- © 1998 American Society of Plant Physiologists
Abstract
The Mi locus of tomato confers resistance to root knot nematodes. Tomato DNA spanning the locus was isolated as bacterial artificial chromosome clones, and 52 kb of contiguous DNA was sequenced. Three open reading frames were identified with similarity to cloned plant disease resistance genes. Two of them, Mi-1.1 and Mi-1.2, appear to be intact genes; the third is a pseudogene. A 4-kb mRNA hybridizing with these genes is present in tomato roots. Complementation studies using cloned copies of Mi-1.1 and Mi-1.2 indicated that Mi-1.2, but not Mi-1.1, is sufficient to confer resistance to a susceptible tomato line with the progeny of transformants segregating for resistance. The cloned gene most similar to Mi-1.2 is Prf, a tomato gene required for resistance to Pseudomonas syringae. Prf and Mi-1.2 share several structural motifs, including a nucleotide binding site and a leucine-rich repeat region, that are characteristic of a family of plant proteins, including several that are required for resistance against viruses, bacteria, fungi, and now, nematodes.
INTRODUCTION
Root knot nematodes comprise a group of endoparasitic roundworms that cause major economic damage to crops around the world (Williamson and Hussey, 1996). These microscopic organisms penetrate the roots of thousands of plant species and migrate to the vascular cylinder, where they initiate a series of changes in the root, resulting in the formation of galls (or root knots) as well as the development of specialized feeding cells, called “giant cells,” in their hosts. These alterations grossly affect nutrient partitioning and water uptake in the host.
Many modern tomato varieties carry a single, dominant gene called Mi. This gene confers resistance to three of the most damaging species of root knot nematodes (Meloidogyne spp). This gene has been a classic example of the use of host resistance to reduce the need for pesticide application (Medina-Filho and Tanksley, 1983; Roberts et al., 1986). Mi was introduced into cultivated tomato, Lycopersicon esculentum, from its wild relative L. peruvianum in the early 1940s (Smith, 1944). With the assistance of linked markers, beginning with the isozyme marker Aps-1 and, more recently, with DNA markers such as Rex-1, Mi has been incorporated into many modern tomato cultivars (Medina-Filho and Tanksley, 1983; Williamson et al., 1994a).
How a single gene can mediate resistance to a nematode and interfere with the establishment of the elaborate changes that the parasite causes in its host has long been a question of interest (Williamson et al., 1994b). Microscopic studies have provided some information on the mechanism of resistance (Dropkin, 1969; Paulson and Webster, 1972; Ho et al., 1992). Nematodes are attracted to and penetrate roots, and they then migrate to the feeding site in a similar manner in resistant and susceptible plants. However, in resistant plants, there is no development of the feeding site. Instead, a localized tissue necrosis or hypersensitive response (HR) occurs at or near the site where feeding normally would be initiated. Nematodes that fail to establish feeding sites either die or leave the roots.
Resistance to diverse pathogens, including viruses, bacteria, fungi, and nematodes, has been shown genetically to be mediated by single, dominant resistance genes (R genes) in the host that are effective only if an avirulence gene is present in the pathogen (Flor, 1955; Keen, 1990). Such gene-for-gene interactions are characterized by commonalities in response, frequently including the presence of an HR (Hammond-Kosack and Jones, 1996). Recently, R genes have been cloned from several different plant species (reviewed in Bent, 1996; Dangl and Holub, 1997; Hammond-Kosack and Jones, 1997; Parker and Coleman, 1997). Most encode proteins that carry a structural motif with a repeating pattern of 20 to 30 amino acids called a leucine-rich repeat (LRR). LRR motifs participate in protein–protein interactions in a wide range of organisms (Kobe and Deisenhofer, 1995; Jones and Jones, 1996). Of the R genes characterized thus far, only Pto does not include this motif. Pto is a tomato gene that encodes a serine/threonine protein kinase and confers resistance to strains of Pseudomonas syringae carrying the avirulence gene avrPto (Martin et al., 1993). However, Pto requires the presence of a second gene, Prf, which does contain a characteristic LRR region, to confer resistance (Salmeron et al., 1996). LRR-containing R genes can be subdivided into two broad classes—those in which the predicted gene product contains an N-terminal, extracellular LRR and a membrane anchor, and those in which the R gene product is predicted to be cytoplasmic. Examples of the former class include Cf-2, which confers resistance to Cladosporium fulvum in tomato (Dixon et al., 1996), Xa21, which confers resistance to Xanthomonas oryzae in rice (Song et al., 1995), and Hs1pro-1, a gene conferring resistance to the sugar beet cyst nematode (Cai et al., 1997). Cytoplasmically located R gene products include those encoded by the virus resistance gene N from tobacco, the Arabidopsis genes RPS2 and RPM1, which confer resistance to specific strains of Pseudomonas spp, the flax rust resistance genes L6 and M, and the Arabidopsis RPP5 gene, which confers resistance to downy mildew fungus (Bent et al., 1994; Mindrinos et al., 1994; Grant et al., 1995; Ellis et al., 1997; Parker et al., 1997). These gene products are characterized further by the presence of a conserved region of ~260 amino acids containing a nucleotide binding site and a C-terminal LRR region. The presence of the nucleotide binding site suggests that binding of ATP or GTP is needed for the function of these genes.
The cloning of Mi has been an important goal for two general reasons. First, the gene would provide a starting point for understanding the basic biology of plant resistance to a parasitic animal and the relationship of Mi to other pathogen R genes. Second, Mi could be introduced into many other crops that can be seriously damaged by root knot nematodes and for which no genetic sources of resistance have been identified. Efforts to localize the Mi gene have been hampered for many years because of the severe repression of recombination near this gene in L. esculentum lines carrying the introgressed L. peruvianum DNA (Messeguer et al., 1991; Ho et al., 1992; Liharska et al., 1996). Recently, this handicap has been circumvented by screening large populations of tomato for recombinants and by identifying recombinants within L. peruvianum populations (Kaloshian et al., 1998). When data from L. esculentum and L. peruvianum recombinant analyses have been combined, Mi has been localized to a region of the genome of <65 kb, as diagrammed in Figure 1A. Here, we address the question of where Mi is located within this region by sequencing 52 kb of contiguous DNA. From the two candidate genes identified, one was shown by complementation analysis to represent the active root knot nematode resistance gene.
RESULTS
Identification of R Gene Homologs by Using Bacterial Artificial Chromosome Cloning and Sequencing
A bacterial artificial chromosome (BAC) library was made from a yeast strain carrying YAC2/1256, which includes the entire region to which Mi was localized (Kaloshian et al., 1998). Three clones, BAC1, BAC2, and BAC3, were identified by hybridization with either of the Mi-flanking DNA probes C32.1 or C93.1. The centromere-proximal end of BAC3 was subcloned to produce the clone B3E. B3E cross-hybridized with BAC1 and BAC2 on DNA gel blots, indicating that the BACs form an overlapping contig spanning the mapped Mi region, as shown in Figure 1B. B3E was then used as a probe to identify BAC4.
Mi Region of the Tomato Genome.
(A) The ~650-kb region of introgressed L. peruvianum DNA to which Mi was localized is indicated by a black bar. Cen and Tel indicate the directions toward the centromere and telomere, respectively. Recombinant analysis (Kaloshian et al., 1998) further localized Mi to the 65-kb region between markers C32.1 and C93.1. The white bar represents the position of YAC 2/1256, which spans the Mi region.
(B) The BAC contig identified using DNA probes C32.1, C93.1, and B3E. The 52-kb region for which the DNA sequence was determined is indicated by the wide black bar. Arrows 1, 2, and 3 represent the positions of the three R gene homologs. RBP, HAT, and A-EST indicate open reading frames (ORFs) with similarity to retinoblastoma binding protein, a transposase of the HAT family, and an Arabidopsis expressed sequence tag, respectively.
Large-scale sequencing was performed with BAC3 and BAC4. In all, 52 kb, including the entire 50-kb BAC3 insert, an adjoining 2 kb from BAC4, as well as 20 kb of sequence common to the two BACs, was assembled into a contiguous sequence (Figure 1). Six open reading frames of at least 400 nucleotides were found in the 52-kb region. Three of these are homologous to each other and to previously identified R genes of the nucleotide binding–LRR class. Highest homology was to the tomato gene Prf (Salmeron et al., 1996). Two of the three open reading frames had ~95% identical residues; these candidate genes were designated Mi-1.1 and Mi-1.2, respectively. The third is an apparent pseudogene because it lacks both the N- and C-terminal coding sequences and contains a deletion and several nonsense codons relative to Mi-1.1 and Mi-1.2. Three additional open reading frames in the region contain sequences with similarity to (1) the Arabidopsis expressed sequence tag PRL2-89B9T7 (Newman et al., 1994), (2) the Smcy/jumonji/XE169/retinoblastoma binding protein family of transcriptional activators (Fattaey et al., 1993; Agulnik et al., 1994; Wu et al., 1994; Takeuchi et al., 1995), and (3) the transposase of the HAT family (Hobo, Activator, and Tam3) of transposable elements (Calvi et al., 1991). The positions and orientations of these genes are indicated in Figure 1B.
Analysis of restriction digests and DNA gel blots of yeast artificial chromosome (YAC) and BAC clones indicated that C93.1 is ~15 kb to the centromeric side of the sequenced 52 kb (Figure 1). To determine whether additional sequences highly similar to Mi-1.1 were present in this region, probe 3-3, containing a 480-bp fragment having the nucleotide binding region of Mi-1.1, was developed. DNA gel blot hybridizations using this probe did not reveal additional bands on BAC1 or BAC2 that were not already represented in the sequenced region (data not shown).
Mi-1.1 and Mi-1.2 Are Transcribed
An RNA gel blot containing poly(A)+ RNA isolated from the roots of nearly isogenic susceptible and resistant tomato cultivars was hybridized with nucleotide binding region probe 3-3 (Figure 2A). Transcripts of ~4 kb were identified in both resistant and susceptible tomato roots.
A cDNA library constructed from mRNA isolated from root tissue of the nematode-resistant tomato line VFNT cherry was screened with probe 3-3, and 24 hybridizing clones were identified. Partial or complete DNA sequence was obtained for eight of the hybridizing clones. Sequence analysis revealed that three clones corresponded to Mi-1.1, two clones corresponded to Mi-1.2, and the remaining three clones corresponded to at least two additional genes with a similar sequence. None of the clones examined corresponded to the third copy, which is consistent with the prediction that it is a pseudogene. The longest clones for Mi-1.1 and Mi-1.2 were 2.9 and 2.5 kb, respectively.
The 5′ ends of the Mi-1.1 and Mi-1.2 mRNAs were obtained by rapid amplification of cDNA ends (RACE) utilizing primers specific for each of the two cDNAs. Each amplification reaction resulted in a major 1.9-kb product on a gel. Several clones from each amplified band were identified and sequenced. All clones sequenced corresponded to the 5′ region of the targeted gene. The partial cDNA clones and their corresponding 5′ RACE products were pieced together to produce deduced transcribed sequences of ~4 kb for both Mi-1.1 and Mi-1.2, corresponding well to the transcript length indicated by RNA gel blot analysis. For each cDNA, an upstream, in-frame stop codon was identified, indicating that the complete open reading frame had been identified. The deduced open reading frame was 1255 amino acids for Mi-1.1 and 1257 amino acids for Mi-1.2. The predicted polypeptides have 91% amino acid sequence identity (Figure 3).
Mi-1.1 and Mi-1.2 Gene Structure
Comparison of the cDNA and genomic sequences revealed that Mi-1.1 and Mi-1.2 each contain two introns at conserved positions near their 5′ end (Figure 2B). Intron 1 interrupts the untranslated region, whereas intron 2 interrupts the coding region. Intron 1 is longer in Mi-1.2 than in Mi-1.1 (1306 and 556 nucleotides, respectively). The intron 1 sequences of Mi-1.1 and Mi-1.2 bear regions of similarity. The 3′ 183 nucleotides are 97% identical between the two copies, and the 5′ 383-nucleotide regions are 70% identical and share regions of homology interspersed with small insertions. An insertion of 738 nucleotides appears to have occurred in the middle of Mi-1.2 intron 1. Intron 2 is 75 nucleotides in length in both R gene homologs, and their sequences are 97% identical. The position of the initiating ATG codon is conserved between the two genes and begins 42 bp 5′ of intron 2. Thus, if our cDNAs are full length, each mature transcript has a 5′ untranslated region of ~86 nucleotides. The lengths of the 3′ untranslated regions for Mi-1.1 and Mi-1.2 are 132 and 108 nucleotides, respectively.
Expression of Mi-1.1 Homologs in Susceptible and Resistant Tomato.
(A) An RNA gel blot with ~5 μg of poly(A)+ RNA per lane from root tissue of nematode susceptible (S) and resistant (R) tomato was hybridized with probe 3-3, which includes the nucleotide binding region of Mi-1.1. Numbers at right indicate the positions of RNA markers in kilobases.
(B) Mi-1.1 and Mi-1.2 transcripts are represented with thick horizontal lines indicating the exons and angled lines indicating the introns. The number of nucleotides comprising each intron is indicated.
Predicted Amino Acid Sequences of Mi-1.1 and Mi-1.2 Gene Products.
The deduced amino acid sequence of the Mi-1.2 gene product is shown, and amino acids that differ in the Mi-1.1 gene product are indicated. Dashes indicate gaps inserted to maintain optimal alignment. The positions of a potential leucine zipper and a heptad repeat motif are underlined. Leucines and isoleucines at the repeat positions of these motifs are in boldface. Boundaries of the conserved region (cons. reg.) and LRR region are indicated. Amino acids comprising the kinase-1a (P loop), kinase-2, and potential kinase-3a motifs of a predicted nucleotide binding site domain are underlined.
The regions 5′ to the putative transcription start site of the two genes, from nucleotides −1 to −446, have 94% sequence identity. In the region 5′ from this point (positions −447 to −1660), the identity drops to 34%. The two genes have an identical TATA box sequence (TATATTT) at −30 bp from the putative transcript start. In addition, Mi-1.1 has a CAAT box sequence at −76 bp.
Mi-1.2 Confers Resistance to Root Knot Nematodes
The nematode-susceptible tomato line Moneymaker was transformed with construct pSM137 by using Agrobacterium-mediated transformation. This construct contains a 14.7-kb insert of tomato genomic DNA in binary vector pCGN1557, including the entire Mi-1.2 coding region, 4.62 kb of sequence 5′ of the putative transcription start site, and 4.77 kb 3′ of the transcription termination site. DNA gel blot analysis indicated that one to four copies of the T-DNA had been transferred to each transformant (Table 1). Four cuttings of each independent transformant were tested for resistance to the root knot nematode M. javanica in assays done in the greenhouse. Of 23 tomato transformants tested, all but three were resistant (Table 1). This experiment demonstrates that the 14.7-kb DNA insert carrying Mi-1.2 is sufficient to produce effective nematode resistance when introduced into nematode-susceptible tomato plants.
Eighteen progeny plants from transformant 143-11, which carries one copy of the introduced T-DNA sequence, were tested for resistance. Three were susceptible and 15 were resistant to M. javanica, which is consistent with the expected ratio for segregation of a single dominant gene. DNA gel blot analysis indicated a complete correlation of nematode resistance with the presence of T-DNA sequences (Figure 4). To compare the resistance specificity to that of Mi, we inoculated six progeny of 143-11 with M. incognita, a second nematode species against which Mi is highly effective. Five of the six plants were resistant to M. incognita. Six additional progeny were inoculated with M. javanica VW5, a strain against which Mi is not effective. All six plants were susceptible to this nematode isolate, indicating that the observed specificity of resistance of the introduced Mi-1.2 gene and flanking sequences resembles that of Mi.
Analysis of Plants Transformed with Mi-1.2
A genomic clone spanning Mi-1.1 was obtained as a 7-kb fragment from BAC3 and inserted into binary vector pPBI-BAG3 to produce pSM152. The 7-kb fragment contained the Mi-1.1 transcribed region, 1.66 kb of sequence 5′ of the putative transcription start site, and 435 bp of 3′ region. The 1.66-kb upstream region contains the intergenic region and the terminal 12 nucleotides of the next upstream open reading frame, which has similarity to the transposase from the HAT family of transposable elements. Twelve transgenic plants, which were shown to contain the introduced gene by polymerase chain reaction and DNA gel blot analysis, were analyzed for resistance to M. javanica. All were found to be completely susceptible, showing that the introduced Mi-1.1 sequences did not provide the Mi phenotype.
Mi Is a Member of a Small Gene Family in Tomato
Blots with DNA from resistant tomato lines Motelle and Sun 6082 and from susceptible line Castlerock II were probed with a 1.8-kb cDNA fragment corresponding to the N-terminal domain of the Mi-1.1 gene product (Figure 5). Hybridization under stringent conditions indicated that there are approximately eight gene family members present in susceptible tomato and approximately six copies in resistant tomato. All or most bands are polymorphic between resistant and susceptible tomato. Because lines Castlerock II and Sun 6082 are nearly isogenic, this polymorphism pattern suggests that all or most copies are clustered in the Mi region. Because the line Motelle shows the same restriction fragment length polymorphism pattern as does the second resistant line Sun 6082, it is likely that all or most of the highly related copies of this gene family are clustered in the 650-kb introgressed region of Motelle. DNA gel blot analysis of additional resistant and susceptible tomato lines and of YAC clones (data not shown) supports the clustering of the family members.
DISCUSSION
Structure of the Encoded Proteins and Similarity to Other Resistance Genes
The predicted proteins Mi-1.1 and Mi-1.2 encoded by Mi-1.1 and Mi-1.2 have 91% identical amino acids. Comparison with other available sequences revealed that these proteins belong to the nucleotide binding–LRR family of plant resistance–associated gene products. The highest similarity is to the putative protein encoded by Prf. Also highly similar, as indicated in Table 2, are the products of the tomato genes I2C-1 and I2C-2, members of a gene family that contains the fungal resistance gene I2, which confers resistance against race 2 of the soil-borne fungus Fusarium oxysporum (Ori et al., 1997), and the Arabidopsis gene RPM1.
The N-terminal regions of the nucleotide binding–LRR family members can be divided into two subclasses. In one subclass, the N-terminal region contains a sequence resembling that of the intercellular signaling domains of the Drosophila Toll protein and the mammalian interleukin-1 receptor protein. This subclass is exemplified by the N gene for tobacco mosaic virus resistance in tobacco and L6, a flax rust resistance gene (Lawrence et al., 1995). Mi-1.2 does not contain this motif. The second class, exemplified by the predicted proteins encoded by Arabidopsis genes RPM1 and RPS2 and the tomato gene Prf, is characterized as containing a potential leucine zipper motif (Hammond-Kosack and Jones, 1997; Parker and Coleman, 1997). Mi-1.1 and Mi-1.2 each contain a predicted leucine zipper motif and thus fit well into the second subclass. The potential leucine zipper domains of Mi-1.1, Mi-1.2, Prf, and RPM1 show sequence similarity and are located at a similar position relative to the conserved region (Figures 6A and 6B, and Table 2). The similarity in amino acid sequence and position of the leucine zipper motif suggests that these genes may all be involved in similar dimerization interactions. I2C-1/I2C-2 do not appear to contain a leucine zipper domain in this region.
Nematode Resistance in Transgenic Plants.
Progeny of tomato line 143-11, which was transformed with Mi-1.2, were assayed for nematode resistance. The presence of the introduced DNA in these plants was detected by electrophoresis and DNA blot hybridization (inset). The expected 4.4-kb EcoRV band was present in all resistant plants (R; sample root mass at left) and was not present in any susceptible plants (S; right root mass). Note the presence of galls or root knots in the susceptible plant. Magnified view shows the presence of egg masses, here stained blue, indicating successful reproduction of nematodes. Lane P contains DNA from the primary transformant 143-11.
A second region containing seven isoleucine/leucine heptad repeats not present in the other R genes spans residues 460 to 502 of Mi-1.2 (Figure 3). These heptad repeats are not likely to form a leucine zipper because they contain two proline residues, which would be predicted to cause a bend in the structure. Although its significance, if any, is not apparent, the sequence LIKEEI is present in both regions. Except for the presence of the leucine zipper, there is little other similarity in the N-terminal regions within this group of gene products. Prf has a considerably longer N-terminal region than does Mi-1.2, whereas the N termini of the proteins encoded by IC-2 and RPM1 are quite short (Figure 6).
The highest similarity among the leucine zipper–nucleotide binding–LRR proteins is in the 260–amino acid central conserved region (Table 2 and Figure 6C), suggestive of a conserved function for this part of the protein. This region contains two motifs, kinase-1a (P-loop) and kinase-2 consensus sequences, that conform in sequence and spacing to those found in known ATP and GTP binding proteins (Traut, 1994) (Figure 6C). A potential kinase-3a motif that differs somewhat from the published consensus (Traut, 1994) but is highly conserved among nucleotide binding–LRR genes is also noted in Figure 6C. Additional conserved regions include a hydrophobic domain containing the sequence GLPL, which is almost invariant among nucleotide binding–LRR genes described to date (Figure 6C; Hammond-Kosack and Jones, 1997).
The C-terminal region of Mi-1.1 and Mi-1.2 can be arranged into ~14 LRRs of ~24 amino acids. This framework is most similar to that of Prf (Figure 6D). The consensus sequence of the LRR of Mi-1.2 is aXXLXXLXXLXa(X)12 (where a indicates an aliphatic amino acid residue and X indicates any amino acid; a consensus is assigned if the amino acid is present in >50% of the residues at a particular position in the repeat). This consensus most strongly resembles that for the cytoplasmic class of nucleotide binding–LRR proteins (Jones and Jones, 1996).
How Does Mi-1.2 Confer Resistance?
Members of the nucleotide binding–LRR family mediate resistance to a broad range of pathogens, including viruses, bacteria, fungi, and now, nematodes. These genes have been proposed to be involved in specific recognition of pathogen products. Based on the lack of a signal peptide, it is likely that Mi-1.2 and other members of the nucleotide binding–LRR class are cytoplasmically localized. If Mi-1.2 does in fact recognize a product from the nematode, it is likely that this nematode product is present inside of the plant cell. Although nematodes are extracellular pathogens, they feed on the cytoplasm of living plant cells and are thought to inject secretions into plant cytoplasm to initiate the development of feeding cells (Williamson and Hussey, 1996). The cellular HR associated with the presence of Mi occurs near the anterior end of the nematode at ~12 hr after inoculation. This corresponds roughly to the time when the nematode would be expected to inject saliva into the cytoplasm of developing vascular tissue cells to initiate giant cell development. This timing is consistent with the hypothesis that Mi-1.2 recognizes something that the nematode injects into the plant cell. Finding such a molecule will be difficult despite the availability of nematode strains that can infect plants containing Mi because this group of root knot nematodes does not reproduce sexually, precluding the use of standard genetic techniques for identifying the associated parasite genes.
Complementation results presented here indicate that the 14.7-kb fragment containing the Mi-1.2 gene and no other open reading frame is sufficient to confer resistance to a susceptible tomato line; that is, no other genes that are specific to the introgressed region in resistant tomato lines are required for resistance to the root knot nematode species M. javanica and M. incognita. The high frequency of success of complementation (20 of 23 independent transformants) is encouraging from several standpoints. It demonstrates that a single copy of the gene confers full resistance, and thus positional effects or gene silencing are not major problems with Mi-1.2 as they have been with some other genes. Also, the phenotype is stably transmitted to progeny plants. These observations increase the likelihood that Mi will be of commercial value and that meaningful analysis of in vitro modifications of Mi can be performed using transgenic plants.
Mi-1.2 Is a Member of a Small Gene Family.
DNA from resistant tomato lines Motelle and Sun 6028 and susceptible line Castlerock II was digested with EcoRV or SpeI. The probe was a 1.8-kb fragment from the N terminus of Mi-1.1. Numbers at right indicate the positions of DNA markers in kilobases.
Percentage of Amino Acid Similarity and Identity of Mi-1.2 to Closely Related Gene Productsa
Transformation of tomato with Mi-1.1 did not result in resistance, suggesting that this sequence does not encode a functional root knot nematode resistance gene. Another explanation for this failure to function is that the transferred sequence did not contain the entire control region required for its appropriate expression even though the 1.6-kb 5′ region contains the entire intergenic region as well as the end of the next upstream reading frame. Alternatively, it is possible that the Mi-1.1 product confers resistance to another nematode species or perhaps has a role in resistance to a different type of organism. These possibilities will be examined in future work, and it is likely that comparison of Mi-1.1 and Mi-1.2 will provide insight into the regions of Mi-1.2 that are important for function and specificity. For example, Mi-1.1 and Mi-1.2 are most divergent in the C-terminal LRR region (Figure 3 and Table 2). The LRR region has been demonstrated to contain determinants of specificity for alleles at the flax L locus, which encodes flax rust resistance genes of multiple specificities (Ellis et al., 1997). However, Prf, the most similar sequence to Mi-1.2 in the database, does not appear to be the determinant of specificity in Pto/Prf–mediated resistance to Pseudomonas spp. Instead, the pathogen elicitor avrPto has been shown to interact directly with Pto, a serine/threonine protein kinase (Scofield et al., 1996; Tang et al., 1996). The availability of Mi-1.2 and its homologs will allow us to investigate determination of specificity in this interaction.
Comparison of Mi-1.1 and Mi-1.2 Gene Products with Those of Most Closely Related Nucleotide Binding–LRR Genes.
(A) Comparison of the positions and sizes (at right in amino acids) of characteristic regions of Mi-1.1/Mi-1.2 (Mi-1), Prf, I2C-1/I2C-2 (I2C), and Rpm1 (RPM1). LZ indicates the position of a possible leucine zipper.
(B) Comparison of the possible leucine zipper regions of Mi-1.1, Mi-1.2, Prf, and Rpm1. I2C1/I2C-2 does not appear to carry a leucine zipper in the N-terminal regions. The positions of the leucines/isoleucines in the zipper are indicated by black diamonds.
(C) Comparison of the central conserved regions. Kinase-1a, kinase-2, and possible kinase-3a motifs are indicated. hd indicates the position of the conserved hydrophobic domain.
(D) Comparison of the LRR domains of Mi-1.1, Mi-1.2, and Prf.
Black boxes indicate that all residues in a column are identical or conservative substitutions. Sequence gaps inserted to maintain alignment are shown as dashes. Sequences were aligned using the Genetics Computer Group program Pileup. The GeneDoc program (http://www.cris.com/~ketchup/genedoc.shtml) was used to display the pileup.
A number of resistance genes to various plant pathogenic nematodes have been identified genetically in several crop plants or their wild relatives, and efforts are under way to clone a number of these genes (Williamson and Hussey, 1996). It will be interesting to determine whether the structure of these genes resembles that of Mi-1.2. So far, the correlation between classes of R genes from different plant species and pathogen types has not been strong, suggesting that the R gene type cannot be predicted easily from the nature of the pathogen (Hammond-Kosack and Jones, 1997). However, evidence suggests that nucleotide binding–LRR genes may mediate nematode resistance in other species. Sequences with these motifs have been identified tightly linked to the cereal cyst nematode resistance gene Cre3 in wheat and the potato cyst nematode resistance gene Gro1 (Leister et al., 1996; Lagudah et al., 1997). However, the product of the only other cloned nematode resistance gene to date, Hs1pro-1, although it contains an LRR-like region, is structurally quite different from that of Mi-1.1 and Mi-1.2 (Cai et al., 1997). This much smaller gene encodes a 282–amino acid protein with an N-terminal LRR and a putative membrane spanning segment. Except for the presence of an LRR region, this gene bears little similarity to other cloned R genes.
Organization and Functions of the Mi Gene Region
Many R genes found in tomato and other plant species are located in clusters (Ellis et al., 1997; Hammond-Kosack and Jones, 1997). The 1-Mb region of the genome where Mi is located is a complex locus that can carry resistance to a number of pathogens and pests. A resistance gene to potato aphids, Meu1, which probably was introduced into tomato along with Mi-1, maps within 650 kb of Mi (Kaloshian et al., 1995). In addition, Cf2/Cf5, which are genes that confer resistance to specific races of C. fulvum, map just centromere distal to the Mi region and are physically located within the same 1-Mb region of the genome (Dixon et al., 1996; Kaloshian et al., 1998). Cf2/Cf5 are representatives of the R gene class characterized by extracellular, N-terminal LRRs, which is a quite different structure than that of Mi (Dixon et al., 1996).
Two expressed copies of the Mi gene family are present in the contiguous 52-kb region that was sequenced to identify Mi. DNA gel blot experiments suggested that approximately six other family members are present in resistant plants and that most or all are clustered in the 650-kb region from L. peruvianum that is present in the line Motelle (Figure 5; I. Kaloshian and V.M. Williamson, unpublished data). In the case of Mi, DNA gel blot hybridization indicates that there may be more copies present in susceptible than in resistant lines. Our cDNA screening identified at least three different genes homologous to Mi-1.2 that are expressed in roots of resistant plants. RNA gel blots showed that homologous genes are transcribed in susceptible tomato. Currently, we can only guess at their function. The availability of the Mi clone will provide an important tool for exploration of this complex R gene region.
METHODS
Plant Materials
Seed of susceptible tomato cultivar Castlerock II (mi/mi) and the nearly isogenic resistant cultivar Sun 6082 (Mi/Mi) were obtained from Sunseed Genetics (Holister, CA). Tomato lines Moneymaker and Motelle were obtained from M. Koornneef (Wageningen Agricultural University, The Netherlands).
Plasmids, Vectors, Yeast Artificial Chromosomes, and Genetic Markers
A yeast strain carrying a 500-kb yeast artificial chromosome (YAC), 2/1256, spanning the region of the tomato genome with Mi, was obtained from P. Vos (Keygene, Wageningen, The Netherlands). Plasmids C93.1 and C32.1 contain subclones of cosmids derived from this YAC (Kaloshian et al., 1998). The bacterial artificial chromosome (BAC) vector pBeloBACII-Spec, obtained from A. Lloyd (Stanford University, Stanford, CA), was a modification of pBeloBACII (Shizuya et al., 1992) that was made by inserting the spectinomycin resistance gene aadA (Hollingshead and Vapnek, 1985) into the BglI site of the vector.
Probe 3-3 is a 480-bp DNA probe spanning the nucleotide binding region of Mi-1.1 and corresponds to amino acids 556 to 713 in Figure 3. Probe 3-3 was produced by polymerase chain reaction amplification of a region from BAC3 with the nucleotide binding region–specific degenerate primers AT (5′-CTGCGTACCAATTCGGNGTNGGNAAAACTAC-3′) and L4 (5′-TGAGTCCTGAGTAAAGNGCNAGNGGNAGCCC-3′) (Shen et al., 1998). The product was cloned into pCRII (Invitrogen, Carlsbad, CA) and used as a probe.
Construction and Screening of Library from BAC3
Agarose plugs containing intact yeast chromosomal DNA, including the 500-kb YAC 2/1256, were prepared as previously described (Schwartz and Cantor, 1984) and then subjected to partial digestion with HindIII. The digest was stopped by adding EDTA to 10 mM, and then the sample was fractionated by contour-clamped homogeneous field gel electrophoresis. DNA migrating between 140 and 180 kb was excised as a gel slice and treated with Gelase (Epicentre Technologies, Madison, WI). DNA fragments were ligated into the HindIII-digested pBeloBACII-Spec vector, which had been treated with shrimp alkaline phosphatase (Amersham Life Sciences), and transformed into Escherichia coli DH10B (Gibco BRL) by electroporation. Approximately 2800 transformants were picked into 384-well microtiter plates and then spotted onto Hybond N membranes (Amersham Life Sciences) by using a Biomek 1000 robot (Beckman Instruments). Membranes were hybridized with 32P-labeled probes C93.1 and C32.1. The average insert size was ~50 kb, and thus the library represented about six yeast genome equivalents.
The ends of BAC3 were subcloned by digesting to completion with ClaI and then religating. ClaI cuts at several sites in the insert but does not cut the vector. The resultant plasmid was cut with NotI plus ClaI, and fragments containing each end were cloned into pBS KS− (Stratagene, La Jolla, CA). The tomato DNA fragment from the centromere-proximal end was called B3E.
DNA Sequencing and Analysis
For large-scale sequencing, BAC DNA was partially digested with Tsp509I, treated with shrimp alkaline phosphatase, and then fractionated on SeaPlaque (FMC Corp., Rockland, ME) agarose gels. DNA migrating between 1.5 and 4.0 kb was excised, purified with a Prep-A-Gene kit (Bio-Rad), and ligated into the EcoRI site of pUC119. Additional random clones were generated through partial digestion of BAC DNA with Sau3AI, complete digests with HindIII or XbaI plus SalI, and ligation into pUC119, pBS KS−, or pCGN1557. Where greater coverage of particular regions was required, additional clones were selected by colony hybridization.
For sequencing, DNA was purified by an ethidium bromide–phenol procedure (Stemmer, 1991). Cycle sequencing reactions were performed using the Thermosequenase kit (Amersham Life Sciences) and were run on a sequencer (model 4200; Licor, Lincoln, NB; or model 377; Applied Biosystems, Foster City, CA). Resequencing of ~20 kb of BAC4 that overlapped with BAC3 revealed eight sequence errors, or approximately one in 2500 nucleotides, which is an estimate of the overall sequencing accuracy.
DNA sequence data were edited and compiled using the Sequencher 3.0 program (GeneCodes Corp., Ann Arbor, MI). Comparisons and analysis of DNA and deduced amino acid sequences were made using the Genetics Computer Group (Madison, WI) software package, version 7.0. Database searching was done with BlastX, BlastN, Beauty, and other algorithms available through the National Center for Biotechnology Information, Bethesda, MD (http://www.nlm.nih.gov/cgi-bin/BLAST/).
The entire sequence of the 52-kb insert is available under Gen-Bank accession number U81378. The DNA and putative protein sequences of Mi-1.1 and Mi-1.2 are available as GenBank accession numbers AF039681 and AF039682, respectively.
DNA and RNA Gel Blot Analyses
Root and leaf tissues were harvested from 7-week-old plants grown in sand. Total RNA was isolated by LiCl precipitation as described previously (Rochester et al., 1986). Poly(A)+ RNA was isolated on paramagnetic beads by using the Poly-A-Track kit (Promega). Five micrograms of poly(A)+ RNA from each sample was fractionated on 1.3% agarose gels containing 6.5% formaldehyde, as described by Sambrook et al. (1989). Ethidium bromide was included in the gel to evaluate loading of samples. The gel was blotted onto a Hybond N membrane. Hybridization with 32P-labeled probe was performed for 16 hr at 42°C in 6 × SSPE (20 × SSPE is 3.6 M NaCl, 0.2 M NaH2PO4, and 0.02 M Na2EDTA, pH 7.7), 1 × Denhardt's solution (100 × Denhardt's solution is 2% [w/v] BSA fraction V, 20% [w/v] Ficoll 400, and 2% [w/v] PVP in water), 0.5% SDS, 0.2 mg/mL salmon testis DNA, and 50% formamide. Washing was performed for 1 hr at 65°C in 0.2 × SSPE and 0.1% SDS. The molecular mass standards used were the 0.24- to 9.5-kb ladder (Gibco BRL).
Genomic DNA extractions from tomato tissue were performed as described (Williamson and Colwell, 1991). Ten micrograms of DNA was digested overnight with either EcoRV or SpeI and fractionated on a 0.7% agarose gel. Gels were blotted to Hybond N membranes and hybridized as described in the above-mentioned RNA gel blot analysis.
Isolation of Genomic Clones for Transformation
BAC3 was partially digested with Sau3AI and then fractionated on a 0.4% agarose gel. DNA migrating between 10 and 15 kb was excised as a gel slice, purified, and ligated into the BamHI site of the plant transformation vector pCGN1557 (McBride and Summerfelt, 1990). The ligation mixture was transformed into E. coli DH10B (Gibco BRL) by electroporation. Colony blot hybridization on ~600 transformants with 32P-labeled probe 3-3 identified ~50 hybridizing clones. Restriction digests and comparison with sequence data allowed identification of clone pSM137, which contains a 14.7-kb insert with Mi-1.2 plus flanking sequences. No clones with intact Mi-1.1 plus flanking sequence were obtained by this procedure. To produce pSM152, a clone containing Mi-1.1 and flanking sequences, we cloned a 7-kb insert as a ClaI/XhoI fragment from BAC3 into the plant transformation vector pPBI-BAG3 (Goldsbrough et al., 1994).
Obtaining cDNA Clones
Approximately 106 clones from a cDNA library made in the HybriZap II vector (Stratagene) from root tissue of VFNT cherry tomato (provided by C. Mau, Center for Engineering Plants for Resistance Against Pathogens) were screened with Probe 3-3, and 24 hybridizing plaques were identified.
Rapid amplification of cDNA ends (Frohman et al., 1988) was performed to obtain full-length cDNAs by using the Marathon cDNA amplification kit (Clontech, Palo Alto, CA). First-strand cDNA was synthesized from 1 μg of poly(A)+ RNA isolated from the roots of the resistant cultivar Sun 6082 by using primer SM7 (5′-GGTCAAGAGGATCAGTGTTCAGCTTTCC-3′), which hybridizes with a sequence common to both the Mi-1.1 and Mi-1.2 cDNAs. After second-strand synthesis and adapter ligation, amplification was performed on separate aliquots of adapter-ligated cDNA with the adapter primer and a primer specific for either Mi-1.1 or Mi-1.2. The Mi-1.1–specific primer was SM9 (5′-TTGTATTCAACAACTTCTTCTCATCAC-3′), and the Mi-1.2–specific primer was SM10 (5′-TTGTATCCAACAACTTCTTGTCGTCAT-3′). Each amplification reaction yielded a 1.9-kb product that was cloned into pCR2.1 (Invitrogen). Sequence analysis verified that each clone represented the 5′ region of the Mi-1.1 or Mi-1.2 cDNA. Mi-1.1 and Mi-1.2 clones were designated pSM101 and pSM109, respectively.
Analysis of Transgenic Plants
Tomato transformation was performed as described by Fillatti et al. (1987). The susceptible tomato line Moneymaker was used for all of the transformation work reported.
The presence of sequences corresponding to Mi-1.1 and Mi-1.2 in DNA from transgenic plants was confirmed by polymerase chain reaction using a common primer and a gene-specific primer. The common primer was C1/2 (5′-CAGTGAAGTGGAAGTGATGA-3′), and the gene-specific primers derived from the 3′ untranslated region were C1S1 (5′-CCCAGCAAAGTACAATCTAC-3′) for Mi-1.1 and C2S4 (5′-CTAAGAGGAATCTCATCACAGG-3′) for Mi-1.2. Fifty-microliter polymerase chain reactions were performed and analyzed as previously described (Yaghoobi et al., 1995). Transgenic plants were screened for the presence of a 1.6-kb polymerase chain reaction product.
To determine the number of copies of the T-DNA incorporated into the genome, we digested plant genomic DNA of transgenic plants with EcoRV, fractionated it by electrophoresis, and transferred it onto Hybond nylon membranes. Membranes were hybridized with a 2.8-kb XhoI fragment of the 35S–Npt gene derived from the vector. This probe hybridizes with a 4.4-kb fragment internal to the T-DNA and to a junction fragment. The size differs for each independent insertion of the T-DNA. Treatment of membranes, hybridization, and washing procedures were as described previously (Yaghoobi et al., 1995).
To assay for nematode resistance, we made four cuttings of each independent transgenic plant in sand in 1-L cups. After 4 to 6 weeks, plants were infected with 3000 second-stage juveniles of Meloidogyne javanica. Six weeks later, roots were examined for the presence of egg masses, as previously described (Yaghoobi et al., 1995). M. javanica VW4 was used unless otherwise indicated. M. javanica VW5 is a variant strain that is able to reproduce on tomato carrying the Mi gene. This strain was selected after transfer of VW4 to a tomato line (VFNT cherry) that carries the Mi gene. VW4 and VW5 are maintained on susceptible and resistant tomato cultivars, respectively, in a hydroponic culture system (Lambert et al., 1992). The M. incognita strain used, VW6, was originally obtained as a field isolate from a cotton plant.
ACKNOWLEDGMENTS
We thank John Gardner for major contributions to the DNA sequencing, Bradford D. Hall for performing the tomato transformation, Chris Mau for construction of the cDNA library, Pieter Vos for YAC clones, Kathy Shen for nucleotide binding region primers, Dan Lavelle for assistance with sequence analysis, Barbara Soots for help with figures, and Amit Bhakta for technical assistance. We are grateful to George Bruening and Richard Michelmore for helpful discussions and comments on the manuscript. We also thank Chin-Feng Hwang and Magdalena Rossi for useful comments on the manuscript. This work was supported in part by National Science Foundation (NSF) Cooperative Agreement No. BIR-8920216 to the Center for Engineering Plants for Resistance Against Pathogens (CEPRAP), an NSF Science and Technology Center, and, in the early stages, by CEPRAP corporate associates Calgene, Inc., Ciba Geigy Biotechnology Corporation, Sandoz Seeds, and Zeneca Seeds.
Footnotes
- Received February 13, 1998.
- Accepted June 1, 1998.
- Published August 1, 1998.
REFERENCES
