Resistance of genotypes of arabica coffee against rust on leaf of coffee plant in different locations in Province of North Sumatra, Indonesia

The most important disease in many coffee production countries is rust on the leaf of coffee plant (RLCP). RLCP epidemics indicated by severity and incidence depend on the host, location and the fungi. The purpose of this research was to study the resistance of genotypes of Arabica coffee against RLCP in different locations. A field experiment with seven genotypes of Arabica coffee in four locations in Province of North Sumatra of Indonesia was carried out using a randomized complete block design with three replications. This research revealed a significant interaction between the genotype and the location on incidence of rust on branch (IRB), incidence of rust on leaf (IRL) and severity of leaf due to rust (SLR). For IRB and IRL, the role of the location was more important than the role of the genotype. For SRL, the role of the interaction was more important than the role of the genotype and the location. Genotype GN-5 had the lowest SRL. Location-3 was not conducive to the growth and development of the fungus H. vastatrix. Pathogen races were very likely to differ between locations. Based on SRL, these pathogens were not aggressive in all locations.


Introduction
Rust on the leaf of coffee plant (RLCP) disease caused huge loss and unsustainability of coffee production.Fungus (Hemileia vastatrix Berk.and Br.) caused RLCP.This disease is an epidemic in Asia, Central America, South America and Africa [1].In Indonesia and North Sumatra Province, data of RLCP incidence and severity is not yet available comprehensively.Some research results in this province reported that this fungus covered some farms of arabica coffee (Coffea arabica L.) [2,3].
There are 24 races of H. vastatrix in Asia and Africa of which ten races are found in Indonesia, and Race II (v5) is the most common and widespread rust race in the world [4].
The theory of complete and incomplete resistance breeding is used [4].In the theory of complete resistance, a host plant with certain resistant gene (s) is full resistant to certain races with certain virulent gene(s).In the incomplete resistance theory, there are many genes or polygenes as resistance controllers so that RLCP indicators are strongly influenced by location and are a continuous variation.In this theory, plants can be expected to have durable resistance, meaning that they are resistant for a long period of time and are resistant to various locational conditions.Potentially resistant genotypes could be found in unknown genotypes of arabica coffee that are being cultivated or growing wild in an area such as North Sumatra Province.The purpose of this study was to examine the resistance of genotypes of arabica coffee to the fungus H. vastatrix in different locations.It was hypothesized that there was a significant interaction between genotype and locations for RLCP disease.The results of this work are expected to contribute to efforts to overcome RLCP disease.

Materials and methods
Experiment was carried out using a randomized complete block design with three replications (blocks) [5].Six plants of each genotype were planted in one row.The distance between plants and rows were 2.5 m and 3.0 m, respectively.During this research, plants grew without shade and no pesticides, insecticides, and fungicides.Spores of H. vastatrix were infected on the experimental plants.These spores were taken from the rust infected leaves of coffee plants that were growing 10-50 m away from the experimental site and were applied directly to the underside of the leaf.The infections of spores were carried out twice (in January and February 2015).Inoculation was given to one leaf in the middle part of each branch.All branches of each plant were innoculated.Four best vigorous plants per genotype per plot were selected as data sources.
Data collection (IRB, IRL, SRL) was carried out eight times (in July and November 2015 and 2016 and August and December 2017 and 2018.Genotype with IRB, or IRL or SRL in amount of >15, >5-15, >0-5 and 0% was interpreted as low, moderate, high and full resistance, respectively. Statistical analysis was carried out using a combined analysis of variance for randomized complete block design (Table 1) and comparison between values were conducted with Fisher's least significant difference (LSD) test [5].

Result and discussion
Analysis of variance showed that genotype difference in RLCP indicators was highly significant (Table 2).A highly significant difference among locations in IRB and IRL was found.Location difference in SRL was nonsignificant.PxL interaction in RLCP indicators was highly significant.
In IRB, location contributed the highest variance (VK = 72.61%)while genotype contributed the lowest variance (VK = 2.82%) (Table 3).In IRL, GxL interaction contributed the highest variance (VK = 52.33%)while the lowest variance was contributed by genotype (VK = 14.24%).In SRL, the highest variance was contributed by GxL interaction (VK = 72.42%)while no variance was contributed by location (VK = 00.00%).Genotype GN-5 performed the lowest IRB and SRL (Table 4).Genotype GN-7 showed the lowest IRL.About half of the genotype showed moderate resistance and the other half showed low resistance based on the SRL.Based on the IRB or IRL, all genotype had low resistance.RLCP = rust on the leaf of coffee plant, df = degree of freedom, KTHS = mean square, IRB = incidence of rust on branch, IRL = incidence of rust on leaf, SRL = severity of leaf due to rust, ns = not significant, * = significant at the 5% level of significance, ** = significant at the 1% level of significance  The interaction between genotype and location (Table 2) shows that genotypes have different changes in the level of RLCP indicators when the location changes so that the ranking of RLCP indicators changes.Because of these interactions, analysis and discussion should be continued for each location (location) using a plant disease triangle (Host-Location-Pathogen) analysis.
As hosts, genotypes that have different levels of RLCP indicators in the same location (Table 4) indicate that these genotypes have different resistance genes against the fungus H. vastarix.Due to the nature of the level of RLCP indicators in the form of continuous variation and significant interactions between the genotype and the location, genotypes might have polygenic resistance genes or incomplete resistance genes (quantitative resistance genes).Compared to other genotypes, genotype GN-1 and GN-4 might have more resistance genes because they did not show an increase in IRB, IRL and SRL throughout the year in all locations indicated by the insignificance of correlation coefficient between RLCP indicators with number of months after incubation in all locations (Table 5a, Table 5b).The performance of GN-1 and GN-4 indicates that coffee plants that have durable resistance may be found.The performances of GN-1 and GN-4 indicate that coffee genotypes that have durable resistance may be found.Such coffee genotypes are resistant for a long period of time and are resistant to various locational conditions.Based on SRL, GN-1 and GN-2 which have durable resistance have moderate to low resistance.Further research needs to be done to find genotypes that perform durable resistance and have high resistance.The chance of finding such a genotype is great because genotypes of existing cultivated arabica coffee in North Sumatra Province were rich in genotypic variabilily in resistance to RLCP [3].Existing arabica coffee cultivars could have a variation in RLCP severity [6,7].Moreover, resistant coffee plant could also have high beans production [7].This durable resistance is able to ensure the sustainability of Arabica coffee production in North Sumatra considering that the 12 districts that produce Arabica coffee have different locational climates, altitute and soil fertility.This sustainability is very necessary.
Loc-3 was not conducive to the growth and development of the fungus H. vastatrix because all genotypes in Loc-3 did not show an increase in IRB, IRL and SRL throughout the year as indicated by the insignificance of correlation coefficient between RLCP indicators with number of months after incubation in the Loc-3 (Table 5).From an agronomic point of view, Loc-3 is an ideal location for coffee cultivation in the context of durable resistance.
Pathogen races were very likely to differ between locations (pathogenic diversity) because RLCP indicators that developed throughout the year were different in different locations as indicated by the correlation coefficient of RLCP indicators with number of months after incubation (Table 5).In Loc-1 IRL developed in genotypes GN-3 and GN-6, in Loc-2 IRL developed in GN-2, GN-5, GN-7, in Loc-3 all three symptons did not develop, and in Loc-4 IRB developed in GN-6.From an epidemiological point of view, these pathogens were not aggressive in all locations as indicated by the insignificance of correlation coefficient between SRL with number of months after incubation throughout the year.Because there are ten races of H. vastatrix in Indonesia [4], in the future it is necessary to research what races are in these four locations..12-0.07x n = 8, RLCP = rust on the leaf of coffee plant, Loc-= location, r = coefficient correlation, IRB = brance rust incidence, IRL = incidence of rust on leaf, SRL = severity of leaf due to rust, r value at the 5% level of significance = 0.707, r value at the 1% level of significance = 0.834, ns = not significant, * = significant at the 5% level of significance, ** = significant at the 1% level of significance, y = RLCP indicator, x = number of month after incubation.
The life cycle of the fungus H. vastatris begins with the development of new spores, then spore dispersal, spore deposition, spore germination, leaf penetration and leaf colonization [8].Spore dispersal, which is the transfer of spores from one part of the plant to another part of the same plant or to another plant, is influenced by the location (wind, rain splashes, insects, coffee harvesters) and plants (leaf size) [5] [8].Based on the comparison of variance components (Table 3), the role of the location was greater than the role of plants in term of spore dispersal as indicated by the variance component of IRB (72.61%) and IRL (15.50%) which were greater than the variance component of plants (2.82% and 11.09% for IRB and IRL, respectively).
After the process of penetration into plant cells through stomata, H. vastatrix absorbs nutrients from plant cells to develop and reproduce by releasing new spores (sporulation) on the leaf surface, and then the cycle process repeats itself.The spores then infect the cells in the leaf, forming a collection of the fungus H. vastatrix.The more the fungus are, the more plant nutrients are absorbed by the fungus, and the wider the leaf surface that is covered or damaged, the more photosynthesis decreases.With this understanding, the severity is measured based on the area of the fungus collection on the leaf surface as measured by SRL.In this phase, fungal development is influenced by the location (temperature and soil moisture) and plant (fruit load and genetic resistance) [8]..04+ 0.37x n = 8, RLCP = rust on the leaf of coffee plant, Loc-= location, r = coefficient correlation, IRB = brance rust incidence, IRL = incidence of rust on leaf, SRL = severity of leaf due to rust, r value at the 5% level of significance = 0.707, r value at the 1% level of significance = 0.834, ns = not significant, * = significant at the 5% level of significance, ** = significant at the 1% level of significance, y = RLCP indicator, x = number of month after incubation.Based on the comparison of variance components (Table 3), the role of the interaction of genotypes with the locations is greater than the role of genotypes or the locations on severity, which is indicated by the variance component of PxL interaction (72.42%) which is greater than the variance component of genotype (18.46%) and the location (0 %).This interaction indicates that genotypes have different changes in the level of SRL indicators when the location (location) changes so that the ranking of SRL indicators changes.A genotype may have a lower SRL (e.g., GN-1) in a given location (e.g.Loc-3) than another genotype that has a higher SRL (e.g.GN-2) in this location (Loc-3), but the first genotype (GN-1) had a higher SRL than the second genotype (GN-2) in other growing location (Loc-4) (Table 2).Hence, genetic improvement of plants for resistance to H. vastatrix must be accompanied by the selection of the right growing location.As locational components consist of biotic components in addition to abiotic components, differences in races in different growing locations can be the cause of the interaction as described above that pathogen races are very likely to differ between locations due to interpretation on the correlation coefficient and regression between RLCP indicators of each genotype and number of months after incubation in each location (locations) (Table 5).
Between RLCP indicators in all locations generally showed a significant to highly significant positive correlation (Table 6a, Table 6b).The higher the dispersal ability was, the greater the severity ability was.IRB and IRL as indicators of dispersal ability could be used to estimate SRL as an indicator of severity.This result was in line with research result conducted by [9] who concluded that IRL could be used to estimate SRL.The research result of [3], however, showed that IRL had no phenotypic correlation with SRL meaning the selection of coffee genotypes for RLCP resistance based on IRB did not guarantee that genes controlling SRL will also be selected.Selection on the basis of IRL was very likely to exclude genes that control SRL.Since SRL is not always phenotypically and genotypically correlated with IRB and IRL [3], selection based on SRL was more appropriate to use.Moreover, the SRL describes the level of damage caused by the fungus H. vastatrix.The larger the SRL is, the higher the damage level will be, and the more fruit production will be lost..65 + 0.66x n = 8, RLCP = rust on the leaf of coffee plant, E = location, r = coefficient correlation, IRB = brance rust incidence, IRL = incidence of rust on leaf, SRL = severity of leaf due to rust, r value at the 5% level of significance = 0.707, r value at the 1% level of significance = 0.834, ns = not significant, * = significant at the 5% level of significance, ** = significant at the 1% level of significance, y = RLCP indicator, x = RLCP indicator.

Conclusion
Based on the plant disease triangle (Host-Location-Phatogen), RLCP epidemics (incidence and severity) depend on the coffee plant (host), locations and H. vastatrix (pathogen).Genotype and location showed a highly significant interaction in IRB, IRL and SRL.Genotype difference in RLCP indicators was highly significant.Genotype GN-5 had the lowest severity (SRL).Genotype GN-1 and GN-4 did not show an increase both in dispersal (IRB, IRL) and severity (SRL) throughout the year in all locations.About half of the genotype showed moderate resistance based on the SRL.Loc-3 was not conducive to the growth and development of the fungus H. vastatrix.Pathogen races were very likely to differ between locations (locations) (pathogenic diversity).These pathogens were not aggressive in all locations (locations).For IRB, location (location) was more important than genotype and GxE interaction.For IRL and SRL, interaction of genotype and location was more important than genotype and location (locations).Between RLCP indicators in all locations generally showed a significant to highly significant positive correlation.
The result of this research might contribute to coffee breeding for durable resistance against RLCP due to the performance of genotype GN-1 and GN-4 showing longterm resistance in diverse growing locations.However, it is necessary to research what races are in these locations.Further research is needed to find genotypes performing durable resistance and high resistance.

RLCP = rust on
the leaf of coffee plant, IRB = incidence of rust on branch, IRL = incidence of rust on leaf, SRL = severity of leaf due to rust, VK = variance component, # ) = negative value of variance component taken as zero.

Table 1 .
Expected mean squares, F-ratios and variance component for randomized complete blocks experiments combined over locations whereby location and genotype were considered fixed variables.
SV = source of variation, df = degree of freedom, EMS = Expected mean squares, KTHS = mean squares; VK = variance component, CV = coefficient of variation, l = number of locations, r = number of blocks, t = number of genotypes.

Table 2 .
Combined analysis of variance for RLCP indicators of seven genotype over four locations

Table 3 .
Variance component of RLCP indicators of seven genotypes over four locations

Table 4 .
RLCP indicators of seven genotypes in four locations RLCP = rust on the leaf of coffee plant, P = genotype, E = location (location), IRB = incidence of rust on branch, IRL = incidence of rust on leaf, SRL = severity of leaf due to rust, E = location.LSD0.05 = Fisher's least significant difference for the 5% level of significance.LSD0.05 = 9.09 for IRB, LSD0.05 = 7.19 for IRL, LSD0.05 = 3.20.The means that followed common letter in IRB, IRL, and SRL were not significantly different for the 5% level of significance based on Fisher's least significant difference test.

Table 5a .
Coefficient correlation and regression between RLCP indicator of each genotype and number of months after incubation in Loc-1, Loc-2

Table 5b .
Coefficient correlation and regression between RLCP indicator of each genotype and number of months after incubation in Loc-3 and Loc-4

Table 6a .
Coefficient correlation and regression between RLCP indicator of seven genotypes in Loc-1 and Loc-2

Table 6b .
Coefficient correlation and regression between RLCP indicator of seven genotypes in Loc-3 and Loc-4 (continued)