Recent progress on the genetics and molecular breeding of brown planthopper resistance in rice

Brown planthopper (BPH) is the most devastating pest of rice. Host-plant resistance is the most desirable and economic strategy in the management of BPH. To date, 29 major BPH resistance genes have been identified from indica cultivars and wild rice species, and more than ten genes have been fine mapped to chromosome regions of less than 200 kb. Four genes (Bph14, Bph26, Bph17 and bph29) have been cloned. The increasing number of fine-mapped and cloned genes provide a solid foundation for development of functional markers for use in breeding. Several BPH resistant introgression lines (ILs), near-isogenic lines (NILs) and pyramided lines (PLs) carrying single or multiple resistance genes were developed by marker assisted backcross breeding (MABC). Here we review recent progress on the genetics and molecular breeding of BPH resistance in rice. Prospect for developing cultivars with durable, broad-spectrum BPH resistance are discussed.


Introduction
Rice is the most important cereal crops in the Asia-Pacific region, particularly China, India, Japan, Indonesia, and Vietnam, where the brown planthopper (BPH, Nilaparvata lugens Stål) has become its most damaging insect pest. In 2005 and 2008 China reported a combined rice production loss of 2.7 million tons due to direct damage caused by BPH (Brar et al. 2009). Currently, the main method of controlling BPH is application of pesticides such as imidacloprid. However, the intensive and indiscriminate use of chemicals leads to environmental pollution, kills natural enemies of the target pest, may result in development of BPH populations that are resistant/tolerant to insecticides, ultimately leading to a resurgence in BPH populations (Lakshmi et al. 2010;Tanaka et al. 2000). Host-plant resistance is therefore most desirable and economic strategy for the control or management of BPH (Jena et al. 2006).
BPH is a migratory, monophagous rice herbivore. According to the length of the wing, adults BPH are biomorphic with varying wing lengths. The short winged cannot migrate, but produces larger amounts of eggs; BPH with long wings are able to fly between regions and bridge gaps in subsequent cropping seasons. The combined effect of the two types makes BPH an internationally explosive and devastating pest of rice. The differentiation of wing type is genetically controlled and a research group at Zhejiang University recently identified two highly homologous insulin receptor genes that play a key role in the wing differentiation (Xu et al. 2015).
Different biotypes (or races) of BPH vary in virulence (or ability to infest) different rice genotypes (Sogawa 1978). Four biotypes have been well known since the 1980s. In China, biotype 2 dominates, from the 1990s has sometimes been mixed with biotype 1 (Tao et al. 1992). However, the current population may be shifting to the more destructive Bangladesh type (Lv et al. 2009). New biotypes arise to overcome resistance genes prolonged use in a single widely used variety or suite of varieties with the same resistance gene (Cohen et al. 1997;Jing et al. 2012). For example, the first resistant variety IR26 possessing the Bph1 gene became susceptible of biotype 2 after only two years of use (Khush 1971). The genetic mechanism of BPH biotype generation in BPH is still not well understood, but there is overwhelming evidence from many plant disease/pest combinations that virulence involves the change or loss of specific effector proteins that are recognized by the plant host to induce the resistance (antibiosis) response.
Rice varieties have different mechanisms of resistance to BPH, classed as antixenosis, antibiosis and tolerance (Alam and Cohen 1998;Painter 1951). Antibiosis is the most commonly studied mechanism (Cohen et al. 1997;Du et al. 2009;Qiu et al. 2010). BPH behavior (host-searching, feeding, mating) is most obviously affected by resistant varieties through antibiosis. After infestation by BPH the rice plant activates its own stress response for defense, including secretion of insect-toxic compounds, activation of expression of genes producing metabolic inhibitors, and formation of physical barrier (such as cuticle thickening, and callose deposition) to prevent continuous feeding by BPH (Cheng et al. 2013). Hao et al. (2008 showed that plants carrying Bph14 undergo quicker deposition of callose on the sieve plate following infestation than those without the gene, suggesting that sieve tube plugging is an important mechanism for defense to BPH. Since the development of molecular markers (SSR, InDel, SNPs) and functional genomics, the genetic studies of BPH resistance in rice have intensified. To date 29 BPH resistance genes have been detected in rice, and four (Bph14, Bph26, Bph17 and bph29) have been cloned (Du et al. 2009;Liu et al. 2015;Tamura et al. 2014;Wang et al. 2015). Both marker-assisted selection (MAS) and conventional breeding have enabled resistance genes to be combined (or 'pyramided') in elite rice varieties to improve BPH resistance and its durability. We review here recent progress on BPH resistance genetics and molecular breeding in rice, aiming to help a wider utilization of BPH resistance genes.

BPH resistant germplasms Evaluation of BPH resistance
A thorough evaluation of BPH resistance in the abundant germplasms is critical for identification and utilization of BPH resistance genes (Jena and Kim 2010). Various evaluation methods were developed to measure response to BPH in rice varieties. Based on the type host:pest interaction, evaluation methods can be divided into two groups. The first directly evaluates host resistance by measuring the degree of damage following BPH infestation. The modified seedbox test (SSST test) is recognized as a standard method. SSST assesses damage to seedlings (leaf yellowing, plant withering and dwarfing) caused by the progeny of an initial infestation with a set number of nymphs (Panda and Khush 1995). It is suitable time-and space-saving assay for testing of germplasm and breeding materials. However, results from this test are affected by temperature, humidity, nymphs instar, density, biotype, population and natural enemies. The second approach indirectly the relative host response by examining the physiological and biochemical reactions of the BPH (feeding rate, fecundity and survival) feeding on different varieties. Parameters measured include honeydew excretion, survival rates, preference settling, and feeding behavior (Pathak et al. 1982;Sangha et al. 2008;Klignler et al. 2005). Some t evaluation methods attempt to address host tolerance using compensation ability and yield loss rate (Dixon et al. 1990;Alam and Cohen 1998). Ultimately, all possibilities for reducing the insect population or its fitness through use host genotype must be reconfirmed in laboratory/greenhouse trials and in the field.

Source of BPH resistance
Since the 1970s, a large number of germplasm accessions have been screened for response to BPH at the International Rice Research Institute (IRRI) by mass screening evaluation (Jackson 1997). After searching the Genesy database (https://www.genesys-pgr.org/zh/ welcome) maintained at IRRI we identified a total of 573 cultivated rice accessions that showed resistance to at least one BPH biotype. Among them, 484 accessions (92.5 %) showed resistance to biotype 1, and only 80 accessions (15.3 %) were resistant to all three biotypes (Fig. 1). Wild rice is a key source of resistant germplasm. Various species commonly show high resistance to all three biotypes. Eighteen species of wild rice, comprising 265 accessions, were highly resistant, Fig. 1 Frequencies of rice accessions resistant to different BPH biotypes. The data were partially selected and summarized from the Genesy database (https://www.genesys-pgr.org/zh/welcome). The total of 573 cultivated rice accessions showed resistance to at least one biotype. Biotype 1, Biotype 2 and Biotype 3 represent the number of cultivars only resistance to biotype1, 2 or 3 of BPH, respectively. 1 + 2 + 3 denotes the number of cultivars resistance to all three biotypes and two species (O. officinalis and O. minuta) accounted for 41 % of the total (Fig. 2).
The first BPH resistance was identified in 1967 (Pathak et al. 1969). Since then genes Bph1, bph2, Bph3 and bph4 have been identified in genetic analyses of various donors (Lakashminarayana and Khush 1977;Khush et al. 1985). These four genes have been used extensively in breeding programs in Southeast Asia (Jairin et al. 2007a), and a large number of BPH resistant varieties have been released by IRRI since 1976. However, some of them have lost effectiveness with the evolution and subsequent increase of new biotypes ( Table 1).

Genetics of BPH resistance
Mapped BPH resistance genes Twenty nine BPH resistance genes have been identified from ssp. indica and wild relatives (Ali and Chowdhury 2014;Wang et al. 2015). Most of these genes were located to specific rice chromosome regions, but the identities of a few (e.g. bph5 and bph8) are confusing because of the lack marker technology in early studies (Qiu et al. 2014). Since the development of molecular markers (such as SSR, InDel, and SNPs) and functional genomics increasing numbers of resistance genes have been fine mapped and some were cloned. To date, more than ten genes have been fine mapped to regions of less than 200 kb ( Table 2). Most of resistance alleles are dominant, but several are a few are recessive (bph4, bph5, bph7, bph8, bph19 and bph29).
All BPH resistance genes identified to date are from indica varieties and wild relatives. Bph1-Bph9, Bph19, Bph25-Bph28 are from indica accessions, wheraes Bph10-Bph18, Bph20, Bph21, Bph27 and bph29 are from wild rice species (Table 2). Introgression lines (ILs) derived from crosses of O. sativa and wild species have been used to map many of the BPH resistance genes (Jena and Khush 1990;Brar and Khush 1997). For example, Bph18, located on 12 L, was identified in IR65482-17-216-1-2, a BPH resistant IL derived from O. australiensis. Up to now, 11 genes have been identified in wild rice, including Bph11-Bph15 were from O. officinalis, Bph10 and Bph18 were from O. australiensis, Bph20 and Bph21 were from O. minuta, and Bph27 and bph29 were from O. rufipogon.
Multiple BPH resistance genes are clustered in a similar way to blast resistance genes (Jena and Kim 2010;Fig. 2 Frequencies of wild rice species accessions with resistance to BPH at IRRI. Data are summarized from a search of the Genesy database; 265 accessions (involving 18 species) showed high resistance to all three BPH biotypes Ramalingam et al. 2003). For example, eight genes (Bph1, Bph2, Bph9, Bph10, Bph18, Bph19, Bph21 and Bph26) are cluster in a 22-24 Mb region on chromosome 12 L, and six (Bph12, QBph4, QBph4.2, Bph15, Bph17 and Bph20) are closely linked in a region of 5-9 Mb on chromosome 4S. Another three genes together are located within a19-22 Mb on chromosome 4 L, and four are concentrated in a 0-2 Mb region on chromosome 6S (Table 2, Fig. 3). These QTLs/gene clusters might involve different genes, different alleles at a single locus, or even the same gene, but mediate different resistance mechanisms or show different response to different BPH biotypes (Qiu et al. 2010). Additional genetic analyses, including allelism tests and gene cloning are needed to resolve these possibilities.
Multiple BPH resistance genes/QTLs with the same names are also located to different positions. For example, Bph1 from three different donors (Mudgo, TKM6 and Nori-PL3) was mapped to different positions on chromosome 12 (Table 2). Bph26 was recently cloned and sequence comparison indicated that it is the same as Bph2 (Tamura et al. 2014). Discrepancies in genetic maps have caused duplicated nomenclature for the same gene. For example, Bph27 and Bph27(t) were fine mapped to the adjacent locations on the long arm of chromosome 4 (Huang et al. 2013;He et al. 2013), and it is possible that they might be different due to their different origins (derived from wild rice and a cultivated relative, respectively). According to the rules of genetic nomenclature for rice, it is necessary for the authors of different reports to rename duplicated genes to avoid confusion to readers. Bph3 and Bph17 each described as single Mendelian factors in the resistant cultivar Rathu Heenati (RH) by different research groups. The rice scientific community has accepted the findings as Bph17 on chromosome 4 (Rahman et al. 2009, Qiu et al. 2012 and Bph3 on chromosome 6 (Jairin et al. 2010, Myint et al. 2012). These reports acknowledged in review papers (Jena and Kim, 2010, Fujita et al. 2013, Cheng et al. 2013) and on the cereal crop GRA-MENE website (http://archive.gramene.org/documentation/nomenclature/) as well as Oryzabase (http:// www.shigen.nig.ac.jp/rice/oryzabase/). However, Liu et al. (2015) reported the gene chromosome 4 cloned from RH as'Bph3' when it actually originally reported as 'Bph17' (Sun et al. 2005). In our opinion the cloned gene on chromosome 4 (Liu et al., 2015) should have been reported as 'Bph17'.

Mapping of minor BPH resistance QTLs
Using different mapping populations (RIL, DH, F 2:3 ) from crosses of susceptible and resistant varieties, more QTLs were detected on all rice chromosomes except 5 and 9 (Alam and Cohen, 1998;Soundararajan et al. 2004;Liu et al. 2009;Ali and Chowdhury, 2014). However, those minor QTLs could not be confirmed due to the complex inheritance of the BPH resistance (Jena and Kim, 2010). Several studies showed that some highly resistant varieties carry many minor QTLs in addition to one or more major genes. Such combinations suggest possibilities for more durable resistance contributed by minor QTLs (Bosque-Perez and Buddenhagen 1992). For example, an elite variety IR64 from IRRI showed more durable and stable resistance than IR26, although The rice varieties or lines was selected from Genesy database (https:// www.genesys-pgr.org/zh/welcome) in IRRI. The information of genes that these lines carry was described as Ali and Chowdhury (2014) and Jena and Kim, (2010). The resistance data of these lines were obtained in our previous study of seeding resistance both carry Bph1. In a further seven minor QTLs were detected on chromosomes 1, 2, 3, 4, 6 and 8 from IR64 (Alam and Cohen 1998). Likewise, the Sri Lankan variety Rathu Henati has shown durable resistance to all four BPH biotypes in Southeast Asia since the 1970s, as it not only carries major genes Bph3 and Bph17, but also minor QTLs on chromosomes 2, 3, 4, 6 and 10 (Jairin et al. 2007a;Kumari et al. 2010;Sun et al. 2005). A recent study showed that indica cultivar ADR52 possesses two major genes Bph25 and Bph26, along with several minor QTLs associated with resistance to BPH, white-backed planthopper (WBPH) and green leafhopper (Srinivasan et al. 2015).

Map based cloning of BPH resistance genes
Gene identity helps to clarify the molecular mechanisms of BPH resistance. Advances in sequencing technology and functional genomics have facilitated BPH resistance gene cloning. To date, Bph14, Bph26, Bph17 and bph29 have been cloned by map-based cloning.
Bph14 is the first cloned BPH resistance gene originated from O. officinalis. Bph14 was originally fine-mapped to a Bph26 was cloned from indica variety, ADR52. Early study showed that ADR 52 carries two genes, Bph25 and Bph26 located on chromosomes 6S and 12 L, respectively (Myint et al. 2012). Like Bph14, Bph26 encodes a CC-NB-LRR protein that mediates antibiosis to BPH. Sequence comparison indicated that Bph26 is the same as Bph2, which was overcome by biotype 2. However, pyramiding of Bph25 and Bph26 could significantly improve BPH resistance, suggesting a valuable application in rice resistance breeding (Tamura et al. 2014).
Bph17 was cloned from Sri Lankan indica variety Rathu Heenati. Initially, Bph17 was fine-mapped to a 79 kb region containing four clustered genes on chromosome 4S. Transgenic tests showed that three genes independently confer resistance to BPH, and gene pyramided transgenic lines showed enhanced resistance. Bph17 is actually a cluster of three genes encoding plasma membrane-localized lectin receptor kinases (OsLecRK1-OsLecRK3), which collectively function to confer broad-spectrum, durable resistance and provide an important gene source for MAS and transgenic breeding for BPH resistance (Liu et al. 2015).
bph29, a recessive gene from O. rufipogon, fine-map to a 24 Kb region on chromosome 6S. Through a transgenic experiment, the bph29 allele from the susceptible variety was transferred into the resistant variety, and the positive progenies were susceptible, whereas the negative progenies retained high resistance. bph29 encodes a B3 DNA-binding protein. Expression patterns analysis showed that bph29 is restricted to the vascular tissue where BPH attacks. Expression of bph29 activates the SA signaling pathway and suppresses the jasmonic acid/ ethylene (JA/Et)-dependent pathway after BPH infestation and induces callose deposition in phloem cells, resulting in antibiosis to BPH (Wang et al. 2015).

Genes and TFs associated with BPH resistance
In addition to the traditional map-based cloning method, some genes and transcription factors (TFs) associated with BPH resistance have been identified through reverse genetics approaches such as T-DNA mutants and genes homology. Bphi008a is a resistance gene that is induced by BPH feeding; it is involved in ethylene signaling. Plants carrying a transgenic Bphi008a allele show significantly enhanced resistance to BPH (Hu et al. 2011b). Another two genes, OsERF3 and OsHI-LOX, are ethylene response factors and lipoxygenase genes, respectively, involved in a JA/Et-dependent pathway and act as inhibitor of the gene Fig. 3 Locations of BPH resistance genes on rice chromosomes expression to improve resistance to BPH (Lu et al. 2011;Zhou et al. 2009). With the development of rice genomics and proteomics, continued screening and validation of genes that are regulated by BPH feeding, and clarification of resistance mechanisms will promote research of BPH-associated genes and offer possibilities for resistance breeding.

Molecular breeding for BPH resistance
Since the 1970s, several BPH resistance genes such as Bph1, bph2, Bph3 and bph4 have been identified and transferred into elite susceptible varieties at IRRI, and a series of improved cultivars (e.g. IR26, IR36, IR50 and IR72) with BPH resistance were developed and released (Jairin et al. 2007a;Jena and Kim, 2010). However, the improved cultivars carrying single resistance gene lose effectiveness due to the evolution of new biotypes (Jena and Kim 2010). Therefore, to develop new varieties with more durable and stable BPH resistance, there has to be use of more genes, preferably pyramided into multiple gene lines or possibly deployed in multiline single gene mixtures such that new biotypes will be hampered or delayed.

Integrating MAS into conventional rice breeding
MAS greatly increases the efficiency and effectiveness of breeding. By determining and developing DNA markers for target genomic regions, desired individuals possessing particular genes or QTLs can be identified in germplasm collections based on genotyping rather than phenotyping (Collard et al. 2005). New strategy that fits breeder requirements should include a planned MAS strategy, MAS-based backcrossing breeding (MABC) and gene pyramiding (Fig. 4).

Construction of a MAS system for using BPH resistance genes
The efficiency of MAS largely depends on the distance between molecular markers and genes/QTLs associated with target traits. The development of useful markers tightly linked to target traits is accomplished by QTL mapping experiments. Generally, the markers are validated in fine mapping studies. Based on the positional information of BPH-resistance genes previously reported (Table 2), SSR and InDel markers adjacent to related genes were designed, and used to track the target genes in the segregating generation, and to test whether these markers were closely linked with genes. Thus, several The primary goal is development of useful markers tightly linked to target QTLs/genes by QTL mapping experiments (primary mapping, fine mapping and QTL validation). MABC include three generations of backcrosses and one generation of selfing, accompanied by positive and negative selection for minimizing the donor segments linked to target gene, and background selection for maximizing the recurrent genome. After phenotype evaluation of BC 3 F 2 lines, NILs containing single target gene are obtained. Multiple NILs that carrying different genes are crossed each other to produce pyramided lines. MAS based conventional breeding include 8-9 generations of selfing, accompanied by multiple cross within three parents, field and MAS selection in a large F 2 population, preliminary and further yield trials in F 3 and F 4-8 population. After phenotype evaluation, the F 8-9 progenies with enhanced target traits and high yield potential could be obtained, designated as 'improved versions' MAS systems with high efficiency associated with these genes were developed (Table 3).

MABC for BPH resistance
It takes a minimum of 6-8 backcrosses to fully recover a recurrent parent genome using conventional breeding methods, but MABC enables the procedure to be shortened to 3 or 4 backcrosses (Tanksley et al. 1989). There are three levels of selection in which markers are applied in backcross breeding (Fig. 5). Firstly, markers are used to select target alleles whose effects are difficult to observe phenotype (e.g. resistance in the absence of actual disease/pest tests), this is referred to as 'positive selection'; secondly, markers are used to select for progeny with the target gene and tightly-linked flanking markers in order to produce chromosomes that harbor the target allele with minimal surrounding DNA from the donor parent (minimizing linkage drag), designated as 'negative selection' or 'recombinant selection'; thirdly, markers that are distributed across all 12 rice chromosomes can be selected for recovery of the recurrent parent genome, known as 'background selection'. A typical example of MABC that used markers for all three objectives was performed by Chen et al. (2000).
MABC has been used to develop multiple BPHresistance introgressions (ILs) or near-isogenic lines (NILs). Using a Bph18-cosegergation marker 7312. T4A for positive selection, and 260 SSR markers across all rice 12 chromosomes for background selection, Bph18 was transferred into an elite japonica variety 'Junambyeo' and ILs with enhanced BPH resistance were developed (Suh et al. 2011). Using negative selection, linkage drag between Bph3 and Wx a alleles was successfully broken resulting in ILs with broad spectrum BPH resistance and good quality (Jairin et al. 2009). In our laboratory, a number of genes (Bph3,Bph6,Bph9,Bph14,Bph15,Bph10,Bph18,Bph20,QBph4,QBph3) were individually incorporated into 9311 (an elite variety in China) using MABC, and a set of NILs was developed with enhanced BPH resistance. These NILs harbor target gene regions of less than 100 kb and the recurrent parent genome (>99.5 %) was recovered with a breeding chip with high-density SNP markers for negative and background selection (data unpublished).

Pyramiding BPH resistance genes
Using MAS, we can simply and easily combine multiple genes/QTLs together into a single genotype simultaneously.  reported development of a new elite restorer line possessing tolerance to BPH, stem borer, leaf folder and herbicide through pyramiding Bph14, Bph15, Cry1C, and glufosinate-resistance gene bar.
In order to develop new cultivars with durable BPH resistance, we should not only use gene pyramiding, but exploit genetic diversity for ecological reasons. Zhu et al. (2000) reported an example of genetic diversity and blast disease control in rice. Furthermore, multiple NILs were developed representing all possible combinations of several blast resistance QTLs/genes from a durably resistant cultivar (Fukuoka et al. 2015;Khanna et al. 2015). Similarly, we have pyramided Bph14 and Bph15 into several different rice hybrids, and experiments indicated that planting resistant pyramided hybrids around conventional susceptible hybrids significantly reduced the overall population of BPH over a large field area, thereby reducing the BPH threat and contributing to sustainable of rice production (Hu et al. 2011a). Moreover, multilines (NILs, ILs, or PLs) carrying different assortments of genes should also help in containing BPH populations to manageable levels.

Conclusion and perspective
In the recent years, significant progress has been made in molecular breeding of rice for yield, quality, biotic and abiotic stress resistances and certain agronomic traits (Rao et al. 2014). However, the genetics rice: BPH interaction and molecular breeding for BPH resistance have been restrained due to the intricacy of interaction between rice and BPH. Host-plant resistance is an effective environmentally friendly approach to control BPH and maintain yield potential of cultivars (Jena and Kim, 2010). Future breeding approaches must focus on developing cultivars with durable, broad-spectrum resistance.
The first objective is to identify and characterize new resistance genes from diverse germplasm resources, particularly wild species. The second objective is to understand the molecular interactions between rice and BPH. We should not only accelerate research on map-base cloning of BPH resistance genes, but also pay attention to and the genome and genetics of BPH itself. The BPH genome was sequenced and genomes of BPH and its endosymbionts revealed complex complementary contributions for host adaptation (Xue et al. 2014). Mapping of the rice resistance-breaking gene of the BPH has facilitated understanding of interactions of BPH and rice Kobayashi et al. 2014). The third objective is to pyramid major genes or QTLs or to deploy NILs or ILs carrying multiple single resistance genes in multilines. Recently, molecular breeding design (MDB) have become popular for molecular breeding in crop improvement and should contribute to future breeding outcomes (Xu and Zhu, 2012). Molecular breeding designs for BPH resistance will involve three steps: (1) map all QTLs for BPH resistance by high-throughput genotyping and reproducible phenotyping; (2) evaluate and reconfirm allelic variation in these QTLs by development of NILs; and (3) conduct design breeding according to a bioinformatics platform and simulation studies. The final objective is to develop new varieties that contain the best genotypic combinations to confer durable resistance.