Characterization and transferability of microsatellites for Gentiana lawrencei var. farreri (Gentianaceae).

PREMISE OF THE STUDY: Microsatellite markers were developed for a medicinal herb, Gentiana lawrencei var. farreri (Gentianaceae), for the future assessment of population genetic structure and potential hybridization events with related taxa. METHODS AND
RESULTS: Using the 454 FLX+ sequencing platform, we obtained 81,717 clean reads with an average length of 291 bp. A total of 3031 primer pairs were designed, and 96 were selected for validation. A set of 20 fluorescently labeled primer pairs was further selected and screened for polymorphisms in three G. lawrencei var. farreri populations and one G. veitchiorum population. Among the four populations, the average number of alleles per locus was 15.2. Finally, a set of 17 unlinked loci were determined to be in Hardy-Weinberg equilibrium after two linked loci were removed.
CONCLUSIONS: The identified simple sequence repeat markers will be useful for genetic diversity and evolution studies in G. lawrencei var. farreri and related taxa.

Gentiana lawrencei Burkill var. farreri (Balf. f.) T. N. Ho (Gentianaceae) is a perennial wildflower that is endemic to the Qinghai–Tibetan Plateau (QTP) and used in traditional Chinese and Tibetan medicine (Ho and Liu, 2001). Gentiana veitchiorum Hemsl., which is a perennial wildflower as well, is closely related with G. lawrencei var. farreri in phylogeny (Ho et al., 1996; Ho and Liu, 2001; Favre et al., 2010). The two species’ plastomes have very similar structures and low sequence variation (Fu et al., 2016; Fu, unpublished data). The morphological distinctions between the two species are primarily in leaf shape and flower color, with G. lawrencei var. farreri having linear stem leaves and a pale blue corolla and G. veitchiorum having narrowly elliptic stem leaves and an intense blue corolla. The two species are common in the QTP alpine meadow and sympatric in the central QTP (Ho and Liu, 2001). According to our field observations and previous studies, both species are outcrossing (Hou et al., 2009). The flowering and fruiting periods of G. lawrencei var. farreri and G. veitchiorum are from August to October and July to October, respectively (Ho and Liu, 2001). Therefore, hybridization and gene flow may occur between them. If simple sequence repeat (SSR) markers developed in G. lawrencei var. farreri could be applied in G. veitchiorum, they could be used as molecular markers for studies of genetic structure, hybridization, species divergence, and gene flow of these two closely related species. Because phylogenetic relationships and species divergences of numerous Gentiana L. species are still unclear (Ho and Liu, 2001; Favre et al., 2010), developed polymorphic SSRs would also aid in studies of population genetics and evolution within Gentiana.

Methods and results

In this study, we tested the amplification and evaluated polymorphisms using four populations: three populations of G. lawrencei var. farreri and one population of G. veitchiorum (Appendix 1). Total genomic DNA was extracted from dried leaves using a modified cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle, 1987). Genomic DNA (600 μg) was fragmented with nitrogen at 45 pounds per square inch (psi) for 2 min; 500–800‐bp fragments were used for further study. The fragments were ligated to adapters using T4 ligase and amplified by PCR with corresponding adapter primers. Genome libraries were constructed using eight biotin‐labeled probes (pGA, pAC, pAAT, pAAC, pAAG, pATGT, pGATA, and pAAAT) and a selective hybridization with streptavidin‐coated beads (Invitrogen, Grand Island, New York, USA) (Armour et al., 1994; Kandpal et al., 1994; Glenn and Schable, 2005). Library quality inspection and sequencing of clones was carried out following the descriptions in Yang et al. (2012). Subsequently, entire libraries were equally pooled and sequenced on a Roche 454 GS FLX+ sequencer (454 Life Sciences/Roche, Penzberg, Germany) using titanium reagents at Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). Processing and analysis of the sequencing data were performed with GS‐FLX+ software version 2.9 (454 Life Sciences/Roche). Using a series of normalization, correction, and quality‐filtering, the sequencing data were processed to remove low‐quality and adapter sequences using EMBOSS (Rice et al., 2000). A total of 87,097 reads were generated from the pooled G. lawrencei var. farreri library, and 81,717 read sequences were used for further analysis after adapter removal. The average read sequence length was 291 bp, with a maximum length of 791 bp. Clean reads were deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (accession number SRP101615).

The MIcroSAtellite identification tool (MISA; Thiel et al., 2003) was used to identify reads that contained SSRs. The minimum motif repeat was defined as six for di‐, and five for tri‐, tetra‐, penta‐, and hexanucleotides. A total of 6381 SSRs were identified. All SSRs with flanking sequence lengths ≥20 bp were assigned into clusters based on 98% similarity of flanking sequences using UCLUST version 1.2.22q (Edgar, 2010). We determined SSR lengths in each cluster with Perl and defined the SSR length polymorphism value as n (1, 2, …) when there were n kinds of SSR lengths in a cluster. Among the 4831 SSR‐containing sequences, a total of 2168 SSRs with flanking sequence lengths ≥20 bp were detected and assigned into 1023 clusters. Of these, 207 clusters (20.33%) had an SSR length polymorphism value ≥2; subsequently, a total of 3031 primer pairs were designed in Primer3 (Rozen and Skaletsky, 1999).

To amplify microsatellite regions, PCR amplifications were carried out in a 15‐μL reaction volume containing approximately 15 ng of template DNA, 1× PCR buffer (with MgCl2), 0.2 μM of each primer, 0.4 μM of each dNTP, and 1 unit of Taq DNA polymerase (TaKaRa Biotechnology Co., Dalian, Liaoning, China). The PCR cycling profile included an initial step at 95°C for 5 min; followed by 35 cycles of 95°C for 45 s, 50–55°C for 30 s, and 72°C for 30 s; and final extension at 72°C for 7 min. We analyzed the PCR products using a 3730XL Genetic Analyzer Sequencer (Applied Biosystems, Foster City, California, USA). Allele sizing was performed by comparing alleles with a GeneScan 500 LIZ Size Standard (Applied Biosystems) in GeneMapper version 4.0 (Applied Biosystems). In this study, 96 primer pairs were selected for validation using G. lawrencei var. farreri samples, and 65 primer pairs produced clear, unique amplification products of the expected size by screening in 2% agarose electrophoresis. Then, G. veitchiorum sample amplification resulted in 47 out of the 65 primer pairs producing clear and unique amplification products. A set of 20 SSR markers that amplified the most easily scorable fragment polymorphisms was chosen for labeling with a fluorescent dye (Table 1) and to evaluate polymorphisms in four populations from the two species.

Table 1

Characteristics of 20 microsatellite loci and primer pairs developed from Gentiana lawrencei var. farreri

Locus	Primer sequences (5′–3′)	Repeat motif	Fluorescent dye	T a (°C)	Allele size range (bp)	Total no. of alleles	GenBank accession no.
Law4	F: TGCAACGGTCACACTTTCTT	(GA)28	FAM	55	226–296	21	MG008318
	R: AAGCTCTTGGCAGATCCTGA
Law5	F: TTTGACCGTGCAATTTCTGA	(CAA)28	FAM	55	199–265	16	MG008319
	R: ATGTGGTCGGCAAAGAATTT
Law12	F: AGTGGCACAAAAACGGACTC	(ACA)22	FAM	55	85–202	3	MG008320
	R: AGCTCGGATTTTGGTTGATG
Law19	F: ATCAGATGGTTCGACAAGGG	(CAA)19	FAM	55	130–202	8	MG008321
	R: GGCCTCTCTCTTCCCAATTC
Law24	F: TGATGCACTCTTCCCATGAA	(CA)17	FAM	55	145–192	15	MG008322
	R: GGGTTTTGTGTGCGAAGTTT
Law25	F: CCGAGGTCGATCCTACAGAG	(AG)17	FAM	55	178–230	6	MG008323
	R: AAACGCTTTTGGTTTGGTTG
Law32	F: CGACGGTACGACCTCAACTT	(CAA)16	FAM	58	164–275	21	MG008324
	R: CGCTGACCTATCGTCATCCT
Law34	F: ATATTTTCGGCCATAAGGGG	(AG)16	FAM	55	120–184	8	MG008325
	R: AATCGAAGAACCGCAAAATG
Law37	F: CCCGGTTTTTACCTCTTTCC	(TCT)15	FAM	55	122–161	12	MG008326
	R: GCCCATTAAGCCATTTCTCA
Law41	F: AATGAATTGTGGTTGTGCCA	(AG)15	FAM	58	281–325	17	MG008327
	R: CAAGGACCCTGAGTGTTTCG
Law43	F: TTGGATTAGTGTACTTGG	(ACA)15	HEX	58	104–251	23	MG008328
	R: TGCCTGATTTTGTTGCTAGG
Law45	F: CCAGTTTTGTAAGCTCTTCTAGGC	(AC)15	HEX	58	167–221	21	MG008329
	R: TATGATCCTGGTCCCAGAGG
Law54	F: GCAGGCAAGATGCAGATACA	(AG)14	HEX	58	122–226	30	MG008330
	R: CAATAGGCGACTGGTTGGTT
Law57	F: GCTAGTTTTTCTTCTTAATTTGGTGG	(TTG)13	HEX	55	189–366	22	MG008331
	R: ATAACTGTTTCGCCATTCGG
Law70	F: CCTCATACCCGGAAAGTGTG	(TTC)12	HEX	55	159–198	13	MG008332
	R: CAGAAGCGTATCGCGTAGAA
Law71	F: CCTTAGGCAAAACCATTCCA	(TTC)12	HEX	58	163–256	19	MG008333
	R: TCAGCAGTTGAGTCTTCCTCC
Law77	F: TCATTTGCTTTAGGATCCGC	(CTT)12	HEX	58	141–201	11	MG008334
	R: CACACGCAAATTCAAATGCT
Law87	F: CCATCAACGCCTACAGGTTT	(TAT)11	HEX	55	136–196	14	MG008335
	R: ATATAGGCGGTCAGAAGGCT
Law88	F: CAGAGGTCGAAAACCAGGAG	(GTT)11	HEX	58	321–357	9	MG008336
	R: CATGGGCAGATTCCTCCTAA
Law95	F: CTTTGGCAGTCACACCATTG	(AC)10	HEX	55	183–221	14	MG008337
	R: CGAACTTTCTGCGTGAAACA

T a = annealing temperature.

Null allele presence, large allele dropout, and scoring errors were checked using MICRO‐CHECKER version 2.2.3 (van Oosterhout et al., 2004). The null allele frequency (r) and linkage disequilibrium (LD) between all pairs of polymorphic loci were calculated using GENEPOP version 4.0 (Rousset, 2008). LD was tested with 10,000 permutations. The number of alleles per locus, observed and expected heterozygosity, and deviations from Hardy–Weinberg equilibrium (HWE) were calculated using ARLEQUIN version 3.5 (Excoffier and Lischer, 2010). HWE was tested with 1,000,000 Markov chain steps. Characteristics of these 20 SSR loci are listed in Table 1. Among the 20 loci, the number of alleles ranged from three to 30 in the two species, with an average of 15.15 alleles per locus. All loci had a small amount of null alleles on average (r < 0.15) among all populations and thus were included in further analysis.

In the three populations of G. lawrencei var. farreri and a single population of G. veitchiorum, there was an average of 15.2 alleles. In the four populations, observed heterozygosity values ranged from 0.222 to 1.000, whereas expected heterozygosity values varied between 0.211 and 0.952 (Table 2). Of the 20 loci, 11 showed significant deviation from HWE in one or more populations, and only one locus (Law25) showed significant deviation from HWE in all four populations (Table 2). Fourteen locus pairs showed significant LD (P < 0.01) across the four populations (Appendix 2). After removing the loci Law12 and Law19, only three locus pairs showed significant LD (P < 0.01).

Table 2

Results of initial primer screening of 20 microsatellite loci developed for Gentiana lawrencei var. farreri in three populations of G. lawrencei var. farreri and one population of G. veitchiorum.a

Locus	Gentiana lawrencei var. farreri									Gentiana veitchiorum
	HY (N = 18)			GZ (N = 18)			XGLL (N = 18)			SP (N = 18)
	A	H o	H e	A	H o	H e	A	H o	H e	A	H o	H e
Law4	8	1.000	0.770	11	1.000	0.833	9	1.000	0.821	16	1.000	0.929
Law5	6	1.000	0.726b	14	1.000	0.913	5	0.900	0.800	9	1.000	0.863
Law12	3	1.000	0.586b	2	1.000	0.517b	2	0.857	0.508	3	0.882	0.533b
Law19	5	1.000	0.692	2	1.000	0.514b	2	1.000	0.514b	4	1.000	0.646b
Law24	6	1.000	0.631b	10	0.941	0.868	5	0.857	0.670	8	1.000	0.738b
Law25	3	1.000	0.603b	4	1.000	0.656b	3	1.000	0.541b	4	1.000	0.624b
Law32	7	1.000	0.754	15	1.000	0.914	7	1.000	0.793	6	1.000	0.800
Law34	6	0.444	0.394	2	0.667	0.457	3	0.389	0.332	4	0.222	0.211
Law37	6	1.000	0.686	5	1.000	0.667	7	1.000	0.709	8	1.000	0.701b
Law41	15c	1.000	0.937	10	1.000	0.843	4c	1.000	0.733	8c	1.000	0.882
Law43	13	1.000	0.913	10	1.000	0.839	14	1.000	0.908	13	1.000	0.884b
Law45	8	1.000	0.911	11	1.000	0.903	9	0.923	0.880	17	1.000	0.938
Law54	15	1.000	0.948	10	0.938	0.893	10	1.000	0.843	15	1.000	0.952
Law57	6	1.000	0.692	9	1.000	0.890	7	1.000	0.779	9	1.000	0.869
Law70	11	1.000	0.903	8	1.000	0.794	8	1.000	0.790	8	1.000	0.816
Law71	12	1.000	0.892	6	1.000	0.617b	9	1.000	0.907	11	1.000	0.883
Law77	4	1.000	0.577b	6	1.000	0.712	2	1.000	0.571	9	0.938	0.851
Law87	9	1.000	0.781	7	1.000	0.762b	7	1.000	0.804	6	1.000	0.726
Law88	4	1.000	0.650b	8	1.000	0.840	3	1.000	0.714	3c	0.333	0.733
Law95	12	0.933	0.897	10	1.000	0.867	5	1.000	0.692	8	1.000	0.806
Mean	7.95	0.969	0.798	8	0.977	0.816	6.05	0.946	0.736	8.45	0.919	0.804

A = total number of alleles per locus; H e = expected heterozygosity; H o = observed heterozygosity; N = sample size for each population.

Locality and voucher information are provided in Appendix 1.

Significant departure from Hardy–Weinberg equilibrium at P < 0.01.

Null allele present.

Conclusions

The 20 validated microsatellite primer pairs in this study will be useful for further studies on population genetics, phylogenetics, and evolution of the large genus Gentiana, especially G. lawrencei var. farreri and G. veitchiorum. The 6381 microsatellites obtained in this study offer a foundation for further research on this large genus, such as marker‐assisted breeding and functional characterization of genes related to trait formation.

Taxon	Population code	Voucher no.	N	Location	Geographic coordinates	Altitude (m)
Gentiana lawrencei Burkill var. farreri (Balf. f.) T. N. Ho	HY	Fu2016022	18	Ganzi, Sichuan Province, China	32°14′58″N, 102°29′21″E	3597
	GZ	Fu2016051	18	Ganzi, Sichuan Province, China	31°39′44″N, 99°44′02″E	3495
	XGLL	Fu2016146	18	Xianggelila, Yunnan Province, China	28°18′59″N, 99°45′09″E	3881
Gentiana veitchiorum Hemsl.	SP	Fu2016029	18	Songpan, Sichuan Province, China	32°59′36″N, 103°41′35″E	3386

N = sample size for each population.

Vouchers are stored in the Herbarium of Luoyang Normal University, Henan, China.

Locus	Law4	Law5	Law12	Law19	Law24	Law25	Law32	Law34	Law37	Law41	Law43	Law45	Law54	Law57	Law70	Law71	Law77	Law87	Law88	Law95
Law4	*
Law5	—	*
Law12	—	—	*
Law19	—	—	+	*
Law24	—	—	—	—	*
Law25	—	—	+	+	—	*
Law32	—	—	—	—	—	—	*
Law34	—	—	—	—	—	—	—	*
Law37	—	—	+	+	—	+	—	—	*
Law41	—	—	—	—	—	—	—	—	—	*
Law43	—	—	—	—	—	—	—	—	—	—	*
Law45	—	—	—	—	—	—	—	—	—	—	‐	*
Law54	—	—	—	—	—	—	—	+	—	—	—	—	*
Law57	—	—	—	—	—	—	—	—	—	—	—	—	—	*
Law70	—	—	—	+	—	—	—	—	—	—	—	—	—	—	*
Law71	—	—	+	+	—	—	—	—	—	—	—	—	—	—	—	*
Law77	—	—	+	+	—	—	—	+	—	—	—	—	—	—	—	—	*
Law87	—	—	—	+	—	—	—	—	—	—	—	—	—	—	—	—	—	*
Law88	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	*
Law95	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	*

+ = significant linkage disequilibrium (P < 0.01); — = nonsignificant linkage disequilibrium; * = no data available.