Selection of Gossypium hirsutum genotypes for interspecific introgression from G . arboreum using ovule culture

Ovule culture is one of the techniques currently used to introgress desirable traits from Gossypium arboreum germplasm into G. hirsutum cultivars. Twenty-six (26) G. hirsutum breeding lines were used as female parents in crosses with five G. arboreum accessions to determine if the G. hirsutum parent influenced the germination and recovery of plants from ovule culture. Variation in boll weight and the number of ovules per boll was observed for crosses with the G. hirsutum lines, but heavier bolls and a greater number of ovules per boll were not associated with a higher germination rate. Ovules derived from crosses with 16 G. hirsutum lines showed germination. Plants were recovered for seven of these lines (Acala GLS, DES 56, DES 119, Deltapine 50, Stoneville 132, Stoneville 506 and Stoneville 825) with vigorous growing plants derived from four crosses (DES 119 x PI 408763, Stoneville 506 x PI 408763, Acala GLS x PI 529779, and DES 119 x PI 615699). The breeding line DES 119 showed a better success rate and typically produced smaller bolls with fewer ovules. However, results would suggest the G. arboreum accessions had a greater influence on the success rate compared to the G. hirsutum lines.


INTRODUCTION
Tetraploid upland cotton, Gossypium hirsutum L., is comprised of over 90% of global cotton production (Zhao et al., 2015).Cultivated G. hirsutum genotypes are considered to have a narrow genetic base, due in part to a monophyletic origin and the use of few genotypes in the breeding of new cultivars (Grover et al., 2012;Zhao et al., 2015).In contrast, the diploid species G. arboreum L. comprises less than 1% of global cotton production, but is considered an important source of genetic diversity for several traits (Liu et al., 2006).Gossypium arboreum is commonly grown on marginal lands with low inputs and is a source of drought tolerance (Basu, 1996;Maqbool et al., 2007) and pest resistance (Yik and Birchfield, 1984;Thengane et al., 1986;Miyazaki et al., 2012).The ability to transfer genes from G. arboreum to G. hirsutum is hindered by incompatibility barriers resulting in abortion of developing bolls (Mehetre and Aher, 2004).Techniques such as exogenous hormone application E-mail: John.Erpelding@ars.usda.gov.Fax: 662-686-5218.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License (Altman, 1988), hexaploid bridging lines (Sacks and Robinson, 2009), in vitro interspecific fertilization (Liu et al., 1992), protoplast fusion (Sun et al., 2006), and ovule culture/embryo rescue (Stewart and Hsu, 1978;Gill and Bajaj, 1987) have been used for interspecific hybrid development in cotton.Ovule culture has frequently been used for hybrid development between incompatible diploid species and for crosses between diploid and tetraploid cotton species (Mehetre and Aher, 2004).Several factors can influence the success of ovule culture such as the species being crossed, age of the excised ovules, cultural media, and environmental conditions for culturing and plant growth (Stewart and Hsu, 1978;Mehetre and Aher, 2004).Research has mainly focused on improving the culture media to increase the success rate of germination and plant recovery (Stewart and Hsu, 1978;Thengane et al., 1986;Gill and Bajaj, 1987;Sacks, 2008).Information is lacking to determine if the G. hirsutum parent used for interspecific introgression has an influence on the success rate.Thus, the focus of this study was to evaluate 26 G. hirsutum breeding lines to determine if the tetraploid parent used in crosses with G. arboreum genotypes influenced the growth and germination of immature ovules and the recovery of plants from tissue culture.

MATERIALS AND METHODS
Twenty-six (26) diverse G. hirsutum breeding lines (Table 1) and five G. arboreum accessions (PI 152088, PI 408763, PI 529724, PI 529779, and PI 615699) were selected from the germplasm collection maintained at the United States Department of Agriculture (USDA) Agricultural Research Service (ARS), Crop Genetics Research Unit in Stoneville, Mississippi.These germplasm lines were planted in the field at the USDA ARS in Stoneville during the 2013 growing season.Each line was planted in a single 3 m row and rows were spaced 0.9 m apart.Standard management practices were conducted for cotton production in Mississippi.The G. hirsutum lines were used as the female parent with the G. arboreum accessions used as the pollen parent.Flowers of the G. hirsutum lines were emasculated by splitting the staminal column with the fingernail and removing the corolla and androecium (Doak, 1934) between 06:30 to 07:30 a.m.Following emasculation, flowers were misted with a single application of a 100 mg L -1 solution of gibberellic acid to prolong boll retention (Miravalle, 1964).Pollinations were conducted the same morning between 10:00 to 11:00 a.m.Each emasculated flower was pollinated using a single G. arboreum flower.Seven days after pollination, the developing bolls were removed for ovule culture.Bolls were weighed, surface sterilized in an aqueous 0.2 M sodium hypochlorite solution for 15 min., transferred to a 95% ethanol solution for another 15 min., and then air-dried in a biological safety cabinet prior to excision of ovules.The number of immature ovules from each boll was counted.The ovules were cultured on MS basal media (Murashige and Skoog, 1962) with B5 vitamins (Gamborg et al., 1968) and supplemented with 20 g L -1 sucrose, 1.9 g L -1 potassium nitrate, and 5 g L -1 gelrite as a solidifying agent (Sacks, 2008).All ovules from a single boll were cultured in the same 100 × 15 mm sterile plastic Petri plate.Ovules were cultured at 30°C in a growth chamber with a 16 h photoperiod.Ovules were transferred to fresh media every 21 days.Seedlings that developed from the ovules were transferred to magenta plant culture boxes containing the same media used for ovule culture and grown under the same conditions.Seedlings that developed roots and shoots were planted in pots containing potting media (Metro-Mix 360, Sun Gro Horticulture, Agawam, MA) and covered with a beaker for seven days to prevent desiccation of the developing seedlings.Pots were placed in a growth room at 27°C with a 16 h photoperiod.At flowering, plants were transferred to a greenhouse for further evaluation.

RESULTS AND DISCUSSION
Boll weight and the number of ovules per boll are presented in Table 1.More variation in mean boll weight and mean number of ovules per boll was observed across the G. hirsutum breeding lines as compared to the G. arboreum accessions.For the G. hirsutum breeding lines, the mean number of immature ovules cultured across the five G. arboreum accessions ranged from 30 to 42.The breeding line Sure-Grow 747 produced the most ovules across the G. arboreum accessions.Mean boll weight across the G. arboreum accessions was also greater for Sure-Grow 747 with the highest mean boll weight recorded for Acala GLS.However, the greater number of ovules per boll and higher boll weights did not result in the production of more plants from culture.The breeding line DES 119 showed the lowest mean number of ovules per boll and low mean boll weight, but produced more plants than breeding lines with a greater mean number of ovules per boll and higher boll weights.Across the G. arboreum accessions no noticeable trend was apparent.A mean of 37 immature ovules were cultured from the crosses with the G. arboreum accessions.Crosses involving accession PI 408763 produced a greater mean number of ovules and heavy bolls across the 26 G. hirsutum breeding lines with ovules derived from crosses with PI 408763 and PI 615699 showing a higher frequency of germination (Table 2).However, accession PI 529724, which had the lowest mean boll weight across the 26 breeding lines, produced a greater number of plants.Altman (1988) showed that exogenous hormone application increased boll weight and suggested that heavy bolls may correspond to better quality seed being produced; however, production of plants from G. hirsutum x G. arboreum crosses was unsuccessful.Flowers for the 26 G. hirsutum breeding lines pollinated with G. arboreum pollen and not sprayed with gibberellic acid aborted bolls within five days after pollination (data not shown).No comparisons were conducted between bolls treated with gibberellic acid and non-treated bolls to determine if the treatment increased boll weight.Rapid abortion of non-treated bolls was observed for crosses with several G. hirsutum lines and the ovules from these bolls failed to grown in culture.The number of ovules per boll was within the range observed for the treated bolls (data not shown); however, no additional data were collected due to poor boll development and ovule growth in culture.
All ovules from the crosses treated with gibberellic acid showed growth in culture.Variation in ovule growth within and between crosses was observed.
Some ovules showed more than a 10-fold increase in size.These ovules that showed greater growth in culture typically did not germinate and produce seedlings.Germination varied among the ovules derived from the crosses (Table 2).Germination was low compared to results reported by Stewart and Hsu (1978), Gill andBajaj (1987), andSacks (2008), but similar to the results reported by Altman et al. (1987) and Altman (1988) Many of the developing seedlings showed little or no additional growth when transferred to media in magenta boxes.Other seedlings showed abnormal growth and failed to survive when transplanted into potting media.Seedlings that germinated from ovules derived from crosses between accession PI 529724 and the breeding lines Deltapine 50, DES 56, Stoneville 132, and Stoneville 825, successfully developed into plants; however, when these plants produced flowers they started to wither and die.These plants failed to develop a vigorous root system to support the growth of the plants.Other researchers have reported abnormal seedling growth and poor plant survival (Thengane et al., 1986;Gill and Bajaj, 1987).In crosses involving three of five G. arboreum accessions (PI 408763, PI 615699, and PI 529779), healthy and vigorous plants have been recovered ( No cultured ovules showed germination from crosses with Deltapine 90; however, four additional crosses between Deltapine 90, and accessions PI 408763 and PI 529724, did result in the recovery of a single plant for each accession.The plant recovered from the Deltapine 90 x PI 529724 crosses failed to survive as was observed for the other plants derived from crosses with this accession.Multiple interspecific crosses were also conducted with Sure-Grow 747, but no plants were recovered from these additional crosses.However, cultured ovules derived from crosses with Sure-Grow 747 and other G. arboreum accessions have resulted in the production of plants at a low frequency with the recovery of one plant from crosses with 24 accessions.Sure-Grow 747 is a source of superior fiber quality traits, but the unsuccessful recovery of plants for the majority of the interspecific crosses would suggest other G. hirsutum lines evaluated in this study would be more desirable for ovule culture.These results would indicate that evaluation of advance breeding lines or recently release G. hirsutum cultivars would be useful for the identification of superior lines for ovule culture.The use of improved cultivars would reduce the time required to recover desirable yield and fiber quality traits.However, the G. arboreum accession used in the cross may have a greater influence on the recovery rate and survival of plants compare to the G. hirsutum parent.Multiple factors such as tissue culture media, parental genotypes, age of immature ovules cultured, and environment conditions (Mehetre and Aher, 2004) can influence the germination for cultured ovules and the recovery of plants.In this study, germination rates were low and modification to the media could increase the recovery of plants.Several media formulations have been published, but similar results across laboratories have not been achieved (Altman, 1988;Sacks, 2008).Selection of a subset of G. hirsutum breeding lines and multiples crosses to the lines would increase the likelihood of generating plants from ovule culture.For some G. arboreum accessions, such as PI 529724, other approaches may be required to recover plants.Crosses with this accession and several breeding lines resulted in plants, but all plants failed to survive.Whereas, the success rate for other G. arboreum accessions, such as PI 615699, was greater with ovule germination observed from crosses with numerous G. hirsutum breeding lines.The successful recovery of plants from three G. arboreum accessions used in interspecific crosses provides breeding lines for further evaluation in the introgression of desirable traits for cotton improvement.

Table 1 .
Boll weight (g) at 7 days after pollination and number of ovules per boll from crosses between 26 Gossypium hirsutum breeding lines used as the female parent and five G. arboreum accessions(PI 152088, PI 408763, PI 529724, PI 529779, and PI 615699)used as the pollen parent to evaluate the germination and recovery of plants from ovule culture.

Table 2 .
Number of ovules showing germination in tissue culture and the number of plants recovered from crosses between 26 Gossypium hirsutum breeding lines used as the female parent and five G. arboreum accessions (PI 152088, PI 408763, PI 529724, PI 529779, and PI 615699) used as the pollen parent.All plants derived from crosses with G. arboreum accession PI 529724 died at the flowering stage.