Analyses on Progenitor Donors of the Cultivated Allotetraploid Cottons Revealed by GISH

GISH (Genomic in situ hybridization) of the mitotic metaphase chromosomes of two cultivated tetraploid cotton (AD)1 (G. hirsutum) and (AD)2 (G. barbadense) with all 3 diploid A cotton gDNA(genomic DNA) as probes, blocking with ssDNA(salmon sperm DNA) respectively. The hybridization signals were dected distribute in the A sub-genome chromosomes of (AD)1 and (AD)2, besides, three pairs of crimson signals were also detected only with the A1-a gDNA probe which named GISH-NORs. GISH of (AD)1 and (AD)2 with all 13 diploid D cotton gDNA as probes, blocking with ssDNA respectively, except the D6 (G.gossypiodies) gDNA probe generated the hybridization signals in all the chromosomes of (AD)1 and (AD)2, the other 12 diploid D gDNA probes only generated the signals on the D sub-genome chromosomes of (AD)1 and (AD)2, the D6 gDNA probe was very specifical.And three pairs of strong GISH-NORs were detected with all 13 diploid D genome species gDNA probes, the intensity of their GISH-NORs were much brighter than the A1-a gDNA probe. These results visually confirmed the amphidiploid origin of the allotetraploid cotton species. DA (distinguishing ability ) values of each gDNA probe generated were calculated basing on the above GISH rsults.It showed that the DA value of A1-a gDNA probe was the biggest in all 3 diploid A genomes both in (AD)1 and (AD)2 GISH, and this indicated that A1-a genome was most likely to be the A sub-genome progenitor donor of (AD)1 and (AD)2, while the D3-d (G. davidsonii ) and D5 (G. raimondii) genome species were most likely to be the D sub-genome progenitor donor of (AD) 1, and (AD) 2 respectively. And this further confirmed that tetraploid cottons are polyphyletic.

The Fluorescence in situ hybridization (FISH) technology was introduced to plant research in 1985 (Rayburn,1985), which has applied to various aspects such as repeat sequences positioning, chromosome identification, cell genetic map construction, the origin of polyploid genome evolution and phylogenetic relationships and so on (Snowdon et al., 1997 ;Tang et al., 2000;Jiang and Gill,2006;Wang k et al.,2008;Nemeth et al., 2015;Melo et al., 2015).

GISH of (AD) 1 and (AD) 2 with all 3 diploid A cotton gDNA as probes respctively
GISH of the somatic metaphase chromosomes of (AD) 1 and (AD) 2 both with all 3 diploid A (A 1 , A 2 and A 1-a ) gDNA as probes respectively, blocking with ssDNA(salmon sperm DNA).The results showed that the red hybridization signals were mainly distributed on the longer 13 pairs of A sub-genome chromosomes of (AD) 1 and (AD) 2 (As shown in Figure 1), and this visually confirmed the allodidiploid origin of the tetraploid cottons (Wendel et al., 2002). Besides, three pairs of crimson signals were detected only with the A 1-a gDNA probe both in (AD) 1 and (AD) 2 (Figure 1-c, f ), which were named "GISH-NORs" (Liu et al., 2005) , of which one in the A sub-genome chromosomes (green arrows), and two in the D sub-genome chromosomes (white arrows), and it significantly different from the A 1 and A 2 gDNA probes, the specific distributions about GISH-NORs were shown in Table 1.
Figure 1 GISH of mitotic metaphase chromosomes of two cultivated allotetraploid cotton species G.hirsutum (AD) 1 and G.barbadense (AD) 2 with all 3 A diploid cotton gDNA probes respectively (all probes were labeled with digoxigenin, and red fluorescence signals were observed on A sub-genome chromosomes in each image). Fig.1-a To 1-c: GISH of G.hirsutum var. zhong 16 with diploid A genome species G.herbaceum var. hongxing (A 1 ), G.arboreum var. shixiya 1 (A 2 ), G.herbaceum wild species arfrium (A 1-a ) gDNA probe rsepectively, blocking with ssDNA (In Fig.1-c, six arrows showed GISH-NORs, of which two green arrows denote the GISH-NORs in A sub-genome chromosomes and other four white arrows denote the GISH-NORs in D sub-genome chromosomes of (AD) 1 ). Fig.1-d To 1-f: GISH of G.barbadense var. Xinhai 7 also with diploid A genome species A 1 , A 2 , and A 1-a gDNA probe rsepectively, blocking with ssDNA (In Fig.1-f, six arrows also showed the signals of GISH-NOR, of which two green arrows denote the GISH-NORs in A sub-genome chromosomes and four white arrows denote in D sub-genome). 1.2 GISH of (AD) 1 and (AD) 2 with all 13 diploid D cotton gDNA as probes respectively Except the D 6 (G.gossypiodies) gDNA probe, GISH of the somatic metaphase chromosomes of (AD) 1 and (AD) 2 both with other 12 diploid D gDNA as probe respectively, blocking with ssDNA (As shown in Figure 2 and 3), the red fluorescence signals were observed on the short D sub-genome chromosomes of (AD) 1 and (AD) 2 . In addition, three pairs of major GISH-NORs were detected, of which one in the longer A sub-genome chromosomes (green arrows showed in each image), and other two in the D sub-genome chromosomes (white arrows in each image),and the specific distribution about GISH-NORs were shown in Table 1. However, there were some differences on the fluorescence signal strength generated by each D genome species probe. GISH of (AD) 1 and (AD) 2 with D 6 gDNA probe respectively, the red fluorescence signals were not only distribute on the D sub-genome chromosomes but also on A sub-genomes (As shown in Figure 2-h and Figure 3-h), and we could not distinguish the A and D sub-genome of (AD) 1 and (AD) 2 , and three pairs of major GISH-NORs were also observed. This showed a great difference between the D 6 genome and the other 12 D genomes, and the result illustrated that diploid D 6 genome may contain plenty of A genome repeat sequences, it is a very special genome species in the diploid D genomes.

DA value analysis based on GISH of (AD) 1 and (AD) 2 with diploid A, D genome species probes
DA (distinguishing ability) value reflect the genetic relationship between the diploid A or D genomes and the tetraploid genomes (Markova et al., 2007). DA values (As shown in Table 2) generated by A 1 , A 2 and A 1-a gDNA probes to (AD) 1 was 0.361, 0.358 and 0.369 respectively. It showed the A 1-a gDNA possess the strongest ability to recognize the A sub-genome chromosomes of (AD) 1 , while the A 1 and A 2 gDNA possess same ability to recognize the A sub-genome chromosomes of (AD) 1 , that's to say the A 1-a genome had closer genetic relationship with the A sub-genome of (AD) 1 . And DA values generated by A 1 , A 2 and A 1-a gDNA probes to (AD) 2 were 0.343, 0.346 and 0.352 respectively. the results also showed the A 1-a gDNA possess the strongest ability to recognize the A sub-genome chromosomes of (AD) 2, therefore, the A 1-a genome also had closer genetic relationship with the A  DA values of (AD) 1 generated by all 13 diploid D gDNA probes respectively were shown in Table 2. Because of the hybridization signals generated by D 6 gDNA probe were distributed on both A and D sub-genome chromosomes of (AD) 1 , therefore the DA value generated by D 6 gDNA probe(DA value was 0.402) was much higher than the other 12 D gDNA probes in (AD) 1 . Except the D 6 gDNA probe, the highest DA value was 0.286 generated by D 3-d gDNA probe, and then followed by D 4 , D 1 , D 5, D 11 , D 2-1 , D 2-2 , D 8 , D 3-k , D 9 , D 10 and D 7 probes. This showed that the D 3-d gDNA probe possess the strongest ability to recognize the D sub-genome chromosomes Molecular Plant Breeding 2016, Vol.7, No.14, 1-10 http://mpb.biopublisher.ca of (AD) 1 , while the other D gDNA probes show weaker ability to recognize the D sub-genome chromosomes of (AD) 1 , therefore the D 3-d genome had closer genetic relationship with the D sub-genome of (AD) 1 . DA values of (AD) 2 generated by all 13diploid D gDNA probes also be shown in Table 2. Except the D 6 gDNA probe(DA value was 0.387), the highest DA value generated by D 5 gDNA probe was 0.263, and then followed by D 3-d , D 1 , D 4 , D 2-1 , D 2-2 , D 11 , D 8 , D 3-k , D 10 , D 9 and D 7 probes. which was also higher than those of the other 12 D gDNA probes. The results showed that the D 5 gDNA possess the strongest ability to recognize the D sub-genome Molecular Plant Breeding 2016, Vol.7, No.14, 1-10 http://mpb.biopublisher.ca 6 chromosomes of (AD) 2 , while the other 12 D gDNA show weaker ability to recognize the D sub-genome chromosomes of (AD) 2 , so the D 5 genome species had closer genetic relationship with the D sub-genome of (AD) 2 .

The A sub-genome progenitor of (AD) 1 and (AD) 2
Since the discovery that allotetraploid Gossypium genomes contain both A and D genomes, investigators had attempted to look for which one of the modern diploid A and D genome species can be best served as the progenitor genome donors of allopolyploid cottons.
Which one of the diploid A genome species was the really donor of the A sub-genome of allopolyploid cottons? Many previous researches suggested that G. herbaceum (A 1 ) was the donor or the similar ancestors of the allopolyploid A sub-genome (Beasley, 1940;Gerstel, 1953;Phillips, 1963;1964). However, the subsequent research had shown that there existed much differences between G. herbaceum (A 1 ) and the A sub-genome of allopolyploid cotton, whether in the chromosome or molecular level (Wendel, 1989;Albert, 1992； Cronn, et al., 1996). And cell cytogenetic and comparative mapping research also revealed that there existed at least two large translocations between their genomes (Gerstel, 1953；Small, et al., 1998；Liu and Wendel, 2001. Branch taxonomy analyses of the molecular sequences had showed that G. herbaceum(A 1 ) was not the actual progenitor of the A sub-genome of allotetraploid cottons (Endrizzi, et al., 1985;Wendel and Cronn, 2002). In the evolutionary process of allotetraploid cottons, G. herbaceum(A 1 ) and G. arboretum(A 2 ) were the phylogenetically sisters between each other and hence were genealogical equidistant to the A sub-genome of the allotetraploid cottons (Cronn et al., 1996;Liu and Wendel, 2001;Wendel, 1989;Wendel and Albert, 1992).
In our experiment, GISH of (AD) 1 and (AD) 2 both with all 3 A genome gDNA as probes, 13 pairs of A sub-genome chromosomes were painted with red fluorescence signals, and the DA value was very similar between A 1 and A 2 gDNA probe, there was no significant difference between A 1 and A 2 both in GISH of (AD) 1 and (AD) 2 , while the DA value of A 1-a gDNA probe was higher than A 1 and A 2 gDNA probe, the A 1-a gDNA had the strongest ability to recognize the A sub-genome chromosomes of (AD) 1 and (AD) 2 , so we considered that A 1-a genome Molecular Plant Breeding 2016, Vol.7, No.14, 1-10 http://mpb.biopublisher.ca 7 species was most likely to be the A sub-genome progenitor donor of allotetraploid cotton (AD) 1 and (AD) 2 .
The A 1-a genome had been divided into the variant of A 1 genome in the taxonomy (Stewart, 1987). Many researches always used the varieties of A 1 and A 2 genome species to study the origination and evolution of the A sub-genome of allotetraploid cottons, but they seldom and nearly didn't use the A 1-a genome as the research materials, and even some researchers took the A 1-a and A 1 genome to lump together (Stewart, 1987;Wendel,1989;Wendel and Albert, 1992;Wendel et al.,1994Wendel et al., , 1995Wendel and Wessler,2000;Wendel and Cronn,2002;Wang et al., 2004). In this experiment, when using A 1-a gDNA as probe we found that the GISH result was obviously distinguished from the A 1 gDNA probe, and Wang et al(1995) had also found the karyotype parameters exist significant differences between A 1 and A 1-a genome. The A 1-a genome made a strong independence from A 1 genome, and we suggested that A 1-a genome should be given the "species" level in the classification of gossypium, that's means the A 1-a genome possessed the same status with A 1 and A 2 genome.

The D sub-genome progenitor of (AD) 1 and (AD) 2
Since the discovery that allopolyploid Gossypium species contain two genomes whose progenitors presently occur in different hemispheres, investigators had attempted to provide pieces to the puzzle of the polyploid origin. A diverse array of tools had been used in an effort to examine this issue, from early study methods by comparative morphology, cytology, cytogenetic, comparative phytochemistry, and protein electrophoretic methods to modern phylogenetic investigations using DNA sequencing of homologous genes. Several investigators proposed that allopolyploid cottons formed more than once, they suggested the best models of the potential D subgenome ancestral donor of allopolyploid cottons were D 1 (G. thurberii), D 3-d (G. davidsonii), D 3-k (G. klotzschianum) and D 5 (G. raimondii), so they considered the allopolyploid cottons are polyphyletic (Kammacher, 1960;Sherwin, 1970;Johnson, 1975;Umbeck, 1985;Stewart, 1987;Da and Bertrand,1995). However, Most authors proposed that allopolyploid cottons formed only once, and that D 5 was more similar to the D subgenome of allotetraploid cottons than other D genome diploids, they considered the allopolyploid cottons are monophyletic (Phillips,1963(Phillips, , 1964(Phillips, , 1966Endrizzi et al., 1985;Wendel, 1989;Cronn et al.,1996;Seelanan et al., 1997;Wendel, 1999, 2000;Wendel and Albert et al., 1992 ;Wendel et al., 1995;Wendel and Cronn, 2002;Admas, 2003).
We compared the GISH signals and DA values of (AD) 1 and (AD) 2 chromosomes generated by different diploid D genome species. The DA values of D 6 (G. gossypioides) were much higher than those of other D genome species both in (AD) 1 and (AD) 2 , but the signals were distributed on both A and D subgenome chromosomes of (AD) 1 and (AD) 2 , and the signal intensity was much weaker than that produced by other D genome species such as D 3-d (G. davidsonii) and D 4 (G. aridum) in (AD) 1 , or D 1 (G. thurberi), D 3-d (G. davidsonii) and G. raimondii (D 5 ) in (AD) 2 . Therefore, the D 6 genome species cannot be the D subgenome progenitor donor of (AD) 1 and (AD) 2 .
The signal intensity in D 3-d probe was more stronger than any other D genome species probes in GISH of (AD) 1 , and the DA value of D 3-d probe was much higher than other D genome species too. This indicated that D 3-d genome species has the strongest ability to recognize the D subgenome chromosomes of (AD) 1 , so we suggested that D 3-d but not D 5 was the possible D subgenome progenitor donor of (AD) 1 .
And on the other hand, the signal intensity in D 5 probe was more stronger than any other D genome probes in GISH of (AD) 2 , and the DA value of D 5 probe was also much higher than other D genome probes.This indicated that D 5 genome species has the strongest ability to recognize the D subgenome chromosomes of (AD) 2 , so we thought that the D 5 is the possible D subgenome progenitor donor of (AD) 2 .
Previous GISH studies had shown that diploid D 1 and D 3-d genome species was the D sub-genome progenitor donor of (AD) 4 (G. mustelinum) and (AD) 5 (G. darwinii) respectively (Wu et al., 2010;, combined with the results of this experiment, allopolyploid cottons maybe formed with different D genome species as the progenitor donor, these results supported the hypothesis that allopolyploid cottons are polyphyletic.
Total gDNA was extracted and purificated from immature leaves using the CTAB method . The purified total gDNA was cut off with 120℃ 10 min, and examined with 0.8% agarose gel electrophoresis for their fragment size, which was generally appropriate in 300-600bp, and then tagged with markers. The gDNA probes were labeled with DIG-High-Prime labeling system (Roche Company, Germany), according to the standard operating procedures. The preparation of mitotic metaphase chromosomes which derived from the root tip cells and the procedure of GISH referenced to the method of Wang .
The hybridization signals were observed using a fluorescence microscope (Ziess Axioskop 2 plus). Images were captured by ISIS (in situ imaging system) software by adjusting their brightness and contrast. And also used this software to calculate the DA (distinguishing ability) value which was described by Markova et al (2007). And used Adobe Photoshop 7.0 software makes the plate.