Introduction

Cotton is an important economic crop worldwide, and cotton fiber is a basic raw material used in the textile industry. In recent years, changes in spinning technology have resulted in the need for unique and often increased cotton fiber quality, especially strength, for processing (Benedict et al. 1999; Deussen 1992). Advances in the use of DNA markers for marker-assisted selection (MAS) have shown promise in streamlining plant-breeding programs. The development of DNA markers linked to fiber quality QTL would allow cotton breeders to identify this important trait at early plant-growing stages or early segregations.

Quantitative trait loci (QTL) conferring fiber quality traits have been described and mapped using molecular markers in inter-specific populations from crosses between Gossypium hirsutum and G. barbadense (Jiang et al. 1998; Kohel et al. 2001; Paterson et al. 2003; Mei et al. 2004; Lacape et al. 2003; Lacape et al. 2005; Park et al. 2005), and in intra-specific G. hirsutum populations (Shappley et al. 1998; Ulloa and Meredith 2000; Zhang et al. 2003; Shen et al. 2005; Ulloa et al. 2005; Shen et al. 2007). More than 100 agronomic and fiber trait QTL have also been mapped in an intra-specific population in G. hirsutum (Shappley et al. 1998).

In our laboratory, we employed bulk segregant analysis to identify a major QTL on Chro.24 (D8) for fiber strength using an F2 segregating population and F2:3 family lines derived from a cross between 7235 and TM-1 in Upland cotton (Zhang et al. 2003). The QTL was detected in diverse environments, including Nanjing and Hainan in China, and Texas in USA, which had an average phenotypic variance (PV) of 18.5–53.8%. The QTL was later further confirmed in (7235 × TM-1) F2, and F2:3 family lines, and (7235 × TM-1) RILs (Shen et al. 2005, 2007). The 95% confidence interval (CI) of this major fiber strength QTL (qFS-D8-1) was identified by Shen et al. (2007) in a recombinant inbred line (RIL) derived from 7235 and TM-1. The 95% CI ranged from 39.1 to 53.9 cM between SSR marker BNL2961 and CIR070a. MAS results revealed that DNA markers linked to this QTL could be used to increase fiber strength of commercial cultivars in early segregating breeding generations (Zhang et al. 2003; Guo et al. 2005). This region was densely populated with markers and QTL, with 8 QTL for 8 traits detected within a chromosome region of 33.4 cM, including 36 SSR loci in Chro.D8 (Shen et al. 2007). In the present study, three recombinant inbred lines (RIL), 7TR-133, 7TR-132, and 7TR-214, developed from a cross between 7235 and TM-1, were chosen and backcrossed with TM-1 to develop three mapping populations to fine-map the major fiber strength QTL derived from 7235 to lay a basis of map-based cloning of this major QTL in cotton.

Materials and methods

Plant materials

Three RILs, 7TR-133, 7TR-132, 7TR-214, derived from a cross between 7235 and TM-1 using a bulk-selfing technique (Shen et al. 2007) were used to fine-map the major fiber strength QTL previously identified from 7235 (Zhang et al. 2003), an introgressive line exhibiting high fiber quality (Qian et al. 1992). These three RIL lines, 7TR-133 from NAU583 to NAU774 within a 19.3 cM region, 7TR-132 from NAU1037 to NAU 780 within an 9.3 cM region, and 7TR-214 from TMD05 to JESPR127 within a 9.4 cM region, had different intervals, but overlapped to cover the major fiber strength QTL region from BNL 2961 and NAU774 in Chro.D8 detected in the (7235 × TM-1) RIL (genetic linkage map shown in Fig. 1) (Shen et al. 2007). The RIL lines were backcrossed as maternal parents with TM-1, Upland cotton genetic standard (Kohel et al. 1970), to generate three mapping populations.

Fig. 1
figure 1

Linkage group (Shen et al. 2007) and chromosome positions of three RILs in Chro.D8

Full size image

Individual plants from 7TR-133,7TR-132 and 7TR-214 were backcrossed with TM-1 in 2005 and their F1 seeds were sent to Hainan Island to produce F2 seeds. In 2006, the F2 seeds were grown for fiber strength scoring, QTL tagging, and self-pollination to produce F3 family seeds in Jiangpu Experiment Station, Nanjing Agriculture University (JES/NAU). In 2007, three F3 family line seeds were divided into two groups, one group planted at JES/NAU, the Yangtze River cotton growing region in China, and the other at the Baotou Lake Experiment Station, Xinjiang Academy of Agriculture and Reclamation Sciences (BES/XJ), China’s Northwest cotton growing region. Nanjing is situated at longitude 118°46′, and latitude 32°03′; and Baotou Lake Experiment Station at longitude 86°07′, latitude 41°46′. It is important to note that cultivation practices differ between these two cotton-growing regions. At JES/NAU, cotton seeds are typically planted in nursery pot seedbeds, and transplanted to the field when seedlings exhibit 3–4 leaves; seedlings are planted 45 cm apart. At BES/XJ, cottonseeds are planted directly in the field under plastic film with plants spaced 15 cm apart. Drip irrigation under film is usually employed at BES/XJ to conserve water. However, at JES/NAU adequate rainfall eliminates the need for irrigation.

Because an inadequate number of seeds were harvested from some F2 individuals to plant F3 progeny at both locations simultaneously, and/or the harvested fiber from the F2 individuals was not enough to conduct fiber tests, different plants and family lines represented the F2 and F2:3 lines at both locations (Table 2). The (7TR-133 × TM-1) F2, (7TR-132 × TM-1) F2, and (7TR-214 × TM-1) F2 included 907, 670, and 940 individuals to be used to construct the linkage map and to produce their corresponding F3 family lines generated by self-pollinating the corresponding F2 individuals, but only 764, 663, and 813 individuals were used in QTL mapping in their F2s because no enough fiber sample was harvested to conduct fiber quality test. Both F2 and F2:3 seeds were first planted in nursery pots and transplanted to the field at JES/NAU during 2006 and 2007; however, F2:3 seeds were directly planted at BES/XJ in 2007. A randomized incomplete block design was employed in the field trials. The traditional cotton growing practices were used in two locations, at JES/NAU, plots were 5 m long with plants spaced 45 cm apart, and at BES/XJ plants were spaced 15 cm apart in 2 m long plots. Fiber samples from F2 individuals at JES/NAU in 2006, and from F2:3 family lines at BES/XJ and JES/NAU in 2007 were collected from bolls in the interior middle of the plant. In 2006 and 2007, fiber quality was tested by HVICC at the Henan Test Center of Cotton Quality, China.

Assay of DNA markers

DNAs were extracted from individual plants of the three F2s as described by Paterson et al. (1993). In the present research, we employed 6123 SSR primer pairs including BNL, JESPR, CRI, etc., presently available in cotton (http://www.cottonmarker.org). Shen et al. (2005) screened the SSR primer pair polymorphisms between the two parents 7235 and TM-1. We subsequently screened the remaining SSR primers from NAU2500 to NAU6123 developed in our lab in this study. The sources of these NAU SSR primers were reported in our prior work (Han et al. 2004; Han et al. 2006; Wang et al. 2006; Guo et al. 2007). SSR–PCR amplifications were performed using a Peltier Thermal Cycler-225 (MJ Research), and PCR product electrophoresis was conducted as described by Zhang et al. (2000, 2002).

Data analysis and QTL mapping

Trait means were calculated using SAS (SAS institute 1989) and linkage maps were constructed with JoinMap Version 3.0 (Van Ooijen and Voorrips 2001). The Multiple-QTL model (MQM) using MapQTL5 (Van Ooijen. 2004) was chosen to determine the QTL likelihood map, gene action, and phenotypic variance (PV) explained by individual QTL. In addition, 1,000 permutation analysis was applied to each putative QTL. Chi-square was used to determine if the allele frequency at each individual locus showed normal segregation patterns. Confidence intervals (90–95%) associated with QTL locations were set as the map interval corresponding to one LOD decline on either side of the peak.

QTL nomenclature was adapted according to the method developed in rice (McCouch et al. 1997), starting with ‘q’, followed by a trait abbreviation of FS designating for fiber strength, followed by F2/F2:3 population, and an abbreviation of the cultivation location (JES/NAU designates Jiangpu Experiment Station, and BES/XJ Baotou Lake Experiment Station, Xinjiang Academy of Agriculture and Reclamation Sciences), and followed by the number of QTL affecting the trait on Chro.D8. Additionally, a/b/c stands for the population names, respectively, for 7TR-133, 7TR-132, and 7TR-214.

Results

Phenotypic values for fiber strength of parents, F2 and F2:3

Significant differences in fiber strength were detected between these RILs and their backcrossed parent TM-1 at JES/NAU and BES/XJ in 2006 and 2007 (Table 1). Fiber strength phenotypic data for the F2 and F2:3 populations at JES/NAU and BES/XJ during 2006 and 2007 are presented in Table 2. F2 and F2:3 fiber strength distributions at JES/NAU and BES/XJ were generated using SAS (SAS institute 1989). To determine if fiber strength were normally distributed, skewness and kurtosis values were calculated for all mapping populations. The skewed values were between 0.24 and 0.94 (Table 2; Fig. 2), indicating they fit normal distributions and can be used to conduct QTL mapping for fiber strength. Low temperature during the late growing stages when the fiber maturation at BES/XJ in 2007 results in approximately 3 cN/tex fiber strength lower at BES/XJ than that at JES/NAU (Table 2; Fig. 2).

Table 1 Distribution of fiber strength (cN/tex) of F2 and F2:3 in three populations at JES/NAU and BES/XJ
Full size table
Table 2 Fiber strength performance of four parents at JES/NAU and BES/XJ
Full size table
Fig. 2
figure 2

The distribution for fiber strength in F2 and F2:3 in three populations in two locations

Full size image

Construction of the three linkage maps

Of the 6123 SSR markers employed in this study, 170 polymorphic SSR markers were detected between 7235 and TM-1, and a total of 25, 12, and 21 polymorphic SSR loci detected between 7TR-133, 7TR-132, 7TR-214, and TM-1 on Chro.D8. The three genetic maps for (7TR-133 × TM-1) F2, (7TR-132 × TM-1) F2, and (7TR-214 × TM-1) F2 were generated by JoinMap3.0, using a LOD = 3.0 (Fig. 3). The SSR genetic map was constructed using 907 individuals from (7TR-133 × TM-1)F2 (Pop A) and included 22 loci covering 13.7 cM, which represented approximately 12.3% of the total 111.7 cM recombinational length of cotton Chro.D8 (Guo et al. 2007). The map constructed from 670 individuals in (7TR-132 × TM-1)F2 (Pop B) included 11 loci covering 19.1 cM, approximately 17.1% of the recombinational length of cotton Chro.D8, and the map constructed using 940 individuals from (7TR-214 × TM-1)F2 (Pop C) included 18 loci covering 10.1 cM, approximately 9.4% of the recombinational length of cotton Chro.D8. Most SSR markers in (7235 × TM-1)RIL and (7235 × TM-1)F2 on Chro.D8 (Shen et al. 2005, 2007) were linked in the present genetic maps (Fig. 3). Eight, 2 and 10 new SSR loci were added to Pops A, B and C, respectively. The average distance between two SSR loci was 0.62 cM in Pop. A, 1.7 cM in Pop B, and 0.56 cM in Pop C. Except for three seriously distorted segregation SSR loci, which interspersed with other loci arrangement in linkage groups, in Pop A, 1 in Pop B, and 3 in Pop C, other distorted SSR markers were included in linkage map construction. Sixteen, 7 and 10 distorted SSR loci were detected in Pops A, B, and C. The percentage of skewed segregation ratios were 72.7, 63.6, and 55.6%, respectively, a great deal of decrease from 100% in (7235 × TM-1)RIL (Shen et al. 2007).

Fig. 3
figure 3

Linkage groups and QTL position in three linkage groups

Full size image

QTL tagging for fiber strength

Fiber strength QTL tagging was conducted by MapQTL 5 software. For Pop A, 764 F2 individuals, and 907 F2:3 family lines at JES/NAU, and 890 F2:3 family lines at BES/XJ were used for fiber strength QTL tagging; for Pop B, 663 F2 individuals, and 670 F2:3 family lines at JES/NAU, and 649 family lines at BES/XJ were used; and for Pop C, 813 F2 individuals, and 940 F2:3 family lines at JES/NAU, and 858 in BES/XJ were used for the analysis (Table 2). We found five QTL (2.5 < LOD < 29.8) for fiber strength on Chro.D8 in these three populations.

QTL tagging in (7TR-133 × TM-1) F2 and F2:3

Two QTL were detected in (7TR-133 × TM-1)F2, and three QTL in (7TR-133 × TM-1)F2:3 in this chromosome region (Fig. 3a; Table 3). The qFS-F 2 -JES-1a and qFS-F 2:3 -JES-1a were detected in the same interval range from TMD05 to BNL3145 in F2 and F2:3, both grown at JES/NAU in 2006 and 2007, and qFS-F 3 -BES-1a from TMD05 to NAU3207 in F2:3 grown at BES/XJ in 2007. Therefore, this is a very stable and common QTL, and we proposed it to be named qFS-1. Another QTL (qFS-F 2 -JES-2a) from NAU2665 to NAU3605 was detected in (7TR-133 × TM-1)F2 at JES/NAU in 2006. However, this QTL was further dissected into two QTL at this region in F2:3 at JES/NAU and BES/XJ. Both qFS-F 2:3 -JES-2a from NAU1534 to NAU3605 and qFS-F 2:3 -BES-2a from NAU1534 to NAU3499 were shown to overlap over a large region, which suggested a new QTL. We proposed qFS-2 as the name for this QTL. In addition, qFS-F 2:3 -JES-3a from NAU4099 to NAU1534 and qFS-F 2:3 -BES-3a from BNL1521 to NAU1534 were identified at JES/NAU and BES/XJ in 2007, and overlapped with three loci, NAU4099, NAU2665. and NAU1534, in (7TR-133 × TM-1)F2:3. Consequently, it suggested that this should be a third QTL, and we proposed to name this QTL as qFS-3. QTL qFS-1 exhibited a phenotypic variance of 5.8–13.6% and a CI of 1.35–1.82 cM with 6–8 SSR markers, and the largest LOD value of 28.82 was detected in a F2:3 grown at BES/XJ. QTL qFS-2 exhibited a phenotypic variance of 6.5–13.4% and a CI of 1.16–2.11 cM with 2–4 SSR markers. The largest LOD value was as high as 29.76 for this QTL in a F2:3 grown at BES/XJ. QTL qFS-3 was detected in the F2:3 at JES/NAU and BES/XJ, showing a phenotypic variance of 7.4–14.8%. The CI of qFS-3 was 0.11–0.66 cM from 3 to 4 SSR markers, and the largest LOD value was 25.32 in the F2:3 from BES/XJ.

Table 3 Characteristics of QTL affecting fiber strength in three populations
Full size table

QTL tagging in (7TR-132 × TM-1) F2 and F2:3

Three QTL for fiber strength were further detected in Pop B (Fig. 3b; Table 3). Among them, both qFS-F 2 -JES-1b and qFS-F 2:3 -JES-1b from NAU1197 to NAU780, and qFS-F 2:3 -BES-1b from NAU3201b to TMD05 were detected in the most regions overlapped in both F2 and F2:3 grown at JES/NAU and BES/XJ in 2006 and 2007. These QTL were anchored in the same position as qFS-1 and bridged with TMD05 locus in (7TR-133 × TM-1) F2 and F2:3(Fig. 3a). Therefore, we considered these QTL as qFS-1. Another QTL, qFS-F 2 -JES-2b from BNL2961 to NAU1534 was detected at JES/NAU in 2006. However, we subsequently detected three QTL, qFS-F 2:3 -JES-2b from BNL2961 to NAU1295 and qFS-F 2:3 -BES-2b from BNL2961 to NAU3562 and qFS-F 2:3 -BES-3b from BNL1037 to BNL1521 in this region. It is likely that QTL, qFS-F 2 -JES-2b and qFS-F 2:3 -JES-2b detected at JES/NAU in 2006 and 2007, was dissected into two at BES/XJ in 2007. The first two QTL overlapped in most regions and bridged with BNL1521 and NAU1534. Therefore, we considered these two QTL are the same as qFS-2. The third QTL, qFS-F 2:3 -BES-3b from BNL1037 to BNL1521 was only detected in (7TR-132 × TM-1)F2:3 at BES/XJ in 2007 and may be the same QTL as qFS-3. qFS-1 exhibited a phenotypic variance (PV) of 6.8–12.9%, a CI of 1.82–3.78 cM with three SSR markers, and the largest LOD value detected was 14.5 in F3 from BES/XJ. qFS-2 showed a phenotypic variance of 5.0–11.2%, and a CI of 1.92–3.86 cM covered by 3–6 SSR markers. The largest LOD value was 14.8 in F3 grown at BES/XJ. qFS-3 was only identified in F3 grown at BES/XJ, and exhibited a phenotypic variance of 10.5%.

QTL tagging in (7TR-214 × TM-1) F2 and F2:3

7TR-214 covered most of the chromosome region where qFS-1 was anchored. However, three QTL were detected in this region (Fig. 3c; Table 3). qFS-F2-JES-1c ranged from NAU5379 to BNL3474, qFS-F2:3-JES-1c from NAU5379 to NAU2926 and qFS-F2-BES-1c from NAU5379 to TMD05 were identified almost in the same interval in both F2 and F2:3 from JES/NAU and BES/XJ in 2006 and 2007. They are bridged with SSR marker TMD05 with Pop A and B, which supported that the QTL might be qFS-1. However, both qFS-F2-JES-2c and qFS-F2:3-BES-2c were identified in the same interval ranged from BNL3474 to NAU1322 in F2 and F2:3 and are bridged with SSR markers NAU827, NAU3207, TMl21 and BNL3145 anchored for qFS-1 in Pop A. The remaining QTL identified in F2 at JES/NAU in 2006 and F2:3 at BES/XJ in 2007 had not previously been detected; however, still included in the CI of qFS-1 in Pop A. It is clear that qFS-1 identified in Pop A and B could be further dissected into three QTL and we propose the names qFS-1, qFS-4, and qFS-5. qFS-1 exhibited a phenotypic variance (PV) of 4.2–9.4%, a CI of 2.31–3.16 cM with 3–5 SSR markers, and the largest LOD value was 12.46 in F2:3 at BES/XJ. qFS-4 and qFS-5 were detected in F2 (JES/NAU) and F2:3 grown at BES/XJ, but not in F2:3 at JES/NAU. qFS-4 showed a phenotypic variance of 6.5–8.0% and qFS-5 a phenotypic variance of 5.7–9.8%. The CI of qFS-4 was 2.41 cM with eight SSR markers, and the largest LOD value was 12.08 in the F2:3 population at BES/XJ. The CI of qFS-5 was 1.95–3.0 cM with 3–5 SSR markers, and the largest LOD value was 11.48 in the F2:3 cultivated at BES/XJ.

Discussion

In our previous reports, one major QTL for fiber strength was identified in (7235 × TM-1)F2/F2:3 and (7235 × TM-1)RILs (Shen et al. 2005, 2007). In the present research, five tightly linked and/or clustered QTL were identified on Chro.D8 in three populations generated using three RIL lines, which overlapped with our previously identified major QTL region (Fig. 1), although slight differences in QTL tagging positions were detected in the data collected at JSE/NAU and BES/XJ, most likely due to their differences in the growing environments and cotton planting practices. QTL qFS-1, identified in all mapping populations of F2 and F2:3 at JES/NAU and BES/XJ, showed a bridge SSR marker-TMD05; therefore, qFS-1 was regarded as a common QTL. The qFS-1 allele increased fiber strength by 0.85–1.13 cN/tex in 7TR-133, 0.9–1.40 cN/tex in 7TR-132 and 0.82–1.06 cN/tex in 7TR-214. qFS-2 had a bridge SSR marker-NAU1534 detected in F2 and F2:3 at JES/NAU and BES/XJ in Pops A and B, respectively. The qFS-2 allele increased fiber strength by 0.78–1.14 cN/tex in 7TR-133, and by 0.88–1.27 cN/tex in 7TR-132. qFS-3 was identified in (7TR-133 × TM-1)F2:3 at JES/NAU and BES/XJ and (7TR-132 × TM-1)F2:3 grown at BES/XJ. Our data suggested that qFS-3 at JES/NAU and BES/XJ might be the same because their confidence intervals overlapped. The qFS-3 allele increased fiber strength by 0.80–1.27 cN/tex in 7TR-133, and 1.16 cN/tex in 7TR-132. The qFS-4 allele identified as qFS-F2-JES-2c at JES/NAU and qFS-F 2:3 -BES-2c at BES/XJ increased fiber strength by 0.76–1.02 cN/tex. The qFS-5 identified as qFS-F 2 -JES-3c at JES/NAU, and qFS-F 2:3 -BES-3c at BES/XJ increased fiber strength by 0.63–0.95 cN/tex. Five QTL for fiber strength exhibited a total phenotypic variance of 28.8–59.6%.

Three overlapped RIL backcrossed progenies and the large mapping populations used for the fiber strength QTL tagging most likely resulted in the five clustered QTL identified for the same trait. The CI of the QTL identified in the present research was decreased, most likely due to the large populations, the RIL backcrossed method and a randomized incomplete block design, might eliminate the influence of the blocks. Such clustered QTL for fiber qualities and yield have been reported in Upland Cotton (Ulloa et al. 2000; Yin et al. 2002).

Skewed segregation ratios have frequently been reported in cotton (Ulloa et al. 2002, 2005; Lacape et al. 2003; Mei et al. 2004). Furthermore, high segregation distortion frequency (49–80%) has been observed in inter-specific crosses, most likely due to divergence between species (Paterson et al. 1988). The SSR markers on Chro.8D for (7235 × TM-1) RIL all exhibited skewed distribution patterns (Shen et al. 2007). In the present research, 55.6–72.7% of the markers in the three genetic maps constructed using TM-1 as recurrent parent were skewed. Furthermore, the frequency of distorted ratios was high, but greatly decreased from 100% in (7235 × TM-1) RIL (Shen et al. 2007). The higher distorted segregation ratio further reflects 7235 introgression from G. anomalum (B-genome). 7235 was developed via a cross between G. hirsutum cv. 86-1 and G. anomalum, followed backcrossing to Xumian 6, Acala 3080 (a G. barbadense introgressed line), and PD4381 (a G. thurberi introgressed line) (Qian et al. 1992). Therefore, distorted markers clustered on Chro.D8 may further indicate that this chromosome contains G. anomalum and G. barbadense introgressed segments, which contribute DNA and/or QTL for fiber qualities.

From a series of QTL tagging experiments using parents 7235 and TM-1, we found 5-clustered QTL on Chro.D8 were major stable QTL. The major fiber strength QTL located on Chro.D8 was used effectively to pyramid QTL by MAS (Guo et al. 2005). Future fine positioning of the clustered QTL for fiber strength, cloning, and sequencing of this region will aid in the investigation of the genetic mechanisms of high fiber strength QTL.