Artocarpus J. R. Forst. & G. Forst. (Moraceae) contains approximately 70 species of monoecious trees with a center of diversity in Malesia (Zerega et al., 2010). Species include several underutilized crops that can improve food security (Jones et al., 2011). In addition to breadfruit (A. altilis (Parkinson) Fosberg) and jackfruit (A. heterophyllus Lam.), Artocarpus contains lesser-known crops like cempedak (A. integer (Thunb.) Merr.) and terap (A. odoratissimus Blanco), and more than a dozen other species with edible fruits whose potential remains largely unexplored (Zerega et al., 2010).

Nuclear microsatellites developed for Artocarpus (Witherup et al., 2013) have been used in characterizing genetic diversity of breadfruit germplasm (Zerega et al., 2015). We present primers for 15 chloroplast simple sequence repeat (SSR) loci from transcriptomes of A. altilis and A. camansi that will complement the nuclear markers in analyzing genetic diversity and population structure. Chloroplast SSRs are usually mononucleotide repeats, and as nonrecombinant, maternally inherited loci (Wheeler et al., 2014), they allow confident identification of maternal and clone lines—often important in vegetatively propagated crops such as breadfruit. These markers were developed from next-generation sequencing (NGS) transcriptome data. This approach enables rapid marker development directly from sequences in the target organisms. Primers were tested in A. altilis (diploid and triploid) and A. odoratissimus. We also constructed an in silico data set from additional transcriptomes of A. altilis, its wild progenitor (A. camansi Blanco), and A. altilis × A. mariannensis hybrids to test for fragment size homoplasy, a common problem with chloroplast SSRs that can overestimate relatedness by masking sequence variations that do not change allele sizes (Wheeler et al., 2014).

METHODS AND RESULTS

Total RNA from two A. altilis accessions and one A. camansi accession (Appendix 1) was extracted using the QIAGEN RNeasy Universal Mini Kit (QIAGEN, Valencia, California, USA). Illumina TruSeq library preparation and sequencing in one lane of HiSeq 2000 (2 × 100, paired-end; Illumina, San Diego, California, USA) took place at Argonne National Laboratory (Lemont, Illinois, USA). Reads were de-multiplexed, quality-trimmed (>Q20 in a 5-bp window), and assembled using Trinity (Grabherr et al., 2011; Bolger et al., 2014). Chloroplast contigs were extracted using a BLAST search seeded with the Morus indica L. (Moraceae) chloroplast genome (GI: 89,574,460). Mono- and dinucleotide repeats were identified, aligned using BLAST, and screened for variability. Initially, primers for 16 chloroplast SSR loci were designed using Primer3 (Rozen and Skaletsky, 1999) (Table 1). Fifteen loci reliably amplified and were subjected to further testing.

To test for variability in A. odoratissimus, all loci were amplified in 105 accessions collected from four districts in Sabah, Malaysia (Appendix 2). PCR reactions were performed in two steps (Schuelke, 2000). For the first step, 10-µL reactions contained 5 µL of MyTaq Master Mix (Bioline USA, Taunton, Massachusetts, USA), 0.5 µL of 10 mg/mL bovine serum albumin (BSA), 0.25 µL of 10 µM forward primer with the M13 tail (5′-CAGGAAACAGCTATGAC-3′), 0.25 µL of 10 µM reverse primer, 3 µL of H₂O, and 1 µL of template DNA. PCR conditions for the first step were 94°C for 3 min; 13 cycles at 94°C for 30 s, 59.8°C for 30 s, and 72°C for 1 min; and 72°C for 10 min. The following were then immediately added: 2.5 µL MyTaq Master Mix, 0.25 µL of 10 mg/mL BSA, 0.125 µL of 2.5 µM MgCl₂, 0.25 µL of 10 µM labeled M13 primer (WellRED Dye D2, D3, or D4 [Beckman Coulter, Brea, California, USA]), and 1.875 µL of H₂O. PCR conditions for the second step were 94°C for 3 min; 27 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min; and 72°C for 10 min. Product was pooled as follows: 2 µL of D4-labeled product, 1 µL of D3, and 0.5 µL of D2. Pooled products were added to 30 µL of HiDi formamide (Azco Biotech, San Diego, California, USA) and 3.3 µL of 400-bp size standard (Beckman Coulter) and analyzed on a Beckman Coulter CEQ 8000 Genetic Analysis System. Alleles were scored using the CEQ 8000 software version 9.0 (Beckman Coulter).

To test for variability in A. altilis, all loci except AALTCP04, AALTCP07, AALTCP11, and AALTCP12 (which were less variable in transcriptomes) were amplified in 73 accessions of A. altilis from Vanuatu (Navarro et al., 2005, 2007), the Caribbean, and India (Appendix 1). Locus AALTCP14 followed the protocol described above. Other loci were amplified at the USDA in reduced PCR reaction volumes (step 1: 5 µL, step 2: 3 µL) without BSA using QIAGEN Multiplex PCR Master Mix (QIAGEN) and analyzed using ABI reagents on a 3730x1 DNA Analyzer and GeneMapper 5 (Applied Biosystems, Foster City, California, USA).

To test for variability in A. camansi and A. altilis × A. mariannensis hybrids and to explore the presence of homoplasy in these markers, loci were amplified in silico from the draft genome of A. camansi, the original four transcriptomes used for developing primers, and 18 additional transcriptome assemblies (Laricchia, 2014) (Appendix 1). Chloroplast contigs were extracted using the BLAST method described above, and amplification in silico took place following Bikandi et al. (2004). Some loci that failed to amplify because the region was split between two contigs or because a priming site was truncated were recovered using BLAST. Sequences were aligned using MUSCLE (Edgar, 2004), and a fragment-length data set was constructed. For both data sets, the number of alleles and a haplotype diversity index for each locus were calculated using GenAlEx (Table 2) (Peakall and Smouse, 2012).

Table 1.

Chloroplast SSRs developed in this study, showing region, primers, motif, melting temperature, suggested pool and dye color for multiplexing, and GenBank accession number for sequences from Artocarpus camansi (NTBG 960,576.001).

Allele sizes were recovered from >60 individuals of A. odoratissimus for all loci but one (37 individuals for AALTCP05), and from >60 individuals of A. altilis for all 11 tested loci (Table 2). In silico capture recovered sequences and fragment sizes from most transcriptomes for all loci except AALTCP13, which tended to be absent from transcriptomes (Table 2, Appendix 3). All loci were polymorphic in the breadfruit complex (A. altilis, A. camansi, and A. altilis × mariannensis hybrids), with A. camansi showing the greatest unbiased haplotype diversity. Although the in silico sample size was small, this finding is consistent with a domestication bottleneck in A. altilis with respect to its wild progenitor, A. camansi (Zerega et al., 2005). The polymorphism in AALTCP04 in A. camansi was not in the repeat motif, but in a 22-bp indel. Six loci (AALTCP03, AALTCP05, AALTCP08, AALTCP10, AALTCP11, and AALTCP12) were monomorphic in A. odoratissimus. Average alleles per locus was 2.5 in A. altilis, 2.3 in hybrids and A. odoratissimus, and 2.2 in A. camansi. For comparison, average alleles per locus in the previously described nuclear markers using the same individuals as our in silico data set (with one parent-sibling substitution in A. camansi) were 2.1 in A. camansi, 5.0 in A. altilis, and 4.6 in hybrids (Zerega et al., 2015).

The in silico data revealed within-species homoplasy due to multiple SSRs in the same amplified fragment in loci AALTCP01, AALTCP09, and AALTCP10. All other loci showed no evidence of fragment-length homoplasy. We also identified single-nucleotide polymorphisms in flanking regions outside the target repeats in loci AALTCP01, AALTCP02, AALTCP07, AALTCP09, AALTCP12, and AALTCP14 (in A. camansi only for AALTCP02, AALTCP09, AALTCP12, and AALTCP14). These loci thus may provide additional resolution when a sequencing approach is used as opposed to a fragment-size approach.

Table 2.

Summary of allele size data for species in the breadfruit complex.

CONCLUSIONS

These chloroplast SSR loci will be useful for rapid and low-cost genotyping in Artocarpus and possibly in other Moraceae species, given the level of conservation typical in chloroplast genomes. By enabling the isolation of maternal lineages, these markers can be applied to characterizing genetic diversity, tracing seed and vegetative dispersal history, and assessing relatedness of germplasm accessions. Even as NGS tools become more widespread, SSRs remain important, as they enable efficient genotyping with common laboratory equipment. This is particularly relevant for nonmodel, underutilized crops, which are often grown in less developed areas where only basic genotyping equipment is available.