Total genomic DNA was extracted from dried leaves of the Japanese local variety of chayote, “Zairai Wase”, by CTAB method [20] . Microsatellites were developed according to a previously described dual-suppression-PCR technique [21] . First, DNA was digested with Alu I, Eco RV, Hae III, and Ssp I restriction enzymes to construct four DNA libraries. The restricted fragments were then ligated by using a DNA Ligation Kit Ver.2.1 (Takara Bio, Japan) with specific blunt adaptors (the first consisted of a 48-mer: 5’-GTAATACGACTCACTATAGGGCACGCGTGGTCGACGGCCCGGGCTGGT-3’ and the second consisted of an 8-mer, with the 3’-end capped by an amino residue: 5’-ACCAGCCC-NH₂-3’).

Flanked regions of the microsatellite were isolated from these four DNA libraries by amplification using com- pound microsatellite primers (AC)₁₀ or (GA)₁₀ and the adaptor primer AP2 (5’-CTATAGGGCACGCGTGGT-3’) designed from the 48-mer adapter. The amplified fragments were then integrated into a plasmid pGEM-T vector using pGEM-T Easy Kit (Promega); the plasmids were then transferred into Escherichia coli DH5α, according to the manufacturer’s instructions. The cloned fragments were amplified from the extracted plasmid DNA of positive clones using M13 forward and M13 reverse primers. The size of inserted fragments were checked by 1% agarose gel electrophoresis and the PCR products were directly sequenced with the M13(-21) primer using BigDye Terminator v3.1 Cycle Sequencing Kit (Life Technology) in the ABI 3130 Genetic Analyzer (Life Technology). Each fragment containing (AC)_n or (GA)_n sequence at one end was chosen for the next step.

The next step consisted of determining the sequence of the other flanking regions for each microsatellite. A pair of primers, IP1 and IP2, was designed from each sequence. IP1 was a nested specific primer, designed from the flanking region of the microsatellites, whereas IP2 was designed based on the sequence between IP1 and the microsatellite. All these primers were designed using Primer3Plus [22] . An adapter-primer for nested PCR (AP1: 5’-CCATCGTAATACGACTCACTATAGGGC-3’, designed from the 48-mer adaptor sequence) was also prepared. Primary PCR reactions were conducted on each constructed DNA library using the combination of IP1 and AP1 primers. Secondary PCR was conducted with a 100-fold dilution of the primary PCR products with IP2 and AP2. Single-banded fragments were then sub-cloned and sequenced. Finally, the primers IP3-forward and IP3-reverse were designed for each locus from the newly identified sequence between the AP2 and microsatellite sequences. The primer pairs IP3-forward/IP3-reverse were defined as the newly designed microsatellite markers.

Amplification of the newly designed markers was tested on 20 Mexican chayote accessions comprising 10 varietal groups with different fruit morphological characteristics (Supplementary Table 1). To obtain fluorescent-labeled products, amplification reactions were carried out using a 10 μl three-primer combination [23] as follows: 10 ng of genomic DNA; 5 μl of RedTaq^® ReadyMix™ PCR Reaction Mix (Sigma-Aldrich); and 0.2 μM of fluorescent-labeled universal primer, 1.5 μM of locus-specific forward primer with 5’ universal tails sequence, and 0.5 μM of locus specific reverse primer. PCR products were amplified using the following parameters: initial denaturation at 95˚C for 3 min; followed by 30 cycles of 95˚C for 30 s, 60˚C for 30 s, and 72˚C for 1 min; and a final elongation step at 72˚C for 1 h. The size of the PCR products was determined by capillary sequencer (3500 ×l Genetic Analyzer, Life Technologies).

General genetic diversity values were calculated at each locus. Pair-wise genetic distances among 20 individuals were calculated using “Codom-Genotypic” option to enable comparisons between codominant (microsatellites) and dominant (PBA) data output. Principal coordinate analysis (PCoA) was also carried out. All these analyses were carried out using GenAlEx 6.5 [24] . Polymorphism information content (PIC) of each marker was calculated using a PIC calculator [25] . The characteristics of the microsatellite markers for Mexican chayote are shown in Table 1.

^aPCR product with Japanese variety “Zairai wase”. Note: N = sample size; Na = Number of effective alleles; Ho = observed heterozygosity; He = expected heterozygosity; PIC = polymorphism information content.

Genetic diversity values from the same set of Mexican chayote samples were evaluated using PBA markers [11] [15] . The PBA primer sets comprised of three forward primers (CYP1A1F, CYP2B6F, and CYP2C19F) and three reverse primers (CYP1A1R, CYP2B6R, and CYP2C19R). Nine combinations of these primer pairs were tested and eight of them produced scorable bands (Table 2). PCR amplifications were performed using 10 ng of template DNA in a total reaction mixture of 10 μl containing 5 μl of RedTaq^® ReadyMix™ (Sigma-Aldrich) and 0.18 μM each of forward and reverse primers. This PCR was run at 94˚C for 5 min, followed by 30 cycles of 94˚C for 1 min, annealing temperature for 2 min, and 72˚C for 3 min, with a final extension of 72˚C for 10 min. Annealing temperature varied for each primer set, as shown in Table 2. PCR products were electrophoresed using 6% of acrylamide gels at 150 V for 3 h, followed by silver staining.

For statistical analysis, amplified DNA fragments were scored in a binary data matrix, where the presence of band was denoted as 1 and its absence as 0. A single data matrix was prepared by combining genotyping results of 8 primer combinations and general genetic diversity values were calculated. Genetic distances among individuals were calculated and PCoA was also carried out. The correlation between estimates of genetic distances based on PBA and microsatellite data was determined by Mantel test [26] . All data analyses were carried out using GenAlEx 6.5 [24] .

A total of 40 clones from the (AC)₁₀ and 24 clones from the (GA)₁₀ libraries were randomly chosen based on amplification size (≥500 bp) and were sequenced. Eighteen (45%) and 9 (37.5%) clones, respectively, from each library, had enough sequence length to design two primers between repeat motif and adapter sequences, and they were used to design IP1 and IP2 primers. In total, 15 primer pairs were designed, 7 from the (AC)₁₀ library and 8 from the (GA)₁₀ library. These primers were initially tested on a subset of S. edule individuals from the Japanese varieties to confirm reliable amplification. Of these, 4 amplified multiple loci while 11 were polymorphic and produced a single band (GenBank accession no.: AB871395 to AB871405).

Specie-specificity of the newly designed markers was confirmed by cross amplification with arbitrary selected other plant species. Four genotypes of heliconias (Heliconia bourgaeana, H. bihai, H. collinsiana and H. uxpanapensis) and four chayote genotypes were tested with primers designed in this study. Amplifications were observed only from chayote DNA samples, and not from heliconias. Equally, cross-amplification of microsatellite markers designed for heliconias (Hac-A116 and Hac-D1) [27] were tested with chayote samples. PCR products were observed only in the four Heliconia spp. DNA samples (Supplemental Figure 1). DNA sequences from the selected libraries were used to do a homology search against NCBI Nucleotide collection database (http://www.ncbi.nlm.nih.gov/), using blastn and no significantly high homology sequences were found (data not shown), confirming that DNA from chayote was used during the primers design and that DNA cross contamination with other species was avoided.

The 11 polymorphic primer pairs were further tested on 20 individuals from a Mexican collection (BANGESe, National Sechium edule Germplasm Bank, Huatusco, Veracruz, Mexico). One primer combination (Sed10) was discarded from further analysis because it produced too much stutter bands and size determination was not possible. Out of 10 loci tested, 7 microsatellites showed polymorphism, ranging from 1 to 7 alleles per locus (Table 1). The mean values for effective allele number, observed heterozygosity (Ho), and expected heterozygosity (He) were 3.1, 0.21, and 0.28, respectively. Polymorphism information values ranged from 0.00 to 0.68. The proportion of polymorphic loci was 70%, whereas isozyme studies reported 59.8% [5] and 57.14% [11] , suggesting that microsatellite markers had higher resolution in detecting polymorphisms within chayote varieties.

Among 20 Mexican chayote individuals tested, only one pair (SE-369 and SE-288) demonstrated identical patterns of allelic combinations, both of them belonged to the same varietal group, albus levis, and share identical fruit characteristics. However, individuals that belonged to other varietal groups (e.g., albus minor, SE-261 and SE-330) exhibited different genotyping patterns. There is still a need for development of additional markers to increase resolution so that microsatellite markers can be used for genotype identification and, in some cases, for marker-assisted selection. However, these results suggest that the DNA profiles based on the application of newly designed microsatellite markers could be used to differentiate chayote individuals and accessions.

Chayote is known as a predominantly cross-pollinated species but its low heterozygosity is reported by isozyme studies [1] [11] . In the present study, 9 out of 10 newly designed microsatellite markers also detected low level of heterozygosity (Ho = 0.00 to 0.20), except for Sed 01 (Ho = 0.85). Other genetic diversity studies on Cucurbitaceae using microsatellite markers demonstrated moderate to high heterozygosity (0.49 - 0.75 [28] , 0.26 - 0.79 [29] , 0.00 - 1.00 [30] , and 0.325 - 0.867 [31] ). Although individual species in the Cucurbitaceae family are generally cross-pollinated, chayote is known to be self-compatible [32] and a single plant of chayote shows a good fruit set and does not demonstrate clear inbreeding symptoms [33] . Therefore, cross-pollination rate in cultivated chayote may not be as high as assumed, since chayote is typically cultivated in home gardens and a small number of plants are used for seed source [11] . Reference [34] also reported low heterozygosity values (range 0.00 - 0.50) in melon samples evaluated with a set of 25 microsatellites and mentioned possible favoring endogamy in cultivated populations. Pollination with self and/or among close relatives during domestication and cultivation of the species may have contributed to the observed low heterozygosity values.

Although the set of Mexican chayote collection consisted of a broad range of varietal group with high diversity in fruit morphology (Supplementary Table 1), the scatter diagram of the PCoA derived by both microsatellite and PBA markers did not show a clear distribution pattern associated with a varietal group or fruit morphological characteristics (data not shown). Fruit morphological characteristics, such as color, shape, size, taste, and presence/absence of spines, are not environmental adaptation-related traits, unless directional selections by humans have strongly influenced chayote domestication. In addition, [35] reported that spines on fruits can be controlled by gibberellic acid application during development. Therefore, a clear relationship between genetic and morphological groupings may not have been observed.

From the 9 primer combinations evaluated for the PBA markers, 8 showed polymorphisms on the same set of Mexican chayote samples. The total number of polymorphic bands was 88, ranging from 2 (CYP2C19F/CYP2C19R) to 17 (CYP2C19F/CYP1A1R) (Table 2). The mean He value calculated with the PBA marker system was almost the same as that obtained by the microsatellite marker system (0.29 vs. 0.28, respectively), indicating that the 10 newly designed microsatellite markers in this study were useful in detecting genetic diversity levels of a Mexican chayote collection set. Mantel test showed that there was no significant correlation between the respective genetic distances estimated by the microsatellite and PBA data (r = −0.051; p = 0.40). This may be due to the existence of a genome-wide distribution of P450-related functional genes, which are detected by PBA markers, whereas microsatellites are detected polymorphisms at a single locus. These results highlight the different properties of each marker system. In a multivariate analysis by using microsatellite and PBA markers in rice, it also revealed different distribution patterns, and mentioned that the different types of molecular markers might contribute in the different aspects of the genetic structure [19] . Therefore, evaluation of genetic diversity combining different assessment methods would be helpful, especially for the species lack number of species specific markers available.