Plant Cell Reports
Plant Cell Reports (1985 ) 4: 300- 303
© Springer-Verlag 1985
Gynogenesis in Beta vulgaris L.: From in vitro culture of unpollinated ovules to the production of doubled haploid plants in soil D. Bossoutrot and D. Hosemans A.D.A.R., Laboratoire d'Am61ioration des Plantes, Bat 360, Universit6 Paris-Sud, F-91405 Orsay, France Received July 7, 1985 / Revised version received September 19, 1985 - Communicated by A. M. Boudet
ABSTRACT
MATERIALS AND METHODS
Our study concerns the different stages involved in the production of doubled haploid beet plants by in vitro gynogenesis. A histological study shows that the embryos obtained could come from the oosphere or the antipodals.Gynogenesis using male fertile plants gives nearly two haploid plants for i000 ovules cultured : in this case, the problem is to avoid plating fertilized ovules. The gynogenetic plants obtained display an endopolyploidy phenomenon at the root meristem level while their shoot meristem remains haploid. Colchicine doubling during in vitro vegetative propagation of buds has been carried out successfully.
Plant material
INTRODUCTION Haploid plants from unpollinated ovary or ovule cultures were obtained for the first time with barley (Hordeum vulgate L.) by SAN NOEUM in 1976. Since then several other species have been successfully cultured in this way - see reviews of YANG and ZHOU (1982) and SANGWAN-NORREEL and DUHOUX (1982). We have previously described the production of haploid plants by the in vitro culture of unpollinated ovules from male sterile plants in Beta vulsaris L. (HOSEMANS and BOSSOUTROT 1983). Gynogenesis is at present the most efficient of the haploidisation techniques tested in this species. Indeed, we immediatly obtained a 0,23% yield of 17 haploid plants for 7237 ovules cultured. Using the most efficient of the other techniques available (pollination of a male sterile Beta vulsaris by Beta vulsaris ssp esculenta i.e. red table beet), SEMAN (1983) obtained only a 0.013% yield of 12 haploids among 93125 plants observed after pollination. This paper describes four areas of further work embryo sac structure in situ was studied to correlate the nearly mature stage with stages of morphological development of the flowers ; the development during in vitro culture of the structures contained in the ovule was studied histologically ; the production of haploid plants led us to look for an artificial means of doubling the chromosomes; the technique of gynogenesis was further developed to adapt it to male fertile plants.
The histological study of the female gametophyte and the structures developing from it was carried out on ovules from four male sterile genotypes grown in the greenhouse. The determination of ploidy level and tests of colchicine doubling were made on gynogenetic plants obtained from male sterile genotypes. The gynogenetic technique was then applied to three genotypes of male fertile field-grown beets. In vitro culture The media used for the present study were those found to be the most productive during our first tests (HOSEMANS and BOSSOUTROT 1983) : ZHU and WU's (1979) basic medium n°l for wheat unpollinated ovule cultures and RAQUIN's (1982) basic medium for petunia anther cultures. These media were complemented by sucrose at 80g/l and 30g/l respectively and by benzylaminopurine (BAP) at Img/1. The pH was adjusted to 5.8 before autoclaving. The ovules were incubated at 27 °± 2°C with a 16 hour photoperiod (2000 lux). Histolo$ical techniques The ovules, taken either in situ at the culture initiation stage or after 15 to 21 days of in vitro culture, were fixed in acetic ethanol (25% acetic acid, 75% ethanol I00 °) during 48hours. They were then included in paraffin and a series of 15 ~ sections was prepared with a microtome. Sections were coloured with alcian blue and safranin red before being mounted in Canadian balm. Ploidy level determination The ploidy level was determined on i00 individual plantlets obtained by vegetative propagation of 14 gynogenetic plants, one month after their transfer to soil, by two techniques : chloroplast counts in the stomatal guard cells, found on the lower epidermis of young leaves, after staining with Lugol Solution (DEUTER 1970, SEMAN 1982). This method of ploidy determination has a high degree of reliability as shown previously (HOSEMANS and BOSSOUTROT 1985) chromosome counts in the root tip cells using t~e Feulgen technique or aceto carmine staining. Observations were made on three different roots for each vegetative propagated plantlet. -
-
301 Colchicine doubling To double the chromosome complement of the haploid plants obtained, colchicine was added at 50mg/l (C I ) or 100mg/l (C 2) to the vegetative propagation medium. The buds were placed on these media during time periods (t) which varied from 1 to 5 days (DORE 1976) : tl:24h, t~=48h, t~=72h, t4=120h. The control, carried ou~ simultaneouslyI consisted of a vegetative propagation medium without colchicine. One month later, in vitro chloroplast counts were made in the stomatal guard cells of young leaves. For each treatment tested, counts were made on the leaves of three adventitious buds taken from 5 different haploid plants. RESULTS AND DISCUSSION Stages for culture initiation of the female gametophyte A histological study of the embryo sacs in situ was made. When the ovules, still white, were plated after anthesis (in male sterile plants, this stage corresponds to the browning of anthers), the embryo sac is mature. At this stage, a longitudinal section showed the embryo sac to be on the side where the micropylar end is located (fig. IA and IB). It extends into a cavity, the caecum (fig.iC).
m
~'
"~ ~s e es ac
h
therefore able to relate the appropriate degree of maturity of the embryo sac to be in vitro ~ l a t e d to a stage in floral development.Thus, in male fertile plants, the effective ovules are removed from anthesis to the browning of the ovules on the floral scape. This stage of ovule development was found in the flowers at a level located from 10 to 30' cm under the end of the scape when it is still lengthening. Histological study of the structures originating from the ovules We also made sections of 30 ovules, taken after about 20 days of in vitro culture. Some of these ovules already contained structures which could be observed under a dissecting microscope. These sections show that in every case the structures observed are well-formed embryos (fig.2) Some of them undergo a secondary proliferation which results in calli surrounding the embryo. Auxin type impurities contained in certain elements of the culture medium particularily sucrose could enhance this callus proliferation. There was only one embryo per ovule. The root pole was oriented towards the micropylar end and the apical pole towards the body of the ovule. The embryos emerged in two ways : - either the root came out through the micropyle as in germination. - or the ovule integument, located over the e m b r ~ sac-caecum group, teared and allowed emergence of the embryo apex.
fF
0,5 L . S . in an o v u l e
Fig. i : A - Longitudinal section in an ovule (m) micropylar end, (h) hilum B - Detail of the embryo sac (es) showing the two synergids (s), the egg cell (e) or oosphere, the three antipodal cells (ac) and the extension into the caecum(c) C - Detail of the caecum (c) containing the fused polar nuclei (fpn). The egg cell and the two very vacuolized or degenerating synergids were present in all the sections observed. The same was true for the three antipodals located laterally near the junction point between the two cavities. In most cases the fused polar nuclei were found in the caecum. In only one case were they located in the embryo sac near the antipodals : this was probably a more immature ovule removed during anthesis. Even though we didn't observe the first embryogenic stages, which probably occur during the first week of in vitro culture, a study of the embryo sacs at culture initiation could explain the origin of the embryos. We confirm ARTSCHWAGER and STARRET's (1933) observations which showed that in most cases the synergids had degenerated and that the two polar nuclei bad fused. The haploid embryos obtained from ovules, therefore, cannot originate from these cells but either from the egg cell or from an antipodal cell. We were
Fig. 2 : Embryo (2) in an ovule Ovule : (M) micropylar end, (H) hilum Embryo : (V) vessels, (RM) root meristem, (SM) shoot meristem. Fig. 3 : A haploid plantlet originating from an ovule
302 Ploidy level determination in the gyno~enetic plants Ploidy levels, as determined by chromosome on the root tips and by chloroplast counts stomata guard cells are shown in Table I. ~
et System idy vel. System ~ ploidy level~
counts in the
Colchicine treatment concentrationduration mg/l hours
l
n n - 2n
C1 : n
2n
50
number observed
5
0
5
16
0
16
41
i
Number oI leaves observed
t 1 = 24
82.4
17,6
0
51
t 2 = 48
59
41
0
61
t 3 = 72
46
48
6
50
t 4 =120
40,5
53.2
6,3
47
t I = 24
35,3
64.7
0
51
70
i0
40
t 3 = 72
20
54,5
25,5
55
t 4 =120
18
42
40
55
0
18
4n
15
1
16
3
0
3
98
2
I00
Control
Table
%
t 2 = 48
18
number observed
4n
%
42
2n - 4 n
n - 2n - 4n
%
20
C 2 : i00 2n
nl2n
Ploidy level in the shoots
1 : Ploidy level in the root and shoot systems of i00 plants obtained by vegetative propagation of 14 gynogenetic plants.
For the root system measurements, 37% of the individuals observed show more than one ploidy level in the three roots taken at random, though each individual root was uniform. Conversely, at the shoot system level, 98% of the plants were haploid and 2% were diploid. Endopolyploidization which is relatively frequent in root meristems in beet, was not found in shoots where the apices were stable haploids. The regenerants originating from male sterile plants did not set seed after open pollination, thus confirming their haploid state. In order to obtain doubled haploid plants, doubling of the haploid buds had to be induced. This was done by colchicine treatment in in vitro culture. Doubling of the haploid plants with colchicine The efficiency of the different colehicine treatments during vegetative propagation in vitro is shown in Table 2. A two factor analysis of variance (concentration of colchicine X time period during which colchicine is applied) was carried out for the number of doubled leaves (2n or 4n). This showed that the effects observed after C. and ~ treatments were significantly differentlat a ~i~o probability level. The C_ colchicine concentration was significantly more efflcient since it induced over 76% (2n + 4 n ) d o u b l i n g against less than 43% for C_ . The effects of the i time periods during which colchicine is applied were also different at a I ~ probability level. It appears from Table 2 that : - for limited periods on media containing colchicine (24 or 48h), the rate of doubling increases with time but there are very few if any tetraploid plants. - for any given eolehicine concentration, from 72 to 120h, some of the treated plants are doubled a second time : the number of tetraploid plants thus increases while the number of diploid plants simultaneously decreases. The C^ ~ or C_tA treatments seem to be the best for the production of homozygous diploid plants. A sample of 38 plants found to be diploid in vitro were transferred to soil after rooting.
: 0
i97.7
2.3
256
Table 2 : Efficiencies of different colchicine treatments on shoots during vegetative propagation in vitro. At this time, a count carried out on their leaves revealed 22 haploid plants, 14 diploids and 2 tetraploids. These two counts show that 58% of the plants observed to be diploid in vitro proved to be haploid after transferring to soil. These very different results arise either from a disturbance in the physio logical functioning of the plantlet due to in vitro culture, or because of a partial doubling of the neoformed shoots resulting in diploid leaves while the apex remains haploid. The chloroplast counts made on leaves taken from in vitro cultured plants are thus not reliable. These counts will have to be made using young leaves from plants transferred to soil. The doubling rate of buds in vitro can probably be improved, especially by significantly reducing the size of the explants cultured. In fact the doubling of entire treated meristem would be easier using very young buds : since the explant is smaller, the colchicine diffuses and acts on a larger percentage of cells. The probability that bud neoformations will occur when starting with doubled cells is greater. It seems that the in vitro colchicine treatment is more efficient when rather high concentrations are used for brief periods. Gyno~enesis usin G male fertile plants The first gynogenetic haploid beets were obtained using male sterile genotypes (HOSEMANS and BOSSOUTROT 1985). The extension of this technique to male fertile genotypes was necessary. In this case the main difficulty consists in culturing unfertilized ovules only. In a first test the closed flowers were emasculated and bagged. Because of the mechanical tensions exerted on the vessels during emasculation, the ovules became necrotic before reaching the stage favourable to culture initiation. This type of emasculation, which is technically difficult to carry out and inefficient under our working conditions, was therefore abandoned. We then bagged the terminal part of the inflorescence in selfing bags without preliminary manual emasculation. Three genotypes having variable auto-incompatibility rates were used. The flowers were protandric with pollen already released when the last ovules on the stalk became mature. Therefore, there was a limited risk of selling in the last flowers. When culture was initiated we distingui$~e~the fertilized
303 ovules from the probably unfertilized ones on the basis of two criteria : the shape of the ovary, which when fertilized, has a conspicuous beak due to the development of the embryo in the embryo sac, and the size of the ovules. Furthermore, we eliminated all the embryos which develop during the first week of in vitro culture. Because of their rapid development they could only have come from self or cross fertilization. A total of 46 embryos was obtained from 1791 ovules cultured (2.56%); of these, 17 (36,9%) produced plants of which three were found to be haploid giving an overall haploid plant yield of 0,17 for one hundred ovules cultured. In fact, because of insufficient genetic markers the origin of 82% of plants which were diploid could not be determined. They could be the result of self-pollination, crosspollination, spontaneous doubling of haploids or development of fused polar nuclei. But these last two hypotheses are less probable because we observed no diploid plants resulting from ovule culture of a male sterile plant, except in one isolated case. For male fertile plants, the number of fertilized ovules cultured must be reduced. A method would be to plate ovules or ovaries well before the time of fertilization. This should produce a substantial increase in the haploid embryo rate if embryo development could be induced from more immature female gametophytes. The production of haploids of Beta vul~aris L. by in vitro culture of unfertilized ovules is already relatively efficient and could be used in breeding programs with this species. Improvements, however, can still be made, especially concerning
the embryos ~ ability to continue their development into a well structured plant. This depends on the conditions in which the donor plants are cultured as well as the sequence of culture media used. As of now the gynogenetic technique detailed here and applied to male fertile genotypes enables perfectly homozygous lines (doubled haploids) usable for breeding purposes to be produced. This work has been carried out within the frameof an A.D.A.R. ELF BIORECHERCHE program at work the Plant Breeding Laboratory of Universit6 ParisSud, Orsay, with the technical assistance of B. DOZOL.
REFERENCES Artschwager E, Starrett RC (1933) Jour.Agr. Research 47 : 823-843 Deuter M (1970) Genetica Polon. II : 219-225 Dote C (1976) Ann. Amelior. Plantes 26 : 647-653 Hosemans D, Bossoutrot D (1983) Z. PflanzenzGchtg. 91 : 74-77 Raquin C. (1982) C.R. Aead. Sc. Paris 294 : 335338 Sangwan-Norreel B S, Duhoux E (1982) Rev. Cytol. Biol~ V~g~t. Bot. 5 : 171-187 San Noeum L H (1976) Ann. Am~lior. Plantes 26 : 751-754 Seman I (1982) Genetika a.Slecht. 18(3) : 169-174 Seman I (1983) Biologia 38 : 1113-1122 Yang H Y, Zhou C (1982) Theor. Appl. Genet. 63 : 97-I04 Zhu Z , W u H (1979) Aeta Genet. Sinica 6 : 181-183
Plant Cell Reports (1985 ) 4: 300- 303
© Springer-Verlag 1985
Gynogenesis in Beta vulgaris L.: From in vitro culture of unpollinated ovules to the production of doubled haploid plants in soil D. Bossoutrot and D. Hosemans A.D.A.R., Laboratoire d'Am61ioration des Plantes, Bat 360, Universit6 Paris-Sud, F-91405 Orsay, France Received July 7, 1985 / Revised version received September 19, 1985 - Communicated by A. M. Boudet
ABSTRACT
MATERIALS AND METHODS
Our study concerns the different stages involved in the production of doubled haploid beet plants by in vitro gynogenesis. A histological study shows that the embryos obtained could come from the oosphere or the antipodals.Gynogenesis using male fertile plants gives nearly two haploid plants for i000 ovules cultured : in this case, the problem is to avoid plating fertilized ovules. The gynogenetic plants obtained display an endopolyploidy phenomenon at the root meristem level while their shoot meristem remains haploid. Colchicine doubling during in vitro vegetative propagation of buds has been carried out successfully.
Plant material
INTRODUCTION Haploid plants from unpollinated ovary or ovule cultures were obtained for the first time with barley (Hordeum vulgate L.) by SAN NOEUM in 1976. Since then several other species have been successfully cultured in this way - see reviews of YANG and ZHOU (1982) and SANGWAN-NORREEL and DUHOUX (1982). We have previously described the production of haploid plants by the in vitro culture of unpollinated ovules from male sterile plants in Beta vulsaris L. (HOSEMANS and BOSSOUTROT 1983). Gynogenesis is at present the most efficient of the haploidisation techniques tested in this species. Indeed, we immediatly obtained a 0,23% yield of 17 haploid plants for 7237 ovules cultured. Using the most efficient of the other techniques available (pollination of a male sterile Beta vulsaris by Beta vulsaris ssp esculenta i.e. red table beet), SEMAN (1983) obtained only a 0.013% yield of 12 haploids among 93125 plants observed after pollination. This paper describes four areas of further work embryo sac structure in situ was studied to correlate the nearly mature stage with stages of morphological development of the flowers ; the development during in vitro culture of the structures contained in the ovule was studied histologically ; the production of haploid plants led us to look for an artificial means of doubling the chromosomes; the technique of gynogenesis was further developed to adapt it to male fertile plants.
The histological study of the female gametophyte and the structures developing from it was carried out on ovules from four male sterile genotypes grown in the greenhouse. The determination of ploidy level and tests of colchicine doubling were made on gynogenetic plants obtained from male sterile genotypes. The gynogenetic technique was then applied to three genotypes of male fertile field-grown beets. In vitro culture The media used for the present study were those found to be the most productive during our first tests (HOSEMANS and BOSSOUTROT 1983) : ZHU and WU's (1979) basic medium n°l for wheat unpollinated ovule cultures and RAQUIN's (1982) basic medium for petunia anther cultures. These media were complemented by sucrose at 80g/l and 30g/l respectively and by benzylaminopurine (BAP) at Img/1. The pH was adjusted to 5.8 before autoclaving. The ovules were incubated at 27 °± 2°C with a 16 hour photoperiod (2000 lux). Histolo$ical techniques The ovules, taken either in situ at the culture initiation stage or after 15 to 21 days of in vitro culture, were fixed in acetic ethanol (25% acetic acid, 75% ethanol I00 °) during 48hours. They were then included in paraffin and a series of 15 ~ sections was prepared with a microtome. Sections were coloured with alcian blue and safranin red before being mounted in Canadian balm. Ploidy level determination The ploidy level was determined on i00 individual plantlets obtained by vegetative propagation of 14 gynogenetic plants, one month after their transfer to soil, by two techniques : chloroplast counts in the stomatal guard cells, found on the lower epidermis of young leaves, after staining with Lugol Solution (DEUTER 1970, SEMAN 1982). This method of ploidy determination has a high degree of reliability as shown previously (HOSEMANS and BOSSOUTROT 1985) chromosome counts in the root tip cells using t~e Feulgen technique or aceto carmine staining. Observations were made on three different roots for each vegetative propagated plantlet. -
-
301 Colchicine doubling To double the chromosome complement of the haploid plants obtained, colchicine was added at 50mg/l (C I ) or 100mg/l (C 2) to the vegetative propagation medium. The buds were placed on these media during time periods (t) which varied from 1 to 5 days (DORE 1976) : tl:24h, t~=48h, t~=72h, t4=120h. The control, carried ou~ simultaneouslyI consisted of a vegetative propagation medium without colchicine. One month later, in vitro chloroplast counts were made in the stomatal guard cells of young leaves. For each treatment tested, counts were made on the leaves of three adventitious buds taken from 5 different haploid plants. RESULTS AND DISCUSSION Stages for culture initiation of the female gametophyte A histological study of the embryo sacs in situ was made. When the ovules, still white, were plated after anthesis (in male sterile plants, this stage corresponds to the browning of anthers), the embryo sac is mature. At this stage, a longitudinal section showed the embryo sac to be on the side where the micropylar end is located (fig. IA and IB). It extends into a cavity, the caecum (fig.iC).
m
~'
"~ ~s e es ac
h
therefore able to relate the appropriate degree of maturity of the embryo sac to be in vitro ~ l a t e d to a stage in floral development.Thus, in male fertile plants, the effective ovules are removed from anthesis to the browning of the ovules on the floral scape. This stage of ovule development was found in the flowers at a level located from 10 to 30' cm under the end of the scape when it is still lengthening. Histological study of the structures originating from the ovules We also made sections of 30 ovules, taken after about 20 days of in vitro culture. Some of these ovules already contained structures which could be observed under a dissecting microscope. These sections show that in every case the structures observed are well-formed embryos (fig.2) Some of them undergo a secondary proliferation which results in calli surrounding the embryo. Auxin type impurities contained in certain elements of the culture medium particularily sucrose could enhance this callus proliferation. There was only one embryo per ovule. The root pole was oriented towards the micropylar end and the apical pole towards the body of the ovule. The embryos emerged in two ways : - either the root came out through the micropyle as in germination. - or the ovule integument, located over the e m b r ~ sac-caecum group, teared and allowed emergence of the embryo apex.
fF
0,5 L . S . in an o v u l e
Fig. i : A - Longitudinal section in an ovule (m) micropylar end, (h) hilum B - Detail of the embryo sac (es) showing the two synergids (s), the egg cell (e) or oosphere, the three antipodal cells (ac) and the extension into the caecum(c) C - Detail of the caecum (c) containing the fused polar nuclei (fpn). The egg cell and the two very vacuolized or degenerating synergids were present in all the sections observed. The same was true for the three antipodals located laterally near the junction point between the two cavities. In most cases the fused polar nuclei were found in the caecum. In only one case were they located in the embryo sac near the antipodals : this was probably a more immature ovule removed during anthesis. Even though we didn't observe the first embryogenic stages, which probably occur during the first week of in vitro culture, a study of the embryo sacs at culture initiation could explain the origin of the embryos. We confirm ARTSCHWAGER and STARRET's (1933) observations which showed that in most cases the synergids had degenerated and that the two polar nuclei bad fused. The haploid embryos obtained from ovules, therefore, cannot originate from these cells but either from the egg cell or from an antipodal cell. We were
Fig. 2 : Embryo (2) in an ovule Ovule : (M) micropylar end, (H) hilum Embryo : (V) vessels, (RM) root meristem, (SM) shoot meristem. Fig. 3 : A haploid plantlet originating from an ovule
302 Ploidy level determination in the gyno~enetic plants Ploidy levels, as determined by chromosome on the root tips and by chloroplast counts stomata guard cells are shown in Table I. ~
et System idy vel. System ~ ploidy level~
counts in the
Colchicine treatment concentrationduration mg/l hours
l
n n - 2n
C1 : n
2n
50
number observed
5
0
5
16
0
16
41
i
Number oI leaves observed
t 1 = 24
82.4
17,6
0
51
t 2 = 48
59
41
0
61
t 3 = 72
46
48
6
50
t 4 =120
40,5
53.2
6,3
47
t I = 24
35,3
64.7
0
51
70
i0
40
t 3 = 72
20
54,5
25,5
55
t 4 =120
18
42
40
55
0
18
4n
15
1
16
3
0
3
98
2
I00
Control
Table
%
t 2 = 48
18
number observed
4n
%
42
2n - 4 n
n - 2n - 4n
%
20
C 2 : i00 2n
nl2n
Ploidy level in the shoots
1 : Ploidy level in the root and shoot systems of i00 plants obtained by vegetative propagation of 14 gynogenetic plants.
For the root system measurements, 37% of the individuals observed show more than one ploidy level in the three roots taken at random, though each individual root was uniform. Conversely, at the shoot system level, 98% of the plants were haploid and 2% were diploid. Endopolyploidization which is relatively frequent in root meristems in beet, was not found in shoots where the apices were stable haploids. The regenerants originating from male sterile plants did not set seed after open pollination, thus confirming their haploid state. In order to obtain doubled haploid plants, doubling of the haploid buds had to be induced. This was done by colchicine treatment in in vitro culture. Doubling of the haploid plants with colchicine The efficiency of the different colehicine treatments during vegetative propagation in vitro is shown in Table 2. A two factor analysis of variance (concentration of colchicine X time period during which colchicine is applied) was carried out for the number of doubled leaves (2n or 4n). This showed that the effects observed after C. and ~ treatments were significantly differentlat a ~i~o probability level. The C_ colchicine concentration was significantly more efflcient since it induced over 76% (2n + 4 n ) d o u b l i n g against less than 43% for C_ . The effects of the i time periods during which colchicine is applied were also different at a I ~ probability level. It appears from Table 2 that : - for limited periods on media containing colchicine (24 or 48h), the rate of doubling increases with time but there are very few if any tetraploid plants. - for any given eolehicine concentration, from 72 to 120h, some of the treated plants are doubled a second time : the number of tetraploid plants thus increases while the number of diploid plants simultaneously decreases. The C^ ~ or C_tA treatments seem to be the best for the production of homozygous diploid plants. A sample of 38 plants found to be diploid in vitro were transferred to soil after rooting.
: 0
i97.7
2.3
256
Table 2 : Efficiencies of different colchicine treatments on shoots during vegetative propagation in vitro. At this time, a count carried out on their leaves revealed 22 haploid plants, 14 diploids and 2 tetraploids. These two counts show that 58% of the plants observed to be diploid in vitro proved to be haploid after transferring to soil. These very different results arise either from a disturbance in the physio logical functioning of the plantlet due to in vitro culture, or because of a partial doubling of the neoformed shoots resulting in diploid leaves while the apex remains haploid. The chloroplast counts made on leaves taken from in vitro cultured plants are thus not reliable. These counts will have to be made using young leaves from plants transferred to soil. The doubling rate of buds in vitro can probably be improved, especially by significantly reducing the size of the explants cultured. In fact the doubling of entire treated meristem would be easier using very young buds : since the explant is smaller, the colchicine diffuses and acts on a larger percentage of cells. The probability that bud neoformations will occur when starting with doubled cells is greater. It seems that the in vitro colchicine treatment is more efficient when rather high concentrations are used for brief periods. Gyno~enesis usin G male fertile plants The first gynogenetic haploid beets were obtained using male sterile genotypes (HOSEMANS and BOSSOUTROT 1985). The extension of this technique to male fertile genotypes was necessary. In this case the main difficulty consists in culturing unfertilized ovules only. In a first test the closed flowers were emasculated and bagged. Because of the mechanical tensions exerted on the vessels during emasculation, the ovules became necrotic before reaching the stage favourable to culture initiation. This type of emasculation, which is technically difficult to carry out and inefficient under our working conditions, was therefore abandoned. We then bagged the terminal part of the inflorescence in selfing bags without preliminary manual emasculation. Three genotypes having variable auto-incompatibility rates were used. The flowers were protandric with pollen already released when the last ovules on the stalk became mature. Therefore, there was a limited risk of selling in the last flowers. When culture was initiated we distingui$~e~the fertilized
303 ovules from the probably unfertilized ones on the basis of two criteria : the shape of the ovary, which when fertilized, has a conspicuous beak due to the development of the embryo in the embryo sac, and the size of the ovules. Furthermore, we eliminated all the embryos which develop during the first week of in vitro culture. Because of their rapid development they could only have come from self or cross fertilization. A total of 46 embryos was obtained from 1791 ovules cultured (2.56%); of these, 17 (36,9%) produced plants of which three were found to be haploid giving an overall haploid plant yield of 0,17 for one hundred ovules cultured. In fact, because of insufficient genetic markers the origin of 82% of plants which were diploid could not be determined. They could be the result of self-pollination, crosspollination, spontaneous doubling of haploids or development of fused polar nuclei. But these last two hypotheses are less probable because we observed no diploid plants resulting from ovule culture of a male sterile plant, except in one isolated case. For male fertile plants, the number of fertilized ovules cultured must be reduced. A method would be to plate ovules or ovaries well before the time of fertilization. This should produce a substantial increase in the haploid embryo rate if embryo development could be induced from more immature female gametophytes. The production of haploids of Beta vul~aris L. by in vitro culture of unfertilized ovules is already relatively efficient and could be used in breeding programs with this species. Improvements, however, can still be made, especially concerning
the embryos ~ ability to continue their development into a well structured plant. This depends on the conditions in which the donor plants are cultured as well as the sequence of culture media used. As of now the gynogenetic technique detailed here and applied to male fertile genotypes enables perfectly homozygous lines (doubled haploids) usable for breeding purposes to be produced. This work has been carried out within the frameof an A.D.A.R. ELF BIORECHERCHE program at work the Plant Breeding Laboratory of Universit6 ParisSud, Orsay, with the technical assistance of B. DOZOL.
REFERENCES Artschwager E, Starrett RC (1933) Jour.Agr. Research 47 : 823-843 Deuter M (1970) Genetica Polon. II : 219-225 Dote C (1976) Ann. Amelior. Plantes 26 : 647-653 Hosemans D, Bossoutrot D (1983) Z. PflanzenzGchtg. 91 : 74-77 Raquin C. (1982) C.R. Aead. Sc. Paris 294 : 335338 Sangwan-Norreel B S, Duhoux E (1982) Rev. Cytol. Biol~ V~g~t. Bot. 5 : 171-187 San Noeum L H (1976) Ann. Am~lior. Plantes 26 : 751-754 Seman I (1982) Genetika a.Slecht. 18(3) : 169-174 Seman I (1983) Biologia 38 : 1113-1122 Yang H Y, Zhou C (1982) Theor. Appl. Genet. 63 : 97-I04 Zhu Z , W u H (1979) Aeta Genet. Sinica 6 : 181-183
