Genetic effect of the Aegilops caudata plasmon on the manifestation of the Ae . cylindrica genome

In the course of reconstructing Aegilops caudata from its own genome (CC) and its plasmon, which had passed half a century in common wheat (genome AABBDD), we produced alloplasmic Ae. cylindrica (genome CCDD) with the plasmon of Ae. caudata. This line, designated (caudata)-CCDD, was found to express male sterility in its second substitution backcross generation (SB2) of (caudata)AABBCCDD pollinated three times with the Ae. cylindrica pollen. We repeatedly backcrossed these SB2 plants with the Ae. cylindrica pollen until the SB5 generation, and SB5F2 progeny were produced by self-pollination of the SB5 plants. Thirteen morphological and physiological characters, including pollen and seed fertilities, of the (caudata)-CCDD SB5F2 were compared with those of the euplasmic Ae. cylindrica. The results indicated that the male sterility expressed by (caudata)-CCDD was due to genetic incompatibility between the Ae. cylindrica genome and Ae. caudata plasmon that did not affect any other characters of Ae. cylindrica. Also, we report that the genome integrity functions in keeping the univalent transmission rate high.


INTRODUCTION
Genetic autonomy of the cytoplasmic genetic system, or plasmon, from the nuclear genome is one of the central themes in eukaryote genetics.Evidently, the plasmon has diversified along with nuclear genome differentiation during the evolution of the Triticum (wheat)-Aegilops complex (Tsunewaki, 1996(Tsunewaki, , 2009)).However, this does not necessarily mean that plasmon diversification took place after the genome differentiation.Ae. cylindrica is a tetraploid species (2n = 28) having the genome constitution CCDD, and is assumed to have originated from spontaneous chromosome doubling of an F 1 hybrid between Ae. squarrosa (synonymous to Ae. tauschii, 2n = 14, genome DD) and Ae.caudata (synonymous to Ae. markgrafii, 2n = 14, CC) (Kihara, 1944), the former being the plasmon donor to Ae. cylindrica (Tsunewaki, 1996).
To obtain evidence of genetic autonomy of the plasmon from a co-existing genome, we attempted to reconstruct an Ae.caudata plant from its own genome (CC) and the plasmon that had passed half a century in common wheat (2n = 42, genome AABBDD).At an early stage of this work, it became necessary to transfer the Ae.caudata plasmon from the alloplasmic common wheat to Ae. cylindrica, because its direct transfer to Ae. caudata was not possible.Therefore, we bred an alloplasmic line of Ae. cylindrica, (caudata)-CCDD, by transferring the Ae.caudata plasmon of the alloplasmic wheat, (caudata)-AABBDD, at its 50 th substitution backcross generation (SB 50 ), to Ae. cylindrica.
In the SB 2 generation of (caudata)-CCDD, we found that some plants exhibited partial male sterility.We carried out an investigation to clarify whether this male sterility was caused by the Ae.caudata plasmon alien to Ae. cylindrica or by the nuclear gene interaction between the C and/or D genome derived from Ae. caudata, T. aestivum and Ae.cylindrica.To clarify this point, we successively backcrossed the (caudata)-CCDD SB 2 to SB 4 plants with the Ae.cylindrica pollen, and the SB 5 F 2 progeny produced by self-pollination of the SB 5 plants were compared to the self-pollinated euplasmic Ae. cylindrica plants.In those successive generations, 11 to 13 morphological and physiological characters of (caudata)-CCDD were observed and compared with those of the euplasmic Ae. cylindrica.

Plant materials
The following designation is used for all alloplasmic lines produced in this investigation: a plasmon donor in parentheses, hyphenated to a genome donor or its genome constitution.For example, (caudata)-Tve SB 50 indicates the 50 th substitution backcross generation of the common wheat strain Tve with the Ae.caudata plasmon.The Ae. caudata strain used as the plasmon donor to this alloplasmic common wheat was Ae. caudata var.polyathera KU6-1.This alloplasmic common wheat line was produced by Kihara (1959) and its backcrossing has been continued until now by the present senior author (Tsunewaki, 2009, unpublished data).The two Aegilops species Ae. caudata var.typica KU6-2 (2n = 14, genome CC) and Ae.cylindrica var.typica KU7-2 (2n = 28, CCDD) were also employed in the present investigation.All three Aegilops accessions originated from the Aegilops-Triticum germplasm collection of Kyoto University, Japan (Tanaka, 1983).
To begin with, (caudata)-AABBDD SB 50 was crossed as female to Ae. caudata, giving rise to the (caudata)-ABCD F 1 hybrid.This hybrid was treated with 0.5% aqueous colchicine solution for five days (Sears, 1941), successfully producing 10 seeds of the amphidiploid (caudata)-AABBCCDD.Next, this amphidiploid was crossed as female to Ae. cylindrica, producing an F 1 hybrid, (caudata)-ABCCDD.The F 1 was backcrossed five times with the pollen of Ae. cylindrica up to the SB 5 generation, which was self-pollinated, giving rise to a fixed progeny, (caudata)-CCDD SB 5 F 2 line.In each generation, 2n chromosome numbers in root-tip cells and chromosome pairing configu- 2) 2n = 30-40 SB 1 plants having 2-12 A and/or B genome chromosomes were obtained.
3) Of these SB 1 plants, 2n = 30-33 plants having 2-5 A and/or B genome chromosomes were used for further backcrossing. 4)The 119 SB 2 plants produced had 2n = 28-33 chromosomes, including 0-5 A and/or B genome chromosomes. 5)Of these SB 2 s, plants with 2n = 28, forming 14" in the first meiotic metaphase (MI) (Fig. 1B), were selected as the female to produce SB 3 offspring by further backcrossing with Ae. cylindrica.Note) (c) in this and other tables means the Ae.caudata plasmon.ration in pollen mother cells (PMCs) were observed by the squash and smear method, respectively.The results are summarized in the rightmost column of Table 1.
All the crossing was made by manual emasculation and hand pollination.Self-pollinated ('selfed' hereafter) seed fertility was estimated by the seed setting in the 1 st and 2 nd florets in all spikelets except one or two underdeveloped basal and apical ones.The percent cross success was calculated as the product of the percent seed setting of the cross and the percent seed germination rate.

RESULTS
Production of alloplasmic Ae. cylindrica with the Ae.caudata plasmon The results of crosses in Steps 3 to 8 (Table 1) are collectively shown in Table 2; the results of the last three steps from (caudata)-CCDD SB 2 to SB 4 backcrossed with Ae. cylindrica pollen were pooled.
The first cross, (caudata)-AABBCCDD × Ae. cylindrica, was difficult as is clear from its low cross success rate of 7.2%.Sixteen seeds selected from 96 seeds produced by hand pollination were sown, and 12 seedlings were grown.An F 1 plant that had 2n = 42 with a probable genome constitution ABCCDD revealed an average meiotic chromosome configuration of 0.98"" + 1.48"' + 10.46" + 12.35' with 0.09 penta-and 0.02 hexavalents.This indicated that the C-and D-genome chromosomes mostly formed bivalents or multivalents with their homologues, whereas the A-and B-genome chromosomes stayed in univalents.
The second cross, (caudata)-ABCCDD F 1 × Ae. cylindrica, was the most difficult of those made in the present investigation, being evident from its lowest cross success rate of 2.5%.All 86 seeds produced, including 15 seeds set in the third florets, were sown, and chromosome checking was successfully conducted with 60 of the 62 viable seedlings (Table 3).
We pollinated more than two thousand florets of the F 1 hybrid, expecting to get 2n = 28 plants with the complete CCDD genome constitution devoid of the A-and Bgenome chromosomes in the SB 1 progeny.However, we did not get any 2n = 28 plants, or even any 2n = 29 plants.This indicated that the transmission rate of the A-and Bgenome univalents was much higher in the ABCCDD hybrid than in the 21 monosomics of common wheat, as discussed later.
As shown in Table 4, we selected 2n = 30 to 33 plants having two to five A-and/or B-genome univalents to increase the chance to obtain 2n = 28 plants with the complete CCDD genome constitution in the SB 2 progeny.
Two SB 1 plants with 2n = 30 and 31 produced no SB 2 progeny.Five SB 1 plants with 2n = 32 and 14 SB 1 plants with 2n = 33 produced 200 and 488 SB 2 seeds, respectively.From each of the SB 2 progeny, 120 seeds were sown and the somatic chromosome numbers of 119 out of the 194 germinated SB 2 seedlings were determined (Table 5).
The somatic chromosome numbers of the 119 SB 2 seedlings ranged from 2n = 28 to 33, of which 89 plants still retained one to five A-and/or B-genome chromosomes.We obtained 30 SB 2 plants having 2n = 28.The meiotic configuration of six of the 30 2n = 28 SB 2 plants was determined (Table 6).
Three of the six plants examined had a modal chromosome configuration of 14", like the euplasmic Ae. cylindrica plant (Fig. 1, A and B).This indicated that the Ae.cylindrica CCDD genomes were restored in those plants.Two of the remaining three plants had a modal chromosome configuration of 13" + 2' and one had 12" + 1"".This indicated that one A-or B-genome chromosome remained in the former two plants and one reciprocal translocation was present in the latter.We backcrossed two of the three SB 2 plants with 14" with Ae. cylindrica pollen.Their SB 3 progeny were further backcrossed twice with the same pollen to produce the SB 5 seeds.The results of all these backcrosses are compiled in Table 7.
In every backcross of the SB 2 to SB 4 parents, the seed set rate was about 90% with a seed germination rate of higher than 65%.Thus, the progeny of (caudata)-CCDD were stably maintained after the SB 3 generation.
Effects of the Ae.caudata plasmon on the genome manifestation of Ae. cylindrica In the (caudata)-CCDD SB 2 generation, some plants had non-dehiscing anthers in spite of their having 2n = 28 chromosomes.They showed a wide variation in selfed seed fertility, from 0 to 97%, with the average of 57.3 ± 30.9%, in contrast to a narrow variation in the selfed seed fertility (92.1 ± 4.2%) of the euplasmic Ae. cylindrica plants (Table 8).
The genetic effect of the Ae.caudata plasmon on various morphological and physiological characters of Ae. cylindrica was investigated with (caudata)-CCDD plants in the SB 2 -5 generations and with euplasmic Ae. cylindrica plants grown in each of the corresponding years (Table 9).No consistent effect of the Ae.caudata plasmon was observed on any of the examined characters, including spike morphology (Fig. 1G), throughout the four backcross generations, except for the selfed seed fertility.
Seeds from hand-pollination of the alloplasmic lines were smaller than those from self-pollination, which might cause some growth retardation.A critical comparison of the alloplasmic and euplasmic lines was made using the self-pollinated seeds of the (caudata)-CCDD SB 5 line (SB 5 F 2 ) and those of the euplasmic line; both lines were grown in the same year.Ten pots containing single plants of each of the two lines were placed in pairs at random on a bench.
The (caudata)-CCDD SB 5 F 2 line showed significant difference from the euplasmic line in five characters: 3 rd internode length, number of spikelets per spike, dry matter weight, pollen fertility and selfed seed fertility.The effects of the Ae.caudata plasmon on three characters (3 rd internode length, number of spikelets per spike and dry matter weight) were inconsistent between the five generations, SB 2-5 and SB 5 F 2 .On the contrary, semi-sterility induction by the Ae.caudata plasmon was consistent in 1) Determined by root-tip cell mitosis.
2) Percentage of the PMCs showing the modal chromosome configuration is given in parentheses.10; 96% pollen and 97% selfed seed fertility), (E) and (F) alloplasmic Ae. cylindrica, (caudata)-CCDD, no.C-6 and no.C-1 plants, respectively, in the SB 5 F 2 generation (Table 10; 95% pollen and 63% selfed seed fertility for the former and 29% pollen and 63% selfed seed fertility for the latter).A high inter-plant variability of the pollen fertility in the alloplasmic Ae. cylindrica is evident.G: Ears of euplasmic Ae. cylindrica (two on the left) and of an alloplasmic (caudata)-CCDD in the SB 5 F 2 generation (two on the right).No remarkable morphological difference was noticed between the ears of eu-and alloplasmic Ae. cylindrica lines.

D
all generations, although the plasmon effect was not statistically significant in the SB 3 generation.The pollen fertility of the SB 5 F 2 line was 34.0% lower than that of the euplasmic Ae. cylindrica.The normal seed setting rate (90.5%) of backcrossed plants in the SB 2-4 generations (bottom row of Table 2) implied their normal female fertility.
As shown in Table 10, the pollen (Fig. 1D) and selfed seed fertilities of the euplasmic Ae. cylindrica were both higher than 90% on average with low intra-line variation.On the contrary, the (caudata)-CCDD SB 5 F 2 line forming 14" in its meiotic metaphase I (Fig. 1C) showed high within-line variability in pollen fertility (Fig. 1, E and F) that was not correlated with the seed fertility (r = 0.375, non-significant).Altogether, we conclude that the caudata plasmon caused male sterility accompanied by its high within-line variation in (caudata)-CCDD.

DISCUSSION
Possible cause of the male sterility expressed in (caudata)-CCDD The Ae. caudata plasmon is known to cause cytoplasmic male sterility in many cultivars of 4x and 6x wheats (Kihara, 1959;Maan and Lucken, 1971;Tsunewaki, 1974).Kihara and Ohta (1970) reported hybrid sterility between two varieties of Ae. caudata var.polyathera and var.typica, and Endo and Katayama (1978) and Endo (1979) found some gametocidal chromosomes in Ae. cylindrica and Ae.caudata, respectively, which caused semi-sterility in the offspring of their hybrids with non-carriers of the chromosomes.All these findings suggest the involvement of different mechanisms in the expression of the male sterility observed in the pedigree of (caudata)-CCDD. 2)Position from the top.
The present investigation indicated that the male sterility observed in the (caudata)-CCDD line was expressed under the fixed genotypic condition with normal bivalent formation, as is evident from the consistent expression of the sterility after many generations of backcross and selfpollination.Therefore, the heterozygosity of nuclear gene(s) and the presence of either alien chromosome(s) or a gametocidal chromosome are excluded as causal agents of the male sterility discovered here, and it can be concluded that its real cause is the cytoplasmic effect of the Ae.caudata plasmon.Gandhi et al. (2005) compared chloroplast and nuclear microsatellite loci of 36 Ae. cylindrica accessions with those of seven accessions of Ae. caudata and 17 accessions of Ae. squarrosa.They found that 35 Ae. cylindrica accessions had chloroplast short sequence repeat (SSR) alleles similar to those of Ae. squarrosa, whereas one accession, CL-TK116, collected in Turkey, had chloroplast SSR alleles similar to those of Ae. caudata.Their result mostly confirmed that Ae. cylindrica originated from a cross between Ae. squarrosa (♀parent) and Ae.caudata (♂ parent), as proposed by Kihara (1944) and Tsunewaki (1996).Murai and Tsunewaki (1986) reported the diphyletic origin of an allotetraploid species, Ae. triuncialis (2n = 28, genome CCUU), from the reciprocal crosses between Ae. caudata and Ae.umbellulata (2n = 14, UU).The presence of one Ae.cylindrica accession with the C-type plasmon may indicate a diphyletic origin of this species from the reciprocal crosses between the two parental species or a misclassification of this accession, similar to a case that happened in a group of Ae. triuncialis accessions (Tsunewaki et al., 2002a).
Effects of the Ae.caudata plasmon on the manifestation of the Ae.cylindrica genome The present investigation revealed that the genetic effect of the Ae.caudata plasmon on the manifestation of Ae. cylindrica genomes was limited to male fertility: the alien plasmon caused pollen and self-pollinated seed sterility, without exerting any detectable effects on the other morphological and physiological characters investigated.Previously, Tsunewaki et al. (2002b) investigated the genetic effects of the Ae.caudata and many other plasmons of the Triticum-Aegilops complex on the genome manifestation of 12 genotypes of common wheat.They found that the Ae.caudata plasmon caused pollen and seed sterility in nine of the 12 genotypes and induced haploid and twin seedling formation in a single genotype, 'Salmon', which carries a 1BL-1RS wheat-rye translocation.They also found that the Ae.caudata plasmon did not affect 14 vegetative characters, including the ten characters studied in this investigation.It was not expected that the (caudata)-CCDD line would produce haploid and twin seedlings, because Ae. cylindrica has no 1RS chromosome arm, and actually no such seedlings were observed during the production of the (caudata)-CCDD line.Thus, the Ae.caudata plasmon exerted a similar effect on the genome manifestation of Ae. cylindrica and common wheat.

Effect of genome integrity on univalent transmission in
Triticum-Aegilops hybrids Univalents formed in the first meiotic metaphase (MI) move in the first meiotic anaphase (AI) toward either opposite pole independently.Although a gamete is expected to receive a specific univalent at the probability of 1/2, the univalent is eliminated at a certain rate in the MI-AI and/or MII-AII (second meiotic metaphase and anaphase) stages, forming micronuclei in the cytoplasm.Univalent transmission rate is given by t, and its elimination rate is by 1-t, where t is estimated by the following formula: where m is the number of univalents in the female parent, N is the number of its next-generation offspring and T is the total number of univalents transmitted to the next-generation offspring.T is obtained by the formula Σij, where i is the univalent number class from 0 to m, and j is the number of plants in each i class.
In the present investigation, the univalent transmission rate was estimated for the two cross combinations ABCCDD × CCDD and [CCDD + i'] × CCDD.In the latter cross combination, two types of female parent were used, namely, CCDD + 4' and CCDD + 5' (Table 11).
Eighty years ago, Kihara and Wakakuwa (1935) reported the transmission of the seven D-genome univalents in pentaploid wheat hybrids that were backcrossed with the tetra-and hexaploid parents.From their results, the average transmission rate of the D-genome univalents was estimated.Tsunewaki (1963) reported the average transmission rate of the 21 monosomes in the Chinese Spring monosomic series.All those results are compiled in Table 11.
The results shown in Table 11 indicate that when all seven chromosomes of a genome or 14 chromosomes of two genomes formed univalents, their average transmission rate became very high (93% for ABCCDD, 88% and 81% for AABBD).In other words, the rate of univalent elimination was very low (7, 12 and 19%, respectively).On the other hand, the average univalent transmission rate was 68% in CCDD + 4', 50% in CCDD + 5' and 54% in AABBDD-1', where part of the chromosome complement of a genome(s) formed univalents.Their elimination rate was close to 50%.
A few investigations have been carried out to clarify the genetic factors that might be involved in controlling meiotic univalent transmission.Kaltsikes et al. (1970) extracted the AABB genomes from three common wheat cultivars and studied female and male univalent transmission in three 5x hybrids between the extracted AABB lines and their original 6x parents.They intended to determine whether univalent elimination was controlled by the D-genome univalents themselves or by the background AB genomes, because the AABB genomes of the three common wheat cultivars and those of their respective extracted AABBs should have been nearly the same.They found a distinct difference in the univalent transmission rate between the three cultivars, but did not find any clear association between meiotic behavior and the number of univalents transmitted.Recently, Silkova et al. (2014) analyzed the effect of six cases of the substitution of one or two wheat chromosomes by their rye homoeologues on the meiotic behavior of univalents, and found that the 2R/2D substitution suppressed the equational univalent division, whereas 6R/6A substitution resulted in a high frequency of univalent elimination.
In conclusion, the complete genome appears to have a function to suppress univalent elimination in meiosis to retain its own integrity.

Table 2 .
Results of repeated backcrossing of (caudata)-AABBCCDD and its offspring with Ae. cylindrica pollen

Table 3 .
Somatic chromosome numbers in root-tip cells of SB 1 1) Pollen parent of all crosses is Ae.cylindrica var.typica.2) [Seed setting rate] × [seed germination rate] in percentage.

Table 8 .
Distribution of selfed seed fertility in the (caudata)-CCDD SB 2 generation, compared to that of the euplasmic Ae. cylindrica 1) Two of the three plants were confirmed to have 2n = 28, forming 13" + 2' in MI.2) Three of the eight plants were confirmed to have 2n = 28, forming 14" in MI.

Table 9 .
Morphological and fertility characteristics of alloplasmic Ae. cylindrica var.typica lines having the plasmon of Ae. caudata in the SB 2 to SB 5 F 2 generations, as compared to those of its euplasmic lines grown in the same years 1) Number of alloplasmics/number of euplasmics.