**3. Results**

### *3.1. Ta-3A1 Was One of the Top Accumulated TRs*

A database of all the raw and filtered nonredundant TRs predicted for the IWGSC RefSeq v1.0 assembly by TRF is available at http://mcgb.uestc.edu.cn/tr [29]. The nonredundant TRs were classified into three classes based on their lengths of period distance (PD), specifically <20 bp, 20–60 bp, and >60 bp. The distribution of different PD sizes on individual chromosomes indicated that the copy numbers of TRs of <20 bp were more abundant than those with 20–60 bp and >60 bp (Figure 1). Interestingly, the TRs of 20–60 bp were highly accumulated within short regions of chromosomes 3A, 5A, 7A, 5B, and 5D. In order to determine the composition of the highly accumulated mini-satellite repeats in the five chromosomes, the TRs with a pattern size ≥20 bp, copy number ≥50, and percent match ≥80 were grouped. We found that a TR array with a length of 44 bp predominated (the consensus sequence was AATAATTTTACACTAGAGTTGAACTAGCTCTA TAAGCTAGTTCA). This predominating array was named TR-3A1 and consisted of 280 nonredundant arrays covering a length of 2,630,397 bp. Across all wheat chromosomes, the estimated total predicted copy number of TR-3A1 was 59,801, which was mainly distributed on the above five chromosomes with copy numbers of 14,987, 11,083, 7828, 8444, and 28,964 on chromosomes 3A, 5A, 7A, 5B, and 5D, respectively (Table 1).

**Figure 1.** The copy number and distribution of tandem repeats in wheat chromosomes 3A (**a**), 5A (**b**), 5B (**c**), 7A (**d**), and 5D (**e**). The copy number was predicted for wheat Chinese Spring genomic sequences v 1.0 at a web server at http://mcgb.uestc.edu.cn/BD2SC with default parameters.

**Table 1.** The predicted copy numbers of Ta-3A1 on sequenced *Triticum urartu*, *Aegilops tauschii*, *T. dicoccoides*, and *T. aestivum* genomes.


The Ta-3A1 sequence was submitted as a query into the website B2DSC using default parameters (pident = 85 and qcovhsp = 80). An example of the entire genomic distribution, detailed chromosomal locations, as well as the sequence directional analysis of Ta-3A1 according to the website B2DSC appears in Figure 2. The predicted copy number and the estimated physical positions of Ta-3A1 (Figure 2a) resembled the 20–60 bp TR localizations in Figure 1, which suggested that Ta-3A1 was actually the principal predicted mini-satellite in wheat genome compared to the results of TR enrichment studies by Lang et al. [29]. The physical organization of Ta-3A1 on chromosome 5DS regions at 146–147 Mb

(Figure 2b) and the head-to-head repeats of the 72.19 Mb region in chromosome 7A (Figure 2c) were indicated in detail.

**Figure 2.** The predicted physical location of Ta-3A1 in wheat genome. The overview of Ta-3A1 on five chromosomes (**a**), copy number in a region of 5D (**b**), and an array of sequences on 7A (**c**) were indicated, respectively. The arrows represent the mini-satellite array, and green and red arrows show the minute and plus strand, respectively.

Similarly, by using the B2DSC website, the copy numbers and physical locations of Ta-3A1 were also predicted for the sequenced genomes of wheat's ancestral species, including the A genome of *T. urartu* [26], the D genome [27], and AABB genomes [28] (Table 1). Based on the comparison of overall genomic copy numbers of Ta-3A1 among those of the A, B, and D genomes, we found that the copy numbers on each of the corresponding chromosomes differed significantly, and the copy numbers increased from diploid to tetraploid to hexaploid wheat. The most abundant sites of Ta-3A1 were located on 5D of wheat, which was significantly higher than those on 5D of *Ae. tauschii*. It is likely that the copy numbers of Ta-3A1 largely increased during the polyploidization of wheat.

### *3.2. Sequence Variability of Ta-3A1 on Different Chromosomal Regions*

The monomer sequences within each of the Ta-3A1 arrays spanning over 200 copies were aligned according to their diagnostic physical positions. A total of 17 clustered physical regions among the 3A, 5A, 7A, 5B, and 5D chromosomes, which represented from 17% to 43% of the nonredundant sequences, were analyzed for sequence variation rates (Table 2). Interestingly, despite their different degrees of abundance, mean *π* values of Ta-3A1 monomers for each of the chromosome regions were roughly similar (from 9.3% in 5B to 19.4% in 3A). The higher nucleotide diversity values (*π*) on different chromosomes and different regions may sugges<sup>t</sup> that the sequence variation and rearrangements of Ta-3A1 were due to recombination events, which may be an important force generating new monomers in chromosomes 3A, 5D, and 7A.


**Table 2.** Nucleotide polymorphism of the repeat features of Ta-3A1 in wheat genome.

Number of monomeric repeats sequenced ( *N*), number of variable sites (*S*), nucleotide diversities (*π*), and haplotype diversity (Hd). Standard deviation (SD), The estimates of the number of segregating sites and the average number of nucleotide differences correlated under the neutral model (Tajima's D). \* showed significant difference at *p* < 0.05.

In order to reveal the evolutionary aspects of the overall Ta-3A1 monomer sequence variation, a phylogenetic tree of the representative Ta-3A1 satDNA in all the chromosome regions was constructed. A total of 216 representative Ta-3A1 monomer sequences predicted from the wheat genome were used for constructing the phylogenetic tree (Figure S2 and Figure 3). The results suggested that the monomers from chromosome 5D were separated first, the sequences of 5A, 5B, and 3A were categorized as the second subgroup, and the 7A sequences appeared in the third group. The sequences of chromosome 3A were mainly clustered in groups 6, 8, and 9. Combining the phylogenetic tree with the physical locations of the Ta-3A1 sequences (Figure 3), we found that the Ta-3A1 sequences from 5D distributed in most groups: However, the sequences in 5B were mainly concentrated on groups 3 and 4. This suggests that the Ta-3A1 on 5D evolved several times during sequence expansion and became more active than that of the 5B chromosome.

**Figure 3.** Phylogenetic tree of Ta-3A1 repeats and their physical location on five wheat chromosomes.

### *3.3. FISH of Oligo-3A1 in Wheat*

Based on the sequences of the Ta-3A1, we produced a labeled oligo probe (Oligo-3A1) for revealing the chromosome organization of Ta-3A1 by ND-FISH analysis. Mitotic metaphase chromosomes from bread wheat, durum wheat, and some related species were prepared to study the hybridization patterns of Oligo-3A1. Sequential ND-FISH with Oligo-pTa535 and Oligo-pSc119.2 was performed for the identification of individual chromosomes of wheat [36]. As shown in Figure 4, the ND-FISH hybridization patterns of Oligo-3A1 in Chinese Spring wheat displayed strong signals on pericentromeric regions on 5DS, intercalaric regions of 5AL, 3AL, and 5BL, and subtelomeric regions of both arms on 7A. The chromosomal distribution of the FISH signals were consistent with the predicted copy numbers of each chromosome (Table 1).

**Figure 4.** Nondenaturing fluorescence in situ hybridization (ND-FISH) patterns of Ta-3A1 repeats in wheat Chinese Spring: The probes Oligo-pSc119.2 (green) and pTa535 (red) (**a**), and sequential Oligo-3A1 (green) (**b**), were used for ND-FISH. The copy numbers predicted by web server page B2DSC and the copy number of Ta-3A1 at the specific region of each 1-Mb scale were illustrated in karyotypes of chromosomes (**c**). Bar indicated 10 μm.

### *3.4. FISH of Oligo-3A1 in Representative Triticeae Species*

In order to shed light on the genetic divergence of Ta-3A1 on representing wheat ancestral species, the chromosome preparations from *T. monococcum* (A genome), *Ae. longissima*, *Ae. searsii*, *Ae. speltoides* (S genomes), and *Ae. tauschii* (D genome) accessions were carried out by ND-FISH analysis (Figure 5). The results showed that the signal strengths varied from faint to strong on 5Am and 3Am to 7Am in *T. monococcum* (Figure 5a,d). Strong hybridization signals appeared on long arms of 5S in *Ae. speltoides* (Figure 5b,e), but conversely on the short arms of 5S<sup>l</sup> in *Ae. longissima* (Figure 5c,f) and 5Ss in *Ae. searsii* (Figure 5h,k). Relatively strong signals of Oligo-3A1 were observed on 5DS of *Ae. tauschii* (Figure 5j,m). Therefore, the strength of signals and the localization of chromosomes varied among the wheat A, B, and D ancestral genomes.

A total of each of 12 *T. durum*, six *T. dicoccoides*, three *T. carthlicum*, and two *T. abyssinicum* accessions were studied by ND-FISH using Oligo-3A1 (Figure 5m–r, Table 3). We observed that the signals on chromosomes 3A and 7A were stable, but the strength of signals on chromosomes 5A and 5B varied extensively. Nine accessions showed strong signals on 5BL and weak signals on 5AL, and the converse was noted in two accessions. The remaining accessions had equally strong hybridization signals on 5AL and 5BL. It is interesting to note that additional discrete signals were observed on chromosome 2BL of five *T. durum* lines, which were derived from International Maize and Wheat Improvement Center (CIMMYT) (Figure 5m,p). The highly polymorphic FISH patterns of Oligo-3A1 were observed among tetraploid wheat accessions.

**Figure 5.** Hybridization of the ND-FISH probes on chromosomes of *T. monococcum* (**<sup>a</sup>**,**d**), *Ae. speltoides* (**b**,**<sup>e</sup>**), *Ae. longissima* (**<sup>c</sup>**,**f**), *Ae. searsii* (**g**,**j**), *Secale cereale* (**h**, **k**), tetraploid *Ae. tauschii* (**i**,**l**), *T. durum* (**<sup>m</sup>**,**p**), *T. carthlicum* (**<sup>n</sup>**,**q**), and *T. dicoccoides* (**<sup>o</sup>**,**<sup>r</sup>**). The probes Oligo-pSc119.2 (green) and pTa535 (red) (**<sup>a</sup>**–**c**,**g**–**i**,**o**–**p**), and sequential Oligo-3A1 (green) (**d**–**f**,**j**–**l**,**p**–**<sup>r</sup>**) were used, respectively.


**Table 3.** The differential hybridization patterns of Ta-3A1 among tetraploid wheat.

The strength of Ta-3A1 hybridization signals indicated by + (weak), ++ (medium), +++ (strong).

Additionally, ND-FISH with Oligo-3A1 was also conducted on chromosome preparations of other related wild diploid or polyploid species and wheat-alien species amphiploids. The identification of individual chromosomes from *Secale* [37], *Dasypyrum* [43], and *Thinopyrum* [44] was achieved by Oligo-pSc119.2 and Oligo-pTa535. Barley chromosomes were completely devoid of any hybridization signals of Oligo-3A1, as were *D. breviaristatum* chromosomes in the wheat—*D. breviaristatum* amphiploid. The hybridization signals of Oligo-3A1 on rye chromosome 5R (Figure 5) and in triticale (Figure 6a,b) were relatively strong, while the signals on *D. villosum* 5V were weak in the durum wheat—*D. villosum* amphiploid (Figure 6e,f). Moreover, only two chromosomes of *Th. ponticum* (2*n* = 10*x* = 70) appeared to carry Oligo-3A1 signals on their long arms (Figure 6c,d), and only two pairs of chromosomes with Oligo-3A1 sites (one in short arms and the other in long arms) were observed for *Th. intermedium* (2*n* = 6*x* = 42) (Figure 6g,h). The distribution of Ta-3A1 was polymorphic for presence/absence in different genomes or species, which suggests that the major loci were probably lost during the evolution of this polyploidy species.

**Figure 6.** ND-FISH on chromosomes of hexaploid triticale (**<sup>a</sup>**,**b**), *Thinopyrum ponticum* (**<sup>c</sup>**,**d**), *T. durum–Dasypyrum villosum* amphiploid (**<sup>e</sup>**,**f**), and *Th. intermedium* (**g**,**h**). The probes Oligo-pSc119.2 and pTa535 (**<sup>a</sup>**,**c**,**e**,**g**) and sequential Oligo-3A1 (**b**,**d**,**f**,**h**) were used, respectively.
