Genomic Analysis of the Snn1 Locus on Wheat Chromosome Arm 1BS and the Identification of Candidate Genes

MATERIALS AND METHODS

Plant Material and SnTox1 Screening

A segregating F₂ mapping population was developed from the cross between the SnTox1-sensitive hexaploid wheat cultivar Chinese Spring and the insensitive CS–T. dicoccoides chromosome 1B disomic substitution line (CS-DIC 1B). The CS-DIC 1B substitution line was maintained and provided by the Wheat Genetics and Genomics Resource Center at Kansas State University (Manhattan, KS). Saturation mapping was performed using 120 F₂ plants, and more than 16,000 F₂ plants were screened for reaction to SnTox1 for high-resolution mapping. Plants were grown in the greenhouse at an average temperature of 21°C with a 12-h photoperiod. The fully expanded secondary leaf was infiltrated with SnTox1 cultures that were partially purified as described by Liu et al. (2004a). Leaves were evaluated 3 d after infiltration and scored as either insensitive or sensitive.

In addition to CS and CS-DIC 1B, eight wheat lines were used to further evaluate the simple sequence repeat (SSR) markers that flank Snn1, including the genetic stock CS N1B T1D, the common wheat lines ‘BR34’ and ND495, the durum wheat lines ‘Altar84’, ‘Divide’, and ‘Langdon’, and the synthetic hexaploid wheat lines TA4152-60 and W-7976. CS, Altar84, Divide, TA4152-60 and W-7976 are sensitive to SnTox1 whereas CS-DIC 1B, N1B T1D, BR34, ND495, and Langdon are insensitive.

Saturation Mapping of the Snn1 Locus

DNA from the parents and the population of 120 F₂ plants was isolated according to Faris et al. (2000). The sequences of 120 ESTs that mapped to the 1BS.sat.18-0.50-1.00 deletion bin were downloaded from http://wheat.pw.usda.gov/cgi-bin/westsql/est_fasta.cgi. Primers for each EST were designed using Primer3 (Rozen and Skaletsky, 2000) and were amplified on parental DNA using the polymerase chain reaction (PCR) conditions outlined in Lu et al. (2006). Amplified products were separated on 6% MDE (Cambrex BioScience Rockland, Inc., Rockland, ME) gels to check for single stranded conformational polymorphisms (SSCPs) as described by Lu et al. (2006). For those EST primer pairs that showed no polymorphism between the parents, the amplified products were digested with the restriction enzyme RsaI and separated on 6% MDE gels. If EST primer pairs yielded no amplification or if no polymorphism was observed using the methods described above, their PCR products were used as probes and mapped as restriction fragment length polymorphism (RFLP) markers as described by Lu et al. (2006). For RFLP analysis, 25 μg of parental DNA was digested with 10 different restriction enzymes (ApaI, BamHI, BglII, DraI, EcoRI, EcoRV, HindIII, SacI, ScaI, and XbaI). Restriction patterns that revealed parental polymorphisms were selected, and the corresponding restriction enzyme was used to map the polymorphic markers in the F₂ population. Restriction digestion and Southern hybridization were performed according to Faris et al. (2000).

In addition to the EST-derived markers from 1BS.sat.18-0.50-1.00, seven RFLP probes (KSUD14, KSUE19, KSU941, BCD22, BCD1124, CDO99, and CDO1340), and two ESTs, BE403631 (mapped to the deletion bin 1DS and 1AS) and BE637553 (mapped to the deletion bin 1DS), were mapped in the population using the methods described above as SSCP and RFLP markers, respectively. Four SSR markers (Xfcp618, Xpsp3000, Xgwm11, and Xgwm273) were also mapped in the population using the PCR conditions described by Röder et al. (1998). Amplified products from SSR primers were electrophoresed through 6% nondenaturing polyacrylamide gels, stained with SYBR Green II (Sigma, St. Louis) and visualized using a Typhoon 9410 variable mode imager (GE Healthcare, Waukesha, WI).

RESULTS

Saturation Mapping and Marker Analysis

Primers were developed for 117 of the 120 ESTs in the 1BS.sat.18 deletion bin. Only 9 (7%) of the 117 ESTs were mapped as SSCP markers with no restriction digestion of the PCR products with RsaI, but the PCR product of one EST (BE406612) exhibited polymorphism when digested with RsaI. The remaining ESTs were surveyed for RFLPs between the parents. Results indicate that 31 of them exhibited polymorphism. Thus, a total of 41 (35%) of the ESTs were mapped in the CS × CS-DIC 1B population either as SSCP or as RFLP markers (Table 1).

Three EST markers (XBE498831, XBF474204, and XBF293222) cosegregated with Snn1 (Fig. 1). Three other EST markers (XBE422980, XBE637568, and XBE605202) cosegregated with each other 0.4 cM proximal to Snn1, and two EST markers (XBF145399 and XBF482862) cosegregated with each other on the distal side of Snn1.

One EST marker, XBE591617, detected two loci (XBE591617.1 and XBE591617.2) (Fig. 1), and both mapped to the Snn1 proximal end. The ESTs BE403631 and BE637553 were reported by the wheat NSF-EST project to detect loci on the short arms of chromosomes 1A and/or 1D, but not 1B. Our results indicate that both ESTs detected fragments on 1BS. Also, although EST BE606221 was reported to be in the 1BS.sat.18 deletion bin, it mapped near the centromere in our population. It is possible that fragments different from those that were physically mapped by the wheat NSF-EST project were polymorphic in our population.

Of the seven RFLP probes tested, KSUD14 detected fragments most closely linked to Snn1. KSUD14 contains a fragment of the Lr21 gene from chromosome arm 1DS (Huang et al., 2003), which confers resistance to the wheat leaf rust pathogen (Puccinia triticina Eriks). It detected multiple loci, two of which were polymorphic and mapped distal to Snn1 as markers Xksud14.2 and Xksud14.1 at genetic distances of 0.4 and 0.8 cM, respectively (Fig. 1). Of the four SSR markers, Xfcp618 and Xpsp3000 were the most closely linked to Snn1. Xfcp618 cosegregated with XksuD14.1 0.8 cM on the distal side of Snn1, and Xpsp3000 mapped 0.4 cM proximal to Snn1.

The genetic linkage map of chromosome 1B generated in the CS × CS-DIC 1B population of 120 F₂s contained a total of 54 markers spanning 64.6 cM (Fig. 1). Forty-two of the markers mapped at LOD ≥ 3.0; the remaining 12 were placed at the most likely positions along the map. The segment of the linkage map corresponding to deletion bin 1BS.sat.18-0.50-1.00 contained 47 markers and spanned a genetic distance of 56.1 cM, representing an average density of 1 marker per 1.1 cM. Within a 1.6-cM region encompassing the Snn1 locus, there were 11 markers, for an average of 1 marker per 0.15 cM.

Searches of the NCBI nonredundant database using the EST–TC and RFLP probe sequences as queries revealed that 29 sequences had significant homology to predicted proteins (Table 2). Of the three ESTs that cosegregated with Snn1, BE498831 and BF474204 had similarity to nucleotide binding site–leucine rich repeat (NBS-LRR)–like genes from Aegilops tauschii and BF293222 had similarity to a protein kinase from rice. The RFLP probes KSUD14 and KSU941 also had similarity to NBS-LRR–like genes from wheat.

Colinearity between 1BS and Rice

To evaluate colinearity between chromosome arm 1BS and rice, the sequences of all mapped RFLP and EST-based markers were subjected to BLAST searches against rice genomic sequences. The sequence for RFLP probe KSUE19 was not available, and KSUD14 and BE591617 each detected two loci; therefore, 47 sequences were evaluated. Of these, 30 had homology to rice genomic sequences on nine different chromosomes (Fig. 1). Among these, 12 marker sequences had similarity to sequences on rice chromosome 5. Comparison of the positions of the markers with similarity to sequences on rice chromosome 5 indicated a good level of colinearity with only a few exceptions (Fig. 2). Eight of the 12 markers were perfectly colinear, but markers XBE313657 and XBF291787 were inverted relative to each other and to the proximal five colinear markers XBE443931, XBE406612, XBG314374, XBF483372, and XBE443103. Also, markers XBF145399 and XBE405749, which flank Snn1, were inverted relative to each other. The sequences of rice chromosome 5 detected by the 10 most distal markers were all within the first 1.0 Mb of the short arm of rice chromosome 5 corresponding to the wheat deletion bin 1BS.sat.18-0.50-1.00.

Five marker sequences had similarity to sequences on rice chromosome 11, but among these, Xksud14, XBE498831, and XBF474204, which all have high degrees of similarity to known NBS-LRR–like genes, all had similarity to the same NBS-LRR–like gene on rice chromosome 11. Therefore, only three unique sequences of rice chromosome 11 were detected by the five marker sequences.

Four marker sequences had similarity to rice chromosome 8 sequences, and the remaining nine detected sequences on rice chromosomes 1 (2), 3 (1), 4 (1), 6 (2), 7 (2), and 10 (1) (Table 2, Fig. 1). Seventeen markers had no significant similarity to any rice sequences. Of these, only five had significant similarities to known proteins (Table 2).

High-Resolution Mapping

The 4255 insensitive F₂ plants were screened with the SSR markers Xfcp618 and Xpsp3000, which closely flank Snn1, to identify recombinants for high-resolution mapping. A population of this size allows the analysis of 8510 gametes and provides a resolution of 0.01 cM. Xfcp618 behaved as a dominant marker and detected a 250-bp fragment in CS, but no amplification was observed in CS-DIC 1B (Fig. 3). Xpsp3000 behaved as a codominant marker and detected a 281-bp fragment in CS and a 252-bp fragment in CS-DIC 1B. Screening of the 4255 plants resulted in the identification of 63 recombinants between Xfcp618 and Snn1 and 14 recombinants between Xpsp3000 and Snn1. Therefore, Xfcp618 and Xpsp3000 flanked Snn1 at genetic distances of 0.74 and 0.16 cM in the high-resolution population (Fig. 4).

We mapped Xksud14.1, XksuD14.2, XBE498831, XBF474204, XBF293222, XBE422980, XBE637568, and XBE605202 in the population of 77 recombinants. The results indicate that Xksud14.1 cosegregated with Xfcp618 at 0.74 cM from Snn1, whereas Xksud14.2 mapped 0.34 cM distal to Snn1 (Fig. 4). Marker XBF293222, which cosegregated with Snn1 in the low-resolution population, mapped 0.12 cM proximal to Snn1. Markers XBE422980, XBE637568, and XBE605202, which cosegregated with Xpsp3000 in the low-resolution population, also cosegregrated with Xpsp3000 in the high-resolution population. It is possible that these three markers actually map proximal to Xpsp3000, but because we used Xpsp3000 to select for recombinants, our recombinant population of 77 plants did not take into account recombination events proximal to Xpsp3000.

These results delimit Snn1 to a 0.46-cM interval flanked by Xksud14.2 and XBF293222 (Fig. 4). Markers XBE498831 and XBF474204 cosegregated with Snn1, indicating that these two ESTs, which have similarity to NBS-LRR–like resistance genes could be considered as putative candidates for Snn1.

DISCUSSION

The wheat pathogen Stagonospora nodorum causes SNB on wheat wherein multiple HSTs are involved in disease development (Friesen et al., 2007, 2008). These toxins are each recognized directly or indirectly by specific genes in the host, leading to compatible interactions and the development of disease. To date, four wheat–S. nodorum HST interactions have been characterized, and it has been shown that SnTox1 interacts with Snn1 (Liu et al., 2004a, b), SnToxA interacts with Tsn1 (Friesen et al., 2006; Liu et al., 2006), SnTox2 interacts with Snn2 (Friesen et al., 2007), and SnTox3 interacts with Snn3 (Friesen et al., 2008), indicating that multiple gene-for-gene interactions govern susceptibility of wheat to SNB. Understanding the molecular mechanism of host sensitivity to the toxins produced by S. nodorum requires the isolation of genes conditioning sensitivity in the host. Here, we developed saturated and high-resolution genetic maps of the Snn1 locus, assessed colinearity with rice, and identified markers closely flanking Snn1, which are prerequisites for positional cloning.

The previous physical mapping of Snn1 to deletion bin 1BS.sat.18 (Liu et al., 2004a) allowed us to exploit 117 ESTs that were bin-mapped by the wheat NSF-EST project. Not only did the use of bin-mapped ESTs provide a rich source of markers targeted to a chromosomal region of interest, but because they are derived from expressed genes, their sequence information can be exploited to obtain evidence of function, and they are ideal for conducting comparative mapping and colinearity studies. Of the 52 markers that mapped to the short arm of chromosome 1B in the CS × CS-DIC 1B population, 47 mapped within a 56.1-cM segment that corresponded to deletion bin 1BS.sat.18, which accounts for about 12% of the physical size of the arm. Five markers spanning only 2.7 cM mapped in the segment of the genetic map that corresponded to the rest of 1BS, which accounts for about 88% of the physical size of the arm. This agrees with previous studies in that a vast majority of recombination that occurs on wheat homeologous group 1 short arms occurs in the most distal portions (Gill et al., 1996b; Sandhu et al., 2001; Erayman et al., 2004). For example, Sandhu et al. (2001) reported that 85% of the recombination along group 1 short arms occurred in the “1S0.8” region, which corresponds to the 1BS.sat.18 deletion bin, and the remaining 15% occurred along the rest of the chromosome. Uneven distribution of recombination has been well documented in wheat and other organisms (Gill et al., 1993, 1996a, b; Tranquilli et al., 1999; Faris et al., 2000; Spielmeyer et al., 2000; Lu et al., 2006).

Studies have also reported that the genes in cereals are present in clusters encompassing physically small chromosomal regions (Gill et al., 1993, 1996a, b; Civardi et al., 1994; Descenzo et al., 1996; Wei et al., 1999) and that the distal regions of the short arms of wheat homeologous group 1 chromosomes are very gene rich (Gill et al., 1996b; Sandhu et al., 2001; Brooks et al., 2002; Erayman et al., 2004; Peng et al., 2004). Sandhu et al. (2001) reported that 75% of markers used in their study derived from expressed sequences mapped to the 1S0.8 region. Analysis of the Lrk gene-containing region at the distal end of 1AS indicated that gene density was as high as one gene per 5 to 20 kb (Feuillet and Keller, 1999).

Therefore, our data indicate that Snn1 lies within a gene-rich recombination hot spot of the wheat genome. The 1BS.sat.18 deletion bin where Snn1 resides represents the distal half of the chromosome 1BS satellite, which is approximately 0.5 μm in length and contains 34 Mb of DNA (Endo and Gill, 1996; Gill et al., 1996b). Taking into account that we have mapped 47 markers to this bin gives an average density of one marker per 723 kb, and because these markers accounted for 56.1 cM of genetic distance, the average physical to genetic distance ratio across the bin is 607 kb per cM, which makes the recombination frequency in this region sevenfold higher than the whole genome average of 4.4 Mb cM⁻¹ (Faris et al., 2000).

Colinearity between Snn1 and rice was studied to determine if information from the rice genome could be exploited to conduct further genomic analysis of the Snn1 region, develop additional markers, and possibly identify candidate genes. Our data indicate that of the 47 marker sequences evaluated, 30 (64%) had similarity to sequences on nine different rice chromosomes, but 12 (40%) of these had similarity to sequences on rice chromosome 5. Other studies have reported mosaic conservation of wheat group 1 chromosomes with rice but also that syntenic relationships exist between wheat homeologous group 1 chromosomes and rice chromosome 5 (Kurata et al., 1994; Van Deynze et al., 1995; Sorrells et al., 2003; Peng et al., 2004; Guyot et al., 2004). Sorrells et al. (2003) evaluated 12 ESTs mapped to 1BS.sat.18 and reported that 4 (33%) of them had similarity to sequences on rice chromosome 5, whereas the other 8 detected loci elsewhere in the rice genome. Peng et al. (2004) evaluated 2212 EST loci mapped using the homeologous group 1 deletion lines and reported that among 37 ESTs that mapped within the homeologous regions of the short arm terminal bins and had matches to rice sequences, 16 (43%) had similarity to sequences on rice chromosome five, whereas the other 21 had similarity to all the remaining 11 rice chromosomes with the exception of chromosome 9. Guyot et al. (2004) compared RFLP probe sequences from wheat 1AS with rice 5S and found good colinearity between the short arm of wheat chromosome 1A and rice chromosome 5. However, in silico comparative analysis of 1500 kb from physical contigs of wheat chromosome 1AS with rice chromosome 5S revealed much lower levels of colinearity. Therefore, our results regarding the colinearity between wheat chromosome arm 1BS and rice agree closely with others in that, overall, there is a low level of conservation between 1BS and rice, but some degree of colinearity between 1BS and rice chromosome 5 is evident.

The disruptions in colinearity between 1BS and rice chromosome 5 are primarily due to inversions, duplications, gene movement, and other rearrangements. Numerous disruptions near the Snn1 locus make it difficult to exploit the rice genome for further marker development and analysis of the Snn1 locus. Within a 2.0-cM region encompassing the Snn1 locus, there are 12 markers mapped, 2 of which are SSR markers (Fig. 1). Of the remaining 10, 4 (XBF482862, XBE422980, XBE637568, and XBE605202) had no similarity to rice sequences, 4 (XksuD14.1, XksuD14.2, XBE498831, and XBF474204) all had similarity to the same NBS-LRR–like gene on rice chromosome 11, 1 marker (XBF145399) had similarity to a sequence on rice chromosome 5, and 1 (XBF293222) had similarity to a sequence on rice chromosome 10, indicating that colinearity with rice is not well conserved near the Snn1 locus.

Our high-resolution genetic map delineated the Snn1 locus to a 0.46-cM interval between the markers Xksud14.2 and XBF293222 (Fig. 4). Given that the average recombination frequency of the 1BS.sat.18 bin is estimated to be 607 kb cM⁻¹, the distance between the closest flanking markers would correspond to approximately 279 kb. Taking into account the two markers that cosegregated with Snn1 gives an average marker density at the Snn1 locus of one marker per 93 kb.

User-friendly PCR-based codominant markers are highly desirable for conducting high-throughput genotyping and marker-assisted selection. Basic attempts were made to convert the EST markers that were closely linked with, or cosegregated with, Snn1 to more user-friendly platforms, but they were unsuccessful. However, Xfcp618 and Xpsp3000, which delineate Snn1 to a 0.9-cM segment, meet the criteria for high-throughput genotyping platforms and should be useful for such purposes. Marker-assisted selection of toxin insensitivity genes is especially efficient in backcrossing schemes because insensitivity is recessive and heterozygous genotypes cannot be distinguished from plants homozygous for the recurrent parent since both classes are sensitive. The use of marker-assisted selection allows one to distinguish these classes and select heterozygous genotypes for conducting further rounds of backcrossing.

The two ESTs BE498831 and BF474204, which cosegregated with Snn1, both had significant homology to NBS-LRR–like proteins (Table 2). NBS-LRR–like genes are viable candidates for genes governing sensitivity to the HSTs produced by S. nodorum. In oat (Avena sativa L.), the Pc-2 gene confers resistance to the rust fungus Puccinia coronata, but Pc-2-containing oats are susceptible to Victoria blight caused by Cochliobolus victoriae and sensitive to the C. victoriae–produced toxin known as victorin (see Wolpert et al., 2002, for review). The rust resistance gene Pc-2 and the victorin sensitivity gene Vb have not been separated genetically, indicating that either Pc-2 has pleiotropic effects or there are two different but tightly linked genes. A gene designated LOV1 was found to confer sensitivity to victorin in Arabidopsis and shown to be a CC-NBS-LRR–like gene (Lorang et al., 2007). Sequence analysis of loss-of-function mutants for LOV1 indicated that the nucleotide binding domain shared by Apaf-1, certain R gene products, and CED-4 (NB-ARC) (P loop) domain and, to a lesser extent, the LRR domain were important for function (Sweat et al., 2008).

Other evidence of toxin sensitivity genes with disease resistance-like features has also been reported. Nagy et al. (2007) constructed a sorghum BAC contig spanning the Pc gene, which conditions sensitivity to the Pc-toxin, and identified an NBS-LRR gene as a likely candidate. Therefore, the evidence to date suggests that disease-resistance genes, particularly NBS-LRR–like genes, are probable candidates for toxin sensitivity genes, and the toxin-producing pathogens may have evolved mechanisms to exploit R-gene receptors to attack the cells and cause susceptibility. Given that two ESTs with homology to NBS-LRR–like genes cosegregate with Snn1, we consider them both viable candidates for Snn1, and they will be a focus of future research. However, it is important to note that, given our estimate of 279 kb for the physical distance between markers flanking the Snn1 locus and the possibility that gene density could be as high as one gene per 5 to 20 kb in this region, it is possible that between 10 and 50 genes could be present within this interval. Although the NBS-LRR–like genes are attractive candidates, other candidate genes likely exist within the interval as well. Additional work, including the development of a BAC-based physical map and sequence analysis of the locus, is needed to determine the identity of the Snn1 gene.

A well-understood system similar to SnTox1 is ToxA from Pyrenophora tritici-repentis. ToxA and SnTox1 are both proteins and light dependent, and both interact with dominant sensitivity genes in the host (Liu et al., 2004a; Faris et al., 1996; Anderson et al., 1999). Manning and Ciuffetti (2005) showed that ToxA is internalized, probably via receptor-mediated endocytosis, and localized to the chloroplasts. In the chloroplast, ToxA was found to interact with ToxABP1 protein required for PSII function and thylakoid integrity (Manning et al., 2007). It is probable that both SnTox1 and ToxA utilize the same, or similar, pathways once internalized within the plant cell. It is possible that NBS-LRR-like genes provide the specificity for recognition of the toxins. The isolation of toxin sensitivity genes such as Snn1 and Tsn1 in wheat will provide more insights regarding the mechanisms of toxin recognition and toxin-induced cell death, as well as how their function is analogous to that of resistance genes and the evolutionary events leading to toxin sensitivity.

CONCLUSIONS

Here, we described the genomic analysis of the Snn1 locus on wheat chromosome arm 1BS. The use of unique wheat cytogenetic stocks such as CS N1B T1D and CS-DIC 1B allowed us to target markers to chromosome 1B and develop and use low- and high-resolution populations segregating for only the target chromosome. The wheat bin-mapped ESTs provided a rich source of markers, and the finished rice genome sequence allowed us to conduct comparative studies using mapped wheat EST sequences. Because colinearity between rice and the wheat Snn1 region appears to be poorly conserved, it will be difficult to use rice as a vehicle for cloning Snn1. However, through high-resolution mapping, we delineated Snn1 to a 0.46-cM region and identified two ESTs, which we consider plausible candidates for Snn1. Therefore, the stage is set for the map-based cloning of Snn1 using an appropriate BAC library, which will be a focus of future research.