THE EMBRYO CULTURE OF INTERSPEGIFIG HYBRIDS OF HORDEUM BY D. ROY DAVIES Department of Agricultural Botany, University College of Wales, Aberystwyth* (Received 2 November 1958) (With I figure in the text) SUMMARY Culturing young hybrid embryos on a synthetic medium proved, in four out of six crosses attempted, to be a successful method of overcoming interspecific incompatibility within the genus Hordeum. Hybrids of diploid H. vulgare and tetraploid H. bulbosum were studied morphologically and cytologically. INTRODUCTION Growing hybrid embryos on synthetic media is a wellknown method of overcoming seed incompatibility in interspecific crosses. In hybrid seeds there is frequently a paucity of endosperm tissue, or its development is abnormal, and the embryo dies at an early stage of development, though it is potentially capable of normal growth. Embryo culture methods have proved very profitable in these cases, and have been utilized for diverse interspecific and intergeneric crosses, though, so far, for only one interspecific cross within the genus Hordeum. Konzak, Randolph and Jenssen (1951) succeeded in growing the triploid hybrid H. vulgare (2x) and H. bulbosum (X). Their success warranted further attempts at overcoming the seedincompatibility which is so prevalent in interspecific crosses within this genus. The results obtained from culturing embryos from six interspecific crosses are discussed in this section. MATERIALS AND METHODS The nutrient medium utilized in the present experiments was that described by Randolph and Cox (193), with one modification. Rappaport (195) recommended using 3 ml. of I % iron citrate instead of the ferrous sulphate used by the former authors. Ferrous sulphate tends to be precipitated from the eulture medium and the developing embryos may sufi^er a deficiency of available iron. The medium utilized was as follows: Calcium nitrate Ca(NO3)2.H2O 236 mg. Potassium nitrate KNO3 85 mg. Potassium chloride KCl 65 mg. Calgon (a sodium hexametaphosphate compound) (NaPOa),! 10 mg. * Present address: Technological Irradiation Group, Atomic Energy Reseaixh Establishment, Wantage Radiation Laboratory, Grove, Wantage, Berkshire.
10 D. ROY DAVIES Magnesium sulphate Ferric citrate i % Agar Sucrose Distilled water 36 mg. 3 rnl 7g 20 g, I 1, Five ml portions of the hot medium were poured, using a graduated pipette, into 6 x inch Pyrex tubes. The latter were stoppered with nonabsorbent cotton wool and sterilized in an autoclave for 20 min at i atmosphere. Hybrid embryos were dissected out 1 to days after fertilization. In the controls, seeds were soaked for days and the embryos then removed. Dissection was done in a small sterilized chamber, and the procedure recommended by Keim (1953) adhered to as closely as possible. The walls of the chamber, bench and dissecting microscope were all liberally soaked with % aqueous solution of Lysol 15 min prior to commencing excising the embryos. Table i. Results of embryo culture Cross {with eiiduspei'ni) 1, H, californicum (zx) X H, vulgare (2x) {without endosperm) 2, H. californicum (2x) X H. vulgare (2x) 3, H, californicum (2x) H. vulgare (2x), H, vulgare (2x) : ' H. bulbosum. (X) 5, H, vulgare (2x);' H, bulbosum (X) 6, H, vulgare (x) : H. bulbosum (X) 7, H. californicum (2x) ' H, bulbosum (X) 8, H, californicum (2x) H, bulbosum (2x) y. H. vulgare (2x) H. bulbosum (2x) 10. H. californicum selfed seed 78 53 I 10 29 139 78 66 55 71 9 12 77 19 107 59 8 22 9 I 2 n cu 53 18 75 19 Ci 37 20 16 16 12 18 23 :6 1 8 23 9 6 I o 'Z. 25 5 10 9 ucl. ^ 12 I 6 9 3 6 3 I 6 I 3 I 7 2 3 3 2 I In cross 2 the single plant died on transferring to soil. In cross 3 the single plant died after 7 months' vigorous growth in the cold weather of JanuaryFebruary, Plant had produced five tillers at this stage. In cross one plant died on transference to soil; other plant grew vigorously. In cross 5 all three plants grew A'lgorously. In cross 6 single plant died on transference to soil. In cross 9 single plant died after 32 months' growth 6 weeks after planting in soil. Had grown to fiveleafed stage by this time. Spikes were immersed in a 50 % aqueous solution of S,T,37 (hexylresorcinol) for I min, hands dipped in 50% alcohol, and needles flamed and stored in ^0% S,T.37. Ovules were then taken out by hand and the embryos dissected out under a binocular microscope with a X 25 magnification. In spite of these precautions there was a very high rate of contamination, especially
InterSpecific hybrids of Hordeum 11 with Rhodotorula, though Keim claimed a contamination of less than i % in his experiments. Subsequently, o.oi % mercuric chloride was utilized as a disinfectant for the spike instead of S.T.37, and the contamination was reduced to an insignificant percentage. In one series the embryos were not dissected out and separated from the endosperm. This was done to determine whether the latter had a deleterious effect on embryo growth. Embryos were placed on the surface of the agar, and the tubes placed under continuous light from a 0 W bulb in a 2 C. incubator. After weeks the tubes were removed and kept at room temperature in diffuse daylight. Seedlings were transplanted at the threeleaf stage when the plants were about 2 inches tall and had strong root growth usually after 12 months. No attempt was made to vary the medium which other workers had shown to be suitable for this genus; it proved quite suitable for the present materials. Other workers have studied this factor intensively and a review of their results has been made by Rappaport (loc. cit.) Embryos were dissected out usually at the latest possible time before the seed began to degenerate. This degeneration was indicated by a browning and discoloration of the seed. In one cross, that between Hordeum viilgare (2x) and H. bulbosum (X) two times of excision were tried, and the later time was naturally more successful in terms of the number of viable plants produced. It does appear tbat all of the six interspecific crosses considered are capable of giving viable hybrid plants, as all showed some amount of growth. That such hybrids were not obtained in all cases is due not to any lethal effects or gross incompatibility of genotypes, but rather to technical difficulties and insufficient replication. The first difficulty, and one which was eventually surmounted, was that of obtaining completely sterile conditions and using an effective disinfectant against the predominant contaminant Rhodotorula. Excising tbe enibryos without damaging them also proved difficult. This is shown by the number of embryos which developed only a shoot or root the latter's meristem especially being exposed and susceptible to damage. Attempts were made to culture the embryos without removing the endosperm, but no mature plants were produced in this manner. This may have been due to the fact that the embryo did not have as good a contact with the medium as the excised embryos the endosperm itself being incapable, or at least less capable, of absorbing or transmitting nutrients efficiently. The concept that the endosperm of these hybrid seeds has a deleterious effect on the embryo cannot be entirely ignored. A lack of differentiation in some embryos cultured may have been due to injury or to an upset in the organization of the tissues in the abnormal environment. Successful transplantation of the seedlings from the sterile agar to sterile soil proved rather difficult. All traces of medium had to be removed by washing, otherwise microorganisms ffourished there. Even then three out of nine hybrid plants died on transplanting, and of the remaining six, one succumbed after 3^ and another after 7 months' fairly strong growth and tillering. The remaining four all crosses of Hordeum vulgare (2x) with H. bulbosum (X) grew vigorously. A particularly unfortunate feature of this technique is that embryo culturing cannot commence until June or July and the seedlings are not ready for transplanting into soil until September or October, By that time the conditions are unfavourable for growth and the seedlings are extremely weak. It would be very advantageous if the hybrids could be produced under artificial conditions in January and February and transferred to soil when optimum conditions for growth are available in spring.
12 D. ROY DAVIES These embryo culture experiments have confirmed the hypothesis, already put forward (Davies 1956), that a frequent cause of seedincompatibility in interspecific hybrids within the genus Hordeum is an abnormal development of endosperm. Such a tissue is incapable of nourishing the developing embryo, and an early degeneration of the latter ensues. ^ Pig. I, Metaphase configurations at meiosis in parents and hybrids, a, H. vutgare (zn = 1) jn h, H. bulbosum {1 = ) 5iv, n: c. Hybrid (2n = ) 2ni, 5ii, 5i: d. Hybrid (20 = ) ini, 611, 61: e. Hybrid (2«= ) 2ni, 5n, 5i: f, Hybrid (zn = ) 2ni, 5ii, Si: g. Hybrid (zn = ) 3iii,.1, i The results indicate that it is indeed possible to overcome some of the barriers which have evolved between the constituent species of the genus Hordeum. Having achieved this, it will now be possible to assess first whether there are any other barriers to their normal growth; secondly, if not, whether the hybrids are of value as sources of new genes; and thirdly how readily available are these genes.
InterSpecific hybrids of Hordeum ij RESULTS Interspecific hybrids o/h. vulgare and H. bulbosum In the present section, the morphology and cytology of the four mature hybrid plants obtained in the present experiment will be considered. The variety of H. bulbosum utilized had low self fertility, was appreciably winter hardy, of perennial habit, and was resistant to powdery mildew [Frysiphe graminis f. hordei) when all the adjacent cultivated forms were very heavily attacked. The cultivated barley. Abed Kenia, was completely selffertile, a nonwinter hardy annual, and very susceptible to mildew. In all stages of their growth, the hybrids resembled the wild species morphologically. At the earlier stages they had the prostrate rosette habit typical of the male parent, Hordeum bulbosum, and later had the thin stems and prolific habit of the wild form. Ear conformation was also more hke that of the male parent. In terms of height, pubescence and bulb formation there was some variation between the hybrids. Two were glabrous, and two very hairy like H. bulbosum. One had very prominent bulb formation whilst the others had thin bases like H. vulgare. Awn size was intermediate as also was anther size anthers were exserted from the fiorets as in the wild parent. Cytological studies The diploid (2M = 1) cultivated parent had a completely normal meiosis (Fig. ia). Tetraploid H. bulbosum again had regular meiotie divisions though it had a very high frequency of multivalents (Fig. ib), up to seven quadrivalents per cell being occasionally observed. Varying numbers of univalents, bivalents, trivalents, and occasionally quadrivalents were observed in the triploid hybrids (Table 2) (Fig. icg). Multivalent frequency did not differ greatly between plants. The number of trivalents varied from 0 per cell, and quadrivalents, though occurring only rarely, from 02. Several cells had more than seven bivalents or bivalents and multivalents, indicating the presence of translocated or duplicated segments within the genomes of the parents, which do not normally pair, but which, in the changed cellular environment and with a lack of completely homologous partners, pair in the hybrid. Table 2. Parental and hybrid forms of H. vulgare X H. bulbosum. Metaphase configurations, chiasma frequency and pollen fertility Plant H. vulgare H. hulbosum Hybrid i 2 3 Chromosome no. 1 Quadrivalents.66 0.0568 O.OI 0.0333 Trivalents 0.681S 0.91 0.9833 0.9333 Bivalents 7.87 5705 585 585 6.0166 Univalents 0.05 6.2 6.56 6.32 6. chiasma frequency per cell 137 26.375 13363 12.98 13233 1.166 No. of cells 2 120 88 100 60 60 % Pollen fertility 93 956 3.00 1597 1.52 035 In the later stages of meiosis, the lagging univalents and half univalents formed micronuclei. The percentage of good pollen was extremely low, varying from 0.35 to 15.97 % No seed was produced on any of the plants, and reciprocal backcrosses on to the H. vulgare parent were unsuccessful. The variation in pollen fertility could not be associated with any variation in multivalent frequency, or mean chiasma frequency.
1 D. ROY DAVIES DISCUSSION Meiotic divisions in the hybrid were characteristic of the triploid progeny of fairly closely related species. The bivalents observed could have been due to autosyndesis of chromosomes of the autotetraploid parent or to allosyndetic pairing of chromosomes of both parents or partly to both. The high frequency of trivalents indicates that allosyndetic pairing probably occurred quite frequently, Trivalents were seen in 62. % of all cells studied there being very little variation in the frequency between plants. Konzak et al. (loc, cit.) stated tbat the frequency of trivalent associations in the hybrids could be correlated with the crossabilities of the parental forms used, but the present data do not agree with this concept. Small differences in the genotype can affect seed setting, endosperm and embryo development, without in any way affecting the pairing behaviour of chromosomes, and it is unlikely that the degree of pairing can be directly correlated with the degree of seed setting, I'he rare production of quadrivalents is probably due to a reciprocal translocation in the H. bulbosutn parent, not represented in tbe female parent. Konzak et al. (loc. cit.) also observed quadrivalents and multivalents of even higher order in their material. In one instance two quadrivalents per cell were observed, indicating that the parents differed to the extent of two reciprocal translocations. In no instance was there any meiotic upset, such as stickiness or chromosome breakage, observed in the hybrids. The very high degree of sterility, the variation between plants, and the absence of any correlation with any behaviour of the chromosomes of this variation, suggests that the sterility could in part at least be attributed to genie factors. It therefore seems likely that, if a sufficient number of genotypes of both parents are explored, some more compatible forms might be found which would produce more fertile progeny. The cytological studies have revealed a high degree of homology between the parents, and gene transfer by backcrosses to the cultivated parent, especially with the aid of enibryo culture techniques, might well be possible. ACKNOWLEDGMENTS I am indebted to Professor P. T. Thomas for advice and guidance during the course of this work, and to the Agricultural Research Council for an award which made the studies possible, REFERENCES DAVIES, D. R. (1956). Cytogenetic studies in wild and cultivated species of Hordeum. Ph.D. Thesis, Univ, of Wales. KEIM, W. F. (1953). An embryo culture technique for forage legumes. Ai;ron. Jour., 5, 50910. KONZAK, C, F., RANDOLPH, L. F. & JENSSEN, N. F. (1951). Embryo cufture of barley species hybrids. Jour, of Heredity, 2, 1253. RANDOLPH, L. F, & Cox, L, G. (193). Factors influencing the germination of Iris seed and the relation of inhibiting substances to embryo dormancy. Proc. Am. Soc. Hort. Sci., 3,, RAFPAPORT, J. (195), In vitro culture of plant embryos and factors controlling their growth. Bot. Rev., 20, 20125,