Combining ability of tropical early maize ( Zea mays L . ) inbred lines for grain yield and resistance to maize streak virus disease

Maize is an important cereal crop in sub-Saharan Africa but production is adversely affected by maize streak virus disease (MSVD). In Ghana, re-occurrence of the disease has been reported in several regions, therefore, necessitating the development of resistant hybrids as the most sustainable and economical option. The objectives of the study were to identify parents and hybrids that combine MSVD resistance with high yield and determine the influence of maternal effect on the inheritance of MSVD resistance. To achieve these, five parental inbred lines namely: TZEI-4, TZEI-7, TZEI-22, TZEI-31 and TZEI-157 were crossed in a full diallel mating design during the major season of 2015. The resulting F1 hybrids were evaluated under natural and artificial infestations during the minor and major seasons of 2015/2016 using 9 x 3 alpha-lattice design with three replications. General combining ability (GCA) and specific combining ability (SCA) mean squares were significant for MSVD severity mean score and only SCA for grain yield. Additive gene effect was preponderant for MSVD severity mean score, whereas grain yield was influenced by non-additive gene effect. Maternal effect had no significant contribution to the inheritance of MSVD resistance. GCA by environment and SCA by environment mean squares were significant for MSVD severity mean score. GCA effects revealed that inbreds TZEI-7 and TZEI-22 were resistant to MSVD. They could be good combiners for grain yield in addition to TZEI-31 and TZEI-157. Hybrids TZEI-4*TZEI-22 and TZEI-4*TZEI-31 showed resistance to MSVD as revealed by their SCA effects and heterotic values. TZEI-7*TZEI-157, TZEI-31*TZEI-157, TZEI-22*TZEI-157 and TZEI-4*TZEI-22 had positive and significant SCA effect, mid-parent heterosis and high parent heterosis for grain yield. Promising hybrids TZEI-4*TZEI-22, TZEI-22*TZEI-157, TZEI-7*TZEI-157 and TZEI-31*TZEI-157 identified in this study should be further tested in multi-locations across Ghana to determine their stability and adaptability.


INTRODUCTION
Worldwide, maize (Zea mays L.) ranks first in production with over one billion tonnes produced in 2014 followed by rice (741 million tonnes) and wheat (729 million tonnes), although the latter ranks first in terms of harvestable area (FAOSTAT, 2015).It is distributed worldwide and serves as a staple crop to most sub-Saharan African countries, providing food and employment to a larger percentage of the entire populace (Magenya et al., 2009).
In Ghana, maize is the second most significant food crop next to cassava and it is produced in all the geographical areas with its production in the transition zone being the highest (MoFA, 2011).FAOSTAT (2015) report indicated a significant reduction in maize produced throughout the country from 1,949,897 tonnes in 2012 to 1,762,000 tonnes in 2014.This reduction has been attributed to frequent biotic and abiotic stresses including pest and disease outbreak, reduced soil fertility and drought (Cairns et al., 2012).
Maize streak virus disease (MSVD) is caused by maize streak virus (MSV) obligately transmitted by leafhoppers in the genus Cicadulina (Storey, 1925).It is a major foliar disease that affects maize throughout the sub-Saharan Africa (Pingali and Pandey, 2001) and its prevalence on farmers" fields has been reported in several regions of Ghana (Oppong et al., 2015).Amongst the diseases that cause economic damage to maize in the world, MSVD ranks third following northern leaf blight and grey leaf spot, besides it has remained the most severe viral disease of maize in Africa resulting in the loss of returns, which ranges from US $ 120 to 480 million yearly (Martin and Shepherd, 2009).With effective MSVD control, no less than half of this loss can be avoided (Martin and Shepherd, 2009).The disease can cause up to 100% yield loss in susceptible varieties under field conditions (Magenya et al., 2008).However, reduction in growth and yield are directly dependent on factors such as time and stage of infection and also varies with the level of resistance (Bua and Chelimo, 2010).
MSVD symptoms are characterized by broken to nearly unbroken chlorotic bands or stripes centered initially on the tertiary leaf veins and these later develop into rectangular tan-coloured lesions that run parallel with leaf veins.As the disease spreads, the lesions merge resulting in blighting of the whole leaf (Agrios, 2005).The density of striping depends primarily on the resistance of the genotypes.In highly susceptible maize plants, the entire leaf lamina shows a severe, uniform white chlorosis which may progress gradually into death of cells and tissues of the plants and afterwards die back, especially when the plants are infested at the seedling stage (Rossel and Thottapilly, 1985).Severe chlorosis in susceptible maize plants leads to stunted growth, scanty ear development, reduced seed setting and ultimately huge yield losses or occurrence of premature death (Mawere et al., 2006;Monjane et al., 2011).
Eleven strains of MSV have been identified and are designated MSV-A to MSV-K.MSV-A strain has been identified to be the most virulent and can cause significant MSVD while others attack cereals such as barley, wheat, oats, rye and millet but not maize (Martin et al., 2001;Shepherd et al., 2010).Oppong et al. (2015) Ige et al. 187 reported that MSV-A1 variant was predominant in the transition and forest zones of Ghana, and it exhibits an increased level of pathogenicity than the other MSV-A variants which are MSV-A2, MSV-A3, MSV-A4 and MSV-A6 (Martin et al., 2001).The incidence and severity of MSVD can be reduced by chemical control of leafhoppers and cultural practices such as crop rotation, irrigation, inter-cropping, application of appropriate fertilizer rate and plant density manipulation but the most economically sustainable option is provided by using disease resistant varieties (Martin and Shepherd, 2009).Despite the successes achieved in breeding for varieties with MSVD resistance, the prevalence of MSVD continues to occur in Africa, causing huge losses in yield due to the erratic changes in climate (Legréve and Duveiller, 2010) which to some extent, makes the epidemiology of the disease complex (Martin and Shepherd, 2009).Commercial varieties in Ghana, which had some degree of resistance to MSVD, have become susceptible over the years.Therefore, it is imperative to identify high-yielding, stable and novel genotypes that can resist or tolerate MSVD outbreak enhanced by drought or erratic rainfall resulting from climate change in the tropical environments.
Therefore, the objectives of the study were to identify maize genotypes that combine high yield with MSVD resistance for sustainable production and determine the effect of maternal inheritance on MSVD resistance.The experimental field was sprayed with rid-out (glyphosate, 360 g/l) at 5.0 l/ha before ploughing and harrowing were done.The 27 genotypes including the checks (Omankwa and Aburohemaa) were planted in a 9 x 3 alpha-lattice design with three replications.A plot consisted of two-rows of 5 m length each.The rows were spaced 75 cm apart while hills were spaced 40 cm apart.Three seeds were sown per hill and later thinned to two plants per hill at three weeks after planting (WAP).Hence, the planting density was approximately 66,667 plants/ha.Recommended crop management practices were applied.Fertilizer equivalent to 90:0:40 kg/ha of N-P2O5-K2O (26:0:4) and sulphate of ammonia fertilizers were applied at two weeks after planting and at ear emergence respectively.Post emergence weeds were controlled with the application of caliherb (2,4-dichlorophenoxy acetic acid, 360g/l) at 4.5 l/ha and manual weeding when necessary.

Artificial infestation of maize genotypes with maize streak virus
Non-viruliferous leafhoppers, Cicadulina mbila Naudé (Hemiptera: Cicadellidae) were collected from maize evaluation fields with the use of a pooter and were reared on pearl millet (Pennisetum americanum L.) in insect proof cages made of galvanized metal with a wooden base (Dimension = 0.9 x 0.9 x 2 m) at the entomology section of the CSIR-CR1, Kwadaso Station.They had an acquisition access period of 48 h from maize plants severely infected with MSV.The 27 genotypes planted in cups filled with loamy soil were infested at two-leaf stage as described by Bosque-Pérez and Alam (1992) but modified.The modification was done by placing the maize seedlings in insect proof cages and after 48 h of feeding period by the viruliferous leafhoppers, they were transplanted nine days after planting after infestation with MSV to the field which has been ploughed, harrowed and laid out using 9 x 3 alpha-lattice design with three replications.

Data collected
Data such as anthesis-silking interval (ASI), plant height (PLHT), total leaf count (TLC), ear leaf area (ELA), plant aspect (PASP), ear aspect (EASP) and 100-grain weight (HGW) were recorded but only maize streak virus disease (MSVD) severity mean score and grain yield are reported.Grain yield (t/ha) was estimated from ear weight per plot, assuming a shelling percentage of 80% and then adjusted to 12.5% moisture content.Plants in a plot were visually scored for MSVD at 3, 6 and 9 WAP according to Beyene et al. (2012) scale; 1 = no symptoms on leaves, 2 = light disease symptoms on 20 to 40% leaf area, 3 = moderate symptoms on 40 to 60% leaf area, 4 = severe symptoms on 60% of leaf area, 5 = severe symptoms on 75% or more of the leaf area.

Statistical analysis
Analyses of variances (ANOVA) were performed separately on MSVD severity mean score and grain yield from the natural and artificial infestations and then combined ANOVA across environments using PROC GLM in statistical analysis system (SAS, 2003) software version 9.1.Genotypes were considered as fixed effect while environments, replications and blocks within replications as random effects.Least significant difference (LSD) was used to determine the significant differences amongst the least square means of the genotypes at the probability level of 0.05.MSVD severity mean score was square root (√x) transformed before performing the analysis, but the original value was reported after back-transformation.
The GCA effects of the parents and SCA effects of the F1 hybrids as well as their reciprocal effects under each and across environments for MSVD severity mean score and grain yield were estimated without the checks by following Griffing"s Method 1, Model l (fixed effects) Griffing (1956) using DIALLEL-SAS program developed by Zhang et al. (2005) adapted to SAS software version 9.1.Effects of GCA, SCA and reciprocal were computed from the mean values adjusted for the block effects under each environment and across environments.T-test was used to detect the significance of GCA, SCA and reciprocal effects.Standard errors were estimated as square root of the GCA, SCA and reciprocal variances (Griffing, 1956).
The least square means for grain yield were used to estimate heterosis in F1 over mid-parent and high parent according to Rai (1979).

Mid-parent heterosis (MPH) =
, MP = High parent heterosis (HPH) = "T" test was then performed to know whether the F1 hybrid means were significantly different from the mid-parent and high parent means as described by Wynne et al. (1970).

Analysis of variance for MSVD severity mean score and grain yield
The combined ANOVA across infestations revealed significant (P<0.001)effects for environment and genotype with MSVD severity mean score and grain yield (Table 2).It was observed that significant variation existed amongst the genotypes under each infestation (Table 3).These implied that the environments were distinct and sufficient genetic differences existed among the genotypes.This would therefore permit effective progress to be made from selection for MSVD resistance and yield.Genotype x environment across infestations revealed significant (p<0.001)differences for MSVD severity mean score but not for grain yield.The significance explained that the response of genotypes to MSVD differed across infestations, implying that there were probably escapes under natural infestation or virus pressure differed across environments.Consequently, higher disease pressure was observed under artificial infestation as compared to natural infestation.Bosque-Pérez et al. (1998) reported that infestation of plant with MSV at early stages leads to greater disease severity.This would then make selection of resistant genotypes difficult under only natural infestation, therefore, stressing the need to evaluate them under artificial infestation thus, enhancing stable performance and productivity of genotypes.
Partitioning the genotypes into general combining ability (GCA), specific combining ability (SCA) and reciprocal components revealed that GCA mean square was significant (P<0.001) for only MSVD severity mean score across infestation and under each infestation (Tables 2 and 3).SCA mean square was significant for MSVD severity mean score and grain yield across infestations and artificial infestation while only grain yield under natural infestation (Tables 2 and 3).Significant GCA and SCA mean squares observed for MSVD severity mean score across infestations showed the relative contributions of additive and non-additive gene effects on the expression of MSVD resistance.However, grain yield was solely controlled by non-additive gene effect as revealed by its significant SCA.Significant GCA by environment and SCA by environment mean squares were detected for only MSVD severity mean score (Table 2).These indicated that the response of both parental inbred lines and hybrids to MSVD differed across environment, suggesting that selection for resistance to the disease should be done in specific target environment.Non-significant reciprocal effect observed for MSVD severity score (Tables 2 and 3) implied that maternal effect had no significant contribution to the inheritance of MSVD resistance, therefore in future maize hybrid breeding programmes targeting MSVD resistance, the choice of a maternal parent is not very important.GCA mean squares to SCA mean squares ratios across infestation and under each infestation for MSVD severity mean score indicated that additive gene effect was preponderant in the control of MSVD resistance in the genotypes evaluated; this suggests that early generation testing may be efficient for selecting resistant genotypes.This result agrees with those of Vivek et al. (2010), Gichuru et al. (2011) and Mutengwa et al. (2012) who found out that additive gene effects were predominant in the inheritance of resistance to MSVD.High GCA mean square implied that the per se performance of the inbred lines used in this study should be a suitable pointer of the performance of their hybrids (Gethi and Smith, 2004;Badu-Apraku et al., 2011).For grain yield, the ratios for all were less than unity indicating that non-additive effect was more predominant, indicating that heterosis could be exploited from crossing the set of parental lines used in the study.It is therefore expedient to assess the parental inbred lines with different testers to be able to identify superior hybrids since the performance of the hybrids cannot be based on GCA alone (Hallauer and Miranda, 1988).This result agrees with Bhatnagar et al. (2004).In contrast, Sibiya et al. (2013) found out that additive gene effect was more predominant in controlling grain yield.Varying gene action controlling grain yield is dependent on the parents and environment under consideration (Gichuru, 2013).
In most of these hybrids, one of the parents had corresponding negative GCA effect except for TZEI-4*TZEI-31, TZEI-31*TZEI-4 and TZEI-157*TZEI-4.Significant SCA effects reveal that the level of resistance of certain hybrids were higher or lower than expected on the basis of the GCA of the two parents involved in the cross (Falconer and Mackay, 1996) and these effects are pinpointing to dominant gene action.Despite the positive GCA effects observed for parents TZEI-4 and TZEI-31, the SCA effect observed for the resultant straight cross hybrid was negative.This could be because of the presence of quantitative trait loci (QTLs) that were too small in effect to be expressed in each of the parents but sufficient to be detected when they are combined.Parental inbred lines TZEI-157, TZEI-7, TZEI-22 and TZEI-31 contributed 0.12, 0.10, 0.06 and 0.01 t/ha to the grain yields observed in the hybrids across infestations.One or both of the parents involved in the following crosses TZEI-7*TZEI-157, TZEI-31*TZEI-157, TZEI-22*TZEI-157, TZEI 4*TZEI-157 and TZEI-4*TZEI-22 had positive GCA effect, suggesting that favourable genes were transmitted to the progenies (Badu-Apraku and Oyekunle, 2012).This implies that these hybrids can be used as testers in subsequent maize breeding programmes.

Mid-parent heterosis (MPH) and high parent heterosis (HPH) for grain yield across natural and artificial infestations
Plant breeders exploit heterosis by crossing distantly related genotypes in order to achieve an increase in desirable traits as compared to the mid-parent or high parent values.All the hybrids showed significant and positive superiority over the mid-parent and high parent except for the non-significance of TZEI-4*TZEI-31 and TZEI-4*TZEI-157 for HPH (Table 6).
This suggests the likelihood of using these crosses for hybrid maize production.The MPH and HPH of all the  (Kara, 2001;Betran et al., 2003;Gissa et al., 2007;Flint-Garcia et al., 2009).The average MPH and HPH estimates for set of hybrids evaluated by Betran et al. (2003) across environments were 171 and 132%, respectively compared closely to approximate estimates of 179 and 139% observed in this study.The significant, positive and high heterosis expressed in F1 hybrids for grain yield revealed the preponderance of dominant gene action.This is buttressed by the significant SCA observed for grain yield.Hull (1945) was of the view that nonadditive effects (dominance and/or epistasis) were of greater importance for the expression of heterosis and that selection should be emphasized for specific combining ability (Sprague and Tatum, 1942).According to Sprague (1983) and Hill et al. (1998), accumulation of good dominant alleles and masking of deleterious effects of recessive alleles by their dominant alleles in the F1 as well as the superiority of F1 heterozygote at a number of its loci to both the homozygous parents have brought about the heterosis.Therefore, the exploitation of heterosis for higher grain yields from these set of single cross hybrids are a breeding advantage.

Conclusion
Important genetic materials, which can be utilized for succeeding breeding programmes, were identified.Across infestations, estimates of GCA revealed that TZEI-7 and TZEI-22 were good combiners for MSVD resistance and also TZEI-7, TZEI-22, TZEI-31 and TZEI-157 can be considered for higher grain yields.TZEI-4*TZEI-22, TZEI-22*TZEI-157, TZEI-7*TZEI-157 and TZEI-31*TZEI-157 were the best performing hybrids in terms of combining resistance or tolerance with high yield based on SCA effects and heterosis.Thus, they can be further evaluated in multi-locations for possible release for commercial production by farmers.TZEI-7*TZEI-157 and TZEI-31*TZEI-157 can be further improved for resistance by using them as females in the development of three-way cross hybrids so that their potential for high yields can be fully exploited.
"t" for MPH = √ "t" for HPH = √ where: F1 = mean of the hybrid, MP (mid-parent) = average of the two inbred parents, P1 and P2 = mean of the inbred parents, HP =

Table 1 .
Characteristics of maize genotypes selected for the study.

Table 2 .
Mean squares from combined ANOVA of 5*5 diallel analysis for maize streak disease virus severity mean score and grain yield across infestations.

Table 3 .
Mean squares from ANOVA of 5*5 diallel analysis for maize streak disease virus severity mean score and grain yield under natural infestation and artificial infestation.
*Significant at p < 0.05, ** Highly significant at p < 0.01, *** Highly significant at p < 0.001.GCA: General Combining Ability, SCA: Specific Combining Ability.mean of the high inbred parent, r = number of replications and EMS = error mean square.

Table 4 .
General combining ability (GCA) effects of parental inbred lines for MSVD severity and grain yield across, and under natural and artificial infestations.

Table 5 .
Specific ability (SCA) F1 hybrids for MSVD severity and grain yield across, and under natural and artificial infestations.

Table
Heterosis for grain yield in 20 hybrids across, natural and artificial infestations.