1. GENERAL INTRODUCTION
This thesis includes the results of research on cytoplasmic male sterility and restoring fertility in petunia. In a general review of the literature different aspects, problems and possibilities of male sterility are discussed.
2. MATERIAL AND METHODS
Different varieties, families and lines of the garden petunia, Petunia
(Hook.) Vilm. were used:
a. The cytoplasmic male-sterile forms MS-Snowball and MS-Blue Bedder (original sterile form received from D. F. Jones), MS-First Lady (received from N.V. Sluis en Groot's Koninklijke Zaadteelt en Zaadhandel), and MS-Rosy Morn (received from R. Frankel).
b. The normal male-fertile forms F-Snowball (the lines used for the maintenance of MS-Snowball), Snowball S. en G., Blue Bedder S. en G. and F-Rosy Morn (from Frankel).
c. Restored fertile inbred lines (R-lines), developed at the Institute of Plant Breeding (I.V.P), Wageningen.
Crosses between the types a, b and c were made and studied (F 1
, F 2
and Back-crosses). In one trial the autotetraploid forms of Snowball S. en G. and Blue Bedder S. en G. were also used.
The methods of growing, grafting, propagation and producing tetraploids are discussed.
3. GENERAL OBSERVATIONS AND VIEWS ON MALE STERILITY IN PETUNIA
The development of the anther of cytoplasmic male-sterile petunia closely resembles that of the cytoplasmic sterile form of such crops as beet (ARTSCHWAGER, 1947), carrot (ZENKTELER, 1962) and onion (MONOSMITH, 1926). Development is normal up to and including meiosis (plate 1). Degeneration first appears in the tetrad stage. The individual microspores do not separate and remain inside the common wall of the pollen mother cell (plate 2). At the same time the tapetal layer expands to a marked extent (plate 3), after which the tetrads merge into a slimy amorphous mass.
Cytoplasmic sterile plants can be distinguished from normal fertile individuals because of the aberrant shape of the anthers (plate 4). Attendant symptoms may sometimes be observed: abnormal corolla, later flowering and abortion of the flower buds ('blindness'). The period of flowering of individual blooms and of the whole plant is longer with cytoplasmic sterile forms than with normal fertile forms because seeds do not form. 'Blindness' is the most limiting factor playing a part in the use of cytoplasmic sterile plants in a crossbreeding program.
GOLDSMITH (see DUVICK, 1959b), HAMILTON (1962, 1965), LANGE (1962), FERWERDA (1963) and EDWARDSON and WARMKE (1967) studied number and nature of genes restoring fertility in petunia. EDWARDSON and WARMKE (1967) found monogenic inheritance of fertility-restoration, the other authors saying that restoration of fertility is governed by more than one pair of genes, although none gave a clear genetic analysis.
A purpose of the present study, started in 1964, was to find out how many genes were involved in the restoration of fertility, whether cryptomeric: restorer genes occur in cytoplasmic sterile and normal fertile plants, and which influence environment has on fertility. The restorer material used was developed by FERWERDA (1963) and LANGE (1962). The parentage of this material is given in fig. 4.1; its fertility, based on the results of fertility estimation in the glasshouse, is in table 4.1. By selfing and crossing plants from table 4.1, and by crossing them with cytoplasmic sterile and normal fertile plants of different varieties, the material for 1966 was obtained (partly included in table 4.4). All this was planted in the field and its fertility was repeatedly judged.
Fertility was judged by estimating the contents of the anthers with the naked eye (and in doubtful cases with a lens, x 5). The degree of fertility or 'fertility value' of the anther (and of the Rower) was expressed in a scale 0 to 8 (0 = no pollen at all, 8 = normal quantity of pollen). The fertility of a plant at a certain observation was assessed from the fertility values of individual flowers, emphasis being placed upon completely sterile flowers. The fertility expression of plants with restorer genes in sterilizing cytoplasm proved very susceptible to external conditions, so that any one plant varied markedly in fertility between observations (cf. tables 4.2 and 4.3; figs. 4.2 to 4.5), especially with temperature (figs. 4.6 and 4.7). Classification of the plants (fert. classes 0 to 8) was based on fertility at all observations (see caption to table 4.4), with bias towards dates on which one or more completely sterile flowers were found (see tables 4.4 to 4.9).
Although the strong influence of environment made it extremely difficult to give a genetic analysis, the data (especially those in tables 4.4 and 4.5) indicated that restoration of fertility was governed by several pairs of genes. Besides one or two major factors, one of which was present in our homozygous lines with restored fertility (R (X)), there are accessory, sometimes cryptomeric, factors for fertility- restoration. Cryptomeric factors occur in cytoplasmic sterile and normal fertile plants in variable frequency and effectiveness, and find expression in differences in fertility within and between F 1
-progenies. Of the cytoplasmic sterile varieties MS-Rosy Morn possessed on average the most cryptomeric genes (in a few cases even a major gene seemed to be present) and MS-Blue Bedder the fewest; MS-Snowball and MS-First Lady were intermediate; compare the F 1
-progeny in tables 4.4 and 4.5.
On determining the correlation of fertility between parents in 1966 and their progeny in 1967 (tables 4.6 to 4.8) it was shown that plants with a similar genotype sometimes had been placed in adjacent fertility classes (e.g. 3 and 4 or 7 and 8) because of environmental variation, but that plants in classes wider apart were invariably different in genotype. As the sterile cytoplasm in MS-Snowball, MS-Blue Bedder, MS-First Lady and MS-Rosy Morn responded to the introduction of restorer genes in much the same way they are probably of the same origin.
The macroscopical assessment of the fertility values 0 and 1 was corroborated by microscopical observations (table 4. 10). The correlation between the microscopically determined contents of the anthers (fertility value) and the quality of the pollen (percentage normally developed grains stainable with acetic carmine) is given in fig. 4.8.
5. TRIALS WITH THE RESTORER MATERIAL UNDER CONTROLLED CONDITIONS
The influence of temperature on fertility expression was further analysed in the phytotron of the Horticultural Department of the Agricultural University, Wageningen, by treating some clones at 9 to 24°C (expt. 1, tables 5.1 and 5.2) and 12 to 24°C (expt. 2, tables 5.3. to 5.5), in steps of 3°C.
Normal fertile clones (F-Snowball) do not show a distinct optimum temperature, for fertility was almost equal at 24, 21 and 18°C. At lower temperatures the fertility of F-Snowball markedly decreased, presumably as a result of inbreeding depression of this material (I 9
). Cytoplasmic sterile clones (MS-Snowball and MS-Blue Bedder) had no pollen at all at any temperature: always complete degeneration. Plants with sterile cytoplasm and fertility-restoring genes invariably displayed a marked fertility optimum (see tables 5.2 and 5.4, figs. 5.1 and 5.2). The optimum temperature was high (18 to 21 °C) for homozygous plants with restored fertility (R (X)) and some fertility was found at all temperatures for these plants. For partially restored plants (all the other progeny types) the optimum temperature was lower, with decreasing number or effectiveness of the restorer genes about 18, 15 or even 12°C, strong or even complete sterility being found for these plants at the higher temperatures.
With plants exposed to alternating high and low temperatures within every 24- hour day (table 5.4, on the right) the average fertility deviated only slightly from that with continuous treatment at the higher temperature, even when the higher temperature was during the 8-hour night.
The pollen was usually of very good quality at the optimum temperature for fertility (plate 6). At temperatures less than optimum the number of elements recognizable as pollen was large, but the quality was bad (plates 7 and 8): incomplete degeneration. As a result of this the percentage 'good' pollen was relatively small at low temperatures (cf. table 5.8). At temperatures above optimum most of the microspores are completely degenerated, the relatively few remaining grains being fairly normal.
The flowers' response to temperature, i.e. the induction of degeneration of the microspores, was at about the meiotic stage (table 5.6). The time required for the development of the flower from the bud stage when the plant is susceptible to temperature to anthesis was markedly influenced by temperature and photoperiod during every 24-hour day in the post-induction period (table 5.7). The length of the daily photoperiod did not influence fertility perceptibly.
6. GENETIC ANALYSIS OF SOME SEGREGATION DATA
The response of 40 plants of the F 2
V9, (R x F-Snowball) (X), to the phytotron temperatures 12 to 24°C was further analysed (table 6.1). The plants could be classified into five groups, A to E, according to average fertility at different temperatures (fig. 6.1). This division into five groups may be explained by the effect of two pairs of genes for fertility restoration, the pair of major genes Rf 1
(complete dominance of Rf 1
at 12 and 15°C, Rf 1
intermediate at 18 and 21°C; complete inactivity at 24°C and the accessory pair of genes Rf 2
(the recessive type rf 2
being most fertile; table 6.3). To account for the total absence of completely sterile plants the effect of a third restorer gene (or gene complex), present in this material in homozygous condition, Rf 3
, had to be assumed.
A clear relation existed between the fertility classification based on field observation of the plants from which the clones derived (table 6.2) and that of the clones in the phytotron; compare table 6.4. Taking into consideration the major factor only, (for Pt 67
V9) the plants of the fertility classes 8, 7 and 6 were homozygous Rf 1
, those of the classes 5, 4 and 3 were heterozygous Rf 1
, classes 1, (0) and 0 including only plants of the genotype rf 1
. Plants of class 2 could be both Rf 1
and rf 1
, the latter probably only when having the genotype rf 1
, This means, that progeny of nearly the same parentage as that of Pt 67
V9) would have to display a segregation into 3 sterile and 13 more or less fertile plants. A good agreement with a 3:13 segregation was found in practically all the tested progenies of the types (R x F-Snowball), (MSSnowball x R) and (MS-First Lady x R) ; see tables 6.5 and 6.6. It is quite possible that fertility-restoring genes other than those described here are present in the varieties Blue Bedder and particularly in Rosy Morn.
7 LINKAGE BETWEEN FLOWER COLOUR FACTORS AND FERTILITY RESTORING GENES
In 1966 a strong correlation was noted in certain back-cross material between practically white flower colour ('white') and a high degree of fertility (table 7. 1). This correlation was further studied on some progeny of self-fertilized and mutually cross-fertilized 'white' and coloured plants (table 7.3). The data from table 7.2 helped to infer the genotype for flower colour of the parental plants from the segregation ratios in the progenies (table 7.4). From tables 7.5 and 7.6 it is quite clear, that there is a rather strong linkage between the major gene for fertility-restoration, Rf 1
, and the recessive allele, b 2
, of the basic dominant gene for flower colour. On account of the influence of minor fertility-restoring genes and of the environment on fertility expression, it was not possible to determine the recombination percentage accurately, but it was found to be ca. 10 %.
8. STUDY OF THE POSSIBLE VIRAL NATURE OF THE AGENT INDUCING STERILITY
The nature of the sterility inducing agent in cytoplasmic male sterile plants is controversial. EDWARDSON and CORBETT (1961) believe, that cytoplasmic male sterility in petunia is caused by a virus, ATANASOFF (1964) adducing arguments in favour of the hypothesis that perhaps all cases of cytoplasmic sterility should be ascribed to viral attacks.
The preliminary trials made by the present author concerning the possible viral nature of the sterility inducing agent in petunia failed to yield any positive result. With the electron microscope no special particles were found in the sterilizing cytoplasm. It was not possible to show transmission of the sterility inducing agent by rubbing leaves of fertile plants with expressed juice of cytoplasmic sterile plants, nor could the agent be inactivated by hot water treatment. Transmission of sterility by aphids or other insects was never observed in the field, nor could sterility be transmitted by grafting (see chapter 9).
9. TRANSMISSION OF STERILITY BY GRAFTING
FRANKEL (1956, 1962, 1964), EDWARDSON and CORBETT (1961) and BIANCHI (1963) described transmission of sterility after grafting between normal fertile and cytoplasmic male sterile petunia plants. The present study also aimed at exploring any possible mutual influence with a large number of graft combinations between fertile and sterile petunia forms (table 9.1). The uninfluenced cutting of each graft component was used as a control. At the same time some homoplastic and heteroplastic control grafts were carried out (table 9.2).
In the graft generation (E 1
; from Dutch 'enten' = to graft) no trace of mutual influence of the graft partners was observed. Also the progeny (E 2
), obtained by self- or cross-fertilizing graft components, did not provide any indication as to mutual influence (tables 9.4 to 9.9). Reduced fertile or even sterile flowers were sometimes observed on some E 2
-plants, but on the whole they were not more frequent than in the lines obtained from the control cuttings.
Critical analysis of the results and conclusions drawn by former authors justifies the inference that they have not shown convincingly sterility transmission by grafting either. It is only in the 1956 article of FRANKEL that somewhat grounded indications as to possible transmission of the sterilizing agent via the graft-joint have been included: some completely sterile E 2
-plants, and cytoplasmic inheritance of 'induced' sterility, although the absence of any sterilizing influence on the fertile graft components themselves is surprising. The results of EDWARDSON and CORBETT (1961) and of BIANCHI (1963) might be interpreted as selection for and 'condensation' of genes reducing fertility in normal, unchanged cytoplasm. The instance of sterility transmission followed by integration of the sterilizing agent in the genome described by FRANKEL (1962, 1964) was attributed to segregation of a pair of genes for fertility restoration in sterilizing cytoplasm, (S)Rfrf, or of a genic sterility factor in normal cytoplasm, (N)Msms.