*2.7. Other Triticeae Species*

Among wild *Triticeae* species, the crested wheatgrass, *Agropyron cristatum*, shows a high crossability with wheat and other species, and represents a potential source of genes controlling resistance to biotic and abiotic stresses. The homoeology relationships of the diploid *A. cristatum* (PP) chromosomes to wheat was analyzed by the FISH locations of 45 wheat full-length cDNA probes in the seven chromosomes of *A. cristatum* [85]. Chromosomes 1P, 3P and 5P seem to be collinear with wheat chromosomes, while the other four chromosomes show a collinearity distortion, such as a pericentric inversion on 4P, a paracentric inversion on 6PL, or a reciprocal translocation between 2PS and 4PL.

The genus *Leymus* includes about 30 allotetraploid species that share genomes Ns and Xm derived from *Psathyrostachys* and an unknown progenitor, respectively. Useful agronomically-important traits have been introduced in wheat, but the homoeologous relationships of the 14 *Leymus* chromosomes to wheat have not been identified. A genetic map containing 799 EST markers was constructed from fertile hybrids between both of the wild rye species *L. triticoides* and *L. cinereus*. Alignments of EST *Leymus* markers to barley EST maps evidenced the presence of a reciprocal translocation between 4NsL and 5NsL, which is absent in the Xsm genome [86].

*Thinopyrum bessarabicum* (JJ) is a diploid perennial maritime wheatgrass with genes for salt tolerance and diseases resistance that can be used in wheat breeding. The physical location of 1150 SNP markers into 36 segments of the seven J chromosomes allowed us to construct a high-density molecular markers map. Synteny analysis with wheat evidenced rearrangements in the J genome, such as a reciprocal translocation between 4JL and 5JL, an interchange between centromeric segments of 2J and 5J, an intra-chromosomal translocation on 6JL, and a paracentric inversion on 7JS [87].

Intermediate wheatgrass (*Thinopyrum intermedium*) is a highly productive cool-season forage grass with resistance to drought, frost and many pests and diseases of wheat and other cereals. In addition, this species was identified as a good candidate for domestication and breeding as a perennial grain. *Th. intermedium* is an allohexaploid (2n = 6x = 42) whose most likely progenitors are the diploid species of *Thinopyrum* and *Pseudoregneria*. There are several proposed genome designations, one of which was JvsJvsJ r J rStSt, where Jvs and Jr represent ancestral genomes of the extant *Th. bessarabicum* and *Th. elongatum* species, respectively, and St the genome of some diploid species of *Pseudoroegneria* from Eurasia [88]. Genotyping-by-sequencing was used to construct a genetic map containing 10,029 markers distributed among the 21 linkage groups of *Th. intermedium* [89]. Comparison of the 21 linkage groups of intermediate wheatgrass with the barley reference genome sequence evidenced three highly collinear homoeologous genomes syntenic with the barley genome with only one exception, a reciprocal translocation between two chromosomes homoeologous to 4H and 5H. This translocation was suggested to belong to the St genome, but the chromosome structure of the *Th. bessarabicum* genome is consistent with a translocation inherited from the donor of the Jvs genome.

### *2.8. Recurrence and Variable Frequency of Chromosome Rearrangements*

Evolutionary chromosome rearrangements vary between the species studied. However, the reciprocal translocation between 4L-5L is present in di fferent diploid and polyploid species. Because of the discovery of this translocation in wheat and rye, a possible monophyletic origin of this translocation in a common ancestor of both lineages was rejected, since the D genome of wheat, which is more closely related to the A genome than the R genome, does not carry this translocation [90]. This translocation is also absent in the genomes of the *Aegilops* species including *Ae. tauschii*, *Ae. speltoides* and *Ae. sharonensis*. This suggests independent origins for translocation 4L-5L in di fferent lineages. Accordingly, although in rye and wheat the breakpoint positions are identical on 4RL and 4AL, they are di fferent on 5RL and 5AL [91]. These sites were suggested to occupy chromosome rearrangements hotspots, as they coincide with inversion junctions produced in the *Triticeae* ancestor [91].

On the other hand, variation in the number of gross chromosome rearrangements among di fferent species suggests that the chromosome structure does not evolve parallel to its genome di fferentiation. For example, despite the theory that *Hordeum* diverged much earlier than *Aegilops*, barley, *Ae. tauschii*, *Ae. speltoides* and *Ae. sharonensis* show a highly conserved synteny. In contrast, rye, *Ae. umbellulata* and *Ae. caudata*, all of which diverged later than *Hordeum*, underwent a considerable number of chromosome rearrangements. Two of these species, rye and *Ae. caudata*, have been shown to undergo introgressive hybridization during their evolution, which is considered the source of their highly rearranged genomes [9,77]. The role of interspecific hybridization on genome reorganization is also apparent in the formation of allopolyploids. Chromosome rearrangements produced in tetraploid wheats contrast the highly conserved synteny of their diploid progenitors. Allopolyploidy is usually accompanied by extensive genome reorganization and changes in gene expression within a short period of time in wheat and other plant lineages. Both the genetic control of pairing and the physical divergence of the homoeologous genomes are considered the genetic systems responsible for the cytological diploidization of polyploid wheat [92,93]. Consequently, interspecific hybridization appears as a releasing factor of genome reorganization, which may accelerate the diploidization of allopolyploids. When recurrent polyploid formation is accompanied with variable genome rearrangements, as in tetraploid wheats, structural chromosome di fferentiation may hinder the genetic flow between the polyploids of di fferent origins and facilitate their speciation.

### **3. An Overview on Meiotic Recombination in Plants**

Meiotic recombination is the cellular process that generates new allele combinations upon which natural or artificial selection can act to favor the establishment of better adapted genotypes, or more useful agronomical traits, respectively. Plant breeders search for novel varieties, which combine valuable traits present in di fferent parental lines. The genetic information provided by both parents is reshu ffled in the F1 hybrid meiosis, and genetic combinations of interest appear in the o ffspring. When traits of interest are controlled by non-linked genes, the searched gene combination is the result of random chromosome segregation at the anaphase I of the hybrid. In the case of joining alleles present in homologous (or homoeologous) chromosomes, intra-chromosomal homologous (or homoeologous) recombination is required. Intra-chromosomal homologous recombination is produced during the first meiotic division as a result of the repairing process triggered at leptotene by a programmed production of double-strand DNA breaks (DSBs) catalyzed by the topoisomerase-like SPO11 protein [94,95]. The repairing process is configured over an intact template that may reside in the sister chromatid or in a chromatid of the homologous partner. There is evidence that both pathways exist in meiosis, although meiotic chromosomes seem to be organized to facilitate the choice of homologous recombination [96]. Nucleo-filaments formed by 3- single-strand DNA overhangs

generated from each DSB bound to recombinases RAD51 and DMC1 to invade a double-strand DNA stretch of the homologous chromosome to find its complementary strand [97]. The use of a non-sister homologous template may culminate with either a reciprocal exchange of large homologous chromosome segments, i.e., a crossover (CO), or the non-reciprocal exchange of a small DNA sequence, i.e., a non-crossover (NCO).

The DSBs repairing process is broadly conserved among plants, and comprehensive insights on the underlying molecular mechanism have been provided in recent reviews [98–100]. Only a minor fraction (5% in *Arabidopis,* or maize) of the homologous interactions are resolved as COs. This is enough to ensure the occurrence of at least one CO, the obligatory CO, per chromosome pair. Most COs are processed in the class I pathway, which relies on a group of proteins called ZMMs (SHOC1/ZIP2, HEI10, ZIP4, MER3, MSH4, MSH5 and PTD) and two additional proteins MLH1 and MLH3 not included in the ZMM group [98]. A feature of the class I COs is that they prevent the occurrence of additional COs nearby. Class II COs are dependent upon structure-specific endonucleases including MUS81, and are insensitive to interference [101]. In addition to these two pathways that lead to the CO production, other mechanisms are involved in DSBs repairing. Three other groups of proteins are involved in the restriction of the CO number [99,100,102]: (i) FANCM and its cofactors MHF1 and MHF2 are thought to unwind post-invasion intermediates to promote NCOs through the synthesis-dependent strand annealing (SDSA) pathway, (ii) the BLM/Sgs1 helicase homologs RECQ4A/RECQ4B and the associated proteins TOP3 α and RMI1, which process probably di fferent recombination intermediates of FANCM, and (iii) FIGL1 and its partner FLIP, which may control the activity of the recombinases RAD51 and DMC1. Mutations that disrupt any of these three NCOs promoter pathways increase the frequency of class II COs dependent of MUS81 in *Arabidopsis* [102]. Recombination frequency increases with the distance to the centromere, and reaches maximum values in distal regions.

Regular segregation of homologous chromosomes at anaphase I is facilitated by the formation of chiasmata, which are visible from diplotene to metaphase I and represent the cytological expression of COs produced in previous stages. The occurrence of a CO between homologous chromatids happens in a context in which each member of the partner and its sister chromatid are held together by a number of ring-shaped cohesin complexes scattered along the chromosomes axes and formed after DNA replication. The cohesin complexes also play an additional role in the DSBs repairing process [103]. Cohesion is released from chromosome arms at anaphase I, but not from the centromere region, where sister kinetochores remain associated and oriented to the same pole [104]. Connections between sister kinetochores are resolved at anaphase II, allowing their segregation to opposite poles.

In most species, COs are non-randomly distributed along chromosomes. They are clustered at recombination hotspots that alternate with poor recombination domains. Initial steps of recombination are to some extent responsible for the CO distribution, but it is not clear whether the CO landscape mirrors the non-uniform DSBs' positioning [105]. High-resolution maps of recombination events in plants show COs located in regions close to gene promoters and terminators [106]. While the DSB and CO maps are rather similar in *Arabidopsis* [107], they are very di fferent in maize [108]. Only a quarter of DSBs produced in maize occur near genic regions, and can be processed as COs, the remaining DSBs situate in repetitive DNA, and do not form COs. Di fferent patterns of CO distribution have been reported in plants. The most common situation is a pronounced localization in a distal euchromatic region. This is the case of maize [109] and *Triticeae* species [26,110–113], with an extremely distal location in *Ae. speltoides* [74]. Some species show a quite di fferent pattern, for example, CO hotspots extend throughout the entire chromosome in *Arabidopsis* and rice [99], while most COs are located in the proximal quarter of the *Allium fistulosum* chromosomes [114].

Preferences for CO location in euchromatic regions sugges<sup>t</sup> a specific role of chromatin structure in the recombination distribution. In fact, CO hotspots have low DNA methylation and transposons in some cereals [98]. In *Arabidopsis,* COs and recombination are correlated with active chromatin features, such as the modified histone H2A.Z, trimethylation of histone H3 on lysine 4 (H3K4me3), low DNA methylation, and low-nucleosome-density regions [115,116]. Epigenetic marks are also responsible for the absence of COs in centromeres and the surrounded heterochromatic regions of plant chromosomes.

Modification of pericentromeric epigenetic marks, such as histone 3 lysine 9 dimethylation (H3K9me2), and DNA methylation in CG and non-CG sequences in *Arabidopsis* defective mutants for the H3K9 methyltransferase genes KYP/SUVH4 SUVH5 SUVH6, or the CHG DNA methyltransferase gene CMT3, increases CO frequency in proximal regions, despite the fact that the total number of COs is not a ffected [117]. In addition to chromatin structure, the initiation of recombination changes across the telomere-centromere axis. DSBs appear initially in sub-telomeric CO-rich regions, and extend later to the intercalary CO-poor regions of barley chromosomes [113], which suggests that early recombination events are preferred for CO formation.

Homoeologous chromosomes are present in interspecific hybrids and allopolyploids, and represent potential partners to be involved in the DSBs' repairing process. Such chromosome partner selection was found to be suppressed in hexaploid wheat by the action of the *Ph1* locus of chromosome 5B [36,37]. *Ph1* was first assigned to a structure composed of a segmen<sup>t</sup> of heterochromatin inserted into a cluster of seven cyclin-like-dependent kinase (CDK) genes [118]. This *Ph1* assignment was later put into question by the meiotic phenotype of a di fferent candidate gene, *C-Ph1,* which was silenced [119]. Included in the heterochromatin segmen<sup>t</sup> of chromosome 5B is the ZMM gene *ZYP4* (*TAZYP4-B2*). This gene was proposed to be *Ph1*, as supported by the high level of homoeologous pairing found in the hybrids of two *Tazyp4-B2* TILLING mutants and one CRISPR mutant of wheat with *Ae. variabilis* [120,121]. In addition, extensive analysis of RNA seq data in wheat, wheat x rye hybrids and triticale with and without *Ph1*, indicated that the silenced C-*ph1* gene corresponded to a copy present on chromosome 5D, since the copy on chromosome 5B is not expressed at any meiosis stage [122]. *Ph1* is the most e ffective suppressor gene of homoeologous recombination in polyploid wheats. *Ph2*, another suppressor located on chromosome arm 3DS, shows an intermediate e ffect [123]. Two main steps of bivalents' formation seem to be under the action of *Ph1*: The suppression of recombination between homoeologous chromosomes, despite the fact that they form synaptonemal complex (SC) during zygotene, and the correction of SC multivalents to form two or more bivalents during pachytene [124]. However, how this can be accomplished is poorly understood.

### **4. Modulating the Meiotic Recombination Landscape for Cereal Improvement**

The reference sequence of the wheat genome gives access to 107,891 genes, with an uneven distribution along the chromosomes [26]. Intercalary CO-poor regions harbor a larger fraction of genes than the highly recombinogenic distal regions. A similar genome organization is present in barley [84]. Increasing recombination in chromosome regions with low CO frequency is upmost in importance from the breeding perspective. A possible strategy would be to increase the overall CO frequency using anti-CO mutants. The TILLING mutants generated in tetraploid and hexaploid wheats, which are accessible in public databases [125], may be very useful for this purpose. Mutation of anti-CO genes greatly increase the CO frequency in *Arabidopsis* and in crops such as *Brassica*, rice, pea and tomato [106]. However, this strategy may fail in promoting CO in low-recombining regions of crops because, in *Arabidopsis*, extra COs fall into the highly recombining regions [102]. Approaches, such as manipulation of epigenetic marks associated with low recombining regions, or the induction of DSBs at specific chromosome sites by a fusion of SPO11 and di fferent DNA-binding proteins, have been proposed to increase the CO number in intercalary or proximal regions [106]. Available stocks of wheat mutants [125] can also be used for the identification of candidate genes involved in chromatin organization, which might have some e ffect on meiotic recombination.

An alternative approach may be based on the environmental e ffect on meiotic development. A number of exogenous factors, including environmental stresses, agrochemical, heavy metals, combustible gases, pharmaceutical and pathogens, are known to a ffect meiosis in plants [126]. Many of these factors produce meiotic abnormalities such as laggards, univalents, bridges, stickiness, or precocious chromosome movements, which detract their practical value to modify the recombination pattern. Special attention has been received on the e ffect of temperature, which is species-specific. In *Arabidopsis*, the response to temperature variation follows a U-shaped curve with increased CO frequencies at low (8º) and high (28º) temperatures relative to a medium-range value (18º) [127]. Extra COs produced at extreme temperatures were of the class I pathway. In contrast, barley plants subjected at a temperature of 30º produce a lower number of chiasmata in male meiosis, which change their positions towards more proximal locations. Repositioning of recombination events a ffects only the class I COs [113,128]. Modification of nutrient concentration also a ffect the recombination frequency. Among *Triticeae*, a high phosphate level increases chiasma frequency in rye [129] and magnesium of the Hoagland's solution increases CO frequency in wheat and the hybrids of wheat lacking *Ph1* with related species [121]. However, it is unknown whether these nutrients modify the CO distribution or not.

Wild relatives are a source of variation for introgressing crops traits, providing tolerance to biotic and abiotic stresses. Introgression can be achieved through meiotic recombination between homologous chromosomes using wild species that share same genome with the crop. A number of genes providing resistance to diseases and pests were transferred from diploid and polyploid wild *Triticum* species to cultivated wheats [34,130]. Genetic variability is much higher in wild species containing genomes that are homoeologous to those of the crops. Transfer of useful genes from such species requires the induction of recombination between homoeologous chromosomes of the cultivated and wild species. This homoeologous recombination is induced in the absence of chromosome 5B, or using the deletion mutant *ph1b*, which lacks *Ph1* [131]. This mutant line has the disadvantage of accumulating translocations between homoeologous chromosomes produced by meiotic recombination [132]. However, other mutants of the *ZIP4-B2* gene in the *Ph1* locus recently obtained, form only homologous bivalents at metaphase I, and preserve better the genome stability [120]. Another way of inducing homoeologous recombination is through the use of genes that suppress the e ffect of *Ph1*. Suppressors of *Ph1* have been reported in *Ae. speltoides* and other species. One of the two major suppressor loci of *Ae. speltoides*, *Su1-Ph1*, which is located distally on the long arm of chromosome 3S, has been introgressed into chromosome 3A of hexaploid and tetraploid wheats [133] and can be used in interspecific gene transfer.

Pioneer works of wild introgression into wheat via homoeologous recombination were those transferring disease resistance genes from *Ae. comosa* and *Ag. elongatum* [134–136]. The transfer of resistance to the yellow rust *Puccinia striiformis* from *Ae. comosa* to wheat was induced by the *Ae. speltoides* genome. Plants homozygous for the resistant allele gave rice to the wheat variety known as Compare [134,135]. Recombination between wheat chromosomes 3D and 7D and their homoeologues of *Ag. elongatum,* produced in the absence of chromosome 5B, made possible the transfer to wheat of resistance to the leaf rust *Puccinia recondita* [136]. After these genetic transfers to wheat, many others have been produced. Species of the genus *Aegilops* represent a valuable source of genetic variability used in wheat improvement. Most *Aegilops* species were crossed with wheat to obtain amphiploids, and addition, substitution, translocation and segmental introgression lines. More than 40 resistance genes from the *Aegilops* species have been introgressed into wheat through chromosome translocation or homoeolgous recombination, and some of them have been used in wheat production [40]. An ample genetic variability present in perennial wild grasses and wild ryes has been incorporated also into the wheat genome in the form of amphiploids or derivatives such as addition, substitution or radiation translocation lines, as well as recombinant lines with segmental introgressions. Transfers involving *Thinopyrum* species have provided valuable contributions in wheat cultivar development [137].

Intercrops transfers have also been employed in some breeding programs. Many e fforts have been made to use the gene pool of rye in wheat improvement. The production of hexaploid triticale AABBRR is a representative example of the success of combining the genomes of both species to produce an excellent feed crop. However, the highest potential of rye introgressions into wheat has been manifested in the production of a number of wheat cultivars carrying the wheat-rye translocations 1RS.1BL or 1RS.1AL, with a noteworthy positive e ffect on yield production [138]. Wheat-barley hybrids and introgressions have been produced also [139]. The transfer to wheat of barley genes controlling agronomical traits, such as drought tolerance, high β-glucan content, salt tolerance or earliness, is conditioned by the low crossability between both species. E fforts should be made to increase the efficiency of the crosses.

Hybrids between wheat and other *Hordeum* species have also been developed. *Tritordeum*, a *H. chilense* × durum wheat amphiploid, is the most successful introgression [140,141]. After the improvement of di fferent agronomical traits in field experiments, *Tritordeum* became a synthetic cereal to some extent comparable to triticale. Recombination between the chromosomes of wheat and *H. chilense* has been induced in the absence of *Ph1*, supporting the idea that the transfer is possible using this approach [142].

In addition to the desirable gene, recombinant chromosomes carrying segments of wheat and alien chromosomes might carry other genes that reduce the agronomical value of the introgression. Therefore, primary recombinant chromosomes should be engineered to remove undesirable genes. This is possible after the production of sets of primary recombinant chromosomes formed by wheat-centromere-wheat/alien genetic material, which di ffer in the translocation breakpoint position, and their reciprocal counterparts, formed by alien-centromere-alien/wheat genetic material. Double heterozygotes containing chromosomes of the types wheat-centromere-wheat/alien and alien-centromere-alien/wheat, carrying the desirable gene from wild species, that undergo a CO in the overlapping alien region, produce secondary recombinants wheat-centromere-wheat/alien/wheat with the desirable gene into a shorter alien intercalary segmen<sup>t</sup> [143,144]. In the absence of such primary recombinant chromosome sets, the size of the translocated alien segmen<sup>t</sup> can be reduced in additional rounds of homoeologous recombination between the translocated alien segmen<sup>t</sup> and the standard wheat chromosome.

### **5. The Impact of Chromosome Rearrangements on Meiotic Recombination**

CO frequency in *Triticeae* was estimated to range between two and three COs per chromosome, or more than one CO per chromosomal arm [112,145]. Distal confinement of the site of the first or only CO in each arm might be conditioned by the subtelomeric location of the initial interactions between homologs imposed by the bouquet arrangemen<sup>t</sup> [146]. In fact, distal recombinational events precede those more proximally located in barley [113]. CO interference could condition the formation and position of additional COs. Given the preferred distal localization of COs in many plant species, an alteration of the standard chromosome structure represents an approach used to understand the molecular basis underlying this recombination pattern. Chromosome structural mutants, such as deletions and inversions, which modify the relative position of the genetic material present in a given chromosome, have been produced in wheat, rye and *Arabidopsis.* In wheat, both types of chromosome mutants are viable in the homozygous and heterozygous conditions, and the same is with rye when the mutant chromosome is introgressed in wheat. Large deletions produced in *Arabidopsis* are viable only in heterozygotes [147]. Recombination in such heterozygotes denotes no major e ffect of the deletions in the total number of COs. The loss of COs at the deletion sites is compensated by increases in recombination frequencies elsewhere on the same chromosome. Thus, changes in the physical structure of a given chromosome redistribute the COs within that chromosome [147].

Studies using deletions and inversions in wheat and rye were aimed at verifying the ability of a given region to form CO after changing its position in the centromere-telomere axis. Heterozygotes for the loss of a long terminal segmen<sup>t</sup> of wheat chromosome arms 4AL, 2BL and 5BL underwent a considerable reduction of chiasma frequency at metaphase I in these arms, but the level of chiasmata became normal in homozygotes for the truncated arms [148]. On the other hand, homozygotes for the loss of the distal 25% of the 1BL arm, or the distal 41% of 5BL, increase the recombination rate of the middle arm region, without modifying that of the proximal region, relative to wild type plants [149,150]. These results sugges<sup>t</sup> that the position of a segmen<sup>t</sup> in the telomere-centromere axis was a decisive factor in its capacity to produce a CO. However, other studies contradict this idea. Regardless, synapsis

is completed, COs are infrequent in the proximal third of the rye chromosome arm 5RL, both in homozygotes for the standard chromosome 5R, and homozygotes lacking the distal 70% of its long arm [151]. Even more convincing is the e ffect of the repositioning from distal to proximal of the highly recombinogenic region of the arms 1RL of rye and 2BS and 4AL of wheat, as a result of large paracentric inversions. COs are restricted to the proximal region in all homomozygotes for the inversion [152–154].

Thus, CO distributions in the arms 4AL and 2BS of wheat and 1RL and 5RL of rye are not dependent of the distance to the telomere; other factors such the DNA sequence and the pattern of chromatin organization should condition the recombination landscape.

Given that the truncation of chromosome arms 1BL and 5BL increases the frequency of recombination in their middle region [149,150], a similar e ffect was expected in other chromosomes of wheat. Progressive shortening of the arms of wheat chromosomes 4A and those of the B genome by terminal deletions showed an apparent shift of the CO site from distal to middle-proximal regions [155]. Chiasmate associations at metaphase I of the truncated homologous arms estimated the CO frequency in the fraction of the arms present in homozygotes for terminal deletions of di fferent size. According to the preferred distal localization of COs, a decrease in the CO frequency accompanied the progressive reduction of the arm length. The recombination level of the standard chromosome was not recuperated in most deletions, but a considerable increase in the middle chromosome arm regions was supported. Deletions with breakpoints in sub-distal sites in chromosome arms del2BS, del2BL and del6BS show higher CO frequencies than deleted arms with more distal breakpoints. Most of the COs produced in the intact chromosome 3B locate in the distal 68 Mb (~ 20% of the total arm) of 3BS and the distal 59 Mb (~ 14% of the total arm) of 3BL [112]. These chromosome segments are smaller than those missing in deletions 3BS-7 (25%) and 3BL-11 (19%). Lines for these two deletions show very high CO frequencies, 66% the 3BS-7 deletion line and 99% the 3BL-11 line [153], suggesting that the shortening of the chromosome arm increases the level of recombination in intercalary regions of chromosome 3B. The intercalary deletion of the *ph1b* mutant maps 0.9 cM from the centromere, which corresponds to a CO frequency of 0.018 [156]. Despite that the interval centromere-*ph1b* deletion is longer than the fraction of the 5BL arm present in the del5BL-5 truncated chromosome, homozygotes for the del5BL-5 deletion form at least one CO in the deleted arm in 62% of meiocytes. The increase of CO frequency in intercalary regions induced by terminal deletions seems to be common to most wheat chromosomes. The 1BS arm represents an exception, since di fferent deletion lines show a low CO frequency. The reason of the CO redistribution in deletion lines is still unknown, but future research in this field can provide valuable information to modify the recombination landscape.

The frequency of association between the homoeologous arms of individual chromosomes of wheat, and between wheat and related species chromosomes, has been reported in interspecific hybrids of wheat with rye, *Ae. longissima*, *Ae. sharonensis*, *Ae. speltoides* or *T. timopheevii* [46,63,67–70,78,79]. The long arms of groups 1 and 2 chromosomes of wheat and rye di ffer in the amount of C-heterochromatin present in their subtelomeric regions. Such di fferences made possible to identify the parental heterochromatin constitution of chromosomes at anaphase I, as well as recombinant wheat-wheat (A-B and B-D) and wheat-rye (A-R, B-R and D-R) homoeologous chromosomes in *ph1b* ABDR hybrids. The recombinant chromosome ratio fits the frequency of association at metaphase I in all the di fferent homoeologous arm combinations [78,157]. Thus the frequency of association at metaphase I between homoeologous arms represents a good estimate of the homoeologous recombination frequency.

Among bread wheat chromosomes, both arms of chromosome 4A, and the arms 5AL and 7BS are involved in evolutionary chromosome rearrangements. The same happens with rye chromosome arms 2RS, 3RL, 4RL, 5RL, 6RS, 6RL, 7RS and 7RL, which are involved in multiple translocations relative to the wheat D genome, as well as with the arms 4SlL and 7SlL of *Ae. longissima*. In both *ph1b* mutant wheat × rye and *ph1b* wheat × *Ae. longissima* hybrids, the frequencies of all possible associations wheat-wheat (A-B, A-D, and B-D), wheat-rye (W-R) and wheat-*Ae. longissima* (W-Sl), are known. Thus, it is possible to assess the e ffect of chromosome structure on homoeologous recombination by a comparison of the level of recombination between chromosome arms showing normal homoeologous

relationships and those rearranged during evolution. Mean frequencies of wheat-wheat, wheat-rye and wheat-*Ae. longissima* associations involving arms with conserved macro-synteny, or arms with gross rearrangements, are given in Table 1.


**Table 1.** Average frequency (%) of association at metaphase I between syntenic and rearranged homoeologous arms in hybrids of the *ph1b* mutant wheat (W) with rye (R) or *Ae. longissima* (Sl).

Although there are di fferences between short and long arms, as well as between homoeologous groups, on average, chromosome rearrangements cause a considerable reduction of homoeologous recombination, and may represent an obstacle in the transfer of genes located in rearranged chromosome arms. Exceptions are the 5RL arm of rye [78] and the 4SlL of *Ae. longissima* [67], which probably carry long translocated segments.

*T. timopheevii* chromosomes evolved a di fferent structure from that of *T. durum* and *T. aestivum*, despite the theory that the donors of the A<sup>t</sup> and A genomes, and of the G and B genomes, are considered to be *T. urartu* and *Ae. speltoides*, respectively, in both lineages [63,64]. All three species share translocation T(4AL-5AL). However, while the inversions in 4A and translocation T(4AL-7BS) occurred in *T. turgidum*, four di fferent translocations involving the arms 3AtL, 4AtL, 6AtS, 1GS and 4GS appeared in *T. timopheevii*. The e ffect of structural di fferences on the frequency of the At-A and G-B associations in interspecific hybrids between these three species is shown in Table 2. There is a considerable reduction of the recombination frequency between arms showing di fferences in the chromosome structure. The e ffect is more apparent among the G-B associations, which, on average, are less frequent than the At-A associations.



Superimposed to the e ffect of chromosome structural di fferentiation on homoeologous recombination is the e ffect of the level of genetic a ffinity between cultivated and wild species, which is a result of their phylogenetic relationships. Chromosomes of species very distant in their evolution have a lower frequency of recombination than chromosomes of species with a higher closeness. This is apparent when the results obtained in wheat × *Aegilops* and wheat × rye hybrids are compared (Table 1). Chromosomes of *Ae. longissima*, *Ae. sharonensis* and *Ae. speltoides*, which are more closely related to wheat than to rye chromosomes, recombine with wheat chromosomes, especially with those of the B genome, much more frequently than with rye chromosomes. Accordingly, At-A chiasmate associations are more frequent than B-G associations in AAtBG and AAtBGD wheat hybrids (Table 2). Both conservation of macro-synteny and genetic di fferentiation are factors conditioning interspecific introgressions.
