Mapping QTL underlying tuber starch content and plant maturity in tetraploid potato

Jianwu Li,Yihang Wang,Guohong Wen,Gaofeng Li,Zhang Li,Rong Zhang,Sheng Ma,Jun Zhou*,Conghua Xie*

a Key Laboratory of Potato Biology and Biotechnology,Ministry of Agriculture and Rural Affairs,Wuhan 430070,Hubei,China

b Key Laboratory of Horticultural Plant Biology,Huazhong Agricultural University,Ministry of Education,Wuhan 430070,Hubei,China

c College of Horticulture and Forestry Sciences,Huazhong Agricultural University,Wuhan 430070,Hubei,China

d Gansu Academy of Agricultural Sciences,Lanzhou 730070,Gansu,China

Keywords:Potato QTL Tetraploid Tuber starch content Plant maturity

A B S T R A C T Tuber starch content and plant maturity are two important agronomic traits of potato.To investigate the complex genetic basis of these traits in the cultivated potato,as well as the relationship between them,we developed a linkage map in a tetraploid population of 192 clones derived from the cross Longshu 8×Zaodabai and mapped quantitative trait loci(QTL)for tuber starch content and plant maturity using data collected in three diverse environments over two years.We detected eleven QTL for tuber starch content distributed on seven chromosomes,of which four,on chromosomes I,II,and VIII,were expressed in at least three environments.For plant maturity,we identified six QTL on chromosomes II,IV,V,VII,and XI,one of which,on chromosome V,showed LOD peaks ranging from 45.2 to 62.5 cM and explained 21.6%-26.6%of phenotypic variation was expressed in five of the six environments.Because the reproducible QTL for plant maturity and tuber starch content mapped to different chromosomes and neither overlapping QTL,nor any genetic interaction between QTL were detected,we infer that tuber starch content and plant maturity are controlled by independent genetic loci.This inference supports the prospect of breeding potato for both early maturity and high starch content.

1.Introduction

Potato(Solanum tuberosum L.)is the world's third largest food crop and is produced on all continents except Antarctica.Potato production has increased dramatically in developing countries in the past two decades,and has now overtaken that in the developed world,underlining the growing importance of potato as a staple food crop to meet the demands of increasing human populations.As a photosynthate storage product,potato starch,which accounts for 10%-25%of fresh tuber weight[1],is used in not only the food but the paper,textile,building materials,pharmaceutical,and chemical industries[2].The tuber starch content of middle European varieties ranges from 10%to 17%for table potatoes,from 14%to 20%for processing potatoes(e.g.,for chips or French-fries)and up to 25%in industrial potatoes[3].Tuber starch content in potato is controlled by multiple genetic loci and is strongly affected by environmental factors[3,4].Because the cultivated potato is a tetraploid(2n=4x=48)with tetrasomic inheritance and intolerance to inbreeding[5,6],genetic analysis of tuber starch content in cultivated potato is difficult.

Quantitative trait locus(QTL)mapping can suggest the minimum number and genomic position of genetic loci controlling a complex trait[4].To elucidate the genetic basis of tuber starch content,QTL analysis has been conducted in various segregating populations.The first analysis[7]was performed in a diploid population and identified 10 QTL for tuber starch content(measured as specific gravity)on chromosomes I,II,III,V,VII,and XI.Subsequently,18 QTL were detected on all 12 chromosomes in two diploid potato populations K31 and LH[8],and QTL for amylose and starch content were identified on five chromosomes each in CE,a diploid population of 249 full-sib progeny[9].Recently[10],12 QTL for tuber starch content were identified on seven potato chromosomes,I,II,III,VIII,X,XI,and XII,with a major QTL located on chromosome I.QTL analyses have also been performed in tetraploid populations.In a population derived from the cross 12601ab1×Stirling,a single QTL for tuber dry matter content was identified on chromosome V[5].Independent QTL for specific gravity and dry matter were identified[11]in a population derived from a cross of the chip-processing cultivar Atlantic and an advanced breeding selection,B1829-5.To investigate the genetic basis of tuber starch content in broad genetic background,association analysis between the candidate loci and phenotype was performed in a panel of 243 tetraploid cultivars and advanced clones and revealed 14 loci for tuber starch content distributed on chromosomes II,III,V,VII,VIII,IX,and XI[1].These diverse results suggest that tuber starch content is controlled by multiple genetic loci that may be genetic backgrounddependent.QTLs identified in different potato populations,especially between tetraploids and diploids,are not directly comparable.

An agronomic trait of potato,plant maturity,is key to adaptation to diverse agricultural systems.It is controlled by both genetic and environmental factors.Classical genetic studies suggested[12]that the inheritance of tuber initiation(or early maturity)was controlled by a few major genes,recessive genes,and some minor genes.An early QTL study[12]mapped 11 QTL for tuberization(including maturity)on seven chromosomes in two diploid populations,BCT and BCB,with a QTL explaining 27%of the variance located on chromosome V.In recent years,more research focused on QTL for maturity in different diploid and tetraploid populations has been reported.QTL have been located on most of the 12 chromosomes,among them major QTL for maturity located on chromosome V[5,14,15].Kloosterman et al.[16]cloned the maturity gene StCDF1(PGSC0003DMG400018408)from a QTL located on the short arm of chromosome V associated with early maturity and tuber initiation in a diploid population.Massa et al.[17]then constructed a SNP-based genetic map in a tetraploid population (MSL603),and identified a major QTL for vine maturity on chromosome V near the StCDF1 gene.In another tetraploid population A05141,QTL for eight traits including citric acid,growth habit,bud-end fry color,early blight score,tuber glucose,tuber shape and Verticillium wilt score co-localized with a vine maturity QTL on chromosome V where the StCDF1 gene was cloned from[18].Knowledge of other or new maturity genes in other diploid and/or tetraploid potato materials is lacking.

Most of potato tuber dry matter is starch.A higher dry matter production usually represents a larger starch yield potential.Tubers of early-maturing varieties tend to have lower dry matter content than late main-crop varieties[19].Tuber fresh and dry matter weight per unit area tended to increase with cultivar maturity date when both microtubers and conventional tubers were used as seed sources[20].However,the genetic relationship between potato tuber starch content and plant maturity has remained unstudied.

Here we report genetic loci for tuber starch content and plant maturity detected in six environments in a segregating tetraploid mapping population.The genetically independent status of these two traits supports the prospect of breeding cultivars for both early maturity and high starch content.

2.Materials and methods

2.1.Plant materials

The tetraploid(2n=4x=48)F1population(LZ)consisted of 192 clones from a cross made in 2010 between cultivars Longshu 8(L8)(female parent)and Zaodabai(ZDB)(male parent).L8 was developed by Gansu Academy of Agricultural Sciences from the cross Atlantic×L9705-9 and showed late(111 to 126 days from plant emergence to physiological maturity)maturity and high tuber starch content(20%-23%of fresh tuber weight)in diverse environments.ZDB was derived by Benxi Potato Institute of Liaoning province from the cross Wuzibai×174-128,and showed early(64-86 days)maturity and low tuber starch content(11%-14%)in different environments.They were evaluated in three diverse agroecological environments over two years to ensure that the phenotypes were stable.

True seeds of population LZ were surface-sterilized and germinated in vitro in plant regulator-free and sucrose-free MS medium[21].Two parents and 232 progeny were propagated in vitro and were transplanted into vermiculite to produce minitubers in years 2011 and 2012.

2.2.Field trails

Two parents and 192 progeny clones were evaluated in years 2012 and 2013 in three ecological zones of Gansu province,including Tianshui(34°61′N,105°65′E,1650 m.a.s.l.,rainfall 403 mm,annual average temperature 11.5°C,loessal soil),Dingxi(35°06′N,103°58′E,2270 m.a.s.l.,rainfall 421 mm,annual average temperature 5.7°C,dark felty soil),and Zhangye(38°50′N,100°22′E,1548 m.a.s.l.,rainfall 171 mm,annual average temperature 7.4°C,gray desert soil).All six field trials were arranged using a factorial randomized complete block design with three replications.Ten minitubers of each progeny and the two parents were planted in a onerow plot for each replication.

2.3.Measurement of tuber starch content and plant maturity

Plant maturity was determined as days from plant emergence to yellowing of 70% of leaves, following Kawakami et al. [22]. In all six environments, maturity was recorded when seven of the 10 plants in each plot reached maturity. All phenotypic data for plant maturity collected in six environments are presented in Table S3.

To avoid possible effects on specific-gravity measurement resulting from changes in air and water temperature,and from tuber physiological changes,tubers were harvested when the plants reached maturity and all tubers of each plot were used for measurement.Water temperature was controlled at 17.5 °C(±0.5 °C).Tuber starch content was determined as specific gravity (weight in air/weight in air-weight in water)on an electronic balance(minimum weight 0.1 g)based on the formula:%starch=17.546+199.07(specific gravity-1.0988)following Sch?fer-Pregl et al.[8].To obtain normalized values for tuber starch content of each clone over years and locations,clone means were calculated over three replications for each site.Despite the fact that starch content is not chemically determined,the gravity method has been applied to breeding material in many studies[5,10,21].All phenotypic data for tuber starch content collected in six environments are presented in Table S4.

2.4.DNA extraction and marker analysis

Leaves of progeny and parents were collected,quickly frozen in liquid nitrogen,and stored at-80°C.DNA was extracted from the frozen leaf samples using a Plant Genomic DNA Kit(Tiangen Biotech,Beijing,China)as following the manufacturer's instructions.

The parents and 14 clones from population LZ were screened with 206 simple-sequence repeat(SSR)primer pairs from the literature[23-25].Amplification of SSR markers followed Feingold et al.[26].The SSR primer sequences and their location information used to identify linkage groups are described in Table S1.

AFLP marker genotyping was performed as described by Vos et al.[27].Amplicons were separated using 6%denatured polyacrylamide gel electrophoresis(acrylamide:bisacrylamide 19:1,80 W for 1.5-2.0 h)with a ΦX174 DNA-Hae III digest DNA Step Ladder(New England Biolabs,Beijing,China)to estimate the allele size,followed by silver staining according to Byun et al.[28].Gels were scored manually,with each band being scored as a locus with dominant(present)and recessive(absent)alleles,and the sizes of amplicons were estimated using Quantity One software version 4.62(Bio-Rad Laboratories,Hercules,CA,USA).For marker screening,265 Eco R I/Mse I combinations,80 Pst I/Mse I combinations,and 64 Sac I/Mse I combinations were used.Pre-amplification primer sequences of Eco R I,Mse I,Pst I and Sac I were respectively 5′-GACTGCGTACCAATTC-3′,5′-GATGAGTCCTGAGTAA-3′,5′-GACTGCGTACATGCAG-3′,5′-GACTGCTACAAGCTC-3′.The name of each AFLP marker consists of the corresponding primer combination followed by the selective bases and the estimated fragment size in base pairs.The polymorphic markers were used in the LZ population for linkage analysis and QTL mapping.

2.5.Linkage-map construction

Linkage analysis was performed with TetraploidMap for Windows[30]as described by Bradshaw et al.[5].This software package was developed for mapping in autotetraploid species[29,30].Four types of markers whose segregation ratios did not differ from expected values were selected for linkage map construction:simplex dominant markers(present in one parent and segregating 1:1)with P＜0.001 from a chi-square test for departures from the expected ratio(in the absence of double reduction),duplex dominant makers(present in one parent and segregating 5:1)with P＜0.01,double-simplex dominant markers(present in both parent and segregating 3:1)with P＜0.01,and SSR(multiallelic)markers.Duplex-simplex dominant markers segregating 11:1(simplex×duplex)or 35:1(duplex×duplex),present in both parents are not useful for estimating recombination and were omitted from linkage analysis.

Initially,the parental dosage of each allele was deduced and the most likely parental genotypes were determined following Luo et al.[31]from the marker band patterns and distribution ratio in the progeny.Simplex markers were then combined into linkage groups by preliminary cluster analysis.Finally,all simplex,duplex,and multiallelic markers were selected and combined into linkage groups by cluster analysis.In each linkage group,simulated annealing was performed to order the markers and to estimate map distances between them[32].Linkage groups or homologous chromosomes were calculated using simplex markers linked in repulsion with duplex markers and multiallelic markers.Double-simplex markers present in both parents with segregation ratios of 3:1 in the full-sib progeny,and SSR markers,were used to associate the parental maps.

2.6.Analyses of heritability,marker-trait association and QTL

The AOV function of QTL IciMapping[33]was used to calculate the broad-sense heritability(H2)of traits for each environment and combined environments.

The distribution of tuber starch content and plant maturity in the LZ population was analyzed for each environment.The correlations between phenotypic data in different environments for each trait were calculated with the same function.

A preliminary analysis of variance(single-point ANOVA)was conducted to identify linkage groups likely to be associated with each trait using TetraploidMap for Windows,and then QTL interval mapping was performed on the linkage groups that were significantly associated with traits(P＜0.01).

QTL analysis was performed with the interval mapping(IM)function of TetraploidMap for Windows[30]using the phenotypic data for tuber starch content and plant maturity collected in all six environments.When a QTL was significant,10 simpler models were contrasted with the full model using a likelihood ratio test to determine whether any was adequate to model the trait means.Permutation tests[34]with 1000 iterations were performed to compute a significance threshold for the presence of a QTL.If one-LOD support intervals overlapped with each other in different environments,the two QTL were considered to be a single QTL and assigned the same name unless they were detected for different traits.Linkage maps and QTL were drawn with MapChart 2.2[35].

The correlations among tuber starch content and plant maturity among environments were calculated to characterize the trait stability under diverse conditions.All of the correlations were significant(P＜0.01)(Table 2),suggesting that these two traits were controlled mainly by genetic loci and affected by environmental factors.The higher correlation coefficients for plant maturity than for tuber starch content suggested higher heritability of the former than for the latter.

3.Results

3.1.Phenotypic data

The broad-sense heritability(H2)of tuber starch content and plant maturity are shown in Table 1.The heritability of both traits were high,with 86.0%(for tuber starch content)and 94.7%(for plant maturity)of the variation in clone means over different environments and replicates due to genetic differences between clones.The results suggested that these two traits were controlled mainly by genetic factors and that a higher proportion of the phenotypic variance of plant maturity than tuber starch content could be explained by genetic differences.

The distributions of tuber starch content for population are shown in Fig.1.In five of the six environments,they were continuous,unimodal,and normal.Only the phenotypic data from Dingxi 2013 did not show a fit with the normal distribution(P-values of the Kolmogorov Smirnov test was 0.043).The male parent ZDB showed a lower level of tuber starch content(11.2%-14.1%)with a mean value of 12.9%in different environments,whereas the female parent L8 showed a higher level from 20.6%-23.0%with a mean value of 21.6%in different environments.The population mean was 16.5%,between the parents.Most of the progeny clones'trait values fell between those of the parents in all six environments.

The distributions of plant maturity in the LZ population are shown in Fig.2.In contrast to the results for tuber starch content,only the values in Dingxi 2013 followed a continuous normal distribution.In the other five environments,they did not fit the normal distribution and showed skewed distributions or double peaks(P-values of the Kolmogorov Smirnov test were respectively 0.001,0.000,0.000,0.000,and 0.000).The male parent ZDB was an early-maturity clone with a mean value of 75 days across environments,whereas the female parent L8 was a late-maturity clone with a mean value of 119 days across environments.The population mean was 96 days,between those of the parents.Most of the progeny clones'trait values fell between those of the parents in all six environments,but a few clones transgressed the parents(Fig.2).

3.2.Linkage maps

The genetic map was constructed using the SSR and AFLP markers.The full population was genotyped with 69 SSR primers generating 159 markers consisting of 12 multiallelic,103 simplex dominant,16 duplex dominant,and 28 doublesimplex dominant markers.A total of 83 AFLP primer combinations resulted in 461 markers,of which only those consistent with expected segregation ratios were used for map construction.

The female parent map was completed with 103 SSR markers and 318 AFLP markers including 9 multiallelic,157 simplex,42 duplex,and 213 double-simplex markers.Two hundred and nine markers were resolved into 12 linkage groups and 212 markers were unlinked.The total map length of the female parent(calculated from the lengths of the overall linkage groups)was 1245 cM and the mean marker interval was 5.96 cM.Eleven of the 12 linkage groups could be assigned to chromosomes I-III and V-XII by SSR markers on the groups,and only one could not be assigned to any chromosome owing to a lack of SSR markers.Ten of the 12 linkage groups contained all four homologous chromosomes,and the other two(chromosome VII and the unassigned linkage group)contained one homolog(Fig.S1).

The male parent map consisted of 385 markers including 301 AFLP and 84 SSR markers,of which 125 were simplex,38 were duplex,213 were double-simplex and 9 were multiallelic markers.Analysis resolved 211 markers into 13 linkage groups and 174 markers were unlinked.Total map length was 1311 cM with a mean marker interval of 6.21 cM.Twelve of the 13 linkage groups could be assigned to chromosomes I-XII by the mapped SSR markers.Nine of the 13 linkage groups contained all four homologs and the other four(chromosomes III,VI,and IX and the unassigned linkage group)contained three homologs.The detailed map can be seen in Fig.S2.

3.3.Marker-trait association

Nineteen markers were significantly associated with tuber starch content(P＜0.05)in three or more environments by single-point ANOVA analysis in TetraploidMap for Windows,including 11 simplex markers and 8 double-simplex markers.Seven of them were located on the L8 map,nine of the remaining were located on the ZDB map,and the other three were unlinked.Eight of these 19 markers were associated with high starch content,including five located on the L8 map(three on chromosome I,one on chromosome III,and one on chromosome X),two located on chromosomes X and XII of ZDB,and one unlinked.The other 11 were markers associated with low starch content and most of them were located on the ZDB map(Table 3).

Table 1-Broad-sense heritability(H2)of traits in each environment and across environments.

Fig.1-Frequency distributions ofpotatotuber starchcontent in192 progeny clonesofthe LZpopulationinsix environm ents.ParentsL8and ZDB are indicated byarrows.

Fig.2-Frequency distributions ofpotatoplant maturityin192 progeny clonesofthe LZpopulationinsix environments.ParentsL8and ZDB are indicated byarrows.

Table 2-Correlation coefficients of tuber starch content and plant maturity in population LZ between environments.

Thirty-eight markers were significantly associated with maturity(P＜0.05)in three or more environments,including 23 simplex,5 duplex,and 10 double-simplex markers(Table 4).Ten of them were located on the L8 map,15 on the ZDB map and the other 13 were unlinked.Eighteen of these 38 markers were associated with early maturity,including three located on chromosomes I,V,and VIII of L8,six located on chromosomes IV,VII,X,XI,and XII of ZDB,and nine unlinked.The other 20 markers were associated with late maturity,including seven located on the L8 map(six on chromosome V and one on chromosome XI),nine located on the ZDB map(six on chromosome V the other three on chromosomes VI,IX,and XI),and four unlinked.

Tables 3 and 4 show that:first,there were both positive and negative alleles affecting these two traits,but no single marker was associated with both traits.Second,most of the positive alleles for tuber starch content were located on themap of the high starch parent L8 and most of the negative alleles were located on the map of the low starch parent ZDB.Finally,maturity was controlled by multiple genetic loci distributed over more than half of the 12 chromosomes in both parents,and chromosome V was critical for maturity in both parents.

Table 3-Markers associated(P＜0.05)with tuber starch content in more than three environments based on single-point ANOVA.

Table 4-Markers associated(P＜0.05)with plant maturity in more than three environments based on single-point ANOVA.

3.4.QTL for tuber starch content and plant maturity

For tuber starch content, 11 QTL were identified in the six environments, of which six were detected on six chromosomes of the L8 map and five were detected on five chromosomes of the ZDB map (Table S2), explaining 9.1% to 27.3% of phenotypic variance. More than half of these QTL were detected in only one or two of the six environments,indicating that tuber starch content was controlled by multiple genetic loci and strongly affected by environmental factors.To identify loci playing stable roles in tuber starch content,only QTL detected in three or more environments were used for further analysis.

Four reproducible QTL for tuber starch content were named TSC_L_01(TSC indicates the QTL for tuber starch content,L_01 identifies the parent L8 and chromosome I in which the QTL is located,similarly hereinafter),TSC_L_02,TSC_L_08,and TSC_Z_08.They were located on chromosomes I,II,and VIII of the L8 map and chromosome VIII of the ZDB map,respectively(Fig.3).Two(TSC_L_01 and TSC_Z_08)were detected in four of the six environments.TSC_L_01 was detected with a LOD peak ranging from 82 to 88 cM and a one-LOD support interval ranging from 76.2 to 89.9 cM on chromosome I of the L8 map.TSC_Z_08 was detected with a LOD peak ranging from 40 to 44 cM and a one-LOD support interval ranging from 27.9 to 45.5 cM on chromosome VIII of the ZDB map.The remaining two(TSC_L_02 and TSC_L_08)were detected in three of the six environments.TSC_L_02 was detected with a LOD peak ranging from 24 to 32 cM and a one-LOD support interval ranging from 17.9 to 34.4 cM on chromosome II of the L8 map.TSC_L_08 was detected with a LOD peak ranging from 52 to 68 cM and a one-LOD support interval ranging from 31 to 77.4 cM on chromosome VIII of the L8 map.The mean proportions in different environments of phenotypic variance explained by TSC_L_01,TSC_L_02,TSC_L_08,and TSC_Z_08 were 17.5%,17.2%,11.0%,and 12.9%,respectively(Table S2).

Fig.3-QTL for tuber starchcontent and plant maturitydetectedinatleast three environments.Labelabove eachlinkage group shows the numberofchromosomes withparent L8orZDB.QTL labeled byenvironment(locationand year)areindicated withone-LOD confidenceintervals.Filledand openbarsrepresent QTL for tuber starchcontent and plant maturity,respectively.Red barsshow the overlapping intervals ofQTL detectedindifferent environments.

Six QTL were identified for maturity,three on the L8 map and the other three on the ZDB map(Table S2),explaining 9.5%-38.4%of phenotypic variance.The only reproducible QTL,PM_Z_05,was detected in five of the six environments and was located on chromosome V of the ZDB map(Fig.3).The LOD peak of PM_Z_05 ranged from 50 to 52 cM and the one-LOD support interval ranged from 45.2 to 62.5 cM.The proportion of phenotypic variance explained by PM_Z_05 was 21.6%to 26.6%in five environments with a mean of 24.9%(Table S2).

No QTL detected for both tuber starch content and plant maturity,implying that these two traits are independently genetically controlled in this tetraploid mapping population.

4.Discussion

Increasing the tuber starch content of cultivars is a means of increasing tuber yield and is critical for tuber quality for most processing purposes.However,a multigene-controlled genetic feature makes starch content increase difficult of potato breeding programs.Identifying the genetic loci controlling starch content has accordingly received great attention in past two decades by researchers aiming at efficient genetic modification or selection of the trait.

The present study revealed 11 QTL for tuber starch content distributed on seven chromosomes,with four of the QTL expressed in at least three environments.Further analysis suggested the high potential of these stable QTL for future breeding applications.TSC_L_01,located on chromosome I of L8,was detected in four environments.In a two-LOD interval in this region,as indicated by the common marker STI043,Sch?nhals et al.[3]also detected a locus responsible for tuber starch content in a population of 282 clones including cultivars,breeding clones,and Andean landraces.QTL TSC_L_02 was identified on chromosome II in three of six environments,which accounted for 17.2%of tuber starch content variation.Although it is not comparable to other reported loci,owing to a lack of common markers,chromosome II has been reported to be a hot spot for starch phosphorylation,starch gelling temperature,and amylose content.Genes SSIV(the starch synthase IV)[3],GBSS II(granule-bound starch synthase II)and SSS III(soluble starch synthase III)[4]controlling granule number and size of transient starch have been identified on this chromosome.QTL TSC_L_08 and TSC_Z_08,detected on both parental maps,are the same locus given that the same double-simplex marker STM1024 was mapped in the one-LOD interval region of both of these QTL.In this region,marker STI022 has been consistently associated with tuber starch content in diploid mapping populations[3,9].In the same region,single QTL for tuber starch content and sucrose content were simultaneously detected on a bi-parental map[10].Consistency of our identified reproducible QTL with reported loci suggests a potential to explore the genes with high impacts on this trait from these genetic loci and the linked markers identified in our study will be useful for selection in breeding lines.

The QTL PM_Z_05 for plant maturity was detected in five of the six environments on chromosome V of the early-maturity parent ZDB.This locus explained 24.9% of phenotypic variation,suggesting that potato plant maturity is controlled mainly by one or more major genes.Chromosome V is known to be associated with the life cycle length of potato plants,and QTL on chromosome V associated with maturity and/or late blight have been identified in many independent reports.Van den Berg et al.[13]identified 11 distinct loci associated with variation in tuberization(maturity)on seven chromosomes in a reciprocal backcross between S.tuberosum and S.berthaultii.Most of the loci had small effects,but a QTL on chromosome V explained 27%of variance.Collins et al.[36]detected one major QTL for maturity associated with late blight resistance on chromosome V in a diploid population.Bradshaw et al.[5]identified a QTL for maturity on chromosome V explaining 56% of phenotypic variance in a tetraploid population.Kloosterman et al.[16]cloned maturity gene StCDF1 from the QTL located on the short arm of chromosome V.The marker STI032 located in the one-LOD interval of PM_Z_05 was associated with maturity in a diploid mapping population[37]and was closely linked to StCDF1.These findings suggest that plant maturity for the LZ population is associated with the reported loci or gene StCDF1.

It has been observed[3,38,39]by many potato breeders that later-maturing cultivars with longer vegetative periods may have higher starch content than early-maturing ones.A physiological explanation could be that a longer growth period allows more time for starch accumulation by photosynthesis.However,QTL mapping for maturity has yielded inconsistent results.Bradshaw et al.[5]identified on chromosome V a major QTL for potato maturity that overlapped with a QTL for tuber dry matter and shape in a tetraploid population.McCord et al.[11]mapped QTL of foliage maturity in a tetraploid potato population at similar positions of QTL for yield on chromosomes II,III,and V,but not for QTL for tuber dry matter,which is more strongly correlated with starch content.D'hoop et al.[40]detected by association mapping a major QTL for starch content located on chromosome II,but it was far from QTL for maturity on the same chromosome.Among the reproducible QTL for plant maturity and tuber starch content detected on different chromosomes in the present study,none overlapped.This finding suggests that tuber starch content and plant maturity of potato are controlled by independent genetic loci,supporting the prospect of breeding early-maturing and high-starch potato cultivars.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2018.12.003.

Acknowledgments

This work was supported by Earmarked Fund for China Agriculture Research System(CARS-09-P07),and the National Natural Science Foundation of China(31160299,31760410).We thank to Dr.Jingcai Li for reading the manuscript and giving kind advice and Tai Lu for growing plants at Tianshui Experimental Station.