Elite local rice varieties resistant to bacterial leaf streak caused by Xanthomonas oryzae pv. oryzicola under field conditions in Burkina Faso

Centre National de Recherche Scientifique et Technologique (CNRST), Institut de l’Environnement et de Recherches Agricoles (INERA), 01 BP 910 Bobo-Dioulasso 01, Burkina Faso. Ecole doctorale Sciences et Technologie, Laboratoire Biosciences, Equipe Phytopathologie et Mycologie tropicale, Université Joseph KI-ZERBO, 03 BP 7021 Ouagadougou 03, Burkina Faso. Institut de Recherche pour le Développement (IRD), Plant Health Institute of Montpellier (PHIM), 911, Av. Agropolis BP 64501 34394 Montpellier Cedex 5, France.


INTRODUCTION
Rice consumption is increasing due to population growth, increasing urban areas, and changes in eating habits. Meanwhile, the global supply of rice is declining due to reduction in the area available for rice cultivation in favour of other crops (biofuels, wood, etc.) and climate change leading to droughts and floods (SNDR, 2011). Although there is ample potential to increase rice production in Africa, rice imports represent a third of the total amount of rice traded on the world market. A number of measures are being implemented, however, by African countries to intensify rice cultivation.
In Burkina Faso, rice ranks fourth after sorghum, millet, and maize in terms of the area of cultivation, the amount produced, and the level of consumption. Indeed, national rice production only covers less than half of the population's consumption needs, which are estimated to be 475,000 tonnes of milled rice annually. Thus, efforts are being made to increase the national production of rice through irrigation schemes, the use of improved varieties, and promotion of the rice sector (Presao, 2011).
In Burkina Faso, BLS is present in the main ricegrowing areas of Bagre and Itenga in the Central-Eastern region, Bama and Banzon in the Hauts-Bassins region, Niassan and Di in the Boucle Mouhoun region, and Karfiguela and Douna in the Cascades region (Wonni et al., 2011(Wonni et al., , 2014Barro, 2015;Barro et al., 2021). BLS symptoms consist of water-soaked lesions that develop into translucent yellow streaks with visible exudates at the leaf surface. BLS develops in the field at any growth stage of rice. Xoc is an intercellular pathogen that enters plants through wounds or by invading open stomata (Ou, 1985). It then multiplies in the substomatal chamber and colonizes the apoplast of the mesophyll cells (Mew, 1987;Niño-Liu et al., 2006). Xoc oozes from natural openings in strands or strings on the leaf surface, and exudates can spread the disease from plant to plant by direct contact or indirectly via irrigation water and by windblown rain (Mew et al., 1993). Xoc is a seed-borne and a seed-transmitted pathogen (Xie and Mew, 1998). Yield losses due to this disease depend on the rice variety being cultivated and the climatic conditions, but typically range from 10 to 20% (Ou, 1985). Although significant yield losses have not yet been observed in Burkina Faso, BLS has a high leaf incidence of up to 100% in certain rice plots of the most irrigated sites (Wonni, 2013;Zougrana, 2017).
In light of the BLS distribution and its prevalence in the main rice-growing sites, there is a need to develop and/or identify resistant rice genotypes that are adapted to different cultivation areas.
Indeed, Wonni et al. (2016), under greenhouse inoculation conditions, identified local varieties of rice with Zougrana et al. 147 a broad spectrum of resistance to the various African Xoc strains. However, these varieties remain to be evaluated in a field environment in order to assess their resistance and stability to BLS. The aim of this study was to identify rice genotypes that are resistant to BLS.

Study sites
The tests were carried out at rice-growing sites known for their previous infestation with BLS disease reported by Wonni et al. (2011Wonni et al. ( , 2014 and Zougrana (2017), which are the irrigated plains of the Kou valley and the Di plains. The Kou valley plain is located at 30 km from Bobo-Dioulasso in the rural municipality of Bama at an altitude of 300 m above sea level between longitude 04°22'W and latitude 11°22'N. It extends over 1,200 ha with a total water control (Sontie, 2006). The climate is typical of southern Sudan, with annual rainfall ranging from 1100 to 12,000 mm (Yameogo et al., 2013).
The Di irrigated plain is located in the northwest of Burkina Faso at 326 km from Bobo-Dioulasso. It covers an area of 2,240 ha with total water control. The area lies at an altitude of 277 m above sea level between longitude 3°20'W and latitude 13°18'N. The climate is typical of northern Sudan, with annual rainfall ranging from 600 to 900 mm.

Rice genotypes tested
Nine varieties of rice whose phenotypes against African Xoc strains were evaluated by Wonni et al. (2016) under artificial inoculation conditions were tested under field conditions. Two rainfed and three irrigated/lowland varieties used by producers were included. The cultivars CG14, WAB56-50, and WAB181-18, which are the parents of NERICA varieties FKR45N and FKR49N, were also tested. The choice of these varieties is justified by their adoption by producers and consumers in Burkina Faso (Table 1).

Experimental design
The tests were set up in the Kou valley and the Di plains from July 15, 2019 to 2020 in one farmer's field per site where BLS infection was observed during the wet seasons in 2017 and 2018. The experimental design was a Fisher block randomized to three replicates separated from each other by a distance of 1 m. The main factor that was assessed was the varieties and the second factor was the disease incidence. Each elementary plot had an area of 4 m 2 , separated from each other by a distance of 0.5 m. The total area of the experimental design was 176 m 2 . The good rice cultivation practices recommended by the national research agency were scrupulously applied.

Data collection
Several parameters were collected at each site to assess the degree of resistance or susceptibility of the various genotypes tested.
(i) The disease incidence was determined for each plant from the 30th DAT, and then every 14 days until maturity. This consisted of ; where n = the number of repetitions, xi = the number of diseased plants per elementary plot, and X = the total number of rice plants per elementary plot.
(ii) The foliar incidence was evaluated for 10 plants chosen at random on the two diagonals in each elementary plot. It was calculated by counting the number of infected leaves out of the total number of leaves according to the following formula: where n = the number of repetitions, = the number of diseased leaves/plant, and X = the total number of rice leaves.
To determine the resistance level of each genotype, IRRI scale (2002) was used.
(iii) The epidemic growth rate (r) was expressed in units per day and assessed using the formula described by Rapilly (1991): where x1 and x2 denote the disease severity expressed as a percentage and t2 -t1 = the days between two observations. (iv) The disease severity (S) was evaluated for the 10 plants chosen to estimate the disease incidence. The severity (S), expressed as a percentage of the total tissue area, was calculated by using the scale of Kauffman et al. (1973) as follows: S = [(n1x1) + (n3x3) + n5x5) + n7x7) + n9x9)] x100 / (n1 + n3 + n5 + n7 + n9) x9; where n1 to n9 are the numbers of leaves denoted from 1 to 9.
(v) Paddy yield: The three central lines of each elementary plot were harvested at maturity. The panicles were dried in the sun and were then seeded and the seeds weighed. The average yield per genotype was determined by calculating the average paddy yield of the three elementary plots of each genotype tested. (vi) Climate data were collected at the meteorological station of the Kou valley and the Di plains. The temperature, hygrometry, and rainfall were recorded from June to November of each year.

Data analysis
Microsoft Excel 2010 software was used for data entry and to calculate the incidence, severity, and growth rate of BLS. Statistica 7.1 software was used for ANOVA tests and to establish the correlation between severity and yield. The comparison of averages was done by the Newman Keuls test at the 5% level.

Incidence per plant
Irrespective of the site and the year of cultivation, the first symptoms appeared as of the 30th DAT with low incidence (5.2%) and progressed over time to reach higher levels (94 to 100%) by the 72nd DAT on all of the susceptible genotypes. Thus, two genotypes groups could be distinguished according to their behavior against BLS. Group 1, which included the WAB181-18, FKR19, FKR45N, and FKR49N genotypes, comprised those that were resistant to BLS. Group 2 comprised the susceptible genotypes, such as the TS2, FKR62N, CG14, IR64, and WAB56-50 genotypes. However, their susceptibility varied according to the site, ranging from 58.33 to 100% (Figure 1).

Leaf incidence
The leaf incidence was significant for all of the susceptible varieties, irrespective of the site and the season at both sites. These comprised the IR64, FKR62N, TS2, and CG14 genotypes. In the Kou valley plot, the highest leaf incidence was recorded with CG14 (91%). In the Di plains plot, the CG14, FKR62N, and WAB56-50 genotypes were the most infected, with 99.63, 95.3, and 100% foliar incidence, respectively. Despite heavy   pressure from BLS, FKR45N,and FKR49N exhibited no symptoms at either site during the two experimental seasons (Table 2).

Average yield
The average yield of the tested genotypes varied from season to season and between the two sites. The lowest yields were recorded for FKR45N, FKR49N, WAB181-18, and CG14, between 2.02 and 4.75 t/ha. However, FKR62N, FKR19, and TS2 had the best yields, varying from 6 to 6.82 t/ha in the Kou valley versus 3.4 to 6.95 t/ha in the Di plains plot. Table 2 shows the average yields obtained by genotype at each site and by study year. The correlation analysis between the disease severity and the yields showed a strong overall negative correlation that was very highly significant (r = -0.74; p = 0.00014) (data not shown). As the severity level of BLS influences the yield, we observed that the yield was low when the severity was high.

Disease severity
The results show that the disease severity correlated with the disease incidence. In general, for all of the susceptible genotypes, the severity was greater in the Di plains plot (31.3 to 68.66%) than in the Kou valley which varied from 14.30 to 63.33%. Of these, WAB56-50 and CG14 were the most severely infected at both sites (Table 2).

BLS growth rate on the susceptible varieties
Interestingly, we noticed that the growth rate of BLS varied according to the vegetative stage of the susceptible genotypes, which were IR64, FKR62N, TS2, CG14, and WAB56-50. At the tillering stage, the growth rate of BLS was low in the Kou valley plot (0.007 ≤ r1 ≤ 0.014) and at Di (0.35 ≤ r1 ≤ 0.6). From maximum tillering to flowering, the growth rate of BLS (r2) increased significantly at both sites (0.18 ≤ r2 ≤ 0.503 in the Kou valley and 0.42 ≤ r2 ≤ 1.49 at Di). At panicle initiation, the growth rate of BLS was greatly decreased for varieties IR64, WAB56-50, TS2, and FKR62N in the Kou valley, except CG14 C, for which the BLS symptoms increased. However, at this phase in the Di plot, CG14 and TS2 exhibited the highest growth rate (r3) (data not shown).

Relationship between temperature, humidity, and BLS incidence
Linear regression analysis between the BLS incidence and the climatic factors revealed a very significant regression. Figure 2A shows a polynomial curve whereby the disease incidence increased from 12.5 to 61.56% as the temperature varied from 13.07 to 20.32°C. Figure 2B shows that the BLS incidence increased when the minimum and the maximum temperatures were close to 20.32 and 40°C, respectively. Figure 2C and D indicates that the BLS incidence increased with humidity, amounting to between 40 and 95%.

DISCUSSION
During the two years of the study, the IR64, FKR62N, TS2, WAB56-50, and CG14 genotypes displayed differential reactions to BLS according to the site and the year, unlike the FKR19, WAB181-18, FKR45N, and FKR49N genotypes, which exhibited resistance. Several factors could be explained the variations of varieties susceptibility observed on the both site. The Di site was developed in 2015, and is full of several weed hosts, including Oryza longistaminata, both within and along the edges of some plots. In addition, this site borders the Sourou River whose banks are mainly populated by O. longistaminata. However, the plain of the Kou Valley was developed in the 1960s and is less invaded by O. longistaminata. Also, producers grow fewer varieties there; in contrast in Di site, where several varieties are produced, sometimes with introductions from neighboring countries such as Mali.
In addition, the variations observed in the behavior of the susceptible varieties relate to one of their intrinsic qualities, which is the absence of an effective resistance gene. These results are consistent with those of Wonni et al. (2015Wonni et al. ( , 2016 who showed that these rice genotypes were highly susceptible to BLS under artificial inoculation conditions. Cultivars WAB56-50 and CG14, which belong to the glaberrima species, were found to be susceptible like FKR62N, which is an interspecific derived from the cross between cultivars TOG5681 and IR64, which are both susceptible to BLS. Therefore, the cultivation of FKR62N and TS2 varieties requires the application of good agricultural practices aimed at mitigating the effect of BLS on their potential yield. These will include the use of healthy seeds, the rational use of nitrogen, and the control of weeds in general and in particular those which are potential reservoirs of Xoc. Indeed, Bradbury (1986) and Wonni et al. (2014)  Moreover, national research should consider an improvement program to develop resistance of these varieties to BLS while preserving their potential productivity.
Interestingly, the FKR19, WAB181-18, FKR45N, and FKR49N genotypes were confirmed to be resistant to BLS, as reported by Wonni et al. (2015Wonni et al. ( , 2016. Despite the diversity of Xoc strains identified at the Di plains and the Kou valley sites (Wonni et al., 2014), these rice genotypes harbored one or more resistance genes. Indeed, these varieties, screened under artificial inoculation conditions, exhibited hypersensitive reactions.
While the WAB181-18, FKR45N, and FKR49N Zougrana et al. 151 genotypes remained immune to BLS infection, the FKR19 genotype nevertheless exhibited symptoms with a very low incidence (≤ 0.03%). These results may indicate the presence of more than one gene responsible for the FKR19 phenotype in regard to BLS. In contrast, the immunity of the resistance genotype could be due to a specific resistance gene. These varieties have a japonica genetic background and are suitable for rainfed rice cultivation, except FKR19. This adaptability to upland ecology may explain the low yields recorded for these genotypes in our study. Therefore, these results are more interesting as they reveal, for the first time, resistant rice varieties in greenhouse and field conditions in Burkina Faso. Zhao et al. (2004) reported that a resistance gene against Xoc had yet to be characterized in cultivated rice. However, to control BLS in Asia, a dominant maize gene, Rxo1, has been isolated and characterized. It confers resistance in maize to Xoc and it also prevents the development of Xoc when it is expressed as a transgene in rice (Zhao et al., 2005). Recently, a recessive resistance gene called bls1 was localized on chromosome 6 of Oryza rufipogon (He et al., 2012). In addition, Triplett et al. (2016) were able to determine that the resistance of the Carolina Gold rice variety is conferred by a single dominant locus, Xo1, located on a fragment of DNA of 1.09 Mbp on chromosome 4.

Conclusion
This study aimed to assess the behavior of different rice genotypes against BLS in natural infection conditions. The results show that four varieties, namely FKR19, WAB181-18, FKR45N, and FKR49N were resistant against BLS. However, these varieties are more suitable for rainfed rice cultivation and are not highly productive, except for FRK19, which is compatible with lowland and irrigated rice-growing systems. On the other hand, the TS2 and FKR62N varieties, which constitute the two most cultivated and consumed varieties in Burkina Faso, were shown to be highly susceptible to BLS. Therefore, identification of effective resistance genes against Xoc strain diversity, and improvement of elite susceptible varieties against BLS, remain essential in light of the spread and incidence of this disease in irrigated rice cultivation in Burkina Faso and in West Africa in general.

CONFLICT OF INTERESTS
The authors have not declared any conflict of interests.

ACKNOWLEDGMENT
This work was carried out with financial support from IRD,