Plant and fungal material.

Mentha longifolia accessions were obtained as stolons or rooted cuttings from the USDA-NCGR, Corvallis, OR. Plants were maintained in a greenhouse at the University of New Hampshire, Durham, in 22-cm-diameter pots and propagated vegetatively. Accessions CMEN17.001 (PI 557755), CMEN585.001 (PI 557767), CMEN501.001 (PI 212314), and CMEN81.001 (PI 557759) were previously classified as verticillium wilt-resistant, whereas accessions CMEN584.001 (PI 557769) and CMEN516.001 (PI 557760) were classified as wilt-susceptible (Vining et al., 2005). For convenience, the “.001” suffix common to all of these CMEN accession numbers is dropped throughout the subsequent text. Peppermint cultivar ‘Black Mitcham’, classified as verticillium wilt-susceptible (Douhan and Johnson, 2001), was also used. The accessions used in this study are diploid (2n = 2x = 24) with the exceptions of CMEN501 (tetraploid: 2n = 4x = 48), CMEN81 (unknown ploidy), and ‘Black Mitcham’ (hexaploid: 2n = 6x = 72) (Chambers and Hummer, 1994). The V. dahliae culture used to generate inoculum was provided by D. Johnson (Washington State University, Pullman, WA).

Molecular techniques.

Genomic DNA was extracted from fresh, unexpanded leaf tissue using a modified CTAB miniprep method as described in Vining et al. (2005). For reverse transcriptase (RT)–PCR, template RNA was obtained from roots of plants that had been incubated in either water or an aqueous V. dahliae spore suspension (≈10⁶ spores/mL) for 16 to 24 h. Total RNA was isolated from 100 to 200 mg fresh or frozen roots using Trizol reagent according to the manufacturer's instructions (GIBCO, Carlsbad, CA). Messenger RNA (mRNA) was separated from total RNA using oligo d(t) cellulose columns (GIBCO).

All PCR conditions are listed in Table 1. Reaction volumes of 25 μL contained 100 ng template DNA, 0.1 U Taq DNA polymerase (Sigma, St. Louis), 2.5 mM each dNTP (Promega Corp., Madison, WI), and either 4.8 μmol of each degenerate primer or 0.8 μmol of each specific primer. RT-PCR was performed with the ProSTAR HF Single-Tube RT-PCR System (Stratagene, La Jolla, CA). Inverse PCR (IPCR) was used to obtain sequences flanking the mint Ve segment initially identified by degenerate primer PCR. One microgram of total genomic DNA from CMEN585 was digested overnight at 37 °C in a 20-μL reaction with 10 U of the restriction enzyme MseI (New England Biolabs, Ipswich, MA). Two microliters of the digest were used directly in a ligation reaction with 400 U T4 DNA ligase (New England Biolabs) to circularize the genomic restriction fragments. The ligation reaction was incubated for 2.5 h at room temperature and then stored at 4 °C overnight. PCR was performed in 25-μL reactions using 5 μL of the ligation products as template, 0.1 U Taq polymerase (Sigma), 2.5 mM of each dNTP (Promega), and 0.8 umol of each inverse PCR primer.

Table 1. Polymerase chain reaction conditions used for amplification of resistance gene analog (RGA) and verticillium wilt-resistance-like (Ve) sequences from Mentha longifolia (L.) Huds.^z

View Fullscreen

Widely used degenerate PCR primers targeted to the P-loop motif (forward primer s2) and hydrophobic domain (reverse primers as2 or as3) (Leister et al., 1996) were used for initial amplification of RGAs from genomic DNA (gDNA) and from cDNA produced by RT-PCR (Table 2). Degenerate primers intended to amplify Ve homologs in mint were designed manually on the basis of tomato Ve1 amino acid and nucleotide sequences aligned with the most similar respective sequences from A. thaliana and Oryza sativa L. obtained by searching GenBank's nonredundant (nr) database using the blastx algorithm (Altschul et al., 1990) submitting the tomato Ve1 amino acid sequence as a search query. Subsequently, specific (nondegenerate) primers intended to amplify segments of putative mint Ve homologs or a 425-base pair (bp) segment of the tomato Ve1 gene itself (as a positive control) were designed using PrimerSelect software (version 5.53; DNAStar, Madison, WI). The tomato-specific forward and reverse primers, Ve5F19 and Ve5R23, bracketed Ve1 nucleotides 1776 to 1794 and 2200 to 2178, respectively (Table 2).

Table 2. Oligonucleotide primers used to amplify resistance gene analog (RGA) and verticillium wilt resistance-like (Ve-like) sequences from Mentha longifolia.^z

View Fullscreen

PCR, RT-PCR, and IPCR products were electrophoretically separated on 2% agarose gels in 1× TBE, pH 8.0 at 4.5 V/cm at room temperature. Gels were stained with ethidium bromide and photographed over ultraviolet light. Amplification products were recovered from gel slices using the Geneclean Spin kit (QBiogene, Irvine, CA) and cloned with the TOPO TA Cloning kit (Invitrogen, Carlsbad, CA). Recombinant plasmids were isolated using the Wizard SV Plasmid Purification kit (Promega).

Cycle sequencing reactions were performed following the University of New Hampshire's Hubbard Center for Genome Studies protocol (Hubbard Center for Genome Studies, University of New Hampshire, 2002) using DYEnamic Terminator ET cycle sequencing premix (Amersham Biosciences, Piscataway, NJ) and standard plasmid vector primers M13F, M13R, T3, and T7. Reaction products were analyzed using an ABI PRISM 377 DNA sequencer (Applied Biosystems, Foster City, CA).

Sequence analysis.

DNA sequence chromatograms were viewed and assessed for base-calling accuracy using SeqEd version 1.0.3 (DNAStar). Similarity searches of the GenBank nr database were conducted using blastx and tblastx algorithms (Altschul et al., 1990). Clustal X (Thompson et al., 1997) was used to perform sequence alignments and to generate an amino acid sequence identity matrix of the RGAs, including top non-mint RGA matches from the GenBank nr database (RGA blast hits). A Clustal X alignment of translated mint RGAs and top blastx hits was used to construct a neighbor-joining tree with 1000 bootstrap replicates.

Resistance gene analogs.

The insert sizes of 52 PCR product clones obtained from genomic DNA of four verticillium wilt-resistant M. longifolia accessions using degenerate RGA primers (s2 + as2 or s2 + as3) ranged from 289 to 680 bp. When these sequences were subjected to blastx and tblastx searches of the GenBank nr database, 20 of them had best matches (i.e., blast hits) that were R genes or RGAs. These M. longifolia genomic RGA sequences are referred to as gRGAs (Table 3). Of the 20 gRGA sequences acquired, seven from CMEN585, six from CMEN17, three from CMEN81, and one from CMEN501, 17 were found to be unique after discounting sequence redundancy. One gRGA sequence was represented by three identical clones: two from CMEN585 (clones 5856 and 58531) and one from CMEN17 (clone 1744). Another gRGA was represented by two identical clones, 1712 and 1725, both from CMEN17.

Table 3. Mentha longifolia resistance gene analogs (RGAs) obtained with degenerate primer polymerase chain reaction from various M. longifolia accessions.^z

View Fullscreen

All but three of the 17 unique gRGA sequences consisted of uninterrupted open reading frames (ORFs). Clone 5857 had an RGA reading frame that was interrupted by one stop codon [position 271–273 (TAG)], and clone 171 had an RGA reading frame interrupted by two stop codons [positions 229–231 (TAA) and 358–360 (TAA)]; otherwise, both of these interrupted RGA sequences were continuous ORFs. gRGA clone 173 had an RGA sequence that shifted the reading frame forward one nucleotide at a point 245 bases into the sequence dividing two nearly continuous ORFs that were each interrupted by a single stop codon [positions 79–81 (TGA), 386–388 (TGA)].

The s2 + as2 degenerate primer combination was also used in RT-PCR with mRNA from both inoculated and noninoculated roots. After RT-PCR, the most densely staining area of the gel lane (in the 250- to 500-bp range), or a discrete gel band, if present, was excised for cDNA recovery. Among 23 sequenced cDNA clones, seven RGAs were identified (Table 3). These cDNA-derived RGAs are referred to as cRGAs. All seven cRGAs were comprised of uninterrupted ORFs. Of the cRGAs from noninoculated root tissue, one, two, and three cRGAs, respectively, were from CMEN17, CMEN585, and CMEN81. Clone 17B9 from CMEN17 was the only cRGA recovered from inoculated roots. The sequence of cRGA clone 1719 was identical to that of gRGA clone 58513 from CMEN585.

Of 31 genomic and 16 cDNA non-RGA sequences obtained using the degenerate RGA primer pairs, one cDNA and four genomic sequences had ABC transporter sequences as best blast hits, and three cDNA and two genomic sequences’ best blast hits were gag-pol or Ty3/gypsy retroelements. Among the other cDNA sequences that had uninterrupted ORFs, blast matches included one chlorophyll a/b binding protein, one L29 ribosomal protein, and two different “senescence-associated proteins.” The rest of the non-RGA genomic and cDNA clones lacked uninterrupted ORFs and had no informative blast hits.

Among a total of 27 gRGA and cRGA sequences acquired using the degenerate RGA primer pairs, there were 23 unique RGA sequences, all of which have been submitted to GenBank (under accession numbers shown in Table 3). All 23 unique RGAs contained the RNBS-A non-TIR motif, and 20 of them had tryptophan in the terminal position of the Kin-2 motif, which is characteristic of the non-TIR RGAs. When each of the 23 unique mint RGAs was used in blastx and tblastx searches of the GenBank non redundant (nr) database, the best blast hits were always mint RGAs from the present data set; but many very good nonmint blast hits were also obtained. In many instances, the same best non-mint blast hit was obtained for two or more mint RGAs.

The 23 unique mint RGAs and a selection of 14 GenBank sequences chosen to represent the diversity of top nonmint blast hits were subjected to a Clustal X analysis that generated an amino acid sequence alignment (Fig. 1) and identity matrix (not shown). Percent amino acid sequence identities of pairs of unique mint RGAs ranged from 18% to 96%. From the Clustal X alignment, a neighbor-joining tree was generated, which organized the 23 mint RGA sequences and best nonmint blast hits into three major groups (I, II, III) that were further subdivided into seven distinct families (A through G) (Fig. 2). Percent amino acid identities of mint sequences within families ranged from 48% to 96%. The two most similar pairs of sequences were in family E, in which gRGA sequences 501AS38A and cRGA 81CRGC4 were 96% identical, and in family B, in which gRGA sequences 5856 and 178 were 89% identical. Mint RGAs in family G were most similar to Phytophthora infestans (Mont.) de Bary R genes from Solanum L. species, whereas mint RGAs in families C and D were most similar to potato and pepper R genes with known pathogen specificities (Table 4). For mint RGAs in families B, E, and F, representatives from several angiosperm species were included among the top nonmint blast hits (Table 4). Top non-mint blast hits for mint family A were RGAs from Coffea L. species.

Table 4. Closest matches (blast hits) for mint resistance gene analog (RGA) sequences obtained by blastx searches of the GenBank nonredundant database.^z

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Ve.

When a blastx search of the GenBank nonredundant database was performed using the tomato Ve1 amino acid sequence as a search query, the most similar sequences (30% to 40% amino acid identities) were those from A. thaliana (NP_187712, NP_198058, NP_180861, NP_187217) and O. sativa (BAA96753, BAB08209). Using both amino acid and nucleotide alignments of the tomato Ve1 and Ve2 sequence with the aforementioned A. thaliana and O. sativa sequences as a basis for identifying conserved regions, degenerate primers Ve2398–2416F and Ve2791–2810R (Table 2) were designed manually. These primers, which targeted sites corresponding to codons 800–805 and 942–948 of tomato Ve1, amplified a 445-bp product from CMEN585, the sequence of which we designate the “original mint Ve.” Excluding from consideration the amino acids coinciding with the primer sites, the original mint Ve predicted amino acid sequence (mint orig) shared 57.1% and 56.5% amino acid identity with the corresponding regions of the tomato Ve1 and Ve2 sequences, respectively (Fig. 3).

View Figure

• Download as Powerpoint
• Download Figure

On the basis of the original mint Ve sequence, mint-specific primers (Ve4–24F and Ve403–386R; Table 2) internal to the degenerate primer sites were designed and tested on four M. longifolia accessions (CMEN numbers 17, 516, 584, 585) and peppermint cultivar ‘Black Mitcham’, in each case producing “second-generation” mint Ve sequences of ≈400 bp that had 94.7% to 100% predicted amino acid identities to the original mint Ve. The predicted amino acid sequence of the second-generation mint Ve product from CMEN585 is shown (as “Mint 2”) aligned with the original mint Ve (Fig. 3).

Also on the basis of the original mint Ve sequence, inverse PCR primers (IPCR138–153F, IPCR325–304R; Table 2) were designed and used to obtain extended Ve sequence from CMEN585 that included and flanked the original degenerate primer sites. The IPCR results extended the original Ve sequence by adding 202 and 431 bases, respectively, to its 5′ and 3′ ends (Mint IPCR; Fig. 3). Examination of the IPCR sequence revealed that the actual mint sequence at the site of the initial Ve degenerate primers was perfectly matched by one permutation of each of the original Ve degenerate PCR primer sequences and that the second-generation “mint-specific” primer Ve4–24F had a two-base mismatch with the actual CMEN585 target sequence (Table 2). The extended mint Ve open reading frame ended with three sequential stop codons (TAA, TGA, TGA) at nucleotides 1013 to 1021. Nucleotides 1022 to 1077 were presumed to be part of the 3′ untranslated region.

Finally, the extended mint Ve sequences provided a basis for design of specific PCR primers (Ve111–133F and Ve974–951R; Table 2), which bracketed 865 nucleotides of the extended mint Ve sequence. The predicted amino acid sequence of the extended mint Ve aligned with the C terminus of the tomato sequences, spanning amino acids 730 to 1053 and 728 to 1079 of Ve1 and Ve2, respectively (Fig. 3). This section of the tomato sequences included the following regions defined by Kawchuk et al. (2001): LRRs 28 to 38 (end of region B); negatively charged extracytoplasmic (C); membrane-spanning (D); and positively charged cytoplasmic (E). Although the extended mint Ve sequence shared ≈57% amino acid identity over the corresponding B region of the tomato Ve sequences, the mint sequence shared only seven of 21 and eight of 21 amino acid identities with the D regions of Ve1 and Ve2, respectively. The C and E regions of the mint and tomato Ve sequences could not be meaningfully aligned.

Resistance gene analogs.

We used degenerate primers targeting the NBS motif in genomic PCR and RT-PCR to identify a set of 23 unique RGA sequences from M. longifolia, of which 20 had uninterrupted ORFs. Best blast hits for these RGAs included angiosperm R genes with a variety of known disease specificities. The 40% to 50% range of amino acid identities to known R genes compares with that seen for RGAs identified using the same degenerate primers in other plant taxa such as strawberry (Martinez Zamora et al., 2004) and coffee (Coffea arabica L.) (Noir et al., 2001). The presence of uninterrupted ORFs in 20 of the 23 unique RGAs is suggestive that at least some of these mint RGAs are derived from functional genes. One gRGA, clone 1719 from CMEN17, was identical to a cRGA (clone 58513) from CMEN585, indicating that this particular RGA is both highly conserved and is expressed, strongly suggesting that it may encode a functional product involved in disease resistance.

Two of the 23 M. longifolia gRGA sequences had stop codons in otherwise uninterrupted ORFs, and one gRGA contained a frameshift that separated two nearly continuous ORFs. Similarly, Martinez Zamora et al. (2004) identified 51 RGAs in strawberry, of which 28 had uninterrupted ORFs and 23 had stop codons or frameshifts. In pine (Pinus monticola Dougl. ex. D. Don), Liu and Ekramoddoullah (2003) found that 18 of 46 genomic RGA clones and 14 of 21 cDNA RGA clones had at least one stop codon or frameshift. RGAs with interrupted ORFs may be parts of pseudogenes. Plant R genes have been shown to occur in complex clusters of related functional genes, pseudogenes, gene fragments, and transposable elements (Lescot et al., 2004; Wei et al., 2002). Whether pseudogenes are evolutionary remnants—once-active genes rendered inactive by accumulation of mutations—or whether they have some function, regulatory or otherwise, remains to be determined on a case-by-case basis.

In two instances, identical RGA sequences were amplified from different accessions. The gRGA clone 1744 from CMEN17 was identical to gRGA clones 5856 and 58513 from CMEN585, whereas gRGA clone 1719 from CMEN17 was identical to cRGA clone 58513 from CMEN585. The amplification of identical RGAs from different accessions suggests that the respective pairs of sequences come from orthologous genes. Four of the mint gRGAs obtained using degenerate primers were represented by multiple clones from the same accession. One pair of identical clones (1725, 1712) was from CMEN17, and another (5856, 58531) was from CMEN585. These identities are probably the consequence of redundant sampling of the same gene. The recoveries of nonidentical RGA sequences from a single accession may result from sampling of different genes or different alleles of the same gene. The closest nucleotide sequence match between nonidentical gRGAs within an accession was 79.8% (clones 8120 and 8116), which is low enough to suggest that these are different genes.

NBS-LRR genes comprise significant proportions of plant genomes. Whole genome sequences are available for: A. thaliana ecotype Col-0 and O. sativa ssp. japonica L. cv. Nipponbare. The A. thaliana genome of ≈157 Mb (Bennett and Leitch, 2005) contains 149 NBS-LRR genes (55 non-TIR, 94 TIR), representing ≈1% of the genome (Meyers et al., 2005).The O. sativa haploid genome size is estimated to be 389 Mb (International Rice Genome Sequencing Project, 2005). O. sativa has ≈500 NBS-containing RGA sequences, all of which are of the non-TIR type, and which, like A. thaliana, represent ≈1% of the genome (Monosi et al., 2004). The M. longifolia genome size of ≈400 Mb is comparable to that of rice (O. sativa) (Bennett and Leitch, 2005). If the total number of M. longifolia RGAs parallels the ≈500 of rice, then the 23 RGAs reported here represent less than 5% of the potential M. longifolia RGAs. The low level of redundancy encountered among the M. longifolia RGAs is also indicative that these sequences represent a small subset of the RGAs in the mint genome.

Studies of resistance gene clusters have revealed evidence of duplication, recombination, and rearrangement events leading to the expansion of clusters, growth of multigene families, and the potential for evolution of new resistance specificities (Chin et al., 2001; Graham et al., 2002; Kruijt et al., 2004). The 23 M. longifolia RGAs grouped into seven distinct families, each of which had identities with different RGAs and R genes from other plant species. Particular amino acid motifs conserved within mint RGA families may indicate some functional specificity. However, closely related R genes do not necessarily confer resistance to the same pathogen. The potato Gpa2 and Rx1 genes are in the same cluster on potato chromosome 12 and are considered to be highly homologous but have very different specificities; Gpa2 is a nematode (Caenorhabditis elegans Miers) resistance gene, whereas Rx1 confers resistance to a virus (van der Vossen et al., 2000). The M. longifolia RGA families C and D each have significant similarity to Gpa2 and Rx1. It would be interesting to determine whether family C and D mint RGAs map to a common genomic region or cluster like their apparent counterparts do in potato.

All the M. longifolia RGAs amplified with degenerate primers based on the P-loop and GLPL domains, or amplified with mint RGA-specific primers, had motifs consistent with the non-TIR class. This result does not lead to a conclusion that TIR-class RGAs are absent from the M. longifolia genome. In general, dicot species have been found to contain varying ratios of TIR and non-TIR classes of RGAs, whereas in monocot species, only non-TIR RGA sequences have been observed (Cannon et al., 2002). However, dicot RGA sequences obtained from degenerate primer PCR have not given any indication that non-TIR and TIR classes are amplified in proportions that reflect their presence in plant genomes. The same P-loop and GLPL degenerate primers used in the present study, or very similar primers, amplified both TIR and non-TIR RGAs in alfalfa (Medicago L.) (Cordero and Skinner, 2002), but only non-TIR sequences from coffee (Noir et al., 2001). In soybean [Glycine max (L.) Merr.], in which Kanazin et al. (1996) found only TIR-class RGAs using P-loop and GLPL-based primers, Peñuela et al. (2002) amplified non-TIR RGAs with a set of degenerate primers pairing a P-loop forward primer with a reverse primer based on the non-TIR RNBS-D motif (CFLYCALFPED). It is possible that the P-loop/GLPL degenerate primer set is biased toward TIR-class or non-TIR-class sequences in a species-dependent manner. The fact that only one RGA was obtained using the as3 reverse primer may also reflect primer bias.

Although the isolation of non-RGA sequences among the PCR products generated by degenerate RGA primers seems inevitable, many RGA reports do not mention whether any non-RGA sequences were obtained with degenerate RGA primers. In the present study, two types of non-RGA sequences were common. Use of degenerate RGA primers generated seven PCR products that had high identity to transposable element-like sequences. Sequences with identity to transposable elements have been mentioned in a few RGA reports (Noir et al., 2001; Timmerman-Vaughan et al., 2000). Five of our mint degenerate primer PCR products had high identity to ABC transporter sequences. Joyeux et al. (1999) amplified an ABC transporter-like sequence from Brassica napus L. using degenerate primers nearly identical to s2 and as2. Amplification of ABC transporter sequences with NBS degenerate primers is not surprising, because ABC transporters have an NBS motif similar to that of NBS-LRR disease resistance genes (Theodoulou, 2000). If the isolation of non-RGA sequences was detailed in more reports of degenerate RGA primer studies, potentially useful insight could be gained into commonalities and patterns of occurrence of such sequences in different species.

Ve.

Because Ve is not an R gene of the NBS-LRR class, it would not be expected to be amplified by NBS-targeted degenerate primers such as those used to amplify RGAs. We used degenerate primers targeting part of the tomato Ve LRR region to isolate an original Ve-like M. longifolia sequence of 445 bp from the verticillium wilt-resistant accession CMEN585. The ≈57% amino acid identity of this sequence to the corresponding regions of the tomato Ve1 and Ve2 sequences is higher than that for any other Ve-like sequence found in GenBank. Design of mint-specific IPCR primers enabled us to extend the acquisition of mint Ve sequence to a 1077-bp segment that corresponds to the distal B region and the C, D, and E regions of the tomato Ve genes as described by Kawchuk et al. (2001). To date, this is the first reported use of degenerate PCR primers to isolate a Ve-like sequence and the first targeted isolation of such a sequence outside of the Solanaceae.

Specific primers based on the original mint Ve sequence detected Ve homologs in several wilt-resistant and wilt-susceptible M. longifolia accessions, and from wilt-susceptible cultivar ‘Black Mitcham’, indicating that this primer pair and others based on the extended Ve sequence will be a valuable resource for use in a broader assessment of the diversity of Ve sequences among Mentha accessions and species.

Thus far, no verticillium wilt-related phenotypes have been associated with any of the Ve-like sequences reported in O. sativa or A. thaliana. The potential rice Ve homologs have been listed as such in GenBank, based on inference, as a consequence of the annotation of the O. sativa genome, but monocot species are not known to be hosts for Verticillium Nees fungi. To date, no expression or other studies have been performed to elucidate any connection between rice Ve homologs and nonhost resistance to verticillium wilt. For A. thaliana, differences in symptom development such as stunting and early flowering have been observed, but no verticillium wilt-resistant ecotypes have been identified (Steventon et al., 2001; Veronese et al., 2003). The latter researchers identified a locus (but not a gene) that was correlated with less severe symptom development in ecotype C-24, but they noted that disease development was difficult to parse from normal plant developmental and senescence features. Distinctly verticillium wilt-resistant and -susceptible M. longifolia accessions have been described (Vining et al., 2005), making this a much more appropriate species than either O. sativa or A. thaliana for identification and characterization of Ve homologs and assessment of their possible roles in conferring verticillium wilt resistance. As yet, the available data do not permit association of the mint Ve-like sequences with resistant or susceptible phenotypes.

The isolation of candidate resistance genes such as RGAs provides a useful foundation for further study of the genetics and genomics of plant disease resistance in mint. This study also describes the development of novel degenerate primers and their use in the isolation of potential Ve homologs in mint. The present work is the first report of R gene-like sequences in Mentha, and in the Lamiaceae, and the first to report the targeted isolation of Ve homologs outside of the Solanaceae. This work advances the development of M. longifolia as a model system for disease resistance in perennials and plants in general and constitutes a step toward identification of associations between candidate R genes and phenotypically assessed resistances against pathogens such as V. dahliae, for which clear resistance/susceptibility is not available in model plant species.

The isolation of a Ve homolog from M. longifolia would open opportunities for introduction of verticillium wilt resistance into wilt-susceptible commercial mint cultivars although the marker-assisted introgression of verticillium wilt resistance from M. longifolia into peppermint may be precluded by the sterility of the latter and by differences in ploidy. Transformation systems mediated by Agrobacterium tumefaciens Smith & Townsend are available for peppermint and other commercial mints (Niu et al., 1998). The Ve-like mint sequences reported here provide a potential tool for isolation of full-length, functional Ve alleles as candidates for introduction into wilt-susceptible commercial mint cultivars.