Genome-wide development of interspecific microsatellite markers for Saccharum officinarum and Saccharum spontaneum

LIU Lei ,WANG Heng-bo ,Ll Yi-han, ,CHEN Shu-qi ,WU Ming-xing ,DOU Mei-jie,QI Yi-yin,FANG Jing-ping,ZHANG Ji-sen,

1 Center for Genomics and Biotechnology/Fujian Provincial Key Laboratory of Haixia Applied Plant Systems Biology/Key Laboratory of Genetics,Fujian Agriculture and Forestry University,Fuzhou 350002,P.R.China

2 State Key Laboratory for Conservation and Utilization of Subtropical Agro-Biological Resources &Guangxi Key Laboratory for Sugarcane Biology,Guangxi University,Nanning 530004,P.R.China

3 College of Life Sciences,Fujian Normal University,Fuzhou 350007,P.R.China

Abstract Sugarcane has a large,complex,polyploid genome that has hindered the progress of genomic research and molecular marker-assisted selection. The user-friendly SSR markers have attracted considerable attention owing to their ideal genetic attributes. However,these markers were not characterized and developed at the genome-wide scale due to the previously lacking high-quality chromosome-level assembled sugarcane genomes. In this present study,744 305 and 361 638 candidate SSRs were identified from the genomes of S.officinarum and S.spontaneum,respectively. We verified the reliability of the predicted SSRs by using 1 200 interspecific SSR primer pairs to detect polymorphisms among 11 representative accessions of Saccharum,including S.spontaneum,S.officinarum,S.robustum,and modern sugarcane hybrid. The results showed that 660 SSR markers displayed interspecific polymorphisms among these accessions. Furthermore,100 SSRs were randomly selected to detect the genetic diversity for 39 representative Saccharum accessions. A total of 320 alleles were generated using 100 polymorphic primers,with each marker ranging from two to seven alleles. The genetic diversity analysis revealed that these accessions were distributed in four main groups,including group I (14 S.spontaneum accessions),group II (two S.officinarum accessions),group III (18 modern sugarcane hybrid accessions),and group IV (five S.robustum accessions). Experimental verification supported the reliability of the SSR markers based on genome-wide predictions. The development of a large number of SSR markers based on wet experiments is valuable for genetic studies,including genetic linkage maps,comparative genome analysis,genome-wide association studies,and marker-assisted selection in Saccharum.

Keywords: Saccharum,genome-wide,simple repeat sequences (SSR),microsatellite,molecular markers

1.Introduction

Sugarcane (Saccharumspp.hybrid) taxonomically belongs to the genusSaccharum,which is placed under the Poaceae family of Andropogoneae tribe and is an important commercial worldwide,providing sustainable economic growth and food security for tropical and subtropical regions of the world (Brand-Miller and Barclay 2017). Around 80% of sugar and 50% of alcohol are produced from sugarcane in the world (Waclawovskyet al.2010). TheSaccharumgenus contains six main interbreeding species: the two wild species areS.spontaneum andS.robustum,and the four cultivated species areS.officinarum,S.sinense,S.barberi,andS.edule(Hunsigi 1994). Dutch breeders in Java made the initial inter-specific crosses between noble caneS.officinarum(female) and wild caneS.spontaneum(male),followed by“Nobilization” of the hybrids by repeated backcrossing toS.officinarum. They thus obtained the disease and stress resistance ability ofS.spontaneumwhile recovering the high biomass and high sugar content ofS.officinarum(Brandes and Sartoris 1936). Modern sugarcane cultivars(2n=100-130) are genetically unstable,highly polyploid,heterozygous,and aneuploid with multi-specific origins and giant genome sizes (～10 Gb). They are interspecific hybrids,with 70-80% of the chromosomes derived fromS.officinarum(Souzaet al.2013),10-15%fromS.spontaneum,and the remaining 5-10% from interspecific recombination (D’Hontet al.1996). All commercial sugarcane breeding populations worldwide share a narrow genetic background and high linkage disequilibrium (LD) due to the limited number of original parental lines involved in sugarcane crosses (Arruda 2012). The genetic complexity,narrow genetic base,and huge genome size of sugarcane create many challenges for breeding. Information about the genetic background of wild progenitors of modern sugarcane cultivars is a vital resource for sugarcane improvement.

In contrast to the labor-intensive and time-consuming procedures of conventional breeding,molecular markers can greatly accelerate the breeding progress by efficiently selecting key agronomic and quality traits. However,sugarcane marker-assisted breeding aimed at further enhancing sugarcane sucrose and biomass production have largely been challenged by the lack of traitassociated makers. For decades molecular markers have had numerous applications allowing important advances in many fields,including cultivar identification (Cordeiroet al.2003),marker-assisted breeding (Panet al.2006),molecular fingerprint (Panget al.2019),genetic population and diversity analysis (Bertaniet al.2021),genetic mapping (Youet al.2019),and the fine mapping and cloning of genes (Wanget al.2017). In sugarcane,various molecular markers are being developed and employed to characterize sugarcane germplasm,including amplified fragment length polymorphisms (AFLPs) (Besseet al.1998;Aitkenet al.2006),restriction fragment length polymorphisms (RFLPs) (Cresteet al.2014),random amplification polymorphism DNAs (RAPDs) (Da Silvaet al.2008;Nawazet al.2010;Ramakrishnanet al.2016),simple sequence repeats (SSRs) (Xuet al.1999;Aitkenet al.2005;Caiet al.2005;Andruet al.2011;Nayaket al.2014;Zaweddeet al.2015),inter-simple sequence repeats (ISSRs) (Almeidaet al.2009;Devarumathet al.2012),and single nucleotide polymorphisms (SNPs) (Yanget al.2018). Among these molecular markers,the userfriendly SSR markers have attracted much attention owing to their ideal genetic attributes (Kantetyet al.2002),such as richness,high-polymorphism levels,multi-allelic nature,and co-dominant inheritance (Amaret al.2011;Baert-Desurmontet al.2016) and enjoyed great success in sugarcane genetic diversity research (Wanget al.2020),genetic mapping (Pan and Liu 2010),and marker-assisted selection (MAS) (Olatoyeet al.2019). Previous studies focused more on the development of expressed sequence tag (EST)-based markers (Oliveiraet al.2007) than genomic ones due to the lack of genomic sequence data.Although EST-SSRs are useful for genetic analysis,they have disadvantages. Specifically,they have relatively low numbers of polymorphisms and high concentrations in gene-rich regions of the genome,which may limit their usage,especially for the construction of linkage maps(Temnykhet al.2000). In addition,considering the high levels of polyploidy in the sugarcane genome,more genomic markers are needed for the advancement in genetics,genomics,and breeding of sugarcane.

Identifying SSRs from available genome sequences has proved to be a robust,rapid,and widely applicable strategy (Zhanget al.2012;Liu Cet al.2016;Cuiet al.2017). The advances in sequencing technologies have taken a significant leap in the last decade. Next generation sequencing (NGS) and third-generation DNA sequencing technologies have enabled the completion of a rapidly growing number of whole-genome sequencing projects. More and more genome sequence information can be obtained from public databases,promoting the development and application of SSR markers (Baraketet al.2011;Giovanniniet al.2012;Pottset al.2012;Emanuelliet al.2013;Hajmansooret al.2013). Recently,a haploidS.spontaneum,AP85-441,was assembled into 32 pseudo-chromosomes comprising eight homologous groups of four members each (Zhanget al.2018).Moreover,the genome ofS.officinarum,Louisiana Purple(LA-purple),was assembled into 80 pseudo-chromosomes in 10 homologous groups (unpublished data in our lab),which provided an opportunity to screen for SSRs across the sugarcane genomes of wild progenitors more thoroughly and efficiently.

The present study used the available dataset of sugarcane genome sequences to mine the SSRs throughout sugarcane assembled genomes using bioinformatics methods (Zhanget al.2018). We identified SSR motifs in a genome-wide manner and characterized these SSR motifs. We also attempted to evaluate genetic diversity and population structure among theS.officinarumandS.spontaneumusing SSR markers. This is the first report on genome-wide interspecific SSR identification and interspecific SSR markers development in sugarcane.The inventory of SSR markers proposed in the present study could facilitate sugarcane genetic and breeding studies. Our study provides a large-scale SSR marker dataset forSaccharumthat could be widely applied in cross-breeding identification,basic research for MAS,and interspecific cross-genetic map construction.

2.Materials and methods

2.1.Plant materials and DNA extraction

In the present study,the available sugarcane genomes ofS.officinarum(LA-purple) andS.spontaneum(AP85-441,1n=4x=32) (Zhanget al.2018) were used for SSR identification and validation. Eleven sugarcane elite accessions (1,Badila;2,Black Cheribon;3,Crystalina;4,Loethers;5,LA-purple;6,SES-208;7,FU89-1-1;8,Guangdong21;9,SC79-2011;10,POJ-2878;11,ROC-22) were used for SSR validation. Besides,39 sugarcane accessions,includingS.officinarum(nine accessions),S.spontaneum(14 accessions),S.robustum(five accessions),and modern sugarcane cultivars (11 accessions),were selected for the diversity analysis(Table 1). Total genomic DNA was isolated from fresh leaves by an optimized CTAB method (Yaoet al.2005).The concentration and quality of all DNA samples were determined using the NanoDrop spectrophotometer (ND-1000 version 3.1.1,USA),and all DNA samples were diluted to 25 ng L-1to enable subsequent PCR-basedinvitroamplification.

Table 1 Sugarcane (Saccharum spp.) accessions used in the study1)

2.2.SSR marker identification and primer development

The LA-purple genome consists of 80 pseudochromosomes with 10 homologous groups of eight members each (http://sugarcane.zhangjisenlab.cn/sgd/html/download.html). The haploidS.spontaneum,AP85-441,is an assembly of 32 pseudo-chromosomes comprised of eight homologous groups of four members each.

The genome sequences were used to identify SSR loci using a Perl program MISA,MIcroSAtellite identification tool (Thielet al.2003) available at thewebsite (http://pgrc.ipkgatersleben.de/misa/misa).The SSR loci containing perfect repeat units of 1-6 nucleotides only were considered. The criteria for identifying the minimum repetition time of mono-,di-,tri-,tetra-,penta-and hexa-nucleotide repeats were 10,7,6,5,4,and 4,respectively. Compound microsatellites were defined as those with the interval between two repeats motifs shorter than 100 nt (Wanget al.2015;Denget al.2016;Liu Fet al.2016).

To find the SSR loci with unique flanking sequences,the BLASTN (Joneset al.2002) search of the genomic sequence was carried out using the repetitive motif and the flanking sequences on both sides of the 150 bp(E-value cut-off of 1e-5) and filtered with 100% identity and 100% alignment length of the flanking sequences.We identified SSR loci with a single hit for marker development. Primer 3 (https://sourceforge.net/projects/primer3/) was used to design PCR primers for all candidate loci to amplify 100-300 bp products with 40-60% GC content. Primer sizes ranged from 18 to 22 bp with an optimum size of 20 bp,and the annealing temperature was set at 60°C.

The precision of the designed primers was estimated through the electronic PCR (e-PCR) program. The genome sequences ofS.officinarumandS.spontaneumwere used as templates for primers to blast with the following criteria: 100 bp margin,100-400 bp product sizes,and 1 bp mismatch. Primers obtained fromS.officinarumspecies were compared to those obtained fromS.spontaneumspecies andvice versa.Through e-PCR,the size of the PCR product for each pair of primers in each template was scored. If the size and position of PCR products were unique in twoSaccharumspecies,this combination of primers was recognized as the interspecific SSR primers forS.officinarumandS.spontaneum.

2.3.Experimental validations of SSR markers

SSR length is an important factor for SSR marker development. According to the length of SSRs,SSRs can be divided into two groups. Class I SSRs (hypervariable markers) are composed of those with ≥20 bp repeats in length,while Class II SSRs (potentially variable markers)contain repeats ＜20 bp in length (Temnykhet al.2001).Only the Class I SSRs and non-mononucleotides were selected for final genome-wide SSR marker development inSaccharumto improve their utilization efficiency.

Mononucleotides were filtered due to the difficulty in distinguishing bona fide microsatellites from sequencing or assembly error and because (A/T)n repeats in the coding region may be confused with polyadenylation tracks. SSR markers derived from longer perfect repeats (SSRs≥20 nucleotides in length) have been demonstrated to show high levels of polymorphisms from the experimental data in many organisms such as human (Iwasakiet al.1992)and rice (Choet al.2000).

To validate the e-PCR results,PCR amplification was performed in a 20-μL PCR volume which was composed of 0.5 μL template DNA (25 ng μL-1),0.5 μL of each primer,10 μL 2×TaqMaster Mix (Boshang Biological,Shanghai,China),and 8.5 μL ddH2O.PCR amplification was completed using the following touchdown program:94°C for 90 s for initial denaturation,then 34 cycles at 94°C (20 s)/59.5°C (20 s)/72°C (30 s),and the last extension step was 72°C for 5 min. The PCR products were identified by polyacrylamide gel electrophoresis.The fragment sizes of the PCR products were examined by identification with the DL50 DNA ladder (TaKaRa,Dalian,China).

2.4.Cloning and sequencing of PCR products

PCR products of SSRs that showed amplification in all sugarcane genotypes were excised from the 5% agarose gel through Gel Extraction Kit (OMEGA Bio-Tek,USA),cloned into the TA cloning vector (pEASYTM-T5 Zero,Transgen Biotech,Beijing,China),and sequenced to examine their conservation and cross-species transferability. The final edited sequences belonging to each locus were compared with the respective original SSR sequence using DNAMAN (Lynnon BioSoft,Vaudreuil,Quebec,Canada) for ascertaining the target domain/SSR conservation (Battilanaet al.2009).

3.Results

3.1.SSR motif content in the sugarcane genome

The genome sequences ofS.officinarumandS.spontaneumwere used for searching the SSRs with various types of required repeating motifs,from single nucleotides to hexanucleotides. The results of MISA showed that the assembled genome size ofS.officinarumwas 6 804 994 kb,in which a total of 744 305 microsatellites were identified with an average interval length of 9.14 kb between SSR loci (Table 2);the assembled genome size ofS.spontaneumwas 3 140 615 kb,in which a total of 361 638 microsatellites were identified with an average interval length 8.68 kb for each SSR loci (Table 2). Based on the assembled genomes ofS.officinarumandS.spontaneum,the distribution of the chromosome of microsatellites was investigated.The distribution of each chromosome showed thatmicrosatellite markers are uniformly distributed throughout the genome (Fig.1;Appendices A and B).

Table 2 Distribution of SSR loci in the genomes of Saccharum officinarum LA-purple and Saccharum spontaneum AP85-441

Of these SSRs,mono-nucleotide repeats accounted for 55.66 and 58.18%,which were the largest proportion,followed by di-nucleotide (19.67 and 17.91%) and tetra-nucleotide repeats (9.89 and 9.21%) in the twoSaccharumspecies,respectively (Table 2). The distribution of SSR motifs inS.officinarumwas consistent withS.spontaneum. The dominance of dinucleotides was attributed to an overrepresentation of AT/TA motifs. TGT/ACA motifs were dominant in trinucleotide SSRs,and AAAT/TTTA and AAAT/TTAA motifs were dominant in tetra-nucleotide SSRs (Fig.2).The total repeat number of T/A was significantly greater than that of C/G in all of the genome-wide SSRs. There were many types of penta-and hexa-nucleotide SSRs,each with low frequencies,ranging from 4.84 to 13.25%across twoSaccharumspecies genomes (Xuet al.2013). The numbers of mono-,di-,tri-,tetra-,penta-,and hexa-nucleotide,and compound-type motifs are listed in Table 2. The frequencies of SSR motifs with regard to the repeat number were examined (Fig.3).We found that for each motif type,the SSR abundance decreased dramatically with an increased number of repeat units. The most significant rate of decrease was for the hexanucleotide motifs.

3.2.Polymorphisms of the interspecific SSR markers between S.officinarum and S.spontaneum

Primers were designed for the candidate SSR loci using Primer3 Software. Amplification primer pairs were designed for 63 414 and 68 214 SSRs inS.officinarumandS.spontaneum,respectively. To test the amplification efficiency,specificity,and polymorphisms of the newly developed SSR primers,the genome sequences ofS.officinarumandS.spontaneumwere used as templates for e-PCR analysis. After SSR detection,a total of 41 725 (65.79%),5 292 (8.35%),3 260 (5.14%),and 13 137 (20.72%) of these primers could generate 1,2,3,＞3 loci inS.officinarum,respectively. A total number of 42 888 (62.87%),8 549 (12.53%),6 140 (9.10%),and 10 636 (15.59%) of these primers could generate 1,2,3,＞3 loci inS.spontaneum,respectively. The amplification efficiency of these primers was 100%.

By comparing SSR sequences between the two genomes,a new dataset containing 1 807 SSR markers was established. These SSRs had unique sizes and specific chromosomal positions for PCR products in bothSaccharumgenomes,indicating that they are valuable as genetic markers (Fig.1). The markers are evenly distributed on each chromosome of sugarcane and can be used as representative candidate markers for subsequent experimental analysis. Among these markers,SSRs with di-,tri-,tetra-,penta-,and hexa-nucleotide motifs accounted for 46.21,15.71,5.53,9.07,4.76,and 18.71%,respectively (Fig.4).

3.3.Development of SSR markers for interspecific identification of Saccharum

Finally,1 200 SSR markers were selected to check their amplification efficiency in theSaccharumgenus(S.officinarum,S.spontaneum,and sugarcane modern cultivars). We randomly selected 15 pairs of primers from each of the 10 homologous groups ofS.officinarumgenome for PCR amplification. Initially,the effectiveness of these primer pairs was estimated for theSaccharumspecies,includingS.officinarum(five accessions),S.spontaneum(four accessions),and modern sugarcane cultivars (two accessions) (Table 1) using polyacrylamide gel electrophoresis (PAGE). A total of 1 152 (96%)SSR markers could amplify at least one band,of which 660 showed interspecific polymorphisms among theSaccharumplant materials (Appendix C). These SSRs were co-dominant markers,and their main bands were clear,stable,and easily distinguished in each accession,while the remaining markers were dominant and displayed a weak band at the expected position (Fig.5-A). One of the reasons might be that the e-PCR would underestimate the complexity of sugarcane and there were more multiple loci amplifications in experimental PCR than e-PCR. The incompleteSaccharumgenome sequences used in this study might be responsible for the underestimation of the genome complexity.

To evaluate polymorphisms in newly developed SSR markers,100 validated microsatellite markers were amplified in theSaccharumgenus,including 39 accessions: nineS.officinarum,17S.spontaneum,fiveS.robustum,and 11 modern sugarcane cultivars (Table 1;Appendix D). All the markers displayed interspecific polymorphisms in these four sugarcane species (Fig.5-B).The highly successful amplification rate inSaccharumdemonstrated that the newly developed 1 807 microsatellite markers for theSaccharumgenome can be used for species identification and extend the sugarcane marker pool for functional genomics and molecular breeding.

To determine the effectiveness of the SSR markers,sequence similarities of the PCR products were compared among fourSaccharumspecies. The bands presented in all species were sequenced. All identical sequences of PCR bands had high similarities (85%) acrossS.officinarum,S.spontaneum,S.robustum,and modern sugarcane cultivars. Marker polymorphisms among the four species were mainly caused by the variation in the number of repeats of SSR motifs. Sequence alignment showed that all bands contained the (CCT)n and (TTTA)n repeat motifs. We found that the same primers were consistent in the amplification of different materials,the accession ofS.spontaneumYN-4 were observed to have six CCT,and four CCT were found inS.officinarumNC20. The sizes of the fragments were also different when amplified by the same primer in the same material.InS.hybridcultivar Co419,the two bands were found to have four and six tri-nucleotide repeat motif CAG,respectively (Fig.5-C). These sequence files were analyzed and unequivocally confirmed the cross-species conservation and transferability.

3.4.Genetic distance-based clustering analysis

Cluster analysis based on Nei’s genetic similarity coefficient using the UPGMA algorithm is shown in Fig.6.The 39 sugarcane accessions were divided into four groups (I,II,III,and IV) according to the clustering results.The similarity among all the accessions varied from 0.51 to 0.85,with an average value of 0.68. When the genetic similarity coefficient was 0.64,39 sugarcane accessions could be grouped into four clusters. Cluster I comprised 14 accessions,which belonged toS.spontaneum.Cluster II and Cluster III grouped two and 18 accessions,respectively,together belonging toS.officinarumand cultivated sugarcane species. Cluster IV had five genotypes withinS.robustum. The UPGMA dendrogram produced clusters separatingS.spontaneum(Cluster I)andS.officinarum,S.hybrid,andS.robustum(Cluster II,III,and IV) into two distinct groups when the genetic similarity was 0.51. The results were highly consistent with the material source information and further indicated that the newly developed sugarcane SSRs can be used to analyze the genetic diversity of theSaccharumgenus.

4.Discussion

Genetic analysis of sugarcane has been hampered because of its complex polyploid genome and lack of sufficiently informative markers. Thus,large numbers of useful polymorphic markers are required for germplasm diversity and analysis of the genetic relationship of sugarcane because of its large genome and complex genetic background. SSRs have increasingly become the marker of choice for these sorts of analyses as they have the advantage of simplicity,high abundance,ubiquity,high level of polymorphisms,stable co-dominance,and multiple-allelismamong genomes (Fraseret al.2004;Chenet al.2006;Hiranoet al.2011). In previous studies,thedenovodevelopment of SSRs was a costly,laborintensive,and time-consuming endeavor due to a lack of genome sequence information (Zaneet al.2010). The increasing availability of genomic sequence data has made it faster and less expensive to characterize and utilize genome-wide SSR markers from many species(Zhanget al.2018). In recent years,considerable progress has been made in sugarcane genomics,including the sequencing ofS.spontaneumAP85-441(Zhanget al.2018) andS.officinarumLA-purple(unpublished data in our lab),which facilitated successful efforts in gene family identification,genetic mapping,population genetic,and genome evolution analysis in sugarcane (Maet al.2020;Liet al.2021;Zhanget al.2021). In this paper,the newly assembled genomes ofS.officinarumandS.spontaneumwere used for largescale genome-wide SSR mining and generated 744 305 and 361 638 candidate SSRs,respectively. The motifs which have more than 7 nt were not considered in this paper because of their small proportion (Xuet al.2017).

4.1.Frequency and characteristics of SSRs in Saccharum genomes

In the current study,the average SSR density was one SSR every 9.14 and 8.68 kb inS.officinarumandS.spontaneum. Based on the survey across other plant genomes,on average,one SSR was found every 1.14 kb inArabidopsis(Lawson and Zhang 2006),3.6 kb in rice(Zhanget al.2007),220 kb in sorghum (Iniguez-Luyet al.2008),15.48 kb in maize (Xuet al.2013),220 kb in sorghum (Iniguez-Luyet al.2008),and 578 kb in wheat(Morganteet al.2002). Compared with other plants,we found that theSaccharumspecies with large genome sizes have a moderate SSR frequency,which agreed well with the generally negative correlation between genome sizes and SSR frequencies in plants (Deguchiet al.1996).

In sugarcane,the most common SSR types were mono-,di-,and tri-nucleotide repeats,which together account for 90% of all SSRs. Mono-,di-,and trinucleotides occupy 63,16,and 11% of the total SSRs inS.officinarumand 62,20,and 8% inS.spontaneum(Table 2),respectively. The frequency of SSRs within certain nucleotide compositions shows species and genome fraction specificity among eukaryotic genomes.The motifs containing A/T were more common than the motifs containing G/C (Table 2). A similar trend has also been observed in other plant genomes such as rice,sorghum,apple,and tobacco (Tonget al.2015;Fuet al.2016) that also exhibit a strong bias to AT-richness in SSR motifs,but this was not the trend inBrachypodium distachyonand maize (Liuet al.2012). AT/TA was the most common repeat unit across sugarcane genomes,representing 56.71% inS.officinarumand 57.92% inS.spontaneum. These results were consistent with the analysis of sorghum,tobacco,and apple,but not other Poaceae plants including wheat (Yuet al.2004),rice(Liuet al.2012),and maize (Duanet al.2006),in which AG motif were the most abundant. The TGT/ACA motif(16.04%) appeared the most frequently in sugarcane genomes. This AT-rich trend was similar toArabidopsis(AAG/TTC) (Lawson and Zhang 2006),tobacco (AAC/TTG) (Limet al.2004),and apple (AAC/GTT) (Heet al.2011) but different from barley,wheat,maize,rice,sorghum,and rye (Choet al.2000;Thielet al.2003;Lvet al.2019;Wanget al.2019),in which CCG/GCC is the most prevalent triplet while ATT/TAA is the least. The strong bias towards AT-rich in motifs especially in 1-3 nucleotide motifs could be attributed to two possible reasons: first,methylated C residue is more easily transformed into T (Levinson and Gutman 1987);second,SSR sites are generated by the high rates of strand slippage during DNA replication (Halsted and Devlin 2001;Andrés and Bogdanowicz 2012),during which time,the SSR motif based on the A-T structure type requires less energy than the G-C type,making the A-T baserich structure type easier to produce than the G-C type.However,with the increased nucleotide repetition motif from 3 to 6,the energy advantage of this hydrogen bond is relatively less.

4.2.Evaluation of genomic SSR markers in Saccharum species

SSR markers have been widely used for genetic analysis,high-throughput genotypic and map construction,and breeding in many crop species. Previous studies have used EST-derived markers inSaccharum,but only a limited number of genomic markers have been developed and evaluated (Da Silva 2001;Pintoet al.2004,2006;Aitkenet al.2005;Paridaet al.2009). The limitation of previously developed genomic SSR markers in trait dissection and discovery of candidate genes restricts their use in sugarcane genetic studies and molecular breeding. In this study,a total of 1 200 interspecific genomic SSR primer pairs were designed,and 1 152(96%) primer pairs proved effective. Due to the selfcomplementary requirement for dimer formation,AT/TA was not used to develop markers in our analysis (Wanget al.2011). Among these primers,100 pairs of SSR primers with polymorphisms were used to analyze the SSR polymorphisms among 39 sugarcane materials(Appendix D),and the total number of amplified bands was 313. The genetic similarity indices among the 39 sugarcane germplasms varied between 0.51 and 0.85 with an average of 0.68,which reflects a substantial level of genetic diversity within the differentSaccharumspecies and cultivated varieties that come from diverse geographic locations across the world. The results of clustering analysis showed that the sugarcane accessions of each group were derived from their respective progenitors and shared genetic intimacy with each other. SSR genetic similarity and cluster analysis of this study revealed that the offspring from the same cross combination had higher genetic similarity with each other,and the same series of sugarcane accessions mostly grouped in one cluster. The clustering results based on genetic similarity of SSR among sugarcane germplasms are largely consistent with the material source. For example,F134 (Co290×POJ2878) and Co419 (POJ2878×Co290)from the same parents had very high genetic similarity coefficients,therefore being located in the same cluster.

However,some results of inheritance similarity and clustering analysis are irreconcilable with the sources of sugarcane materials,which may be closely related to the cross-breeding process of sugarcane or the genetic composition of sugarcane. Sugarcane is an allopolyploid plant with high heterozygosity,and cross-breeding,backcrossing,and self-crossing will lead to high-level separations in offspring,so the offspring from the same cross combination may have a large genetic difference,while the offspring from different combinations may also have a small genetic difference. In addition,genealogical records of some cultivars failed to reflect their genetic relationships due to the phenomenon of self-crossing,powdering,and seed mixing.

The cluster analysis revealed a distant genetic relationship amongS.spontaneum,S.officinarum,cultivated sugarcane species,andS.robustum. One of the main reasons is the northern India origin of the diversity ofS.spontaneum,where the lowest chromosome number of accessions of this species (2n=40) were found(Zhanget al.2013).Saccharumofficinarumlikely evolved fromS.robustumas they share the same center of origin and diversity in Papua New Guinea (Skinneret al.1987).High genetic similarity between theS.officinarumclones and commercial cultivars indicates a large amount of introgression fromS.officinaruminto modern cultivars(Selviet al.2003;Singhet al.2011;Chandraet al.2013),confirming the consensus that modern sugarcane cultivars are aneuploidy,interspecific hybrids with chromosomes ranging from 2n=100 to 130,comprising 70-80%S.officinarum,10-20%S.spontaneum,and 10% recombinant chromosomes between the two species(Daugroiset al.1996).

The PCR-based DNA markers facilitate the determination of the genetic variability that occurs among conventional progenitor species and commercial cultivars (Singhet al.2019). The genomic constitution of different plant genetic resources can be accurately and finely evaluated by determining the population structure with different DNA markers (Nachimuthuet al.2015;Singhet al.2017). Based on the advances in genomic resources,SSR markers are preferred to be applied in rapid analysis of the genetic structures of breeding germplasms (Liu Cet al.2016;Singhet al.2019).

5.Conclusion

Our study is the first systematic and comprehensive analysis of the SSR markers from the sugarcane genome.We predicted 744 305 and 361 638 candidate SSRs in the genomes ofS.officinarumandS.spontaneum,respectively. Out of 1 200 SSR markers,660 displayed interspecific polymorphisms among the 11 representative accessions,proving the reliability of the examined SSRs.We also used 100 SSRs to detect the genetic diversity for 39 representativeSaccharumaccessions. Our study provides a valuable resource for the genetic analysis of sugarcane.

Acknowledgements

This work was supported by the National Key Research and Development Program of China (2021YFF1000101-5),the Science and Technology Planting Project of Guangdong Province,China (2019B020238001),the Natural Science Foundation of Fujian Province,China(2019J05066),the National Natural Science Foundation of China (41906096),and the Guangdong Laboratory for Lingnan Modern Agriculture and the State Key Laboratory for Conservation and Utilization of Subtropical Agrobioresources,China.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm