Phenotypic, Cytological and Molecular (AFLP) Analyses of the Cotton Synthetic Allohexaploid Hybrid (G. hirsutum × G. longicalyx)² ()
1. Introduction
Cotton, from the genus Gossypium, is the most important natural fiber source for the textile industry in the world [1]. The Gossypium genus is composed of 53 species among which 7 are tetraploid (2n = 4× = 52) and 46 are diploid (2n = 2x = 26) [2]. Diploid cottons are classified into 8 genome types, denoted A-G and K, based on chromosome pairing relationships [3] [4]. The tetraploid cottons have a genome designated by AD, which resulted from the ancestral allopolyploidization of progenitor A-genome and D-genome diploids about 1-2 million years ago [2] [5].
Two cotton diploids species (G. arboreumn L and G. herbaceum L) and two tetraploids species (G. hirsutum L and G. barbadense L) are cultivated for their spinnable fiber [6] [7] [8]. G. hirsutum, known as Upland cotton, provides more than 90% of the world’s cotton production due to its high yield [8]. The remaining cotton fiber supply is produced from the other three cultivated cottons [9]. Apart from these four species, the other 46 species of Gossypium are wild.
Genetic improvement of the main cultivated cotton G. hirsutum can be done using wild species as donor of traits of interest [10]. Indeed, wild diploid species are important sources of several desirable genes to improve fiber quality, resistance to diseases and insect pests, or tolerance to abiotic stress of Upland cotton [4] [11]. The African wild species G. longicalyx Hutchinson and Lee (F1 genome) represents an interesting source of genes that can potentially be transferred to the main cultivated cotton species. This wild diploid species could be used as donor of the desirable trait of fiber fineness, which is very important to textile industry [4] [12] [13], and also as donor of the resistance to reniform nematode [14] [15].
Technically, the use of wild diploid species to improve cultivated tetraploid cotton faces the problem of the sterility of the triploid (2n = 3x = 59) obtained, mainly due to ploidy. To overcome this problem, a strategy used involves the creation of an allohexaploid hybrid as a bridge genotype. This method begins with the hybridization of G. hirsutum with a diploid wild species to obtain a sterile triploid (2n = 3x = 59). The next step consists in doubling the chromosome number of the sterile triploid with colchicine to give a fertile allohexaploid (2n = 6x = 78). Such a synthetic allohexaploid hybrid could subsequently be used to introgress the alien genes through trispecific hybrid or monosomic addition lines [16] [17] [18]. In cotton breeding, the successful use of this method has been demonstrated by the effective introgression of alien chromosomic material into upland cotton and the expression of the beneficial effects of exotic genes [4] [11] [19] [20].
At the Laboratory of Tropical Agro ecology of Gembloux Agro-Bio Tech (Liège University, Belgium) a synthetic allohexaploid hybrid (G. hirsutum × G. longicalyx)2 has been created. As a prelude to the use of this hexaploid in a breeding program, the present study aimed at its phenotypic, cytological and molecular analysis (AFLP markers) in order to confirm its hybridity and check its expression of the desirable traits from G. longicalyx.
2. Material and Methods
2.1. Plant Material
Plant material included plants of a synthetic allohexaploid hybrid (G. hirsutum x G. longicalyx)2 and its parental species G. longicalyx and G. hirsutum. These plants were obtained from seeds coming from the collection of the Tropical Agro Ecology Laboratory of Gembloux Agro-Bio Tech (University of Liège, Belgium). In the crossing scheme used to generate the allohexaploid hybrid (G. hirsutum × G. longicalyx)2, the tetraploid G. hirsutum (A1A1D1D1, 2n = 4× = 52) was crossed with the diploid G. longicalyx (F1F1, 2n = 2× = 26) to give triploid hybrid (A1D1Fl, 2n = 3x = 39) seeds. Hybrid seedling plants were then treated with 0.15% colchicine for chromosome doubling and an allohexaploid (A1A1D1D1F1F1) with 78 chromosomes was obtained. This putative synthetic allohexaploid hybrid produced flowers and set bolls normally. Its seeds were used to produce the plants that are investigated in the present study. The plants were grown in greenhouse, in 5 liter pots filled with a 3:2:1 (v:v:v) sterile mixture of compost, sand and peat.
2.2. Cytological Analysis
To check the chromosome numbers of the hexaploid and its parental species, mitotic chromosome preparations were carried out using root tips. Fast-growing root tips were collected in 0.04% 8-hydroxyquinoline for 4 hours at room temperature (RT) and fixed for 48 h in a fresh fixative fluid (3:1 ethanol: acetic acid) at 4˚C. After washing in distilled water (2 × 10 min), treating in 0.25 N HCl (10 min), rinsing in distilled water (10 min) and treating in a 0.01M citrate buffer (10 min), root tips were digested in an enzyme solution (5% cellulase Onozuka R-10, 1% pectolyase Y- 23 in citrate buffer) at 37˚C for 1 hour. The enzyme mix was removed by rinsing in distilled water for 10 min, and on a clean glass slide a single root tip was spread in one or two drops of fresh fixative (3:1 ethanol:acetic acid) using a fine-pointed forceps. After staining with 4,6-diamino-2-phenylindole/Vectashield, mitotic metaphase plates were visualized and the chromosomes were counted under fluorescent light with a Nikon Eclipse E800 microscope (Nikon, Tokyo, Japan) equipped with a JVC KY-F 58E camera (JVC, Yokohama, Japan).
2.3. Phenotypic Observations
Morphological observation and plant fertility evaluation
The appearance, shape and size of the hexaploid seeds were observed and compared to those of its parental species and their germination rates were assessed. The test of germination was conducted on 44 seeds of the hexaploid and 30 seeds of each of its parental species. The seeds were placed to germinate in Petri dishes with moist filter paper at 28˚C. The criterion for germination was a radicle length of > 1 mm. The germination rates were calculated as the percentage of seeds that germinated from the total number of seeds placed in the Petri dish.
The morphological observations carried out on the plants concerned: the aspect of seedlings (presence of malformations or not), the shape and the size of the leaves, the number of main stem nodes, the final plant height, the flower aspect and the capsule size.
For pollen fertility evaluation, about 300 pollen grains per plant were analyzed. Pollen grains collected in the morning on the day of anthesis were dipped in a drop of 1.5% acetic-carmine solution on a slide for 30 minutes and were analyzed under a stereomicroscope Nikon Eclipse E800 (Nikon, Tokyo, Japan). Only fully stained and large pollen grains were scored as viable and non-aborted. The quantity of viable pollen was estimated as the percentage of stained pollen.
The self-fertility was assessed by determining the average number of seeds obtained per self-pollinated flower. The cross-fertility was assessed by counting the average number of seeds obtained per cross-pollinated flowers.
Fiber fineness analysis
For fiber fineness analysis, cotton fibers were harvested at full maturity and used for the 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. We let one or two drops of 18% NaOH solution penetrate by capillarity into the fibers. The NaOH solution swells the fibers. 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 [4] [13] [21]. 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 differences between means at P = 0.05.
Evaluation of resistance to reniform nematode
The resistance to reniform nematode (Rotylenchulus reniformis Linford & Oliveira) was assessed following the protocol of [15]. Briefly, 30-days seedlings planted in 20-cm-diameter plastic pots, were inoculated with 6000 eggs of R. reniformis. Sixty days after inoculation, the soil was gently removed, the roots carrying the nematode eggsacs were weighed and the reniform nematode eggs were then extracted and counted according to a NaOCl-Blender-Sieving-Centrifugation-Flotation method. For each plant, the number of eggs per gram root was determined. These numbers were used to assess the relative resistance of each plant compared to the susceptible control G. hirsutum by calculating the percentage of eggs per gram root for each plant considering the 100% level for the susceptible control. The scale of relative resistance used contains the following classes: 0 - 10% = highly resistant, 11 - 25% = resistant, 26 - 40% = moderately resistant, 41 - 60% = low susceptible, 61 - 100% = susceptible as control, and above 100% = very susceptible. As this technique for evaluating resistance to reniform nematode was destructive, before their evaluation the plants were grafted onto vigorous seedlings of G. hirsutum in order to keep a copy of each of them.
2.4. Molecular Analysis
DNA isolation
Total genomic DNA of G. hirsutum, G. longicalyx and two synthetic allohexaploid (G. hirsutum x G. longicalyx)2 plants were isolated from young fresh leaf tissues following the CTAB method as described by [22].
AFLP Analysis
AFLP was performed by Automated Laser FLuorescence (ALF) analysis. Three AFLP primer combinations were used: E-ACC/M-CAG, E-ACT/M-CTG and E-ACT/M-CAG. AFLP was carried out using the “AFLP Analysis System I / AFLP starter primer kit” (Invitrogen, Belgium) following the protocol proposed by Invitrogen. Briefly, genomic DNA (250 ng) was double digested with EcoR I and Mse I restriction endonucleases. The digested DNA fragments were ligated to EcoR I and Mse I adaptors with T4 DNA ligase to generate template DNA for amplification by PCR. Two consecutive PCR were performed: a pre-selective and selective PCR. In the pre-selective reaction, DNA was amplified using an AFLP pre-amp primer pair complementary to the adaptors and each having one selective nucleotide. Pre-selective PCR amplification was used as template for the selective amplification using AFLP primers, each containing three selective nucleotides. The PCR amplification products were run on 6% denaturing polyacrylamide gel using the ALF-Express (Pharmacia Biotech, Freiburg, Germany), which is an Automated Laser FLuorescence DNA sequencer. The obtained digital image of the profiles was analyzed. The scoring of bands was done as present (1) or absent (0) for AFLP marker loci and data were entered in a binary data matrix as discrete variables.
3. Results and Discussion
3.1. Mitotic Chromosome Analysis
Analysis of the mitotic metaphase plates showed 52 chromosomes for G. hirsutum, 26 chromosomes for G. longicalyx and 78 chromosomes for the putative (G. hirsutum × G. longicalyx)2 hexaploid hybrid (Figure 1). This number of 78 chromosomes proves the hexaploid status of the material studied because it is in
![]()
Figure 1. Chromosomal configuration at mitotic metaphase: (a) 52 chromosomes of G. hirsutum (×1500); (b) 26 chromosomes of G. longicalyx (×1000); (c) 78 chromosomes of the (G. hirsutum × G. longicalyx)2 hexaploid hybrid (×1500).
agreement with the number of chromosomes expected for a synthetic hexaploid cotton (2n = 6x = 78) resulting from the doubling of the chromosomes of a triploid issued from the cross between a tetraploid cotton (G. hirsutum, 2n = 4x = 52) and a diploid cotton (G. longicalyx, 2n = 2x = 26). This result confirms that artificial somatic allopolyploidization can be successfully accomplished in vivo through the use of antimitotic reagents, i.e. metaphase inhibitors such as colchicine [23] [24]. This reagent causes the depolymerization of the microtubular cytoskeleton in the early phases of metaphase, blocking the separation of chromosomes in mitoses, consequently leading to polyploidization of the cells [25].
3.2. Seed Aspect, Germination Rate and Seedling Abnormalities Analysis
The seeds of the (G. hirsutum × G. longicalyx)2 hexaploid hybrid had normal appearance and shape but they were all larger than the seeds of the parental species (Figure 2). This result is in accordance with [24] and [26] who reported that polyploids have usually larger seeds than parental species. It is probably a direct consequence of large cell size in polyploids [27], since genome duplication increases cell volume by increasing genome size [28]. This is in line with the expectation that the sizes of seed are a function of cell size, which is larger in polyploids [29].
In the germination test, the seeds of the hexaploid hybrid showed the relative lowest germination rate with 38 germinated seeds on 44 (86.36%) compared to its parental species G. hirsutum and G. longicalyx, which respectively presented 100% and 96.67% (29 germinated seeds on 30) germination rates. The germination rate gives an estimate of the viability of the seeds. The result obtained suggests a problem of viability of about 13% of the seeds produced by the allohexaploid hybrid. [24] working on synthetic polyploidy in Hylocereus monacanthus also reported problems of seed viability. The genome of newly formed polyploid plants usually undergoes extensive genetic and epigenetic changes which can alter gene expression and generate physiological changes that can affect seed viability [28] [30] [31].
![]()
(a) (b) (c)
Figure 2. Seeds of the hexaploid hybrid and its parental species: (a) smaller seed of G. longicalx; (b) intermediate seeds size of G. hirsutum; (c) larger seeds of the hexaploid (G. hirsutum × G. longicalyx)2.
Some of the hexaploid seedlings obtained from the 38 germinated seeds, presented some abnormalities at cotyledon leaves stage (Figure 3). Three types of abnormalities were observed: i) cotyledonary leaves welded on the petioles (Figure 3(a), 4 seedlings), ii) cotyledonary leaves welded on the entire length of the leaves (petioles and blades) giving the impression of a unique leaf (Figure 3(b) and Figure 3(c), 3 seedlings), iii) and seedlings with a normal cotyledonary leaf associated with a progressive necrosis of the other cotyledonary leaf (Figure 3(d), 3 seedlings). This necrosis consisted in a prolonged chlorosis of the cotyledonary leaf. It leaded to the premature death of the concerned leaves but not of the plant. However, [32] reported such an anomaly in polyploid wheat, where it ultimately caused the death of the whole plant in certain wheat hybrids. All abnormalities observed in the hexaploid seedlings were not observed in the parental G. hirsutum and G. longicalyx species. Out of 38 germinated hexaploid seeds, 10 seedlings showed one or other of the anomalies observed, i.e. 26.31% malformed seedlings. According to [33], such abnormalities could result in lethal or semi-lethal condition that imposes a great barrier when trying to transfer desirable traits between species. Actually, after polyploidization the architecture of the cell is modified, the cell must adapt to the new nuclear DNA content and deal with changes in the homology of the chromosomes, gene expression and epigenetics [25]. Changes in gene expression and physiological processes due to genome duplications can generate anomalies that can be deleterious for the development of the polyploid individual [25]. Fortunately, only 26.31% of the hexaploid plants in the present study presented anomalies, which were not necessarily fatal for the plants. Indeed, the malformations observed seemed to be minor and they had no detrimental consequences on the development of the plants except for a seedling with the abnormality of the fully welded cotyledonary leaves. This seedling remained at cotyledon leaves stage for a long time until death.
3.3. Analysis of Morphological Observations of the Plants
The height of the hexaploid plants varied from 152 to 236 cm with an average of
![]()
Figure 3. Pictures showing abnormalities observed in some seedlings of the hexaploid at cotyledon leaves stage: (a) a hexaploid seedling showing cotyledonary leaves welded at the level of the petioles; (b), (c) a hexaploid seedling having cotyledonary leaves welded on their entire length (petiole + blade); (d) a hexaploid seedling with a normal cotyledonary leaf (red arrow) and a cotyledonary leaf showing progressive necrosis (white arrow); (e) a hexaploid seedling with normal cotyledonary leaves.
192 cm, while the height of G. hirsutum and G. longicalyx were on average 82 cm and superior to 300 cm respectively. The hexaploid had 12 to 34 nodes (25 on average) on the main stem. The numbers of nodes on the main stem were on average 11 and 33 for G. hirsutum and G. longicalyx respectively. The hexaploid plants mostly carried tri-lobated leaves. The size of these leaves varied from 7.4 cm to 14.9 cm with an average of 10.54 cm. G. hirsutum presents the largest leaves (23.5 cm on average) with 3 - 5 lobes, while G. longicalyx had the smallest leave (6 cm on average) with no lobe (simple leaves). The flowers of the hexaploid plants were a bit smaller than those of G. hirsutum but their general aspect was similar. G. longicalyx had the smallest flowers. The hexaploid produced relative smaller capsule than G. hirstum but bigger than G. longicalyx. These results showed that most of the morphological characteristics exhibited by the synthetic allohexaploid hybrid were intermediate between G. hirsutum and G. longicalyx. This hexaploid hybrid did not generate transgressive morphological characters (characters which show values beyond the range of parental species). [20] studying the synthetic allohexaploid hybrid (G. hirsutum x G. anomalum)² also found that most of morphological characteristics of the hexaploid plants were intermediate between the two parental species. The intermediate morphology of hybrids compared to their parental species is a characteristic observed in several hybrids, since these characters are generally under polygenic control with simple additive effects [34] [35].
3.4. Fertility Analysis
The mean proportion of stainable pollen grains (pollen fertility) of the hexaploid plants was 83% while the pollen fertility of G. hirsutum and G. longicalyx approached 100%. The pollen fertility level of the hexaploid is quite good, even if it is relatively lower than that of the parental species. By selfing, the hexaploid gave a mean of 4.39 seeds per pollinated flowers while G. hirsutum and G. longicalyx produced respectively 34 and 6 seeds per capsule on average (Table 1). These results show the restoration of fertility at the hexaploid level by the doubling of the chromosomes of the sterile triploid hybrid, even if the self-fertility of the parental species was higher. The results of cross-pollinations between the hexaploid and G. hirsutum (Table 1) gave practically no seed per capsule, i.e 0.018 and 0.3 seed per capsule when the hexaploid was used as female and male respectively. This very low success rate of cross-pollination, despite the good level of pollen fertilities (for both hexaploid hybrid and G. hirsutum) is probably due to the presence of incompatibility barriers between the hexaploid hybrid and G. hirsutum [36] [37].
3.5. Expression of Fiber Fineness and Resistance to Reniform Nematode Traits
The results of the fiber fineness analysis are presented in Table 2. The hexaploid (G. hirsutum × G. longiclayx)2 showed a mean value of ribbon width of 12.53 µm. G. longicalyx had the finest fibers with 5.94 µm of ribbon width against 17.765 µm for G.hirsutum. These results show the interesting fiber fineness of
![]()
Table 1. Selfing of the hexaploid (G. hirsutum × G. longicalyx)2 and its parental species, and backcrossing of the hexaploid to G. hirsutum.
![]()
Table 2. Ribbon width of the hexaploid (G. hirsutum × G. longicalyx)2 and its parental species.
G. longicalyx as reported by several authors [4] [12] [13], and its remarkable potential to improve the fiber fineness of G. hirsutum, with regard to the expression of this interesting trait in the hexaploid (G. hirsutum × G. longiclayx)2.
The data concerning the evaluation of the resistance to the reniform nematode are presented in Table 3. The parental species G. hirsutum presented the greatest number of eggs per gram root (205.8 eggs/g root) while the number of eggs per gram root of G. longicalyx and the hexaploid were very low, 0 and 3.8 eggs per gram root respectively. Compared to the susceptible G. hirsutum species, G. longicalyx (0% eggs/g root) and the hexaploid (1.96% eggs/g root) were very resistant. This finding confirms the high resistance to the reniform nematode of the wild African cotton species G. longicalyx [14] [15] and shows the inheritance and expression of this interesting trait in the hexaploid.
3.6. Molecular Analysis with AFLP Markers
The results of the AFLP analysis are presented in Table 4. The AFLP electrophoretic profile obtained (Figure 4) revealed a total of 143 loci distributed in 52, 45 and 46 loci for respectively the primer pairs E-ACC/M-CAG, E-ACT/M-CTG and E-ACT/M-CAG. Of the 143 loci revealed, 57 were common to G. hirsutum and G. longicalyx and therefore monomorphic between these two parental species. All of these 57 monomorphic loci were detected in the hexaploid. In addition to the monomorphic loci, 44 loci were specific to G. hirsutum and 42 others were specific to G. longicalyx, which makes 86 polymorphic loci in all, representing a percentage of polymorphism of 60.14%. Thus, the AFLP markers
![]()
Table 3. Results of the assessment of the resistance to the reniform nematode of the hexaploid (G. hirsutum × G. longicalyx)2 and its parental species.
aS: susceptible; bHR: highly resistant.
![]()
Table 4. Results of the analysis of the AFLP electrophoretic profile of the hexaploid (G. hirsutum × G. longicalyx)2 and its parental species.
![]()
Figure 4. AFLP electrophoretic profile of the hexaploid (G. hirsutum × G. longicalyx)2 and its parental species using the three primer pairs E-ACC/M-CAG, E-ACT/M-CTG and E-ACT/M-CAG: 1 = hexaploid#1; 2 = G. longicalyx; 3 = G. hirsutum; 4 = hexaploid#2. The red arrows show some G. longicalyx specific bands; the blue arrows show some G. hirsutum specific bands; the black arrows show specific bands G. longicalyx that are missing in the hexaploid hybrid.
efficiently discerned differences between the two parental species and distinguished them distinctly from each other. This is consistent with [38] who also highlighted such differences between G. hirsutum and G. longicalyx, using SNP markers.
All 44 specific loci of G. hirsutum and 38 of 42 specific loci of G. longicalyx were found in the hexaploid. In total, all the loci revealed in the hexaploid come from G. hirsutum and G. longicalyx, which confirms the hybrid status of the hexaploid as indicated by cytological analysis. However, 4 of the 42 specific loci of G. longicalyx were missing in the hexaploid (Figure 4). [20] studying the synthetic allohexaploid (G. hirsutum × G. anomalum)2 also found missing SSR alleles of the wild species G. anomalum in the hexaploid. Generally, the main reasons used to explain parental band missings in hybrids are loss of chromosomes or rearrangements of chromosomes. In the present study, the loss of chromosomes cannot explain the missing AFLP loci because the synthetic allohexaploid had the expected number of chromosomes (2n = 78). The explanation of chromosomal rearrangements is also unlikely since the two different hexaploid plants used for the molecular analysis had exactly the same missing loci. It is unlikely that there will be exactly the same recombination in two different plants. A slight genetic differentiation between the G. longicalyx plant used to develop the hexaploid hybrid and the G. longicalyx plant used in this molecular analysis could likely be the reason for the missing loci in the hexaploid hybrid. Indeed, intraspecific differentiation exists in wild cotton species. For example, [39] using AFLP data, reported intraspecific variation (in terms of percentage of polymorphic fragments) in four species of cotton (G. aridum, G. laxum, G. lobatum, and G. schwendimanii). Anyway, phenotypic analysis showed that the current synthetic allohexaploid hybrid had the characteristics of interest (resistance to reniform nematode and fineness of the fibers) sought in G. longicalyx. Therefore, this hexaploid material is perfectly suited for use in a cotton improvement program.
4. Conclusion
The present study demonstrated the hybridity and hexaploid status of the genotype studied using cytological and molecular marker methods. Moreover, this synthetic allohexaploid hybrid exhibited the useful traits of the African wild diploid species G. longicalyx with regard to the fineness of the fibers and the resistance to reniform nematode. This synthetic allohexaploid hybrid represents very interesting genetic stocks that can be used as a bridge for the transfer of useful agronomic traits from wild species to upland cotton varieties.