The development of cotton fiber appears to be a highly regulated process of fiber cell differentiation, expansion, and cellulose synthesis. The genetic determinants important for this development would involve cellulose production, fiber elongation, and fiber strength. Identifying those genes that control aspects of fiber strength would offer the great potential to impact fiber quality.
Presented are efforts to identify genetic determinants important to the development of fiber strength in cotton using interspecific crosses of plants known to vary considerably in fiber strength characteristics.
Development Of A Cotton F2 Population
A cross between G. hirsutum acc. TM-1 and G. barbadense acc. 3-79 were chosen because they are the genetic and cytogenetic standards for their respective species (Kohel, 1973). TM-1 is a long-term inbred (for greater than thirty years) derived from 'Deltapine 14'. 3-79 is originated from doubled haploid Pima cottons before any known introgression with G. hirsutum occurred. Of 400 original F2s produced, 137 individuals of (TM-1 x 3-79)F2 were maintained in the greenhouse. These plants served as a resource for DNA sampling, verification of phenotypic traits, and a fiber source for fiber strength determination. Among these, 70 individuals of (TM-1 x 3-79)F2 were chosen to survey with RAPD PCR analysis.
Aneuploid Cottons
A collection of cotton monosomic and monotelodisomic F1 cottons were collected for DNA extraction (Endrizzi and Ray, 1991; Stelly, 1993). The cottons were F1 plants of monosomic or monotelodisomic plants in a G. hirsutum TM-1 background crossed with G. barbadense 3-79. Of the aneuploids used in crosses, monosomic plants were available for chromosomes 1, 2, 3, 4, 6, 7, 9, 10, 12, 16, 17, 18, 20, and 25; and monotelodisomic (short (S) and long (L) arm) plants were available for chromosomes 1, 2, 3, 4, 5 (L only), 6, 7, 9 (L only), 10, 12 (L only), 14 (L only), 15 (L only), 16, 17 (S only), 18 (L only), 20, 21, and 26.
COTTON TRAITS
Bundle Fiber Strength
Fiber samples collected from individual F2 plants were analyzed at least twice for each sample. Bundle fiber strength of each individual was measured by Stelometer in the fiber laboratories of Chembred Inc.(Maricopa, AZ) and Star Laboratory (). Measurements were also obtained for length, uniformity and elongation (not shown).
Phenotypic Characters
The F2 plants were scored for segregating phenotypic characters of G. hirsutum TM-1 and G. barbadense 3-79. G. barbadense 3-79 had dominant alleles for yellow pollen (P(1)), petal spot (R(2)), brown lint (Lc(1)), leaf shape (L(2)o), and yellow petal (Y(1)). Thus the phenotype for G. barbadense 3-79 was P(1)R(2)Lc(1)L(2)oY(1) and that for G. hirsutum TM-1 was p(1)r(2)lc(1)l(2)oy(1). Chromosome assignments have been determined for these observed phenotypes: P(1), chromosome 5; R(2), chromosome 7; Lc(1), chromosome 7; L(2)o, chromosome 15, and Y(1), A genome. Available F3 plants were planted and evaluated to verify scored phenotypic traits as homo/heterozygous.
PCR and RAPD Screening Of Populations
RAPD-PCR methods were used to identify polymorphism between G. hirsutum and G. barbadense parents. The RAPD-PCR methods used 10-mer oligonucleotide primers (Operon Technologies, Alameda, CA or Genosys Biotechnologies, Inc, The Woodlands, TX) and thermostable DNA polymerases (AmpliTaq and Stoffel fragment, Perkin-Elmer Corp., Norwalk, CT) for amplifications.
PCR Amplification reaction mixtures were in volumes of 25 ul that contained 2.5 ul of 10X PCR-II buffer (MgCl2 free); 0.2 mM each of dATP, dCTP, dGTP, and dTTP; 0.2 uM oligonucleotide primer; 0.5 Unit AmpliTaq DNA polymerase; and 10 ng of cotton genomic DNA. Concentrations of MgCl2 were varied from 1.5 to 4.5 mM in the reaction mixture (Park and Kohel, 1994). The 10X PCR-II buffer (50 mM Tris, pH 8.3, 500 mM KCl), dNTP, MgCl2, and AmpliTaq DNA polymerase were purchased from Perkin-Elmer Corp. (Norwalk, CT). DNA was prepared from lyophilized young cotton leaves of adult plants grown in the greenhouse (20-25 C).
Amplification was performed in a thermal cycler (Perkin-Elmer GeneAmp PCR System 9600 or MJ Research PTC-60) programmed for 2 min at 95 C and 45 cycles of 15 sec at 94 C, 30 sec at 40 C, 90 sec at 72 C, followed by 5 min at 72 C (Park and Kohel, 1994). Reaction products (12 ul) were resolved by electrophoresis (12 V/cm) in TBE for 90 min. in 1.4 % agarose gels and visualized by ethidium bromide staining.
The RAPD-PCR methods were used 1) to identify polymorphism between the G. barbadense and G. hirsutum parents, 2) to identify chromosome linkage of polymorphic bands using monosomic and monotelodisomic F1 plants, 3) to identify segregation of RAPD bands within an interspecific F2 population, and to 4) generate polymorphic bands for subcloning as RFLP probes.
Of the original 400 TM-1 X 3-79 F2 plants generated, 129 plants yielded bundle fiber strength information. A histogram of fiber strength frequencies showed approximately a normal distribution for those plants sampled from the population of interspecific F2 plants. The bundle fiber strength (cN/tex) of 116 individuals of (TM-1 x 3-79)F2 ranged from 17 to 34.6 (r2=0.96; Fig. 1). The average of the bundle fiber strength (cN/tex) of the F2 plants (116 individuals) was 24.7+-0.3 (mean+-standard error). The variations in fiber strength appear to represent a normal distribution between the extreme measurements for fiber strength in TM-1 and 3-79. Mean+-standard errors of bundle fiber strength (cN/tex) of the of the parent lines TM-1 and 3-79 (more than 20 replicates) were 20.2+-0.4 and 30.2+-0.8, respectively. The bundle fiber strength for the sampled plants used in RAPD analysis is also shown, and also has a distribution represented by the larger sampled population.
Figure 1. Fiber Strength distribution among interspecific F2 population. For the parents, bundle fiber strength (cN/tex) ranged from 27.9-30.3 to 18.9-20.22 for G. hirsutum TM-1 and G. barbadense 3-79, respectively.
Phenotypic Markers
Phenotypic markers, with known linkage to four different chromosomes, were scored on 113 F2 plants. All five of the phenotypic markers appeared to segregate as expected with a dominantly expressed single gene trait (X2(3:1)= 0.36 to 1.84, Table 2.).
Most RAPD-PCR amplifications were done with AmpliTaq DNA polymerase; however, differences were observed with the Stoffel fragment (Figure 2.). An identical set of plant DNAs amplified with a single 10-mer primer, but amplified with different thermostable DNA polymerases yielded diffferent ranges of DNA fragment sizes. The DNA bands amplified with Taq polymerase holoenzyme amplified products in the 500 to 3500 bp range, whereas the Stoffel fragment amplified products within the 300 to 1100 bp range. The same cycling program was performed with each series. The Stoffel fragment consists of a shorter enzyme product lacking 5' to 3' exonuclease activity when compared to the holoenzyme. Polymorphism is visualized in both parent lines and can be seen with both enzyme series. Some of the same products appear to be amplified in the overlap in size of the DNA products. In screening for polymorphic markers, it appears that the use of at least two different thermostable DNA polymerases may yield additional markers for mapping purposes.
Figure 2. Comparison of Taq Polymerase holoenzyme and the Stoffel fragment for RAPD detection. Amplified DNA products are from plant DNAs of G. hirsutum TM-1 (H), G. barbadense 3-79 (B), and a collection of F1 plants of mono-/monotelodi-somic plants in a TM-1 background crossed with 3-79. The DNA size marker is phage Lambda DNA digested with restriction enzyme BstNII (M).
Approximately 220 10-mer primers were screened for polymorphism between G. hirsutum and G. barbadense. Over 1400 amplified DNA fragments were observed ranging in size from about 200 to 3500 base pairs. The average size of the amplified DNA fragments were 1.3+-0.5 kb (Mean +- standard error). A study of 1296 of the amplified DNA fragments resulted in 442 (34%) of the amplified DNA fragments appearing polymorphic. On the average, 6 to 8 DNA fragments were amplified per reaction mixture with at least one band appearing polymorphic.
Having identified several polymorphic DNA bands distinguishing the G. hirsutum and G. barbadense backgrounds, the segregation of RAPDs among a segregating F2 population may be used identify linkages (Figure 3.). At present, 85 primers have been screened on 40 F2 plants, and 32 primers have been screened on 70 F2 plants. Of the 117 primers screened on the F2 population, polymorphism was observed with markers associated with each of the parental genomes.
Figure 3. Screening F2 population with RAPDs. Amplified DNA products are from plant DNAs of G. hirsutum TM-1 (H), G. barbadense 3-79 (B), and a collection of 70 F2 plants. The DNA size marker is phage Lambda DNA digested with restriction enzyme BstNII (M). In the example presented here a G. hirsutum TM-1 marker known to be associated with chromosome 25 is shown along with two other markers present in G. barbadense 3-79 (chromosome assignments not yet determined).
To identify RAPD markers associated fiber strength determining genes, 85 primers and the generated polymorphic bands in an F2 screen were compared fiber strength of the plants tested. A total of 234 RAPD fragments were generated from 85 primers. The average fiber strength in the F2 plants tested for the 234 RAPD fragments was 24.7 +- 0.5 cN/tex. This value was not significantly different from the mean sampled F2 plants (P <= 0.05). Polymorphic DNA fragement scores were entered into Mapmaker 3.0b.
To establish anchor positions in the cotton genome for QTL analysis and to identify linkages of the RAPD DNA markers, the RAPDs were screened against the interspecific F1 monosomic and monotelodisomic cotton lines with tracable parental backgrounds. When polymorphism was observed between the two parent lines, a RAPD screen was performed on a series of monosomic lines to try to attempt identification of the associated chromosome of the marker. Preliminary screens were conducted on monosomic lines (Figure 4.). Because the monosomic lines were developed in a TM-1 background, only RAPD fragments generated in the TM-1 parent may be scored. Once linkage to a chromosome was determined, a followup screen was conducted including monotelodisomic F1 plants (Figure 5.). This helped to reconfirm the association of an amplified band with a chromosome, and with the monotelodisomic F1 plants included, judgements may help associate the marker with the short or long arm of a chromosome.
Figure 4. Screening of RAPDs on monosomic cotton lines. Amplified DNA products are from plant DNAs of G. hirsutum TM-1 (H), G. barbadense 3-79 (B), and a collection of F1 plants of monosomic plants in a TM-1 background crossed with 3-79. The chromosome number for the respective monosomic line is indicated. The DNA size marker is phage Lambda DNA digested with restriction enzyme BstNII (M). In the example provided, primers E011, 5020, and 5017 demonstrated polymorphic patterns between the two parents. When screened against the monosomic F1s, selected polymorphic DNA bands appeared to be associated with chromosomes 4, 16, and 25, respectively.
Figure 5. Screening of RAPDs on monosomic and monotelodisomic cotton lines. Amplified DNA products are from plant DNAs of G. hirsutum TM-1 (H), G. barbadense 3-79 (B), and a collection of F1 plants of mono-/monotelodi-somic plants in a TM-1 background crossed with 3-79. The chromosome number for the respective mono-/monotelodi-somic line is indicated. The DNA size marker is phage Lambda DNA digested with restriction enzyme BstNII (M). For instance, in the example provided, two seperate primers amplified products which were associated with chromosome 12 in monosomic crosses. In these examples, the indicated DNA bands appear associated with the short arm of chromosome 12, since the band was missing for the chromosome 12 monotelodisome missing the long arm.
Presently, 64 primers have been used to uncover possible linkage of the DNA markers to selected chromosomes of G. hirsutum. Approximately 25 RAPD fragments have been identified which are associated with 16 of the chromosomes screened for in F1 monosomic/monotelodisomic cotton lines (Table 2.). All of these fragments are for TM-1 associated RAPD fragments. These markers were useful for screening F2s by RAPD methods; however, the some of the amplified products were subcloned into vectors for archiving purposes and for possible use as RFLP probes. At least 87 RFLP markers have been made available for screening which may help confirm assignments to some of these markers (Reinisch et al, 1994).
The influence of MgCl2 concentration on products in PCR to generate RAPD fragments was evaluated between two highly polymorphic cotton lines (Park and Kohel, 1994). The optimum concentration of MgCl2 was within the range of 1.5 to 4.5 mM in the PCR reaction mixtures for the random primers tested.
Most of the RAPD screens used Perkin-Elmer Taq DNA polymerase for amplification of DNA fragments. Potentially, the Stoffel fragment may be used to identify a whole other set of polymorphic DNA bands because of the different size range of amplified DNA fragments. Because Stoffel lacks a 5'-3' exonuclease, the smaller DNA fragments may be due to inability to remove annealed primers to the template. Another attribute may be that the Stoffel fragment of Taq DNA polymerase is not as sensitive to MgCl2 concentrations in the reaction mixture: however, some differences in the quality of fragments generated were noted (not shown). With a large number of variations of thermostable DNA polymerases available, it appears to be very important to limit a series of evaluations based on the individual enzyme used.
The collection of the interspecific F2 population appears to be normally distributed for the phenotypic markers, DNA markers, fiber strength, and other quality traits. We have identified some of the RAPD markers as anchors to known chromosomes of cotton. The fiber strength analysis from 129 of the F2 population sets a starting point for conducting quantitative trait loci analysis to associate RAPDs with fiber strength characteristics. With the recent availability of RFLP probes (Reinisch et al, 1994) the confirmation of RAPD assignments may be verified. The RFLP markers hybridize to at least 123 mapped loci of the 41 assembled linkage groups. Since the RAPD assignments identified to date are from the G. hirsutum background, the RFLP probes will help to identify other markers from both G. hirsutum and G. barbadense. More certainty about the phenotypes of the F2 population, such as fiber strength measurements, will allow for RAPD testing by bulk, or selected pool, analysis.
We have constructed a 173 plant population of interspecific F2 plants which appears normally distributed for phenotypic markers, RAPD markers, fiber strength, and other quality traits. A collection of DNA markers are available which may serve as anchor sets to help determine genomic loci important toward contributing to cotton fiber strength and possibly other characteristics.
The use of DNA markers for marker-facilitated selection in cotton offers extraordinary promise for streamlining many plant breeding efforts (Tanksley et al., 1988; Paterson et al., 1991), particularly for introgression of valuable genes from exotic germplasm (Paterson et al., 1991) and breeding for traits affected by many quantitative trait loci (QTLs) (Lande and Thompson, 1989; Dudley, 1993). Thus, the DNA marker(s) that is diagnostic of an agriculturally important trait, can accelerate breeding by allowing selection for the markers rather than the trait (Lander and Bostein, 1989). Numerous advantages in reduced time and population size can accrue to the breeder by use of marker-facilitated genotype selection rather than classical phenotypic selection (Paterson et al., 1991; Tanksley et al., 1988). The breeder might practice selection among seedlings too young to express a trait, or practice selection on individuals in environments where the trait is not expressed normally, such as in the greenhouse or winter nursery. These benefits and others can reduce the time to develop new cultivars in cotton. Further, identification of DNA markers near a gene of interest has proven to be a valuable starting point for gene cloning (Rommens et al., 1989; Wallace et al., 1990) when no protein product is available and the gene affects a complex agricultural trait (Paterson et al., 1990).
