Potential of ten wild diploid cotton species for the improvement of fiber fineness of upland cotton through interspecific hybridization

Cotton is the highest source of natural fiber in textile industry worldwide. With the modern spinning technologies, the demand for cotton fiber with higher quality has increased, making the genetic improvement of fiber quality one of the main challenges for cotton breeders. In cotton breeding, wild species are important source of several desirable genes for genetic improvement of the main cultivated cotton Gossypium hirsutum L (Upland cotton). Besides length and strength, fineness is one of the most important criteria associated with cotton fiber quality. In this study, ten wild diploid species of cotton were investigated for their fiber fineness and potential to improve fiber fineness of G. hirsutum L. The method was measuring of ribbon width after caustic swelling. The results showed the potential of four wild species (G. longicalyx Hutch. & Lee, G. anomalum Wawra & Peyr., G. thurberi Todaro and G. stocksii Mast.) to significantly improve the fiber fineness of upland cotton in a hybrid configuration. Among them, G. longicalyx stood out for its exceptional fiber fineness, and its remarkable impact on reducing the fiber fineness of G. hirsutum L. The wild species highlighted in this study constitute an interesting genetic resource for the development of upland cotton varieties with improved fiber fineness.


INTRODUCTION
Cotton fiber is the major commercial product from cotton and the most widely used natural fiber in the world's textile industry (Ayubov et al., 2018). This important fiber crop belongs to the genus Gossypium which includes 46 diploid (2n = 2× = 26) and 7 tetraploid (2n = 4× = 52) species (Fang et al., 2017). All the diploid Gossypium species originated from a common ancestor and diversified into eight genome groups from A to G, and K (Wu et al., 2018). All tetraploid cotton species are allotetraploid and have a genome designated by AD; they come from a natural hybridization event between an Agenome species and a D-genome species, followed by a doubling of the chromosome number 1 to 2 million years ago (Wendel and Grover, 2015;Fang et al., 2017).
Among the 53 Gossypium species, only four species including two diploids (G. arboreum L and G. herbaceum L) and two tetraploids (G. hirsutum L and G. barbadense L) are cultivated for their spinnable fibre (Gallagher et al., 2017;Wang et al., 2018;Ijaz et al., 2019). The remaining 46 species are wild.
G. hirsutum L, which is also known as Upland cotton, Long Staple cotton or Mexican cotton, is extensively cultivated due to its wide adaptability to the environment, high production, and better yield potential. It fulfils over 90 % of the output of global cotton fiber yield (Shim et al., 2018;Konan and Mergeai, 2020). G. barbadense L, otherwise known as Sea Island cotton, Pima cotton or Egyptian cotton, is known for excellent fiber quality with long, strong, and fine fibers (Avci et al., 2013). It contributes to 8% of the global cotton production (Shim et al., 2018). The cultivated diploid species provide approximately 2% of the world's cotton and are cultivated in the more traditional growing areas of India, Pakistan, China, Bangladesh and Iran (Kulkarni et al., 2009, Wendel et al., 2010Shim et al., 2018).
Based on genetic hybridization properties, Gossypium species are grouped into the primary, secondary and tertiary gene pools. Both the cultivated (G. hirsutum L and G. barbadense L) and wild allotetraploids (G. tomentosum Nuttall ex Seemann, G. mustelinum Miers ex Watt and G. darwinii Watt) comprise the primary gene pool of cotton. The secondary gene pool includes the diploids having the A, B, D and F genomes, whereas the tertiary gene pool is composed of species with C, E, G and K genomes (Campbell et al., 2010).
Previously, cotton breeders primarily emphasized yield and agronomic characteristics, but with the recent development of high-speed spinning technologies, the demand for cotton fiber with higher quality has increased, making the improvement of fiber quality highly crucial in Upland cotton (Islam et al., 2016;Shang et al., 2016;Ayubov et al., 2018). Faced with this existing demand and the dynamics of modern textile industry, the perpetual need of genetic improvement in fiber quality is one of the main challenges for cotton breeders today. Biologically, cotton fibers are single-celled trichomes that grow from the epidermal cell layer of the ovule in a boll (Miao et al., 2017;Ayubov et al., 2018;Ijaz et al., 2019). Besides the length and the strength, the fineness is one of the most important criteria associated to cotton fiber quality (Bradow and Davidonis, 2000;Konan and Mergeai, 2020).The fineness of mature fiber is critical for fiber processing. It influences the fabric lustre, dye appearance, fabric stiffness, spinning performance, and yarn strength (Rodgers and Thibodeaux, 2012). The better the fineness of cotton, the more would be the number of fibers per cross-section. This would result in higher yarn strength, which improves spinning efficiency and yarn evenness (Ahmad et al. 2003;Islam et al., 2016).
Cotton fiber fineness can be expressed as the perimeter, diameter or ribbon width (RW), cross sectional area, and standard fiber weight (Rodgers and Thibodeaux, 2012). The indirect methods used for its measurements are Advanced Fibre Information System (AFIS), Fibre Maturity Tester (FMT), and Near Infrared (NIR) spectroscopy, Vibroscope, High Volume Instrument (HVI) for micronaire etc; the most common direct measurements of fiber fineness include cross-sectional image analysis and ribbon width measurement after caustic swelling (Rodgers and Thibodeaux, 2012). The most effective way to improve cotton fiber fineness is through breeding (Nacoulima and Mergeai, 2014;Islam et al., 2016). Previous progress in the improvement of fiber quality of upland cotton has been mainly achieved using the genetic diversity present in the primary gene pool of cotton (especially G. barbadense L), but currently, this available diversity has been exhaustively utilized (Gotmare et al., 2000;Ayubov et al., 2018). Accordingly, it has become a necessity to exploit useful genes of wild species from the two other gene pools. Indeed, in cotton breeding, wild species constitute an important resource with several useful traits which can be introgressed into the main cultivated species for improvement (Konan and Mergeai, 2020). The objective of the present study is to detect donor parents for fiber fineness by determining the fiber fineness of a collection of wild diploid species using ribbon width measurement and evaluating their potential to improve fiber fineness of upland cotton through interspecific hybridization.

Plant material
The plant material included plants from the living cotton collection of the Laboratory of Tropical Agro ecology of Gembloux Agro-Bio Tech (Liège University, Belgium). It was composed of eleven diploid cotton species, their bi-species hybrid with G. hirsutumL, one cultivar of the tetraploid species G. barbadense L, four cultivars of the tetraploid species G. hirsutum L and fifteen second back-cross (BC2) progenies of the HTL tri-species hybrid (G. hirsutum L × G. thurberi Todaro)² × G. longicalyx Hutch. & Lee ( Table 1). The crossing scheme used to generate the bi-species hybrid and the BC2 progenies of the HTL tri-species hybrid are presented in Figures 1 and 2, respectively. The crossing procedures used are presented in detail by Konan et al. (2007) and Konan and Mergeai (2020). The plants were maintained in a ventilated greenhouse where the growing conditions during capsule maturation period were 55-60% relative humidity and 35-26°C day-night air temperatures. The plants were grown in 5 L pots filled with a 3:2:1 (v:v:v) sterile mixture of compost, sand and peat. Cotton fibers were harvested at full maturity and used for the analysis of their fineness.

Fiber fineness analysis
Fiber fineness analysis was conducted on all the genotypes studied. For this analysis, the fibers were combed and a tuft of parallel fibers was cut from the seed. Their free points were also cut and the median region was placed on a slide and covered with a cover glass.

G. hirsutum L A h A h D h D h 2n=4x=52
A diploid species of Gossypium X genome 2n=2x=26

Sterile triploid hybrid A h D h X 2n=3x=39
Fertile allohexaploid 2(A h D h X) 2n=6x=78 x Doubling of chromosomes by colchicine  One or two drops of 18% NaOH solution was allowed to penetrate into the fibers by capillarity. The NaOH solution swells the fibers ( Figure 3). The diameter of at least 100 fibers was then measured with the software NIS-Elements BR 2.30 (Nikon, Japan) using the Nikon Eclipse E800 microscope (Nikon, Tokyo, Japan) equipped with a digital JVC KY-F 58E camera (JVC, Yokohama, Japan). The ribbon width was determined by dividing the mean of the diameters measured by the 1.3 Summers coefficient (Roehrich, 1947;Nacoulima et al., 2016;Konan and Mergeai, 2020).

Statistical analysis
All the data collected were subjected to the analysis of variance (ANOVA) using the software Statistica 7.1 (Stat Soft, France). The least significant difference (LSD) was used to establish the differences between means at P=0.05.

Analysis of fiber fineness of studied diploid and tetraploid cotton species
The results of the analysis of fiber fineness for the studied diploid and tetraploid cotton species are presented in Table 2. The ribbon width of the ten wild diploid species varied from 5.940 µm (G. longicalyx Hutch. & Lee) to 15.533 µm (G. thurberi Todaro), while that of the cultivated species ranged from 17.765 µm (G. hirsutum L cv. C2) to 24.374 µm (G. arboretum L). All the wild diploid species had finer fibers than the cultivated species. Their fibers were even finer than the Sea Island cotton (G. barbadense L), which is known for its fine

Trispecies hybrid (HTL)
A   (Demol et al., 1978;Nacoulima et al., 2016;Konan et al., 2020). The results also highlighted another African wild species, G. anomalum Wawra & Peyr., which presented good fiber fineness close to that of G. longicalyx Hutch. & Lee, with no significant difference. The good fiber fineness of G. anomalum Wawra & Peyr. has also been reported by Mehetre (2010). The American wild species G. harknessii showed finer fiber than G. raimondii Ulbr, but it is rarely cited as a good source of fiber fineness like G. raimondii Ulbr (Gotmare et al., 2000;Islam et al., 2016).G. harknessii Brandegee is most often cited for its resistance to Verticillium wilt and Fusarium wilt, and as source of cytoplasmic male sterility and fertility restorer (Ano et al., 1982;Gotmare et al., 2000). Among the cultivated species, the Upland cotton varieties G. hirsutum L (cv. C2) and G. hirsutum L (cv. NC8) had the finest fibers with 17.765 and 18.294 µm ribbon width respectively; while G. barbadense L presented a ribbon width of 19.117 µm. Although G. barbadense L is recognized as having finer fiber than Upland cotton (Avci et al., 2013), the present results showed finer fibers for these two varieties of G. hirsutum L. Actually, several varieties of upland cotton resulting from breeding programs for fiber quality have gained in fiber fineness comparable to that of G. barbadense L; this is the case for these two varieties of G. hirsutum L (cv. C2 and cv. NC8) in the present study.
Of the results presented in Table 2, the cultivated diploid species G. arboreum L had the highest ribbon width value. This result showed that not all diploid species produce fine fibers, even if all the other (wild) diploid species studied had finer fibers than the tetraploid cotton studied. It again stresses that wild diploid species can be a source of desirable genes for the genetic improvement of cultivated cotton (Konan and Mergeai, 2020).

Analysis of fiber fineness of the bi-species hexaploid hybrids
To evaluate the influence of the studied diploid genomes on the fiber fineness of upland cotton, hybrids including each of these genomes and genome of G. hirsutum L cv C2 or cv NC8 were examined for their fiber fineness. The results of this analysis are shown in Table 3. The mean values of ribbon width of the different hybrid ranged from 12.526 to 26.072 µm. The bi-species hexaploid hybrid (G. hirsutum L cv. C2 × G. longicalyx Hutch. & Lee)² showed the finest fibers with a mean value of ribbon width of 12.526 µm. It was followed by (G. hirsutum L cv. C2 × G. anomalum Wawra&Peyr)² with on average 15.833 µm of ribbon width, and then (G. hirsutum L cv. C2 × G. thurberi Todaro)² and (G. hirsutum L cv. NC8 x G. stocksii Mast.)² with mean value of 16.835 and 16.852 µm of ribbon width, respectively. The highest value of ribbon width  Todaro and G. stocksii Mast, all the other diploid species did not bring an interesting improvement in fiber fineness of G. hirsutumL. Even some wild diploid species such as G. harknessii Brandegee (genome E3) and G. raimondiiUlbr. (genome D5) which had good fiber fineness (ribbon width <10 µm) could not reduce the ribbon width of G. hirsutum L when combined to it in bispecies hybrids. These results suggest that the genes that control the fineness of the fibers in the different wild diploid species did not have the same action when they are confronted with the genome of upland cotton in a hybrid configuration. The diameter of the cotton fiber is primarily a genetic trait and the genetic mechanisms of fiber traits are complex (Matic-Leigh and Cauthen, 1994;Bradow and Davidonis, 2000;Zhang et al., 2013;Islam et al., 2016). According to Ijaz et al. (2019), cotton fiber quality traits are controlled by multiple genes (polygenic inheritance) with different mechanisms and complex genetic architecture. For instance, in the past decades, studies on cotton fiber quality traits on G. hirsutum L and G. barbadense L found a significant association between SSRs and fiber quality traits and identified 70 stable loci for target traits including 30 for fiber length, 27 for fiber strength, and 13 for fiber fineness (Zeng et al., 2009;Cai et al., 2014). Later, several studies, on cotton fiber quality traits that focused on both G. hirsutumL and G. hirsutum L×G. barbadenseL populations, have mapped fiber QTLs in large genomic regions that may include hundreds or thousands of genes (Said et al., 2013;Fang et al., 2014;Shang et al., 2015;Tang et al., 2015;Ma et al., 2017Ma et al., , 2018Ijaz et al., 2019). QTLs are chromosomal regions which contribute cumulatively to a trait with varying percentages of phenotypic variance from each QTL (Said et al., 2015). According to Ijaz et al. (2019),the number of fiber quality trait QTLs over the chromosomes of the cotton genome is not identical, and QTLs associated with cotton fiber quality obtained from Cotton QTL database (http://www.cottonqtldb. org) are distributed unevenly across the 26 chromosomes of the cotton genome.

Analysis of fiber fineness of the tri-species hybrid and its BC2 progenies
The HTL tri-species hybrid (G. hirsutum L × G. thurberi Todaro)² × G. longicalyx Hutch. & Lee (Konan et al., 2007) with a ribbon width of 12.65 µm (Table 3) Konan and Mergeai (2020), working on twelve BC1 progenies of the trispecies hybrid HTL, reported ribbon width ranging from 13.039 to 16.276 µm with four BC1 plants having the lowest ribbon width (13.039 -13.416 µm). This fiber fineness segregation among the HTL/BC plants may be due to the segregation of G. longicalyx alleles among the BC plants. This suggests the differential presence or absence of this diploid species chromosomes and/or chromosome recombinants as shown by Konan and Mergeai (2020) with genomic in situ hybridization (GISH) analysis. The persistence of the outstanding fiber fineness of G. longicalyx Hutch. & Lee, in the bi-species hybrid with G. hirsutum L, in the HTL tri-species hybrid and in the HTL/BC1 and BC2 derivative plants demonstrates the inheritance of this trait through the crossing scheme (Figure 4)  for fiber fineness.In addition, according to Demol et al. (1978), the fibers of G. longicalyx Hutch. & Lee have exceptional fiber strength and a high molecular weight. Such finer and stronger fibers than those of G. barbadense L would undoubtedly be much appreciated by spinners. These results therefore make G. longicalyx Hutch. & Lee an interesting source that deserves more attention from breeders for the improvement of cotton fiber quality

Conclusion
The results obtained in the present study show the potential of four wild cotton diploid species (G. longicalyx Hutch. & Lee, G. anomalum Wawra & Peyr., G. thurberi Todaro and G. stocksii Mast.) to significantly improve the fineness of the fibers of upland cotton in a hybrid configuration. However, among these wild species, G. longicalyx Hutch. & Lee stood out for its exceptional fiber fineness, and its remarkable impact on improving the fiber fineness of G. hirsutum L. This wild African diploid species seems to be a good donor for the introgression of this useful trait into upland cotton. In view of the results of this study, the species G. longicalyx, and to a lesser extent the three other highlighted wild species, constitute interesting genetic resources for the development of cotton varieties with outstanding fiber fineness.