Mapping and Analysis of QTL for Early Maturity Trait in Tetraploid Potato (Solanum tuberosum L.)

Maturity is one of the important traits of potato. In order to get the genetic segment of potato early maturity trait, a tetraploid potato maturity segregation population of Zhongshu 19 × Zhongshu 3 was used for genetic analysis through the combination of high throughput simplified genome sequencing (2b-RAD) and bulked segregation analysis (BSA). A genetic segment related to the early maturity trait at the 3.7~4.2 Mb locus on the short arm of chromosome 5 was obtained and eight markers were developed based on this segment, while five of them were closely linked to the early maturity trait loci. Moreover, 42 SSR markers were developed based on the reference sequence of DM. Finally, a genetic map of chromosome 5 contained 50 markers was constructed using the Tetraploidmap software. The total map length was 172 cM with an average genetic distance of 3.44 cM. Combining with phenotypic data of the segregation population, we mapped the early maturity trait QTL with the contribution of 33.55% on the short arm of chromosome 5, located at 84cM between the flanking markers SSR5-85-1 and SCAR5-8 with the physical interval of 471kb. Gene annotation showed that there exist 34 genes in this region, 12 of them are unknown function. Among the other 22 annotated genes, E3 ubiquitin ligase gene PUB14 may be related to maturity and regulate tuber formation. Our fine mapping of the early maturity QTL made a solid foundation for cloning of the early maturity controlled gene or genes. Key message Early maturity site was mapped using a tetraploid potato segregation population derived from cv. Zhongshu 19 and Zhongshu 3. One major QTL with 33.55% contribution to early maturity was fine mapped in physical interval of 471kb on chromosome 5.

linked to molecular marker BA47f2t7 (P1) on chromosome 5 (Sliwka et al. 2007). Later, scientist found that the 73 major QTLs that control maturity on chromosome 5 were closely related to the QTLs controlling late blight 74 resistance but were independent of each other and closely linked to molecular marker GP21 (Danan et al. 2011). 75 Using tetraploid materials, a major QTL was also obtained to be associated with maturity on chromosome 5, 76 which showed a contribution rate of 54.7% (Bradshaw et al. 2004). Later, by increasing the number of progeny 77 populations, some minor QTLs with contribution rates between 5.4% and 16.5% were identified (Bradshaw et al. 78 2008). In recent years, a large number of single nucleotide polymorphisms (SNPs) have been developed, and a 79 high-density genetic linkage map containing 3,839 SNPs has been constructed by analyzing dose information of 80 SNPs in tetraploid potato. One QTL associated with maturity closely linked to molecular marker c2_476095 has 81 also been mapped on chromosome 5, with a contributing rate of 55% (Hackett et al. 2014 populations (Wang et al. 2012). At present, this technique has been applied to construct high-density genetic 102 linkage maps of animals, plants, and marine organisms. The high-density genetic linkage map of Chlamys Farrer 103 includes 3,806 molecular markers, the average distance between markers is 0.41 cM with genome coverage of 104 99.5%, and is it mapped to growth-and-sex-related QTLs (Jiao et al. 2013(Jiao et al. , 2014. 105 In this study, we intend to use the tetraploid potato maturity segregation population, combine 106 high-throughput simplified genome sequencing 2b-RAD with the BSA method, to identify the genetic segment 107 related to the early maturity trait and develop molecular markers based on the genetic segment. Linkage Map 108 was constructed by using software TetraploidMap for Windows (Hackett et al. 2007), which was designed 109 for calculating linkage maps from the maker phenotypes of the parents and segregating progeny of a cross 110 in an autotetraploid species. Combined with 4 years foliage maturity phenotypic data, QTL analysis was 111 performed using this software, which will provide a foundation for cloning and functional verification of the 112 maturity gene or genes. 113

Plant materials 115
The mapping population, consisted of 221 individuals, was from a cross between two tetraploid varieties： 116 'Zhongshu19'×'Zhongshu 3'. The female parent 'Zhongshu 19' with a pedigree of (CIP92.187×CIP93.154) 117 ×(Bierma ×Colmo) is a late maturing variety with a growth period of 110 days after emergence, and the male 118 parent 'Zhongshu 3' with a pedigree of (Jingfeng 1×BF77A) is an early maturing variety with a growth period 119 of 75 days. They were bred by the Institute of Vegetables and Flowers, Chinese Academy of Agricultural 120

DNA extraction and progeny mixed pool construction 139
Two-hundred twenty-one samples were collected from young leaves of Zhongshu 3, Zhongshu 19, and F1 plants. 140 The genomic DNA was extracted by the CTAB method, and DNA quality was detected with 1% agarose gel and 141 BioDrop for marker development and validation. 142 We used a punch to take one young leaf from the same part of each F1 generation early and late maturing 143 material with a growth period of less than 70 DAE and greater than 110 DAE, respectively, mixed them in equal 144 amount, extracted their DNA, and constructed an early and late maturing mixed DNA pool for simplified 145 genome sequencing. The DNA extraction method and quality testing method are the same as mentioned above. 146

Simplified genome sequencing analysis and marker development 147
A high-throughput simplified genomic 2b-RAD sequencing technique was used to construct a tag sequencing 148 library of parental DNA and extreme progeny DNA pool. The single-terminal sequencing was performed on the 149 Hiseq 2500 v2 platform. After removing from original reads of the sequences that contain no BsaXI recognition 150 sites, low quality sequences and sequences with more than 10 consecutive identical bases, the individual high 151 quality reads were mapped to the potato DM reference sequence using SOAP software. The number and depth 152 of specific tags that can be used for typing were obtained, and genome-wide SNP screening and typing was 153 conducted. According to the classification results, we constructed a chromosome tag density distribution map of 154 four samples with Matlab software, further, constructed specific tag density distribution and absolute value 155 distribution of the specific tag density between the two mixed pools, respectively, and then screened out 156 chromosome segments with a large difference in tag densities. 157 Based on the tag information obtained in the different segments, the synthetic primers were designed 158 according to the genomic sequences on the potato genome sequence website 159 (http://solanaceae.plantbiology.msu.edu/cgi-bin/gbrowse/potato/). The quality and specificity of the primers 160 were detected by using the genomic DNAs of the late maturing parent Zhongshu 19 and early maturing parent 161 Zhongshu 3 as the templates. The amplified products were detected by agarose gel electrophoresis with a 162 concentration of 1.2%. The PCR reaction system consisted of: 5.6 μL ddH 2 O, 1.0 μL buffer (10 × PCR), 0.8 μL 163 dNTPs (10 mmol L −1 ), 0.2 μL forward primer (10 μmol L −1 ), 0.2 μL reverse primer (10 μmol L −1 ), 0.2 μL Taq 164 enzyme (2.5 U μL −1 ), and 2 μL DNA (25 ng μL −1 ). PCR was performed as follows: 94°C for 3 min, followed by 165 35 cycles of 94°C for 30 s, 59°C for 30 s, and 72°C for 50 s, and finally, 72°C for 10 min. 166 The PCR products were digested by BsaXI restriction endonuclease if the primers could amplify clear and 167 non-different bands in the parents (Zhongshu 3 and Zhongshu 19). The enzyme digestion system was 8 μL PCR 168 product, 1.5 μL CutSmart® buffer, 0.2 μL BsaXI endonuclease (0.5 μL/U), and 5.3 μL ddH 2 O. The primers 169 whose PCR products can be digested only in one parent and not digested in the other parent were used to 170 develop CAPs markers. The primers that amplified a band in one parent but not in the other parent, or amplified SSR markers on chromosome 5, was conducted using the genomic DNAs of Zhongshu 3, Zhongshu 19, and 16 176  Table S1. Consequently, combination with the mean value of 4 years progeny phenotype 192 data, details see Table S2. the maturity QTL was mapped by using TetraploidMap. over the entire scoring range (Fig. 1). Detail phenotypic data among four years was shown in Table S1. Genomic 206 DNA was extracted from 35 very early and early maturing genotypes and 33 very late maturing genotypes, and 207 the corresponding early maturing DNA pool and late maturing DNA pool were constructed. 208 Screening and marker development of early maturity traits higher than 90%, and finally 125,556 unique tags in average of each sample were obtained. 214 SNP marker typing was performed according to the obtained sequencing tag sequences and the tag 215 distribution and density profiles on 12 potato chromosomes were got, and the results showed that the tags were 216 evenly distributed on all 12 potato chromosomes, and there was no large-scale tag information missing (Fig. 2). 217 Furthermore, according to the specific tag density of early maturing and late maturing pools, the density 218 difference map of the specific tag between early maturing and late maturing pools was obtained, and the results 219 showed that the segments with a large difference in tag density between early and late maturing pools were 220 mainly concentrated on chromosomes 4 and 5, where in the segment with the most significant difference in tag 221 density was found at 3.68-6.19 Mb on chromosome 5, followed by 18.6-20.9 Mb and 27.4-30.3 Mb on 222 chromosome 4 (Fig. 3). Moreover, the specific tag density of the early maturing pool among the differential 223 segments of chromosome 5 was relatively high, suggesting that this segment may be related to the early 224 maturity trait. But all the specific tag densities of the late maturing pool were relatively high among the 225 differential segments of chromosome 4, which might be related to the late maturity trait (Fig. 3). were further used to test the F1 generation. In this study, we focused on early maturity, and the genetic linkage 244 map for early type parent 'zhongshu 3' was constructed (Fig 4). The map including a total of 50 makers, of 245 which 33 were simplex, 4 were duplex, and 13 were double simplex makers (TetraploidMap generate 246 automatically combined with the segregation ratio) ( Table 3). The total length of the map coverage was 172 cM, 9 and 17 polymorphic SSR markers. The 25 single markers were come from the division of 11 polymorphic SSR 250 markers, and they were SSR5-22-1, SSR5-22-2, SSR5-36-1, SSR5-36-2, SSR5-38-1, SSR5-38-2, SSR5-40-1, 251 SSR5-40-2, SSR5-40-3, SSR5-55-1, SSR5-55-2, SSR5-85-1, SSR5-85-2, SSR5-85-3, SSR5-85-4, SSR5-100-1, 252 SSR5-100-2, SSR5-103-3, SSR5-103-4, PM0333-2, PM0333-3, STI049-2, STI049-3, STG0021-1, and 253 STG0021-2. Actually, among the 32 polymorphic SSR markers only 28 markers were used to construct the 254 genetic linkage map. 255 QTLs for foliage maturity type were identified using TetraploidMap. Initially, the significance of individual 256 markers for each year trait was tested by analysis of variance (ANOVA) and the analysis procedure 257 Kruskall-Wallis in TetraploidMap. This two tests were used to compare the differences of the mean value of 258 different makers genotype (Simonsen and McIntyre, 2004). A P-value of less than 0.01 was used as a threshold 259 criterion for QTL detection. Results from above two tests suggested the existence of QTLs for foliage mature 260 type on chromosome 5. There are 22 makers on chromosome 5 may associated with maturity. The maker cluster 261 including maker CAPS 5-21-2, CAPS5-24, CAPS5-3-2，SCAR5-5，SCAR5-8 were closely linked to the QTL 262 locus, since not only their P value are 0.000 but also their smallest SED with 0.0937 (Table 3). Therefore, the method can be used for the development of markers of complex genomes such as potato. 304 Combined with published potato reference genomes, this method can greatly shorten the marker development 305 cycle and map the markers. Which will also provide a reference for marker development of other polyploid 306 complex genomic organisms. 307 In this study, we found that not only was there a maturity related genetic segment on chromosome 5, but there 308 were also some segments with large difference in specific tag density on chromosome 4. Although there are a 309 few reports that mention the existence of a maturity-related genetic segment on the chromosome 4, the tag 310 validation results in this study showed that the genetic segment on chromosome 4 was not related to maturity. 311 This may be due to several reasons: First of all, the tetraploid potato genome is highly heterozygous, more 312 complicated than diploid potato in heredity, and in the high-throughput sequencing process, may be due to 313 sequencing depth not being enough or BsaXI recognition sequence preferences; some information was not 314 measured. Secondly, due to the filtering out some of the low quality reads resulted in partial loss of genetic 315 information, and eventually the differential tag interval appears. Thirdly, when performing specific tag filtering, 316 removing the tags with fewer occurrences may also cause a difference in tag density. 317

Genetic linkage map construction and QTL mapping 318
In this study, a 50-marker genetic linkage map was constructed using tetraploid potato segregation population, 319 and an early maturity trait QTL was mapped at the position of 84 cM near the marker SCAR5-8 (or SCAR5-5, 320 CAPS5-3-2, CAPS5-24, CAPS5-21-2) on the short arm of chromosome 5. Which indicates that the QTL for the 321 early-maturity trait obtained by genetic map mapping is consistent with the maturity genetic segment results 322 obtained by high-throughput simplified genome sequencing, and also proved the feasibility and accuracy of each other, and no significant separations between the markers in the genetic map. Furthermore, the five 327 molecular markers were clustered on the 3rd homologous chromosome and belonged to the same linkage group. 328 The other three molecular markers, CAPS5-16, SCAR5-18, SCAR5-25 probably linked to the late maturity trait 329 loci, were clustered on the 1st homologous chromosome, and belonged to another linkage group (Fig. 4). 330 Previous studies on the maturity QTL mapping showed that the genetic interval of maturity QTL was 331 relatively large. The major QTL for maturity was mapped at 0-6 cM on chromosome 5 in diploid materials, and 332 the genetic distance between two flanking markers was 6 cM (Visker et al. 2003). In tetraploid materials, the 333 major QTL for maturity was also mapped at 14-22 cM on chromosome 5, and the genetic distance between two 334 flanking markers was 8 cM (Hackett et al. 2014). In this study, based on simplified genome genetic segment 335 mining and marker development, the QTL for the early maturity trait was mapped in a physical interval of 471 336 kb, which greatly approached the early maturity trait loci and the average distance between two makers was 3. StCDF1 with the reference genome showed that the StCDF1 was located in 4.537-4.542Mb, but the maturity 367 locus in our study is located in 3.7-4.2Mb. Pedigree investigation showed that the StCDF1.2 gene was cloned 368 from CE3130 that should originated from phureja (Kloosterman et al., 2013), the early maturity genes of 369 Zhongshu 3 maybe come from cv. Katahdin. All the above indicated that there should exist other different 370 maturity-related genes in our maturity segregation population of this study from the gene StCDF1. 371 The tetraploid mapping software TetraploidMap (Hackett et al. 2007) used in this study only can identify 372 three marker types. For a single dominant marker, the parental genotype is AOOO × OOOO, and the segregation 373 ratio of progeny is 1:1. For double dominant markers, the parental genotypes are AAOO × OOOO and AOOO × 374 AOOO, and the segregation ratios are 5:1 and 3:1. There are still two types of markers the software does not 375 recognize, i.e. markers of parental genotypes AOOO × AAOO and AAOO × AAOO with the segregation ratios 376 of 11:1 and 35:1. Therefore, among the 32 polymorphic markers selected from 152 SSR markers, four markers 377 belonged to the 11:1 marker type, and the software did not recognize them. Therefore, only 28 polymorphic SSR 378 markers could be used in the study to construct the map. Because the tetraploid mapping software limits the 379 recognition of marker types, some markers are not available, and made the number of markers less than normal 380 in the genetic map. Nevertheless, we have mapped the QTL for the early maturity trait in the physical region of 381 471 kb and analyzed the genes within this interval. Which provides a solid foundation for the cloning of the 382 major gene or genes that control the early maturity trait. 383 384 Table Captions   Table 1 Information of the 50 primers used in the genetic map Table 2 Basic analysis of sequencing data Table 3 Makers associated with physiological maturity KWSig: the significant of Kruskal-Wallis test.
AVSig: the significant of the analysis of variance. Mean (0):the mean when the maker is absent. Mean (1): the mean when the maker is present. SED: the stand error of difference between the means. Table 4 Gene annotation and location in the physical interval