.9 under infection by avirulent M. incogita, virulent M. incognita and M. javanica, respectively.

In 2014, the full set of 48 test genotypes was screened for root-galling response on the same infested field sites at SCREC (Fig. 2). There was significant effect by genotypes, nematode isolates and their interaction for root-gall indices (p < 0.0001). Significant differences among genotypes for mean root-galling induced by each nematode isolate were detected at GI = 1.4. All test genotypes except one (FN-2-11-04, GI = 3.9) were resistant to root-galling induced by avirulent M. incogita (G ≤ 3), and all controls, except those lacking resistance genes (CB-46-Null and UCR779, GI = 5.2 and 6.0, respectively), were resistant to avirulent M. incogita (Fig. 2). The average root-galling phenotypes induced by avirulent M. incognita infestation ranged from 0 to 6, and most of the differences in root-galling response among test genotypes were not significant (p > 0.05).

Figure 2

Several of the test genotypes that were resistant to M. javanica, were also resistant to virulent M. incogita (Fig. 2), including FN-2-9-04 and VAR-3A. These two genotypes were also resistant to virulent M. incogita in the 2012 test (Fig. 1). Of the controls, CB27, NIL-3 genes and CB3-gg were resistant to virulent M. incognita, confirming their response in the 2012 test. The control NIL-2 genes were resistant to virulent M. incognita in 2014 (Fig. 2), but only moderately resistant in the 2012 test (Fig. 1). Root-galling responses of resistant controls CB27, NIL-3 genes, and CB3-gg were not different (p > 0.05) from resistant test genotypes. However, the root-galling phenotype of the resistant control NIL-2 (GI = 2.6) was significantly higher (p < 0.05) than that of FN-2-9-04, Gile-K-Local, VAR-3A, CB27, and NIL-3 genes (GI = 0.5, 0.9, 0.6, 0.8, and 1.1, respectively). Genotypes INIA-41, Maputo, and Muinana-Lawe were also resistant to virulent M. incognita, but their response to root-galling did not differ (p > 0.05) from that of CB3-gg, CB27, NIL-2 genes, NIL-3 genes, and resistant test genotypes FN-2-9-04, Gile-K-Local, VAR-3A, INIA-5A, FAEF-14-INE, Namuesse-D, and VAR-11D. Root-galling phenotypes with virulent M. incogita ranged from 0.5 to 5.2 (Fig. 2). Susceptible test genotypes included SP-860, SP-866 and FN-2-11-04. As expected, CB46, UCR779, and CB46-Null were also susceptible.

In the test with M. javanica, eight genotypes (VAR-3A, FN-2-9-04, Namuesse-D, INIA-5A, Gile-K-Local, FN-1-14-04, VAR-11D and FAEF-14-INE) were resistant (GI ≤ 3) to root-galling, of which four (VAR-3A, FN-2-9-04, Namuesse-D and FAEF-14-INE) were also resistant in the 2012 field test (Fig. 1). Of the controls, CB27, NIL-2 genes, and NIL-3 genes were also resistant to M. javanica, but controls carrying only the Rk gene (CB46 and CB50), the root-galling gene (CB3-gg) or no gene were susceptible, as expected. The responses of the resistant genotypes FAEF-14-INE, FN-1-14-04, FN-2-9-04, Namuesse-D, and VAR-3A, were similar, and lower (p < 0.05) than those of the Rk-gene controls. Root-galling phenotypes (ranging from 0.3 to 5.5) induced by aggressive M. javanica were in the same range as those induced by virulent M. incogita. At KARE, these genotypes were planted in separate fields infested with avirulent M. incognita isolate “Project 77” and M. javanica, and at SCREC screening was done on sites infested with avirulent M. incognita “Beltran,” virulent M. incognita and M. javanica each. The results from these field experiments were consistent with the 2012 and 2014 experiments (data not shown).

The test genotypes were screened for resistance to root-galling by M. javanica in pots under greenhouse conditions. Based on the ANOVA, genotypes differed (p < 0.0001) in M. javanica root-galling response. Mean root-galling index per genotype ranged from 1 to 8 (Fig. 3). Consistent with the results observed in the field (Figs 1, 2), except for FN-1-14-04, genotypes FN-2-9-04, FAEF-14-INE, VAR-3A, Namuesse-D, Gile-K-Local, INIA-5A, and VAR-11D exhibited resistant M. javanica root-galling phenotypes. The controls CB27, NIL-2 genes, and NIL-3 genes also showed consistent resistant root-galling phenotypes as in the 2014 field test. The susceptible phenotypes observed for controls CB46, CB3-gg, UCR779, and CB46-Null were also consistent with the results observed in the field experiments under M. javanica infestation. Significant differences in root-galling phenotypes were detected at GI = 1.75 (p < 0.05).

Figure 3

Root-galling phenotypes observed among the resistant test genotypes (FN-2-9-04, VAR-3A, Namuesse-D, INIA-5A, VAR-3A, FAEF-14-INE, and Gile-K-Local) and controls CB27, NIL-2 genes, and NIL-3 genes were not different, but root-galling phenotypes of the resistant test genotypes were lower (p < 0.05) than those of CB46, CB46-Null, CB3-gg, and UCR779, as expected (Fig. 3). The root-galling responses among the resistant test genotypes were not different, nor were root-galling responses between controls CB27, NIL-2 genes, and NIL-3 genes.

The cowpea subset evaluated in greenhouse inoculations with virulent M. incognita included genotypes FN-2-9-04, VAR-3A, Namuesse-D, INIA-5A, VAR-3A, FAEF-14-INE, and INIA-41, plus all controls. The genotype Gile-K-Local was not tested due to seed shortage. The results from this experiment (root-galling responses) were not consistent with those observed in the 2012 and 2014 field experiments. Of these genotypes, FN-2-9-04, VAR-11D, and INIA-41 showed moderate resistance to root-galling similar to CB27, CB46, and NIL-2 genes (p > 0.05) under virulent M. incognita greenhouse inoculation (data not shown).

The test cowpea genotypes were also evaluated for the ability to suppress reproduction of avirulent M. incognita and M. javanica assessed by production of egg-masses (EM) per root system in controlled inoculations (Fig. 4). There was significant effect by the nematode isolates, genotypes and their interaction for EM (p < 0.0001). The mean EM per root system for avirulent M. incognita and aggressive M. javanica ranged from 0 to 64 and 1.31 to 105, respectively (Fig. 4). Significant differences in EM production per root system between genotypes were detected at EM = 20.7 and 18.4 (p < 0.05) for avirulent M. incognita and M. javanica, respectively.

Figure 4

Avirulent M. incognita reproduced poorly on most test genotypes compared to M. javanica, with the exception of FN-2-11-04 and INIA-76 (EM = 47.9 and 47.5, respectively) (Fig. 4). Among the control genotypes, CB3-gg, CB46-Null, and UCR779 were susceptible to avirulent M. incognita (EM = 47.1, 56.5, and 64.0, respectively) (Fig. 4). Most of the test genotypes (FN-2-9-04, VAR-3A, Namuesse-D, INIA-5A, VAR-3A, FAEF-14-INE, and Gile-K-Local) had very low EM by avirulent M. incognita and M. javanica, indicating they were highly resistant, and the EM phenotypes among them were not different (p > 0.05). Also, their phenotypes were not different from control genotypes carrying the Rk genes (CB46, CB27, CB50, NIL-2 genes, and NIL-3 genes). These control genotypes supported fewer EM than controls lacking Rk (CB3-gg, CB46-Null, and UCR779) (p < 0.05) (Fig 4).

The root-galling and egg-mass production responses under infestation by M. javanica were moderately correlated (r = 0.60, p < 0.0001) (Fig. 5), and a small proportion (R ² = 0.36) of the variability in this relationship was explained by the response of the test genotypes to root-galling and egg-mass production. Based on this relationship, three classes of genotype responses were identified (Fig. 5): (i) genotypes with low root-galling and low EM production (GI from 0 to 3; e.g., FN-2-9-04); (ii) genotypes with moderate root-galling and moderate EM production (GI from 3 to 5; e.g., Timbawene-Monteado); and (iii) heavily galled genotypes with high EM production (GI > 5; e.g., INIA-76).

Figure 5

Root-galling phenotypes induced by virulent M. incognita and M. javanica were highly correlated (r = 0.98, p < 0.0001) (Fig. 6A), and a large proportion (R ² = 0.97) of the variability in root-galling was explained by the response of the genotypes under infestation by both nematodes.

Figure 6

Root-galling responses to avirulent M. incognita and to M. javanica also were highly correlated (r = 0.72, p < 0.0001) (Fig. 6B), and the relationship was largely explained (R ² = 0.97) by the observed variability in root-galling among the test genotypes. In this relationship, the test genotypes could be classified into three groups: (i) M. javanica resistant genotypes (GI = 0 to 3) with low or no noticeable root-galling symptoms under avirulent M. incognita infestation; (ii) moderately resistant genotypes that were resistant to root-galling by avirulent M. incognita (GI = 3-4); and (iii) genotypes that were susceptible to both RKN isolates.

Virulence index (VI) for each nematode isolate was estimated as the ratio between RG or EM production of a test genotype and susceptible control lacking RKN resistance. The genotype UCR779 was used as the reference susceptible control to estimate VI values because it had the highest average RG and EM production. Estimates of virulence presented in Figure 7 are expressed based on average RG and EM production per root system of all test genotypes. As expected, the avirulent M. incognita isolate was less virulent to the test cowpea genotypes compared to the virulent M. incognita and aggressive M. javanica isolates (Fig. 7), with VI values based on RG and EM data of 12 and 5%, respectively. Based on RG data, M. javanica (VI = 50%) and virulent M. incognita (VI = 42%) were about fourfold more virulent than the avirulent M. incognita isolate (Fig. 7A). Estimated virulence index for virulent M. incognita using field RG data was lower (VI = 42%) than that estimated based on greenhouse data (VI = 75%) (data not shown). The RG levels between M. javanica and virulent M. incognita were not different, but their RG levels were lower (p < 0.05) than that of avirulent M. incognita.

Figure 7

Using EM data, the average virulence index of aggressive M. javanica was 18% compared to 5% for avirulent M. incognita (Fig. 7B).

Analysis of the spectrum of resistance to root-galling induced by avirulent and virulent M. incognita and aggressive M. javanica in field experiments showed that seven test genotypes (FEAF-14-INE, FN-2-9-04, Gile-k-local, INIA-5A, Namuesse-D, VAR-11-D, and VAR-3A) exhibited resistance to all three isolates. In contrast, INIA-41 and Maputo were only resistant to avirulent and virulent M. incognita, respectively.