GI L.--

1 "l

RECOMBINATION MECHANISMS AND TEE CONTROLS OF GENE-CONVERSION IN

ASCOBOLUS IMMERSUS

by

AGLAIA GHIKAS

B.Sc. (NATURAL SCIENCES, UNIVERSITY OF ATHENS, GREECE)

M.Sc. (GENETICS, WEIZMANN INSTITUTE OF SCIENCE - TEL-AVIV UNIVERSITY, ISRAEL).

A thesis submitted in part fulfilment of the requirements for the degree of Doctor of Philosophy of the University of London.

Department of Botany and Plant Technology,

Imperial College of Science and Technology,

London S.W.7.

JANUARY, 1978 Dedicated to

My Parents

CONSTANTINOS & IRENE AT1IANASIADIS for they cherished the life they passed to me 3 ABSTRACT In this work with Ascobolus immersus, gene conversion at the w-78 site and of various newly isolated mutations was studied.

For the w-78 site two new narrower ratio classes were detected, namely 6:2(2), 2:6(2), with numbers in parentheses giving the number of pairs of non-identical sister-spores contained in the respective octal'. The 6:2(2), 2:6(2) as well as the 4:4(2) octads detected here were used in the estimation of corresponding-site interference coincidence coefficients which showed that in agreement,with previous findings, corresponding- site interference in Ascobolus is weak.

The relative frequencies of symmetrical and asymmetrical hybrid DNA at the w-78 site were investigated. It was concluded that hybrid DNA in this site occurs mainly symmetrically, in two chromatids, without excluding the possibility of asymmetrical hybrid DNA occurring, but to a smaller extent. A new method was developed for carrying out this kind of project, which will allow the study of all possible arrangements of the w-78 site and its gene conversion controlling factors.

The gene conversion spectra of 40 mutants were studied in crosses to different wild-type strains. It was found that the type of conversion spectrum shown by a mutant often differs drastically from cross to cross, to the degree that there are mutants showing the spectrum of a frameshift in one cross and that of a base substitution in another cross.

Moreover it is found that there is an intergradation of the conversion spectra in connection with the relative frequencies of postmeiotic segregation, with no sharp distinction between frameshift and base substitution mutations.

The data are discussed within the framework provided by the current literature, on the bearing of both the chemical nature of the mutant and of other genetic factors on the type of spectrum shown by the mutant in a cross. 4 CONTNTS

Page No.

ABSTRACT 3 TABLE OF CONTENTS 4 LIST OF PLATES 8 LIST OF TABLES 9 LIST OF FIGURES 11 GENERAL INTRODUCTION 12 LITERATURE SURVEY 17 1. Genetic recombination 17 2. Gene conversion 18 3. Fidelity of gene conversion 19 4. The models: their evolution and predictions 20 i)"Dual hybrid chromatid" models 25 ii)"Single hybrid chromatid" models 27 5. Recombination related phenomena 29 0. The genetic control and factors affecting, recombination 32 7. Corresponding-site events 40

GENERAL MATERIALS AND METHODS 45 I General materials 45 1. The organism 45 2. The stocks 49 3. The Chemicals 52 4. Incubators 52 II General methods 53 1. Media 53 2. Sterilization, inoculation, crossing techniques collection of spores, scoring, germination of ascospores and photography 54 3. Calculations 55

PART A Gene conversion at the w-78 locus of Ascobolus i mme rs us 56 SECTION I The detection of 4:4(2), 6:2(2) and 2:6(2) segregations and their use in corresponding-site interference studies 57 5 Page No. 1. INTRODUCTION 57 a) Statement of problems Thvestigated 57 b) The markers 57 i) Properties of pfr-1 and fpr mutants 58 ii) Properties of gr-3 and w-78 mutants 62 c) The detection system 64 i) The detection of 4:4(2) 69 ii) The detection of 6:2(2) 75 iii) The detection of 2:6(2) 78 iv) The detection of other unique narrower ratio classes 81

2. RESULTS 83 a) 4:4(2) 86 b) 6:2(2) and 2:6(2) 89

3. DISCUSSION 93 a) aberrant 4:4 octads 93 b) The detection of 6+:2m(2) and 2+:6m(2) 96 c) Wider ratio octads 97 d) Corresponding-site events 98

4. CONCLUSIONS 99

SECTION II Analysis of the 5:3 and 3:5 classes in IICF x

LCF monohybrid crosses 101

1. INTRODUCTION 101

2. THEORETICAL CONSIDERATIONS 102 a) Cross: LCF wild-type x IICF w-78 102 b) Cross: IICF wild-type x LCF w-78 105

3. RESULTS 109

118 4. DISCUSSION

5. CONCLUSIONS 129 6

Page No.

SECTION III Background Studies 130 1. The nature of w-78 phenotype 130 a)The development of the "collapsing" character 130 b) w-78 in crosses to other white mutants 134 2. The re-isolation of "K" derived wild-types 135 a) Method of isolation 135 b) Tests of the "K" derived wild-types 139 3. Germination' and Fertility tests 143 a) Germination 143 b) Fertility 145 4. Search for outside markers 148 5. Attempts at higher temperature crosses 150

PART B The Study of white-spored mutants of different mutagenic origin 154 I INTRODUCTION 155 1. Problems investigated 155 2. Isolation of mutants 155 3. Induction of mutations 156 a) UV induced mutations 156 i)Induction with UV 156 ii) Frequency of UV induced mutations 157 b) NG induced mutations 160 i)Induction with NC 160 ii) Frequency of NG induced mutations 161 c) ICR induced mutations, 161 i)Induction with ICR 161 ii) Frequency of ICR induced mutations 162 4. Spontaneous mutations 162

II RESULTS 164 1. The mutants 164 2. Linkage relations 165 a)Locus I 170 b) Locus II 170 c)Locus III 171 d) Locus IV 171 7 Page No.

3. Conversion spectra 172

4. Aberrant asci with wicHr ratios 193

III DISCUSSION 199 Locus I 199 Locus II 205 Locus III 206 Locus IV 208 UV induced mutations 209 NG induced mutations 214 ICR170 induced mutations 214 Spontaneous mutations 216 Mutagen specificity and interchangeability of conversion spectra 217 Intergradation of conversion spectra 219 Wider ratio octads 223

IV CONCLUSIONS 224 OPEN PROBLEMS AND SUGGESTIONS FOR FURTHER WORK 1. The detection of 5:3(3)s and 3:5(3)s 22G 2. A new method for assessing the relative frequencies of symmetrical and asymmetrical hybrid DNA at a mutant site 229 3. Studies with the mutant UVKw8 236 4. Studies with some UV induced mutants 237 S. Studies with mutants of locus II 238 6. Studies with mutant BBm 239

LITERATURE CITED 240

ACKNOWLEDGEMENTS 251 8 LIST OF PLATES

Page No. Plate 1 46 Plate 2 46 Plate 3 61 Plate 4 61 Plate 5 67 Plate 6 67 Plate 7 68 Plate 8 68 Plate 9 72 Plate 10 72 Plate 11 73 Plate 12 73 Plate 13 74 Plate 14 74 Plate 15 131 Plate 16 131 Plate 17 132 Plate 18 132 Plate 19 133 Plate 20 133 Plate 21 152 Plate 22 152 Plate 23 153 9

LIST OF TABLES Page No. Table No.l. 50 Table No.2. 59 Table No.3. 60 Table No.4. 66 Table No.5. 85 Table No.6. 91 Table No.7. 109 Table No.8. 112 Table No.9. 113 Table No.10. 114 Table No.11. 115 Table No.12. 117 Table No.13. 137 Table No.14. 141 Table No.15. 142 Table No.16. 144 Table No.17. 145 Table No.18. 157 Table No.19. 161 Table No.20. 162 Table No.21. 164 Table No.22. a) 167 b) 168 c) 169 Table No.23. 175 Table No.24. 176 Table No.25. 178 Table No.26. 179 Table No.27. 180 Table No.28. 182 Table No.29. 184 Table No.30. 185 Table No.31. 186 Table No.32. 187 Table No.33. 188 Table No.34. 189 Table No.35. 190 Table No.36. 191 Table No.37. 194 10

Page No. Table No.38. 196 Table No.39. 198 Table No.40. 218 Pedigree Chart 51 11

LIST OF FIGURES Page No. Figure 1 102 Figure 2 104 Figure 3 106 Figure 4 107 Figure 5 120 Figure 6 122 Figure 7 138 Figure 8 140 Figure 10 158 Figure 11 225 Figure 12 226 Figure 13 227 Figure 14 228 12 GENERAL INTRODUCTION

The present work was carried out with the Discomycete fungus Ascobolus immersus and attempts to investigate problems concerning the mechanism and control of gene conversion.

Specifically it deals with gene conversion at. the w-78 locus as well as at sites of other newly isolated spore colour mutants. Conversion at w-78 is controlled by the closely linked factor (P), which, when homozygous, (in monoallelic crosses) confers high conversion frequency (8-18%) while, when hetero- zygous, low conversion frequency (2-6%). Two more factors (K) and (91) also closely linked to w-78, confer low conversion frequency in all cases for w-78. Work on this locus has been done in the past by Emerson and Yu-Sun (1967), who first established the high and low converting strains and partly the controlling factors involved, by Lamb (19724LambgWickramaratne (1973) who detected the occurrence of genuine wider ratio octads namely 7:1, 1:7, 8:0 and 0:8, and Wickramaratne & Lamb (1975), who studied the determinants of gene conversion properties in Ascobolus immersus and induced changes in the gene conversion spectra of w-78. These studies have established, among other things,that a) conversion events at corresponding sites in both pairs of non-sister chromatids of a single bivalent may occur in Ascobolus immersus and b) that environmental and genetical factors may alter the gene conversion frequency and spectrum of a single mutant, so that it may be of spectrum class C or D, as these classes were defined by Leblon (1972a), according to the conditions involved. In most of these studies mutant site w-10, closely linked to w-78, was used as well,

Site w-10 was not used here except for the re-isolation of the "K" wild-types. 13

Other complex questions need answering, such as "what is the mechanism of gene conversion at the w-78 locus?" or

"what is the difference in the control of gene conversion between high and low converting strains?", and "how faristhe conversion spectrum of a mutation . characteristic of its nature?"

Such questions, because of their complexity, have to be resolved into other simpler ones, which can be dealt with experimentally and where the results can be related to the current theoretical considerations. So the main questions that this work is trying to answer are as follows:

1. Do unique narrower ratio classes occur at the w-78 locus?

These classes, whose origin is given in Table 1, of

Lamb and. Wickramaratne (1973), are only obtained from hybrid DNA formation at the same site in both pairs of chromatids. Their detection in unordered octads is not possible unless a special system is developed, whereby auxiliary markers (visual or biochemical) are used, whose tetratype segregation defines the sister-spore pairs. Attempts were made to detect 6:2(2) and 2:6(2) classes - with numbers in parentheses giving the number of non-identical sister-spore pairs. The detection of these classes would be further evidence that all four chromatids pair intimately, two by two, at corresponding sites, and this would have a bearing on the synaptinemal complex structure.

2. Do aberrant 4:4s occur at the w-78 locus?

Their detection, if they exist, would serve three purposes: 14 a) They would be used in the calculations of the corresponding site interference coincidence coefficient along with the 6:2(2) and 2:6(2) and the wider ratio classes.

b) Their detection would imply that symmetrical hybrid DNA does occur at the w-78 locus, though not revealing its

extent or its proportion in relation to asymmetrical hybrid DNA formation, if the latter occurs at all.

c) It would provide a point of comparison between high and low converting strains of w-78, particularly after considering the results from the investigation of question 3. These

two types of strains of w-78 may differ either in the amount of hybrid DNA formed (with high converting strains forming

hybrid DNA more often than the low converting ones), or in the efficiency of correction of mispairs (with high converting

strains showing higher efficiency of correction than the low

converting ones), or in both. If we consider the simpler case, where high and low conversion strains differ in only one of these two factors, then:

i)with equal efficiency of correction any difference in the

frequencies of aberrant 4:4's between crosses involving

high and low conversion strains of w-78 would imply a difference in the amount of symmetric hybrid DNA formed

in the two strains, and

ii) a lower efficiency of correction of mispairs in the low conversion strains would show as significantly higher

frequency of aberrant 4:4's, relative to other conversion

classes, in crosses involving the strains compared with

the ones of the high conversion strains. 15

The detection of aberrant 4:4's was attempted using the same system as with the detection of 6:2 (2) and of 2:6 (2).

3. Does hybrid DNA at the w-78 locus form symmetrically or

asymmetrically or both and, in the last case, what are the

relative frequencies of the two kinds?

This is a basic question and its investigation was attempted through the analysis of the 5:3 and 3:5 ratios at the w-78 locus, with the help of conversion-controlling factors (P) and (K), which are closely linked to w-78,but do not often co-convert with it.

4. Do Leblon's (1972a) mutant types (A,B,C,D) occur in the

Pasadena strains of Ascobolus immersus and in either case,

to what extent do genetic factors, as opposed to the mutant's nature./ influence the conversion spectra of the

mutants?

This question was dealt with by isolating a number of white spore mutations, spontaneous or induced by different mutagens (UV, NG, ICR170) and crossing them to different wild- types of the (P), (K) and (91) types in order to obtain their conversion spectra. This way the mutants may have the chance to show different gene conversion spectra,which may or may not be accounted for by their presumed nature. Also by using (P), (K) and (91) wild-types, in the crosses with the mutants, there is a possibility of detecting other loci apart from w-78 which may be controlled by the same factors.

The investigation of these questions was undertaken 16

during this work as an attempt of providing answers to them, but also in the hope that through steps of this sort a better understanding of the gene conversion at the w-78 locus of

Ascobolus immersus and probably of the phenomenon as a whole will be achieved. 17

LITERATURE SURVEY

1. Genetic Recombination

Genetic recombination, as given by Whitehouse (1973), is the process by which new combinations of parental characters may arise in the progeny. Recombination of non-synnemal genes is basej on independent assortment of chromosomes in meiosis. Recombination of linked and synnemal genes usually takes place by crossing-over between homologous chromosomes and involves the reciprocal transfer of DNA segments. Recombination within the same gene is in most cases the result of interaction between alleles in meiosis (Whitehouse, 1973), a process known as gene conversion (Winkler, 1930). The term was re-introduced by

Lindegren, (1953), to account for aberrant ratios in the products of meiosis apparently arising from such interaction. Whitehouse and Hastings (1965) have suggested that if gene conversion is due to the correction of mispairing of bases in DNA, it may give rise to reciprocal as well as non-reciprocal recombinations.

Genetic recombination is a fundamental property of all living systems starting from the RNA-containing viruses and ending with higher plants and animals (Kushev, 1974), and as a source of variation is one of the principal factors in evolution. Recombination mechanisms, according to some investigators, are a basic factor in variation within the species

(Kushev, 1974), and they have been discussed as playing a significant part in the dynamics of evolution, particularly as it applies to the evolution of coadaptive gene clusters, (Fogel

& Mortimer, 1969; Chovnick, 1973). 18

Crossing-over is considered to proceed by breakage and reunion, after it was shown to be so for prokaryotes (Meselson & Weigle, 1961; Fox & Allen, 1964) and after Taylor, (1965), demonstrated that chiasmata correspond to points of breakage and rejoining. The unravelling of the mechanism of crossing- over and gene conversion is the target of, the models of recombination which are discussed later.

2. Gene Conversion In a wild-type (+) X mutant (m) fungal cross meiosis results in 2+:2m segregation in tetrads and in 4+:4m segregation in octads. Occasional deviations from the expected ratio 2+:2m may occur - mainly of the 3+:1m or 1+:3m type - and they are attributed to gene conversion. In octad-producing fungi, gene conversion may produce the following kinds of octad segregations: narrower ratio classes (Lamb, 1972) 6+:2m, 2+:6m, 5+:3m, 3+:5m, and wider ratio classes (Lamb, 1972): 8+:0m, 0+:8m, 7+:1m, 1+:7m.

After Lindegren's (1953) report of gene conversion,

Mitchell, (1955), demonstrated the phenomenon in Neurospora crassa, in studies with pyridoxine-requiring mutants and it has since been detected and extensively studied in other Ascomycetes like Sordaria and Ascobolus, using ascospore colour mutants

(references in Fincham & Day, 1971).

Several aspects of fungal intragenic recombination were demonstrated in Drosophila by Smith, Finnerty & Chovnick, (1970), in studies with the maroon-like cistron, i3allantyne & Chovnick,

(1971), with the rosy cistron and Carlson, (1971) with the 19 rudimentary locus in Drosophila melanogaster. This is an important advance showing that conclusions from fungal studies are likely to apply to Eukaryotes in general.

3. The fidelity of gene conversion

The question of fidelity of gene conversion at a heterozygous site refers to whether the process is a conservative one, utilizing the pre-existing genetic input or it generates novel information at this site. It was Olive, (1959), who stated that mutant alleles resulting from conversion, behave like their parents in m X + crosses. Case and Giles (1964) in tetrad analysis studies of the pan-2 locus of Neurospora crassa concluded, among other things, that in tetrads where mutant or wild-type alleles are represented more than the expected number of times, such alleles have so far proved indistinguishable from parental ones on the basis of mutation,

recombination and complementation tests. In 1970 Fogel & Mortimer, in studies with yeast, were able to show that a specific codon in the excised region was actually replaced during conversion with information identical to that carried

in the corresponding homologue,which was used as a template.

In 1969 Stadler & Kariya in their studies on intragenic

recombination at the mtr locus of Neurospora, concluded that

conversion is nota process generating novel codons, neither is there any special mutagenesis involved in conversion that

takes place as a non-reciprocal event. McCarron, Gelbart &

Chovnick (1974), working on intracistronic mapping of electrophoretic sites, confirmed the fidelity of information

transfer by gene conversion in Drosophila melanogaster. 20

However, there are some rare cases in which alleles, which had probably undergone conversion, had altered conversion properties. Such is the case of the h2a isolate of Sordaria , reported by Kitani & Olive (1967), whose altered behaviour

Lamb (1975) attributed to loss of nearby cryptic heterozygosity.

It is evident therefore, that gene conversion normally operates with complete fidelity in respect to the genetic constitution of the affected site.

4. The models: their evolution and predictions During the last fifteen years, molecular models were proposed which attempted to explain both recombination at the fine structure level and gene conversion. In the basic working model a recombination event starts with single strand break(s), followed by the formation of lengths of hybrid DNA, or heteroduplex. The resulting heterozygosity may be resolved by the removal of a length of either nucleotide chain and replaced by replication that uses the other chain as a template. Failure to remove the heterozygosity results in post meiotic segregation.

The configuration is resolved so that it may lead to crossover only, crossover associated with gene conversion, or gene conversion only. The hybrid DNA theory for genetic recombination is concisely reviewed by Pukkila (1977).

There are models of recombination in which one or more of these steps are missing. So Stahl (1969) and Paszewski (1970) presented models in which gene conversion does not involve repair mechanisms. These models fail to explain a big body of genetic data concerning the inequalities of conversion to wild- 21 type and to mutant, and they do not account for marker effects like map expantion. These objections were formulated by Holliday and Whitehouse (1970), and they seem to be generally accepted,though Hastings (1975) speculates that since some marker effects appear to act at a level other than repair of heterozygosity, it would seem that excision repair models may have the same problem. However, he concludes that the work of Leblon and Rossignol (1973) makes excision repair the most plausible hypothesis. Similar objections may be raised against

Moore's (1974) mechanism of genetic recombination by dynamic unwinding of DNA helices rather than initial strand breakage, where the critical role of the base mispair to promote repair is not considered.

The evolution of models of recombination is reviewed by Hotchkiss, (1974a), as leading from symmetrical complexes (Whitehouse, 1963; Holliday, 1964; Sigal & Alberts, 1972) to asymmetrical complexes (Boon & Zinder, 1969; Hotchkiss, 1971;

1974b), emphasizing the non-reciprocality•in the way in which the two parental DNA components enter into the formation of an intermediate complex. The key to model making, according to

Hastings (1975) seems to be to recognise the separate species - locus -, region - and allele-specific effects, and to allow the parameters of heteroduplex occurrence and length, the occurrence and length of excision repair and the recombination of outside markers to vary independently.

The hybrid DNA theory as proposed by Holliday (1964) and Whitehouse & Hastings (1965) provided the basis for experimental approaches. The essential features of a workable 22 model that attempts to explain gene conversion are, as recognised by Kitani and Olive (1967), the hybrid DNA formation and a repair mechanism.

Hybrid DNA formation is assumed in most models to be a necessary component of recombination and the evidence for the involvement of hybridity has been reviewed by Davern (1971). Broker and Lehman (1971) were able to observe recombination intermediates by electron microscopy, using polynucleotide ligase-deficient and DNA polymerase-deficient mutants of phage T4. The origin of the heteroduplex DNA is postulated differently in the different recombination models. Whitehouse (1969) and Hastings (1972) suggested that initiation points are provided by single strand gaps left at the ends of replicons after the premeiotic S phase. Holliday (1968) suggested that a nuclease binds to specific base sequences distributed at random along the chromosome. The enzyme has two binding sites and cuts at the same time strands of like polarity at homologous points. Sobell (1972), considering stabilized Gierer loops (Gierer, 1966, cited by Sobell, 1972), suggested that they may occur at operator regions and endonuclease may cut single strands at the loops' ends. Finally Hotchkiss, 1971, and Benbow,

Zuccarelli and Sinsheimer, 1974b, suggested that recombination is initiated by single strand "aggression" on a recipient duplex. This assumes that some kind of transient homologous association between three strands occurs, which is followed by a nick in one of the recipient strands.

Whatever the mechanism for initiation of heteroduplex, 23 it is now widely accepted that a likely intermediate is the

'half chiasma', in which polynuclectide chains of like polarity switch pairing partners at homologous points to give two reciprocal regions of hybrid DNA (Holliday, 1977).

The second necessary component of the recombination

models is the involvement of a repair mechanism. The evolution

of a mechanism for the repair of mismatched bases may serve the general repair of abnormal or damaged bases in DNA strands (Holliday 1974) since it bears no selective pressure for or

against elimination of mutants. Such removal of abnormal bases was reported by Lindahl (1974) where uracil is released from spontaneously deaminated cytosine. It is also possible that the repair of mismatches occurs during or immediately after

replication of nucleotides by DNA polymerase (Wagner and

Meselson, 1976).

It was suggested by Fincham & Holliday (1970) that the repair is promoted by the base mismatch itself and it is

now accepted that the pattern of conversion and postmeiotic

segretation is greatly dependent on the type of mutation used in the cross (Leblon 1972a; Yu-Sun, Wickramaratne & Whitehouse 1977). Leblon & Rossignol (,1973), have shown that conversion

at one site can decrease postmeiotic segregation at a closely linked site, which provides strong evidence in support of the hypothesis that gene conversion proceeds by the correction

of heterozygosity. Another piece of evidence for the correction hypothesis was the discovery of an enzyme in Ustilago maydis,

which can recognise DNA containing a single mismatch, reported

by Ahmad, Holloman & Holliday (1975). 24

The correction is not a strictly local event but involves degradation of about 1000 nucleotides and resynthesis

(Fogel and Mortimer, 1969).. This type of repair is necessary if two mutant sites closely linked co-convert, by being included in the same region of hybrid DNA.

For the purpose of this work two types of hybrid DNA model will be considered for their predictions (i) those where hybrid DNA is formed symmetrically in two homologous non- sister chromatids (Whitehouse, 1963; Holliday, 1964; Whitehouse & Hastings, 1965) called 'dual hybrid chromatid' models in

Wickramaratma & Lamb, (1978), and (ii) those where hybrid DNA may form in only one chromatid at a site although two chromatids are involved in its formation - called 'single hybrid chromatid' models in Wickramaratna & Lamb, (1978) - as in Whitehouse &

Hastings, 1965; Whitehouse, 1967; Stadler & Towe, 1971; Sobell, 1972).

There is experimental evidence from Sordaria and Ascobolus for both kinds of model. So the existence of aberrant 4:4 class, reported for several mutants (Kitani, Olive & El-Ani,

1962; Leblon & Rossignol, 1973; Kitani & Whitehouse, 1974 Ghikas & Lamb, 1977) and of a double aberrant 4:4 (with four pairs of non-identical sister-spores, Kitani & Whitehouse, 1974) support the operation of a'dual hybrid chromatid' mechanism. On the other hand Stadler & Towe, (1971), and Whitehouse (1974b) obtained evidence for postmeiotic segregation derived from a recombination intermediate with only one region of hybrid DNA. 25

There are crucial differences in the formulae and the interpretation required according to the two types of model in order to estimate the conversion parameters during octad or tetrad analysis of conversion events. The quantitative implications of the DNA repair model of gene conversion were analysed by Emerson, (1966), and Gutz, (1971a). A detailed estimation of conversion parameters is given by Wickramaratne & Lamb, (1978), as follows, for both 'dual' and 'single' hybrid chromatid models:

i) 'Dual hybrid chromatid' models Here, according to Wickramaratne & Lamb (1978) changes in

the conversion parameters are considered as follows: a) With all other factors (e.g. efficiency and direction

of correction) remaining constant, a change in the frequency of hybrid DNA formation would result in a

parallel change in all narrower ratio aberrant classes. This would show as a change in the total conversion frequency, but since it would not affect the frequencies

of the different aberrant classes in relation to each other,the direction of conversion as well as meiotic

segregation/postmeiotic segregation will not change,

unless amounts of correction enzymes are limited.

b) On the 'dual hybrid chromatid' models no correction gives aberrant 4:4; correction in one chromatid gives 5:3 or 3:5; correction in both chromatids in opposite

directions gives correction 4:4.

Two factors are affecting the correction system: the 26

efficiency and the direction of conversion. On the 'dual hybrid chromatid' models a change in the efficiency of correction would be followed by a parallel change in the

6:2, 2:6 and correction 4:4 octads. The 5:3 and 3:5 octads and the total observed conversion frequency may increase, decrease or remain constant depending on the extent to which the increased efficiency caused correction where there was no correction, and correction in both hybrid chromatids where there was origin'ally correction in one only.

The efficiency of correction ratio for dual hybrid chromatid events was given as follows:

2(6:2 + 2:6 + correction 4:4)+5:3 + 3:5 octads 2(aberrant 4:4)+5:3 + 3:5 octads

Undetected aberrant 4:4 and correction 4:4 octads could result in over and underestimation respectively of this ratio.

Referring to changes in the direction of correction on dual model events, the relative frequencies of correction to the wild-type and to mutant may be obtained by the formula derived by Lamb (1968), with the addition of the correction 4:4's to both parts of the ratio, (Wickramaratue & Lamb, 1978) direction of correction ratio on dual models: 2(6:2)+5:3 + correction 4:4 2(2:6)+3:5 + correction 4:4 27 ii) Single hybrid chromatid models

Here again according to Wickramaratne & Lamb, (1978) the situation is as follows: a) Changes in the frequency of hybrid DNA formation

have the same consequenc es as on dual hybrid

chromatid models, but there are two more variables involved in the frequency of hybrid DNA according to the single hybrid chromatid models: the chromatid

invasion frequency and the chromatid strand preference. These two variables have a bearing on the direction of conversion and they are taken into consideration for the estimation of relative correction to + and m ratio.

b) On single hybrid chromatid models no correction gives 3:5 or 5:3; correction to wild-type gives 6:2 or

correction 4:4; and correction to the mutant gives 2:6 or correction 4:4.

A change in efficiency of correction should result in a parallel change in the 6:2, 2:6 and correction 4:4 octads. The efficiency of correction ratio on the single hybrid chromatid models was given as follows:

6:2 + 2:6 + correction 4:4 octads 5:3 + 3:5 octads

Undetected correction 4:4 octads could result in underestimation of this ratio. The estimation of the direction of correction ratio, on the single hybrid

chromatid model, is more complex. On the assumptions 28 that both strands of a chromatid may invade the homologous non-sister chromatid with equal frequency and that the same mispairs have the same correction properties in either chromatid, Wickramaratne & Lamb

(1978) obtain the relative frequencies of Correction to wild-type and to mutant as:

(6:2) (3:5) (2:6) (5:3)

This estimate is unaffected by undetected correction 4:4's, because of its derivation, which takes into consideration the frequencies of invasion of a + chromatid by a strand of a m chromatid and vice-versa, as well as the frequencies of non-correction, correction to + and correction to the m. The consequences of the two types of invasion event, resulting alternative events and their frequencies of occurrence are given by Wickramaratne (figure 1-3.3, 1974).

The ratios given for the estimation of the efficiency and the direction of conversion, are mainly useful for comparing different crosses and conditions rather than for absolute determinations in any one cross.

The observed total conversion frequency of a specific site is partly determined by the frequency with which this site is included in a hybrid DNA region, and partly by the efficiency and direction of correction: therefore it most probably represents a rough minimum estimate of hybrid DNA formation at this site. 29

5. Recombination related phenomena..

The recombination models are also trying to account for phenomena usually accompanying recombination and gene conversion, namely exchange of flanking markers and marker effects like map expansion, polarity, interference.

The marker effects were reviewed by Stadler (1973), Catcheside, (1974), and Hastings, (1975), and may be strictly local, acting only on the site itself and not causing dependent effects in the rest of the locus, as is the case with Ascobolus in studies reported by Rossignol (1969), or they may extend to other parts of the locus as in the case of S. pombe, reported by Gutz (1971b). In this last case they may be due to cryptic heterozygosity and expressed in the way described by Lamb (1975). The marker effects reported in prokaryotes - Ephrussi-Taylor, (1966); Norkin, (1970) - were concluded by Stadler (1973) as failing to confirm the proposal of Leblon & Rossignol that mutants that arise from the same kind of event should behave alike in recombination. Yet this may not be an appropriate comparison due to the difference in the materials of study. Stadler also concluded that the specific marker effects are real and that their molecular explanation should involve a sequence larger than the molecular composition of the mutant codon alone. Hastings (1975) felt that the study of the marker effects necessitates the assumption of a precursor of the heteroduplex, with no evidence of its nature as yet.

Map expansion is the term used to describe the phenomenon in which the length of a long interval exceeds the 30

sum of lengths of the short intervals of which it is comprised. It was first described by Holliday (1964) and presented as evidence that heterozygosity interfered with the process of gene conversion and crossing-over. In 1970 Fincham & Holliday suggested an explanation of fine structure map expansion in terms of excision repair: in short intervals, recombination was caused only by the ends of this heteroduplex because inclusion of both markers in a length of heteroduplex would always lead to co-conversion and hence could not cause recombination. Another possible cause of map expansion comes from Ahmad & Leupold, (1973), who proposed that crossovers can be included in long intervals but not in short' intervals and this could be a cause of expansion on the assumption of a marker effect that reduces conversion but does not affect induced reciprocal events.

Polarity is seen in different ways by different authors

(Lissouba, Mousseau, Rizet & Rossignol, 1962; Murray, 1963), but it is generally viewed as preferential conversion of a marker on one site rather than a marker on another site in an interval. Polarity has been explained in the original models of recombination - Holliday, (1964); Hastings and Whitehouse

(1964) - by postulating that events originate between genes and extend to a variable length of heteroduplex, in which correction can occur. Recombination involving the heteroduplex ends contributes to the appearance of polarity, and it will• not show polarity if it results from independent conversion of two markers. Hastings, (1975), reviewing relevant data, concludes that polarity appears to be a property of the distribution of 31 free conversion heteroduplex and not of the conversion process, which extends in both direction after mispairs arise, nor of the opportunity to convert, which is much more common than free conversion heteroduplex.

Interference, which may be negative or positive and may occur between chromosomes, chromatids, or corresponding sites in the two pairs of non-sister chromatids, is not explained by the molecular models of genetic recombination.

Holliday (1977) offered a mechanism to account for the widespread phenomenon of positive chiasma or chromosome interference by postulating the existence of a DNA binding pairing protein which could form the "recombination nodules" reported by Carpenter, (1975) and whose depletion in the vicinity of an exchange point makes it unlikely that there can be stabilization of any other crossover nearby.

Exchange of flanking markers has been scored in many studies of intragenic recombination in Neurospora, Aspergillus, Ascobolus and Sordaria with frequencies showing a wide range.

It is among intragenic recombinants in yeast that the frequency of marker exchange has been consistently observed to hP about 50% by Hurst, Fogel & Mortimer (1972), and a theoretical basis for this observation was provided by the molecular hypothesis of strand isomerization developed by Sigal & Alberts (1972). This observation implies a random resolution of a half chromatid chiasma which in turn would imply that crossover and non-crossover events do not differ in any other way. However, there are several reports of 32 differences between the two types of event, as by Whitehouse & Hastings, _(1965); Bond, (1973); Kitani & Olive,

(1963); and Stadler & Towe, (1971). Stadler, (1973) and

Hastings (1975), review the relevant evidence and the latter suggests that a non-crossover event may be a unidirectional transfer of information or it may be caused by the occurrence of two half chromatid chiasmata close together, both of which become crossovers.

6. The genetic control of and factors affecting recombination There is ample evidence for a variety of genetic controls in recombination in different organisms (e.g. yeast, Neurospora, Podospora, Schizophyllum, Aspergillus, Ustilago, Ascobolus, Drosophila) and the relevant evidence is reviewed by Catcheside (1974, 1977) and Baker, Carpenter, Esposito, Esposito & Sandler (1976).

Two classes of mutants that control recombination are known: one of general effect and one of local effect.

Mutants of the first class have been reported by Badman, (1972), and Holliday, (1967) in Ustilago maydis and by Simonet, (1973) and Simonet & Zickler, (1972) in Podospora anserina. Genes of general effect on recombination are reported to yield recessive mutants, with increased sensitivity

to UV light and to result in the reduction or elimination of recombination and probably to sterility. In Neurospora crassa mutants of general effect in recombination, when homozygous in a cross, bring about complete infertility

(Schroeder 1970a,b). Equivalent to genes of general effect 33

are the genes of coarse control (Simchen and Stamberg, 1969)

in Schizophyllum commune, considered to have extreme effects on recombination ('off-on' type) throughout the entire genome, with only rare variants in nature and presumably controlling

steps (synapsis, breakage, repair and separations) which must operate sequentially.

Genes of local effect on recombination may be linked or unlinked to the target region and yield mutants that may

increase or decrease recombination. Mutants of this kind were first discovered in N. crassa, by Jessop & Catcheside (1965), between alleles at the his-1 locus, and the study of the

control of recombination by Catcheside's group (Catcheside, 1966; Smith, 1966; Catcheside, 1968; Catcheside

& Austin, 1969; Angel, Austin & Catcheside, 1970; Smyth, 1971) provides unambiguous evidence that, in this system, a major part of the heterogeneity in recombination frequencies is caused by gene functions that regulate exchange in specific regions of the genone. Two classes of regulatory elements covering recombination in this region have been found: rec+ mutants which decrease recombination and they are generally

Unlinked to the site of action and cog+ mutants, that increase recombination and are very closely linked to Ahe site of action. A third element, con, has been postulated as the

recognition site for the action of the rec+ gene product.

The rec and con systems appear to correspond to the systems proposed by Simchen & Stamberg (1969) to be the fine

controls of recombination reported for Schizophyllum 34

(Stamberg 1968, 1969a,b; Stamberg & Koltin, 1971). These genes of fine control in Schizophyllum had small effects on small highly specific region of the genome and they are of two kinds, one producing the controlling materials and the other constituting the recognition sites situated in the region in which control is exercised.

The existence of a third category of local control of conversion and recombination was described in Ascobolus, immersus by Emerson & Yu-Sun, (1967), with the w-62 locus, where the heterozygous condition of a closely linked control system suppressed recombination. A similar situation is the one reported by Rizet, Rossignol and Lefort, (1969), and by Girard & Rossignol, (1974), with the European strains of Ascobolus immersus, whereby genetic factors, closely linked to the affected loci and acting in the heterozygous condition, modify recombination by a factor of several hundreds, leading to nearly total suppression.

Finally, another example of locus-specific control of conversion seems to be the control of conversion at the w-78 and w-10 sites of Ascobolus immersus (Pasadena strains), as studied by Emerson & Yu-Sun, (1967), Wickramaratne & Lamb, (1978) and Lamb & Helmi, (1978). The characteristic features of the controls of conversion here are their incomplete dominance and their cis-trans position effects.

In all these cases variability in recombination at a specific site is clearly attributed - at least to a great extent - to other parts of the genome, that is other genes, 35 which directly or indirectly control the recombination process of the site under study.

Variability of recombination however may well be due partly to environmental factors or the influence of unknown and not easily recognized parts of the genome.

Recombination normally takes place in meiosis but the first preparatory stages evidently begin in S-period or at its end (evidence reviewed by Kushev, 1974). It is therefore expected that factors not directly connected with the conversion process may influence the conversion results.

These may be simply artifacts as in the case of differential maturation and bursting of asci reported by Lamb, (1967), or factors that truly affect the conversion process and not only the conversion results. So variations of conversion and/or recombination frequency following temperature changes prior to or during meiosis were reported in Neurospora by McNelly,

Lamb, & Frost, (1966), Landner (1970), and Lamb (1971), and in Sordaria by Lamb (1969a,b). It was suggested by Landner

(1970), that temperature does not necessarily act directly on the recombination process but can partly work by inter- mediary substances.

Induced changes in genetic recombination by temperature

treatment during premeiotic interphase were also reported

for maize (Maguire, 1968) and Drosophila (Grell, 1967).

Premeiotic irradiation was also found to have an

effect on recombination in Drosophila (Chandley, 1968) and 36

Chlamydomonas (Lawrence 1965, 1967). The latter also reported that cold-shock and metabolic inhibitors could alter recombination frequencies.

Apart from such environmental factors that influence recombination there are others which seem to originate from the genome itself. There are several reports showing that the nature of a mutant and its surroundings seem to have a great effect on the recombination or conversion given by the specific site.

Krumewska & Gajewski, (1967), reported the case of two mutants in Ascobolus immersus with probably the same position in the map (they do not show recombination in interallelic crosses) that have completely different conversion frequencies, and it is possible that this is due to differences in the nature of the mutants involved. However, it is the work done by Leblon (1972a,b) and Leblon & Rossignol (1973) that gave much insight in the role that the nature of the mutant plays in the type of conversion pattern shown. The conversion spectrum was defined by two criteria: the amount of postmeiotic segregation and the direction of conversion (predominantly to the wild-type or to the mutant). The study involved several white spore colour mutants of Ascobolus immersus, spontaneous or induced by a mutagen like ICR170, EMS or NG. Leblon (1972a,b) found a strong correlation between the nature of the mutant and its conversion spectrum: frameshift mutations had low postmeiotic segregation with deletions showing predominantly conversion to the wild-type (A class) and additions showing predominantly conversion to the mutant 37

(B class). Base substitutions had high postmeiotic segregation and conversion may be either predominantly to the wild-type (C class) or to the mutant (D class). This was done with the help of reversion mutations and information on the types of mutants induced by the different mutagens, (Auerbach, 1976).

A similar report came from Yu-Sun, WickramarattAe & Whitehouse, (1977) with mutants in Sordaria brevicollis, where similar classes of mutants were identified but with less difference between the low and high postmeiotic classes. From these reports it seems reasonable to conclude that the gene conversion spectrum of a mutant gives a definite evidence for the nature of the mutant.

However, this does not seem to hold true in all cases. So classes C and D seem to integrade as is revealed by the work of Lamb & Wickramaratue (1975), Yu-Sun et al (1977) and Leblon (1972a), while the latter reported also on a number of mutants being of the AB type (intermediate class between classes A and B). Earlier in 1969, Paszewski and Prazmo, working also with Ascobolus immersus, confirmed that the pattern of intragenic recombination shows mutant specificity as well as cross specificity. And recently a study was carried out by Moore & Sherman (1977) with 16 mutants of Saccharomyces cerevisiae, containing either the addition, deletion or substitution of single base-pairs, located within a defined segment of the gene that corresponds to the 11 amino acid residues at the amino terminus of iso-l-cytochrome C; approximately half of these mutants had alterations of the 38

AUG initiation codon, some at the same base-pair. The wild-type nucleotide sequences, corresponding to the mutant alleles used in this study, were also known. From their data with crosses of mutants in the same base pair to other mutants it is apparent that different base-pair changes at the same position can result in different frequencies of recombination, although the differences are dependent upon the mutants to which they are crossed. These results indicate that the variation in recombination frequencies is not simply related to the type of mutations, and that two mutant sequences may interact to increase or decrease each other's influence on recombination frequencies. It is concluded by the above authors that the specific factors responsible for the variability among diploid strains containing the same heteroallelic cyc 1 pair are not presently understood, except that a large part appears to be due to the influence of undefined genetic background.

This undefined genetic background may well contain a certain degree of cryptic heterozygosity, whose effects on gene conversion were discussed by Lamb, (1975), in an attempt to explain data such as those of Kitani and Olive (1967), that seem to contradict the fidelity of gene conversion, or those of Emerson & Yu-Sun (1967) with the w-62 locus of Ascobolus immersus giving cis-trans effects in crosses to different wild-types. Cryptic mutations - that is, undetected changes in the genetic material - could affect gene conversion results because,when heterozygous,they cause mismatched base- pairs in hybrid DNA in the same way as known mutations, with 39 a possible bearing on polarity and similar effects. It is noticeable however that the marker effects described by

Moore and Sherman (1977) for the cyc 1 gene of S.cerevisiae are distinct from polarity in gene conversion that Lamb (1975) suggests may be due to cryptic heterozygosity in that they derive from the mismatch involved and perhaps the surrounding sequence, whereas polarity effect appears to derive from the position of the marker within the gene.

Nevertheless, cryptic heterozygosity, acting as.the detected one, namely by triggering excision-repair mechanisms, is bound to affect conversion results, if correction is needed for conversion.

There are several indications that heterozygosity promotes excision and restricts chiasmata formation. So Frost (1961) and Stadler &Towe (1962) found that increased heterozygosity reduced crossing-over in Neurospora, while Stadler & Kariya (1969) found that intragenic recombination was reduced by non-homology, again in Neurospora. Inbreeding or back-crossing however did not increase crossing-over frequency in studies in Schizophyllum by Simchen & Connolly

(1968) as it did in Neurospora (Stadler & Towe, 1962).

In summary it may be said that there is strong evidence that the nature of the mutant plays a vital role in defining the gene conversion spectrum of a mutant, but it is rather unsafe to turn the argument round and decide about the nature of the mutant solely on the basis of its gene conversion spectrum. 40

7. Corresponding-site events

Corresponding-site events are considered the events that involve hybrid DNA formation at corresponding sites in all four chromatids of a bivalent (Lamb & Wickramaratne 1973).

Correction in each chromatid to wild-type would result in a 8:0 octad and to mutant in a 0:8 octad. Failure of repair in one chromatid would result in the production of 7:1 or 1:7 octads. Failure of repair in more than one chromatid would produce narrower ratio classes, namely 6:2 (2) or 2:6 (2),

5:3 (3) or 3:5 (3) and 4:4 (4), with numbers in parentheses giving the pairs of non-identical sister spores contained in each class. The events in each of the pairs of non-sister chromatids that lead to the formation of the above classes are given in Table 1 of Lamb & Wickramaratne, (1973).

From all classes resulting from corresponding-site events, 5:3 (3), 3:5 (3) and 4:4 (4) could not arise from single hybrid chromatid occurrence, since they require a

2:2 (2) segregation in one or both pairs of non-sister chromatids, which is not predicted from the single hybrid chromatid hypothesis. Wider ratio octads as well as 6:2 (2), and 2:6 (2), may result on both, dual and single hybrid chromatid hypotheses when hybrid DNA formation involves all four chromatids. Aberrant 4:4 (2) octads are the result of a two chromatid event according to the dual hybrid chromatid models,but of a four chromatid event according to the single hybrid chromatid models.

Fig. 1.7- 1 in Wickramaratne (1974), gives the octad ratios resulting from the involvement of both pairs of 41 chromatids in hybrid DNA formation, according to the single hybrid chromatid hypothesis.

The detection of what might be corresponding-site events goes as far back as 1953 when Lindegren observed

4:0 and 0:4 ratios in Saccharomyces cerevisiae, and their occurrence has since been confirmed by Lamb (1972), Leblon (1972a), Lamb & Wickramaratne, (1973), Kitani and Whitehouse,

(1974), Yu-Sun et al (1977), Ghikas & Lamb (1977). From all possible corresponding-site events only the 5:3 (3) and 3:5 (3) have not yet been discovered.

It is generally accepted that two of the four meiotic chromatids are involved in a single genetic exchange or recombinational event at any one interval. However, Emerson (1969) suggested the possibility of occasional involvement of more than two chromatids, and all authors that reported the detection of wider ratio octads or of 6:2 (2), 2:6 (2) and 4:4 (4), attempt to analyse them as involving events in more than two chromatids, as for example in Table 6 in Kitani & Whitehouse (1974). The detection of genuine corresponding site events shows that both pairs of chromatids can be involved in hybrid DNA formation - and therefore have intimately paired - at exactly the same point in a bivalent, (Lamb & Wickramaratne, 1973).

The possibility of having both pairs of non-sister chromatids involved in recombination events at corresponding sites poses certain difficulties to some models of chromo - 42 some pairing and of the formation of synaptinemal complex

(Moses, 1968; Westergaard & von Wettstein, 1972), a structure which holds the paired homologues in register and from which the chiasmata originate. The latest review on the assembly of the synaptinemal complex is that of von Wettstein, (1977), where all the findings upon which the step by step formation of the synaptinemal complex is based are discussed in relation to mechanisms of chromosome pairing and chiasmata formation.

A feature of the synaptinemal complex, which has been shown to occur in many Ascomycetes, including Ascobolus (Zickler, 1973), as it is presented up to now is the participation in it of only one pair of non-sister chromatids, at any one point. The function of the synaptinemal complex is generally thought to consist of holding the homologues in register throughout the length of the bivalent containing it - so that crossover sites can then somehow be established along it, with their distribution not being controlled by the complex itself - and perhaps to provide machinery for the crossover process. It is also generally accepted that gene Conversion and crossing over originate from the same basic process, and in this context the synaptinemal complex cannot account, as it is, for the occurrence of four chromatid events, which have been shown to occur, and, as it was pointed out by Lamb (1977a), they predict complete positive corresponding-site interference. This interference in Ascobolus immersus, where four chromatid events have been shown to occur, is weak, if present at all,

(Lamb & Wickramaratne, 1973). 43

Lamb (1977b) suggested that the arrangement for chromosome pairing which is most compatible with the data on four-chromatid events is a compact one, in which none, one, or both pairs of non-sister chromatids can pair at any point. The lack of strong negative corresponding-site interference (:interference between the two pairs of non- sister chromatids of a bivalent in hybrid DNA formation at exactly corresponding sites) provides evidence against the extensive occurrence of a linear or linear/compact arrangement of chromosome pairing (Lamb, 1977b), which do not allow for more than two-chromatid events at corresponding, sites.

On the other hand the very function of the synaptinemal complex along with other commonly accepted aspects of homologous chromosome pairing is challenged by

Maguire (1977), who provides evidence suggesting that the synaptinemal complex may somehow function in the crossover process but its complete deployment throughout each normal bivalent may serve some other role. This suggestion may be especially significant in view of the existence of strains with high levels of mitotic recombination, e.g. the isolation of a mutant,(a duplication) in Aspergillus (Parag & Parag, 1975), that shows mitotic recombination, mostly by conversion in a frequency slightly higher than in the equivalent diploid, while the existence of synaptinemal complexes have been reported to occur almost exclusively in meiosis.

In summary it may be said that corresponding-site events have been shown to occur and actually in organisms 44 where well-developed synaptinemal complexes have also been shown to occur. The generally accepted form of the synaptinemal complex does not allow for such events and either this has to be modified or a variety of mechanisms regulating recombination may exist to account for species- specific adaptations. 45

GENERAL MATERIALS AND METHODS General Materials

1. The organism

The organism used throughout this work was the fungus Ascobolus immersus (Pasadena strains). Classification

(Ainsworth & Bisby, 1971).

Kingdom: Fungi

Division: Eumycota

Subdivision: Ascomycotina Class: Discomycetes

Sub-class: Pezizales

Family: Ascobolaceae

Genus: Ascobolus

Species: Ascobolus immersus Etymology (van Brummelen, 1968)

Ascobolus from .Greek: (;:akcis: leather-sack (ascus)

and faAL: to throw

immersus from Latin: immersus: plunged, below the surface

Most of the information on "th rganism" comes from Van

Brummelen (1968), unless another reference is given. Ascobolus

immersus is a coprophilous fungus that has been isolated from

dung of cow, horse, sheep, goat, nilgai, antelope, elephant, dog, hare and rabbit. It seems to be cosmopolitan and the commonest

member of the genus, its distribution being limited by the

occurrence of herbivorous,animals. The mycelium of pale, almost

white colour, is coenocytic and hyphae are septate (septa with

pores), and branched. Apothecia are scattered or gregarious, at 4 6

br

O at,

%Iv

4

I .4 •

PLATE 1 Apothecia with intact asci from a + x m cross in

Ascobolus immersus

PLATE 2 Spore octads from a cross between two unlinked white

spore mutants, dehisced on collecting lid. 47 first often immersed, then erumpent or superficial, about 10mm diameter and 0.7mm high, with receptacle globular or pear shaped, at first closed, then irregularly opening near the top, Plate No.l. Asci are cylindrical and show an extraordinary increase in volume during ripening, when their walls are stretched enormously and their tips curve towards the maximal light intensity: they are positively phototropic. Development of the asci is asynchronous and dehiscence period may cover two weeks.

The uninucleate ascospores, usually 8 per ascus and oblong shaped, are about 601A long and 344 wide, unordered and covered with a gelatinous envelope. The contents of the ascus can be shot away over a distance 500 times the length'of the stretched ascus. In this way the ascospores are collected in octads on a collecting lid, Plate No.2, and studied by means of unordered tetrad analysis. Wild-type'ascospores carry a dark red pigment, but mutant strains with colourless ascospores have been isolated, giving a typical 4:4 segregation in crosses with the wild-type. In abnormal cases binucleate spores are produced, larger than usual. Two such larger spores were germinated during this work, but were infertile in crosses among themselves and to several wild-type and mutant strains. Groups of spores containing such large spores were left out of consideration during scoring.

In Ascobolus immersus more than 2,000 ascospore mutants are known, only differing from wild-type by a single gene

(Lissouba et al 1962). 48

The species is heterothallic, the plus and minus mating types being morphologically alike, distinguishable only through mating behaviour. Crosses can easily be controlled by this simple incompatibility.

Ascobolus immersus has a haploid set of chromosomes whose number vary among the authors from 16 (Esser & Kuenen, 1967) to 12 (Zickler 1973) to 14 (Rossignol, personal communication to Lamb). Linkage relationships are not very well known (Yu-Sun,

1964 and Emerson & Yu-Sun, 1967).

The advantages of Ascobolus immersus as an experimental organism for genetic research are its rapid life cycle (12-14 days), its large and easy to handle spores, and its shooting of whole octads which can be collected without the need to dissect apothecia. But there are disadvantages too, such as the lack of asexual spores which hinders the use of selective methods, like filtration enrichment, for isolating auxotrophic mutants

(Fincham & Day, 1971). The presence of unordered octads rather than ordered tetrads, necessitates the use of indirect methods as in Whitehouse (1957), for obtaining information on the second division segregation frequencies and centromere distances. Linkage relations can be determined through two- factor crosses by comparison of the ratios of PDs NPDs and Is and from recombination frequencies amongst random spores. Further disadvantages, especially during this work (PART A, section

III), have been the loss of fertility in the strains and the

bad germination of the ascospores. The Pasadena strains of 49

Ascobolus immersus used here can be cultured in completely defined nutrient media (Yu-Sun, 1964).

2. The Stocks All stocks were of the Pasadena strains of Ascobolus immersus (Yu-Sun 1964) and their origin is given in the pedigree chart.

Strains P5- and K5+ (standard wild-types), w-78+, w-78-, w-10+ w-10-/I 577w-78+ 549w-78-, 415w-10+ 421w-10+$ and 91-2R+ were all kindly supplied by Dr. B.C. Lamb, who had obtained them from Professor Sterling Emerson. 91-2R+ is a derived wild-type strain of uncertain origin (Emerson, personal communication to Lamb). w-78 and w40 are closely linked white-spored spontaneous mutants, presumed to be allelic (Emerson 1969), and located in linkage group VIII (Yu-Sun 1969). IWT3+, EC11w-78-7-,- EC2w-78.5+ were re-isolates made by R Wickramaratne (1974) Sl-w-10-, BBR1-, BIIR2-, BHR1- are re-isolates made by Lamb. KIV+, 92-, 61-, 42-, 352+, 1w3KEC+, and 5wSKEC+ are re-isolates made by S. Helmi 9r+, 3r-, 25-5R+, A-4R, 9KS+, 2K1-, A-w-10.4+, 9-4w-78-, 22-w-78-, 1Kw-78.3- are re-isolates made by the present author.

The (P) strains will be referred to as "high conversion frequency" (IICF) strains: they give conversion frequencies of 8% to 18% in monohybrid crosses wild-type (P) x mutant (P) (TABLE 1). 50

The (K) strains will be referred to as "low conversion frequency" (LCF) strains: they give conversion frequencies of 1.25% to 6% in monohybrid crosses wild-type (P) x mutant (K)) wild-type (K) x mutant (P) or wild-type (K) x mutant (K), (.TABLE 1).

The (91) strains are LCF, often lower than (K) and they carry control factors Affecting conversion at w-78 derived from 91-2R (TABLE 1).

Unless otherwise stated the term "conversion frequency" includes both meiotic and postmeiotic aberrant segregations, and all octad ratios are given with the wild-type strain first and the mutant strain second (+:m).

TABLE 1 The (P), (K) and (91) strains in crosses defining their conversion frequency characteristics

Cross Octad classes % total 4:4 6:2 2:6 5:3 3:5 NRCF* octads BHR1-(wt,P)xEC2w78-5+(m,P) 82.48 10.30 3.55 1.39 0.25 15.50 3,152 9r+(wt,P)x22-w78-(m,K) 97.64 0.98 0.93 0.24 0.29 2.45 6,102 A-4R+(wt K)xEC11w78.7-(m,P) 95.85 3.46 0.20 0.38 0.07 4.11 4,130 K1V+(wt,K)x22-w78-(m,K) 98.75 0.68 0.13 0.22 0.21 1.25 6,723 91-2R+(wt,91)xEC11w78.7-(m,096.78 2.08 0.54 0.30 0.28 3.20 5,322 91-2R+(wt,91)x22-w78-(m,K) 97.09 1.25 0.83 0.52 0.29 2.89 6,861

*narrower ratio conversion frequency Pedigree chart of the Pasadena strains of Ascobolus immersus P5- x KS+

91-2R+(91) U+(P) w10(K) w- 7Y(K ) 1:7 2.61, f 2:6 577w-78+(P) 549w-78-(P KIII- XIV+ 1wT3+(P) 16:2 179wt 2:6 2:6 415,w-10+(P) 421w-10+(P) 1 6:2 Slw-10-(P) J ?r+(P) 3r-(P)

A-w-10.4+(P)

t2:6 1 2:6 2:6 1 6:2 6:2 6:2 EC11w-78.7-(P) EC2w-78.5+(P) 9-4w-78-(P) 25-5R+(K) A-4R+(K) 22-w-178- 1 BBR1- (P) 1Kw-78.3- t 2:6 2:6 1w3KEC+ (P) 5wSKEC+ (K)

B11R,2-(P) BI1111-(P) 1 92- (P) 61- (P) 42-(91) 352+(91)

Figures in parentheses give the conversion control characteristics of the strains in relation to w-78 and w-1C The octad ratios above certain strains give the type of octad in which the respective strain arose through a conversion event. Exception to this is strain S1-w10- (isolated in a 6:2, but not arisen by conversion) Wavy Line: details unknown

FOOTNOTE: letter w denotes a white spore mutant 52

3. Chemicals The chemicals used and their respective producing companies were as follows:

- All chemicals for the salts stock solutions, Vogel trace

element solution and sucrose: Analar - Biotin, pronase, thiamine, phenylalanine, vitamins (for the vitamins solution), nucleic acids from yeast, monopotassium

phosphate, di-sodium phosphate, and L-asparagine: BDH chemicals.

Bacto-agar, bacto-peptone, bacto-casamino acids, and bacto yeast extract: Difco - Methyl 4-hydroxybenzoate (mould inhibitor) and sodium

desoxycholate: Hopkin & Williams - DL-p-Fluorophenylalanineand chloramphenicol Sigma

- Mycological peptone and agar No.3: Oxoid - Fibrous cellulose powder: Whatman

- N-Methyl-N'-nitro-N-nitrosoguanidine (NC): Aldich Chemicals Co.

- Acridinc mustard ICR170, was provided by Dr. H.J. Creech to to Dr. B.C. Lamb. - Furfural was provided by Dr. M.R.T. Wickramaratne

4. Incubators as described by Wickramaratne (1974) 53

GENERAL METHODS

1. Media All media were prepared with distilled water and all concentrations, unless otherwise stated, were in weight/volume percentages. Vogel's micronutrient solution, phosphate salts plus biotin and magnesium sulphate salt were all kept in separate sterile stocks, which were renewed approximately every three months. The minimal, crossing, complete and germination media were those of Yu-Sun (1964), as modified by Wickramaratne (1974).

The germination medium was recently replaced by a horse dung germination medium, introduced by S. Helmi (personal communication). It was prepared by shaking (at 150. r/m) or stirring in a warm (50°C) magnetic plate 60g/1 of air dried horse dung for 30 minutes. The dry matter was then filtered off through double muslin cloth and the remaining solution centrifuged at 5,000 RPM, for 30 minutes. The supernatant with 2% bacto-agar was autoclaved and used fresh.

The spore collecting medium was a 1.5% water agar (oxoid No.3) plus 7% (v/v) of the mould inhibitor stock solution, which was a 10% solution of methyl 4-hydroxybenzoate in ethanol (96% or 98%). This inhibitor was to prevent spontaneous germination of ascospores on the collecting lids (method devised by Lamb, Personal communication).

The detailed, composition of nutrient media was as follows: 54 Complete medium: 10m1 of phosphate and biotin solution (25g. KH2PO4, 30g' K2004,

1Mg biotin in 480m1 distilled water), 5m1 of a solution containing 25g MgSO4.71120 in 250m1 of distilled water, 0.1m1 Vogel's trace element solution (Vogel, 1956), 3g yeast extract, lOg glucose, 0.5g casaminoacids and 18gr Difco bacto-agar in

980m1 distilled water.

The crossing medium was of the same composition as the complete medium but contained in addition 2g cellulose powder.

The minimal medium in addition to salts, glucose and bacto-agar at the concentration of the complete medium contained 0.5% asparagine and 0.01% (mg/1) thiamine.

These media, when necessary, were appropriately supplemented to satisfy experimental conditions or growth requirements of the strains under study.

The germination medium used before the introduction of

the horse dung.one consisted of 4% water agar (Difco bacto-agar) plus 1.25% bacto peptone. When big samples of spores (over 50 per dish)were handled after treatment with pronase, bacto-peptone was replaced by mycological peptone, which helped reduce bacterial

contamination.

2. Sterilization, Inoculation and crossing techniques, collection of spores, method of scoring, and photography were as described by Wickramaratne (1974). Germination of ascospores

was basically the same as in Wickramaratne (1974) but sometimes 55 longer periods of heatshock were given to red spores (up to

2 hours), while granular ones were not heat-shocked at all.

3. Calculations were in part with a Sony Sobax 1CC-2550 programmable electronic calculator and mainly with a Commodore SR4120D electronic calculator. -56-

PART A

Gene conversion at the w-78 locus of Ascobolus immersus 57 SECTION I

The detection of 4:4(2), 6:2(2) and 2:6(2) segregations and their use in corresponding-site interference studies.

1. INTRODUCTION a) Statement of problems investigated

The main investigation under this section is concerned with i) the existence or not of the aberrant 4:4's (4:4(2)), in

crosses of wt(P)x mutant (P) and wt(P)x mutant (K), and their

absolute and relative frequencies; ii) the detection of 6:2's and 2:6's with postmeiotic segregations - 6:2(2) and 2:6(2) - in the above two types of cross, and the estimation of their

frequencies, which along with those of the 4:4(2) and the

wider ratios, could be used in studies on corresponding-site interference, as defined by Lamb & Wickramaratne, (1973). b. The markers The detection system used here includes four markers, all

unlinked to each other. These are:

1. w-78, a spontaneous, white ascospore mutation (Emerson and Yu-Sun, 1967) which, depending on closely linked genetic controlling factors, gives high conversion

frequencies (HCF) or low conversion frequencies (LCF).

2. gr-3, a UV-induced mutation giving ascospore pigmentation in prominent granules outside the spore wall instead of the wild-type uniform distribution over the wall. 58 3. pfr-1, an NG-induced mutation, para-fluorophenylalanine

(FDA) resistance, confering the ability to grow in the presence of 100mg/1 of this chemical.

4. mt, mating type, + or -.

Using Whitehouse's method (1957) the centromere distances have been calculated, from tetratype frequencies, for w-78, ar-3 and pfr-1 as 30, 1.5 and 23.5 recombination units respectively. For the centromere distance of mating type (mt), Yu-Sun (1966) gave the value 2.

i) Properties of pfr-1 and fpr mutants

The relation between our pfr-1 and Stadler, Towe and Rossignofs (1970) are unknown. The fpr strain is resistant to FPA at 15mg/1. When this concentration was used with complete medium (general materials & methods) in this work for isolation of new mutants, it was found

that our wild-types were already resistant to this

chemical. Its concentration in the selective media was then raised to 100mg/1. It was found then that our wild- type strain and the pfr+1 mutant, both growing well on

minimal medium, (general materials & methods) could not grow well on minimal +15mg/1 of PFA. This is shown

clearly in Table 2. 5y

TABLE 2

The rate of growth of pfr-1, fpr and wt(9r+) strains in different media

Strains Media

Complete Minimal Complete+15*1 Minimal+15* Complete+100- 24 hours after inoculation wt(9r+) 4.0** 3.0 3.0 0.1 0.5 fpr 2.5 2.8 2.6 1.5 3.0 pfr-1 2.5 2.4 3.0 0.5 3.0 84 hours after inoculation wt(9r+) 9.0' 9.0 9.0 0.1 0.5 fpr 9.0 9.0 9.0 8.0 4.5 pfr-1 9.0 9.0 9.0 0.5 5.5

* concentration of FPA in mg/1

** colony diameter in cm. Each number is the mean of at least two test values.

71 colonies of 9cm diameter had reached the edge of the Petri

dishes.

In Table 2 it is shown that the fpr strain is resistant to both concentrations of FPA in either complete or

minimal medium, while pfr-1 mutant is resistant only in complete medium. The test was not done in Minimal + 100mg/i

of FPA, since 9r+ and pfr-1 are sensitive to the much lower concentration of 15mg/1 in minimal medium. To

investigate the cause of this difference of behaviour of wt and mutant strains in complete and Minimal medium with the same concentration of PFA the following tests were done: 60 1. The above strains were cultured in

a) Minimal medium + 23mg/1 phenylalanine (concentration

calculated from amount of phenylalanine contained in casamino-acids, Fieser & Fieser, 1957) + 15mg/1 FPA, and b) Minimal medium + 23mg/1 phenylalanine + 100mg/1 FPA.

This was done to investigate the effect that ordinary

phenylalanine might have in complete medium, namely a possible interference of the normal amino acid with the drug's uptake or mode of action. The results show no

such interference, (Table 2).

2. The same strains were cultured in

a) Complete medium without casamino acids + 15mg/1 FPA

and b) Complete medium without yeast extract + 15mg/1 FPA.

The results show (Table 3) that wt(9r+) is suppressed in b) medium, while the growth of pfr-1+ mutant is

about the same in both a) and b) media.

TABLE 3 The growth of the fpr-, pfr-1+ and wt(9r+) strains, in

different media

Strains Media *Minimal +23mg/1 phenylalanine **Complete +15mg/1 PFA

+15mg/1 PFA +100mg/1 PFA -Casamino acids -Yeast extract wt(9r+ 0.5/ 0.1 3.0 0.1 fpr 2.0 1.0 2.0 0.7 pfr-1 1.0 0.1 1.8 1.4 measured at 40 hours after inoculation ** measured at 18 hours after inoculation / colony diameter in cm PLATE 3 Octad with twin spores, giving a false 5:3

segregation, in a + x vr-3 cross.

+ PLATE 4 Tetratype segregation in a cross w-78 , gr-31- x w-78 ,

gr-3 , showing epistasis of the w-78 over the L1-3

marker (see text) 62

It is therefore the presence of yeast extract that helps the wild-type 9r+ to be resistant to PFA at the level of 15mg/1, in complete medium.

Further tests showed that vitamins and nucleic acids were not supporting growth in the presence of FPA. The

investigation on the nature of the pfr-1 mutant was not continued beyond this pOint, as it was outside the scope

of the main work.

ii) The properties of the gr-3 and w-78 mutants The relations between our L1-3 and Emerson's (1969)

2.11-1 or Yu-Sun's zr-2 are unknown. Dominance relations + + between gr-3 and gr-3 , and between w-78 and w-78 , can not be established by conventional tests in diploids as

these alleles are only expressed in the haploid ascospores. Thereis however indirect evidence that the two wild-type alleles are dominant to their respective recessive alleles.

Occasionally two spores in an ascus are fused by their + walls and in crosses of w-78 x w-78 or gr-3+ x gr-3 , such asci often show phenotypic ratio of 5 wild-type: 3

mutant for w-78 or LI-3. Plate No.3. Such "twin spores" were often very difficult to separate with needles without destroying one spore. The single spores remaining intact

after separation in six such pairs were germinated and crossed to establish their genotype. Of these two spores that were red proved on crossing to have been genotypically w-78 and three out of the four non-granular

ones were genotypically E1-3-. Phenocopies were 63 generally very rare for w-78 or zr-3 yet these 5:3 asci with "twin spores" had frequent phenocopies, with genotypically mutant spores being given a wild-type phenotype from direct contact with genotypically wild- type spores. This strongly indicates dominance of w-78+ + to w-78 and of fir-3 to ur-3 .

Any octads with "twin spores" whatever their phenotypic ratio, were excluded from all the main experiments, to avoid any biases from phenocopies.

Marker w-78, whose segregation is studiedi is epistatic to gr-3. This means that spores of the genotype w--78,

Lr-3+ and w-78 , gr-3 , cannot phenotypically be distinguished from each other, because they are both white, Plate No.4. This epistasis of the white over the granular locus does not affect the frequency of detectable octads of any class, for w-78 in the present work, though it would affect the detection of certain aberrant ratios for the granular marker, if these were studied at the same time. If however, w-78 were hypostatic to any of the other markers used here, the study of 6:2(2), 2:6(2) and

4:4(2) segregations would be much more difficult.

Gene conversion at the granular site is extremely rare with a frequency of 0.004% in 23,969 octads in control + + + crosses of w-78 , LI-3 x w-78 , gr-3 . This is of great importance as most of the aberrant ratios occuring in the + R crosses w-78 , gr-3 , pfr + x w-78 Rr-3+, pfrS,- used 64

here, could be attributed mainly to conversion of the

w-78 site, whose conversion frequencies are much higher (3-14% in different crosses, LCF and HCF). c) The detection system

Crosses between strains of Ascobolus immersus produce unordered octads. This means that visual identification of the sister spores, as it is done in ordered octads through the

position of the spores in the ascus, is not possible here

without the aid of additional markers. When additional markers are present, te tratype segregation of any pair of

markers would define the spore pairs in an octad. This is necessary if postmeiotic segregation, giving non-identical sister-spores is to be identified.

The detection system used here includes: two visual markers,

namely w-78 whose gene conversion is studied, and gr as an auxiliary visual marker, helping in the preliminary identification of 6:2(2), 2:6(2) and 4:4(2) segregations for

w-78, and two non-visual markers, pfr and mt which were used

in confirmatory tests, all three- markers being unlinked to each other.

The visual markers were used in repulsion in these crosses and Table 4 gives the phenotypic and genotypic ratios of

all classes produced in such a cross. FOOTNOTES FOR TABLE 4 R: is red, non granular, G: red granular, and

W: white. The two genotypic classes containing w-78 are both white, because of the epistatic effect of the w-78 over

Narrower ratio unique classes (Lamb &

Wickramaratne, 1973) distinguishable as such.

** Aberrant 4:4, 4:4(2) distinguishable as such.

Numbers in parentheses show the numbers of pairs

of non-identical sister spores, contained in the

respective octad. 66 TABLE 4 Classes produced in all types of octad in the cross w-78+, qr-3- + x w-78 , LE-3 , and their phenotype and genotype ratios. Conversion events are considered only for site w-78, while LE-3 site is constantly 4:4

PHENOTYPE/ R G W TYPES OF UNORDERED OCTAD for GENOTYPE w-78, Li-3 w-78+ w-78 w-78 ar-3+ r-3 PD NPD T 4 4 8:0 8:0 8:0 4 3 1 7:1 7:1 4 2 2 6:2 , 6:2(2) 6:2 , 6:2(2) 4 1 3 5:3 4 4 4:4 3 4 1 7:1 7:1 *3 3 1 1 6:2(2)+ 6:2(2) 6:2(2) 3 2 2 1 5:3(3) 5:3 , 5:3(3) **3 1 3 1 4:4(2) 4:4(2) 3 4 1 3:5 2 4 2 6:2 , 6:2(2) 6:2 , 6:2(2) 2 3 1 2 5:3(3) 5:3 , 5:3(3) 2 2 2 2 4:4(4) 4:4(4) 4:4 , 4:4(2),4:4E0 2 1 3 2 3:5(3) 3:5 , 3:5(3) 2 4 2 2:6 , 2:6(2) 2:6 , 2:6(2) 1 4 3 5:3 **1 3 1 3 4:4(2) 4:4(2) 1 2 2 3 3:5(3) 3:5 , 3:5(3) *1 1 3 3 2:6(2) 2:6(2) 2:6(2) 1 4 3 1:7 1:7 4 4 4:4 3 1 4 3:5 2 2 4 2:6 , 2:6(2) 2:6 , 2:6(2) 1 3 4 1:7 1:7 4 4 0:8 0:8 0:8 67

+ + PLATE 5 Aberrant 4:4 segregation in cross w-78 , Lr-3 x w-78 L1-3: 3R:1G:4W

+ + PLATE 6 Aberrant 4:4 segregation in cross w-78 , Li-3 x w-78 A

L1- 3- : 1R:3G:4W 68

PLATE 7 Parental ditype segregation (4:4) in cross w-78 ,

x w-7,8 +

PLATE 8 Non-parental ditype segregation (4:4) in cross w-78 gr-3+ x w-78+, gr-3

-

bJ i) The detection of 4:4(4_1

Octads detected as 4:4(2) can be of two phenotype ratios: 3R:1G:4W, Plate 5, or 1R:3G:4W, plate 6. Tetratypes (T)

are giving rise to both these ratios while NPDs produce

only the first one and PDs only the second one. This is

so, irrespective of whether the 4:4(2) octads were due to

hybrid-DNA formation in one pair of non-sister chromatids

(a 2:2(2) and a 2:2 event), or in both pairs of non-

sister chromatids (a 3:1 and a 1:3 event). nth the

present system all 4:4(2)s occurring at the w-78 locus in

either PDs or NPDs are visually detectable. Of the ones

occurring in Ts only half are detectable,•giving either of

the two phenotype ratios above, while the other half of

them would pass unnoticed since their phenotype would be

identical to that of a normal T. This is due to the

following considerations.

In order for a 4:4(2) to be visually detectable the two

pairs of non-identical sister-spores for the marker

under study (w-78) must not be accompanied by the same

allele of the auxilliary marker (gr-3). In PDs (4 w-78+,

u-3, 4 w-78, or-3+), the occurrence of 2 pairs of non-

identical sister-spores would always produce an ascus with

the following genetic constitution:

A pair of identical sister-spores both spores w-784- ,,gr-3-(G) + 1 spore w-78 cTr-3 (G) A pair of non-identical sister-spores 1 spore w-78 (W) + 1 spore w-78+ ,rr-3 (R) A pair of non-identical sister-spores 1 spore w-78 , (W)

A pair of identical sister-spores both spores w-78 (W) 70

In PD therefore, the two pairs of non-identical sister- spores are accompanied by a different allele of the gr-3 marker each and the overall phenotype of an 4:4(2) octad is 1R:3G:4w.

In NPD (4 w-78, LI-3:4 w-78+, AL-3+) the occurrence of 2 pairs of non-identical sister-spores would always produce an ascus with the following genetic constitution.

A pair of identical sister-spores both spores w-78,R1-3(W) 1 spore w-78 ,L11-3 (W) A pair of non-identical f sister-spores 1 spore w-784-,z1-3(G)

A pair of non-identical ' 1 spore w-78-,z1-34-(W) sister-spores 1 spore w-78+,g1-3+(R) A pair of identical sister-spores both spores w-78+,R1-34.(R)

Here again the two pairs of non-identical sister-spores are accompanied by a different allele of the .gr-3 marker each. The phenotype ratio of all 4:4(2)s occurring in NPD is 3R:1G:4w.

Each of these two phenotypes of 4:4(2) can similarly arise in T as well. In T however, (2 w-78+, gr-3 : + + + 2 w-78, gr-3 :2 w-78 ,gr-3 :2 w-78 ,gr-3 ) the two pairs of non-identical sister-spores are not always accompanied by different alleles of the second visual marker. In fact half of the 4:4(2) may have either the phenotype appearing in PD or the one appearing in NPD. In the other half of them the two pairs of non-identical sister-

71

spores will be accompanied by the same allele of gr-3 (either gr-3 or gr-3 ) as shown below:

Either

1 pair of non-identical 1 spore w-78,z1-3(11) sister-spores [ 1 spore w-78+,z1-3(G) 1 pair of identical sister-spores both spores w-78 ,fir-3+(W)

1 pair of non-identical 1 spore w-78“Yr-3(W) sister-spores [ 1 spore w-78+,.-3 (G) 1 pair of identical sister-spores both spores w-784.,L1-34.(R)

where both pairs of non-identical sister-spores are carrying the allele gr-3-, Or 1 pair of identical sister-spores both spores w-78,a1-3(11)

1 pair of non-identical 1 spore w-78,ar-3+(W) sister-spores 1 spore w-784- ,RE-34.(R) 1 pair of identical sister-spores both spores w-78+,gr-3-(G) + + 1 pair of non-identical 1 spore w-78 (R) sister-spores 1 spore- w-78 ,a--3+(W)

where both pairs of non identical sister-spores are + carrying the allele gr-3 .

In both these 4:4(2)'s occurring in T's the overall phenotype ratio is 2R:2G:4w, which is identical to that of the normal T1 Plate No.4.

From all these it follows that the proportion of 4:4(2)s

21 T detected here - PD+NPD+TPD+NPD+ 72

PLATE 9 A 6:2(2) octad in the cross w-78 , gr-3 + x w-78+ gr-3 ;

3R:3G:2W

+ + PLATE 10 A 6:2 octad in the cross w-78 ,gr-3 x y-78 , gr-3 ,

4R:2G:2W 73

+ + PLATE 11 A 6:2 octad in the cross w-78 gr-3 x w-78 gr-3 ,

2R:4G:20[

,+ + PLATE 12 A 2:6(2) octad in the cross w-78, Pr-.5 x w-78 or-3

1R:1G:6W 74

+ PLATE 13 A 2:6 octad in the cross w-78, gr-3+ x w-78 , gr-3 ,

2G:6W

+ + PLATE 14 A 2:6 octad in the cross w-78 , gr-3 x w-78 , ar-3

2R:60i 75 ii) The detection of 6:2(22

The 6:2(2)s, Plate 9, may he due either to a 3:1 ratio in both pairs of non-sister chromatids in 'a bivalent or to a 4:0 ratio in one pair of non-sister chromatids and a 2:2(2) ratio in the other. In any case therefore, the

occurrence of the 6:2(2) presupposes the formation of

hybrid-DNA at corresponding sites in both pairs of non- sister chromatids of a bivalent. With the first of these

origins, 6:2(2)s have the following possibilities of

detection in a repulsion cross for the visual markers used in this system:

A 6:2(2) in PD would have the following genotypic constitution:

2 similar pairs of identical + sister-spores four spores w-78 ,L1-3 (G) ++ 2 similar pairs of non- ' 1 spore w-78 (R) identical sister-spore; +(W) each pair containing 1 spore w-78,z1-3 TOTAL: 2R:4G:2w, Plate No.11.

Conclusion: 6:2(2)s occurring in PD and due to two 3:1 events, are not distinguishable from normal 6:2's. A 6:2(2) in NPD would have the following genotypic constitution:

2 similar pairs of identical + + sister-spores four spores w-78 ,LI-3 (R)

2 similar pairs of non- ' 1 spore w-78+,ar-3 (G) identical sister-spores 1 spore w-78,gr-3(W) each pair containing TOTAL: 4R:2G:2W, Plate No.10.

6:2(2)s from two 3:1 events occurring in NPD are not

distinguishable from normal 6:2's. 76

The 6:2(2)s occurring; in T are all of the genotypic constitution:

1 pair of identical sister-spores both spores w-78+,.g1-3 (G) 1 pair of identical sister-spores both spores w-78+,z1-34.(R) + 1 spore w-78 ,02-3(G) 1 pair of non-identical f sister-spores [ 1 spore w-78,21-3(W) 1 spore w-78+,221-34-(R) 1 pair of non-identical + sister-spores [ 1 spore w-78 ,pr-3 (10 TOTAL: 3R:3G:2w

All 6:2(2)s of this origin occurring in T are detectable.

For 6:2(2) segregations therefore from two 3:1 events the ratio of the detected segregations to those occurring, T is PD+NPD+T

6:2(2)s due to a 4:0 event in one pair of non-sister chromatids of a bivalent and a 2:2(2) event in the other have the following possibilities of detection here.

In PDs and NPDs they have the genotypic constitution that they have when they are in T of 6:2(2) of two 3:1 events. Therefore, here all the 6:2(2) occurring in either a PD or an :VPD are detectable, since in both they will give the phenotype: 3R:3G:2w. In T half of the 6:2(2) will be of the genotypic constitution that gives the phenotypic

3R:3G:2w, and the other half of them may be either of phenotype 4R:2G:2w, or 2R:4G:2w, both the two lather ones indistinguishable from normal 6:2(2). 77

For 6:2(2) segregations from a 4:0 plus a 2:2(2) event the ratio of the detected segregations to those occurring PD+NPD+1T. is therefore PD+NPD+T Lamb and Wickramaratne (1973) provided a formula by which the frequency of 6:2(2)s expected to occur in a cross, in the absence of corresponding-site interference, can be calculated. This 2 is 6:2(2) = 2ae+c where a: is the frequency of 4:0 in one pair of non-sister

chromatids in a bivalent e: is the frequency of 2:2(2) in one pair of non-sister chromatids in a bivalent c: is the frequency of 3:1 in one pair of non-sister chromatids in a bivalent.

The same authors give partly corrected formulae for the values of a, e and c. a(=4:0)=16+:2m+8:0+17:1 e(=2:2(2))=14:4(2) c(=3:1)=15:3+17:1

From the frequencies of these classes one can calculate the frequency of 6:2(2)s expected in a cross, using the above formula.

There are some four strand events not accounted for by this formula, but they are considered negligible, and they are missed because they are not detectable.

Taking into consideration the fractions giving the ratio of the detected 6:2(2) to the formed ones in each of the - -

78

two kinds of origin, one can estimate the frequency of 6:2(2)s expected to be detected with this system. This would be very useful for comparison with the actually

observed frequency of this class, and for the estimation

of the corresponding-site interference coincidence

coefficient.

6:2(2) formed = 2ae+c2

o:2(2)expected to be PD+NPD+1T 2 T detected here = 2aexPD+NPD+T +c x PD+NPD+T iii) The detection of 2:6(2) A 2:6(2), Plate 12, octad may be due to

1) a 1:3 event in each of the two pairs of non-sister chromatids

2) a 2:2(2) in one pair of non-sister chromatids and 0:4 in the other. 2:6(2)s from the first of the two kinds of origin have

the following possibilities of detection in these crosses.

In PD a 2:6(2) from two 1:3 events would contain

2 similar pairs of identical sister-spores 4 spores w-78- ,01--3+(W) + 1 spore w-78 ,E1-3 (C) 2 similar pairs of non- identical sister-spores 1 spore w-78 ,.,2._gr-3 (W) TOTAL: 2G:6W, giving a phenotype undistinguishable from normal 2:6 , Plate No.13. In PD a 2:6(2) from 2 1:3 events would contain:

2 similar pairs of identical sister-spores 4 spores w-78 r-3 (W) + + 1 spore w-78 ILI-3 (R) 2 similar pairs of non- identical sister-spores 1 spore w-78 ,fir-3+(N)

TOTAL: 2R:6W, giving a phenotype indistinguishable from

normal 2:6. Plate No. 14. 79

In T a 2:6(2) from 2 1:3 events would contain: 1 pair of identical sister-spores both spores w-78,gr-3(W)- - 1 pair of identical sister-spores both spores w-78- ,E1-34-(W) +,L1-3(G) 1 pair of non-identical 1 spore w-78 sister-spores 1 spore w-78,R1-3(4)

1 pair of non-identical 1 spore w- 7 8+ ,RI- 3 (R) sister-spores 1 spore w-7 8 ,21-34-(W) TOTAL: 1R:1G:6W visually distinguishable from normal 2:6

The ratio of detected 2:6(2)'s from 2 1:3 events to the formed ones is Pll+NPD+T

2:6(2)s due to a 2:2(2) event and a 0:4 one have the following possibilities of detection by the present system.

In PD a 2:6(2) of this origin would contain: 1 pair of identical sister-spores both spores w-78,Lr-3(W) 1 pair of identical sister-spores both spores w-78,Gr-3- 4.(W) 1 spore w-78+,-3 (G) 1 pair of non-identical [ sister-spores 1 spore w-78,L1-3(W) + + 1 spore w-78 (R) 1 pair of non-identical sister-spores , 1 spore w-78 ,fir-3 (W) TOTAL: 1R:1G:6W, visually distinguishable from normal 2:6.

In NPD a 2:6(2) of this origin would have the same genotypic constitution as the 2:6(2) occurring in PD. 80

So all 2:6(2)s in NPPs are visually detectable. In T only half the 2:6(2)s give the phenotype 1R:1G:6W whereby they are visually detectable. The other half contain the same kind of spore pairs as the 2:6(2) occurring in

PD and NPD of the first type of origin, namely:

2G:6W or 2R:6W, therefore being undetectable with this system.

The ratio of detected 2:6(2)s due to a 2:2(2) and a 0:4 event to the formed ones is equal to PD+ NPD+1 T PD*NPD+T ' According, to Lamb and Wickramaratne(1973)2:6(2)s expected = 2be+d2, with no corresponding-site interference, where b = the frequency of 0:4 in one pair of non-sister chromatids in a bivalent e = the frequency of 2:2(2), in one pair of non-sister chromatids in a bivalent d = the frequency of 1:3, in one pair of non-sister chromatids in a bivalent b(=0:4)=12:6+0:8+;1:7 c(=2:2(2))=14:4(2) d(=1:3)=13:5+11:7

In this way one can calculate the expected frequency of detected 6:2(2)s in the absence of corresponding-site interference, from the frequencies of other.classes in the cross. For the purpose of this work the formula giving

2:6(2) = 2be+d2 has been modified as follows: 81

PD+NPD+1T+d x2 T 2:6(2)=2 bexi„D- +NPD+T PD+NPD+T

In this way the frequency of 2:6(2)s expected to be detected is estimated. This will be used, in the case

of 6:2(2), in the estimation of corresponding-site

interference coincidence coefficients, along with the observed frequency for the same class.

With the present system, it is not possible to

distinguish between the two types of origin of the 2:6(2) (namely 2 1:3 events or a 0:4 and a 2:2(2)) or those of 6:2(2) (Namely 2 3:1 events or a 4:0 and a 2:2(2)). This

would only be possible if linked markers were used for

w-78. iv) The detection of other unique narrower ratio classes, The remaining unique narrower ratios are 5:3(3), 3:5(3) and 4:4(4). The 5:3(3) arise by a combination of a 3:1

event in one pair of non-sister chromatids and a 2:2(2) in the other pair. The phenotypes by which 5:3(3) appear in these crosses are: 2R:3G:3W in PD and T and 3R:2G:2W,

in NPD and T. These phenotypes, and their genotypes are

exactly the same as the ones of the 5:3(1) occurring in T, thus being undectable as 5:3(3). The 5:3(1)s occur in two

more phenotypes in the cross, namely 4R:1G:3W in NPD and 1R:4G:3W in PD. The only way therefore that the 5:3(3)s

could be detected with a similar system is if the

phenotypes of 5:3(1) that are identical to those of

5:3(3) were eliminated, so that any octad appearing with

this phenotype could be attributed to 5:3(3) rather than 82 to 5:3(1). This could only be done with a system suggested by lamb (personal communication) showing no, or a very low frequency of tetratypes between w-78 and a second visual marker. Further markers would have to be used, which in tetratypes would define the four spore pairs, so that the postmeiotic segregation at w-78 could be confirmed.

The 3:5(3) could be identified under the same circumstances as 5:3(3). The phenotypes of 3:5(3) that are identical to those of the 3:5(1) occurring in T are

1R:2G:5W (coming from PD or T in the case of 3:5(3)) and 2R:1G:5W (coming from NPD or T in the case of 3:5(3)).

Finally, the detection of the "doubly aberrant 4:4",

4:4(4), is not possible with the present system. This is due to the fact that 4:4(4) give the phenotype 2R:2G:4W, which is exactly the phenotype of normal tetra- types and as the octads here are unordered, 4:4(4)s go undetected. However the "very low or no T" system discussed above could be of some help here again. As the visual markers would give very few tetratypes, a good number of the octads with tetratype phenotype, namely. 2R:2G:4W, would be due to 4:4(4). They could also be due to 4:4(2), but only half the 4:4(2) occurring in T would have this phenotype, and since T in general would be very low, these 4:4(2) would be expected to be very few. 83

Probably the best system to use in order to detect 5:3(3),

3:5(3) and 4:4(4) octads would be one with three visual

markers, including the one segregating with the above ratios. Other non-visual markers would be useful for confirmatory tests, whereby it could be shown whether or

not the aberrant segregation occurred in the marker under

study or in any of the other two visual ones. It would

also be desirable to have the two auxiliary visual

markers showing low frequencies of gene conversion,

compared to the ones of the marker under study.

2. RESULTS

The two spore pigmentation markers w-78 (white spore colour) and gr-3 (granular distribution of pigment) have made possible the visual detection in spore collecting lids of octads whose phenotype ratios correspond to 6:2(2), 2:6(2) and 4:4(2)

genotypic segregations at the w-78 locus.

The investigation includes both high and low conversion + R frequency strains of w-78 in crosses of w-78 , gr-3 , pfr , + x w-78 +, pfrS,

The results of these two kinds of cross with all

segregation classes detected for w-78 are given in Table 5.

All the eight spores of each octad detected with 6:2(2),

, 2:6(2) and 4:4(2) ratios were isolated on germination medium and given the germination treatment (general materials and methods).

Cultures from the germinated spores were checked for all the

markers involved. This analysis included culturing on complete 84

S medium containing PFA, where a pfr wild-type strain was used as control, as well as crossing to + and - mating-types of a wild-type strain to check w-78, E1-3 and mt genotypes.

Of 96 octads visually identified as 4:4(2) in both high and low conversion frequency crosses, only 10 octads were satisfactorily analysable. These proved genuine: nine were fully analysed, while one spore failed to germinate in the tenth one. Of nine octads identified as 6:2(2), only two were fully analysable and of the eight octads identified as 2:6(2) only one was fully analysed: all three were genuine.

This very low proportion of fully analysed octads obtained was primarily due to the extremely bad germination that any kinds of spore of all crosses were then showing, which was well before horse dung germination medium was introduced. With the old germination medium only one or two spores from each octad usually germinated, thus rendering the bulk of the octads non- analysable. In about one every ten octads though, six or seven spores would germinate after the first germination treatment

(heat shock at 50°C and incubation at 37°C). The few remaining ungerminated spores were transferred to fresh germination medium and given a second treatment and in some cases a third one, until ten octads were obtained fully analysed.

A second obstacle was the frequent lack of fertility, which hindered analysis of the single spore cultures by not permitting crosses for checking the various markers. 85

TABLE 5 Segregation classes for w-78, and unordered tetrad types

HCF' LCF/ Cross Class No. 1, No. % 6+:2w 2,511 7.04 406 1.47 5:3 378 1.06 94 0.34 3:5 258 0.72 129 0.47 2:6 1,016 2.56 344 1.25 4:4(2) 3R:1C:4W 28 0.08 22 0.08 4:4(2) 1R:3G:4W 26 0.07 20 0.07 0:2(2) 7 0.02 2 0.007 2:6(2) 4 0.01 4 0.01 8:0 12 0.03 2 0.007 7:1 4 0.01 1 0.004 1:7 5 0.01 1 0.004 0:8 6 0.02 8 0.03 4:4 T 21,040 58.99 17,818 64.55 4:4 PD+NPD 10,373 29.08 8,750 31.71 Total 35,668 27,596

HCF, high conversion frequency crosses, total narrower ratio conversion frequency, 11.86%; LCF, low conversion frequency crosses, total narrower ratio conversion frequency 3.56%. 86

The result is that only thirteen (13) octads in all have been analysed fully from all three kinds of segregation. Of the remaining octads, only 1, 2 or 3 spores have been analysed in each, but in no instance was there any evidence that the octad concerned was not genotypically what had been detected phenotypically. Phenocopies are generally very rare for w-78 (Emerson & Yu-Sun, 1967, Lamb & Wickramaratne, 1973, and present work) and only one phenocopy for granular has been detected, in another cross, one granular phenotype being genotypically g1-3+.

a) 4:4(2) Of the 4:4(2)s analysed, six were of the phenotype 3R:1G:4w

and four were of the phenotype 1R:3G:4W. The detailed genotypes of these octads are given below. The pairs of non- identical sister-spores, with postmeiotic segregation at

w-78 are designated by the brackets.

Octad No.10 is one where one spore did not germinate. It is

likely that the ungerminated spore was a sister-spore of the

designated (*) one, since its phenotype was G, (w-78+,

Octad No. PHENOTYPE GENOTYPE

No. of spores w-78 gr pir nit

1 3R,1G,4W 2 + + S + 2 - - R - 1 + + S + 1 - + S + 1 + - R - 1 - - R -

coritinue:virage 88 87

Octad No. PHENOTYPE GENOTYPE

No. of spores w-78 pfr mt 8 & 9 1R,3G,42 2 + - R + (identical) 2 - + S - 1 + - S + 1 - - S + 1 + + R - 1 - + R -

10 1R,3G,4W *1 + - S + 2 - + S + 1 + + R - 1 - + R - 1 + R - 1 - - R -

In octads No. 4, 5, 6, 8, 9 and 10 each of the four spores involved in the two pairs of non-identical sister-spores is

unique in the octad and these two pairs can be identified from the segregation of Lr, pfr and mt.

In the rest of the 4:4(2)s analysed, one spore of each pair of non-identical sister-spores is unique, clearly belonging to a pair of non-identical sister-spores, but the other spore

in each case has two similar spores in the octad and it is not known which of the three belongs to the non-identical sister-

spore pairs.

The frequencies of the two phenotypes of 4:4(2) are about equal in both types of crosses (high conversion frequency -

IICF, and low conversion frequency - LCF).

COntillike on page 89 88

Octad No. PHENOTYPE GENOTYPE

No. of spores w-78 jZr pfr mt

3R,1G,4W 2 + + R - 2 - - S + 1 + + R - 1 - + R - 1 + - S + 1 - - S +

3 3R,1G,4W 2 + + S - 2 - - R + 1 + S - 1 - + S - 1 + - R + 1 - R +

4 3R,1G,4W 2 + + S - 2 - S + 1 + + R - 1 - + R 1 + - R + 1 - - R +

5 3R,1G,4W 2 + + R + 2 - - S - 1 + + S - 1 - + S - 1 + - R + 1 - - R +

6 3R,1G,4W 2 2 1 1 1 1-

7 1R,3G,4W 2 2 1 1 1 1

continue on poig-e el 89

The proportion of 4:4(2) detected is equal to PPD+NPD+1T D+NPD+T ' since half the 4:4(2) occurring in T, as already stated, are not detectable visually, because their phenotype is identical to that of a normal tetratype. No attempt was made to

identify these last 4:4(2) among the normal tetratypes. b) 6:2(2) and 2:6(2)

All spores from octads with phenotype ratios expected for

6:2(2) and 2:6(2) segregations were isolated for testing

in the same way as for 4:4(2) octads. For reasons explained

in the previous section, only two octads from the nine 6:2(2)s detected and one octad from the eight 2:6(2)s detected were fully analysed. Their detailed genotypes are given below,

with brackets showing the pairs of non-identical sister- spores.

Octad No. PHENOTYPE GENOTYPE No. of spores w-78 gr pfr mt 1 3R,3G,2W 2 + + R (6:2(2)) 2 + - S + 1 + + R - 1 - + R - 1 + - S + 1 - - S +

2 3R,3G,2W 2 + + R + (6:2(2)) 2 + - R - 1 + + S - 1 - + S - 1 + - S + 1 - - S + 90

Octad No. PHENOTYPE GENOTYPE No. of spores w-78 fr pfr mt

3 1R,1G,6w 2 - + S + (2:6(2)) 2 - - R - 1 + + S + 1 - + S + 1 + - R - 1 - - R -

In octad No.2, the spores that are the result of the same mitotic division - and therefore the non-identical sister- spores - are unambiguously identified, especially as pfr and

mt as well as gr and mt showed tetratype segregation. In

octads No.1 and No.3, the existence of two pairs of non- identical sister-spores is apparent, due to the occurrence of

two different unique spores in each octad.

As with some 4:4(2)s, it was not possible to know with the

present system, which of the three similar spores belonged

to a non-identical sister-spore pair.

These unique narrower ratio classes were used for

corresponding-site interference estimation 'fable 6; wider ratio classes were also considered here. 91

TABLE 6 Corresponding-site interference data

HCF LCF

Cross Class No. No. 6:2(2) Observed * 7 0.02 2 0.007 Expected * 2.5 0.007 0.4 0.001 CSICC / 2.8 5.0 2:6(2) Observed * 4 0.01 4 0.01 Expected * 1.1 0.003 0.4 0.001 CSICC / 3.6 10 8:0 Observed * 12 0.03 2 0.007 Expected* 46 0.13 1.5 0.006 CSICC / 0.3 1.3 7:1 Observed 4 0.01 1 0.004 Expected 14 0.04 0.71 0.003 CSICC 0.3 1.4 1:7 Observed 5 0.01 1 0.004 Expected 3.8 0.01 0.9 0.003 CSICC 1.3 1.1 0:8 Observed 6 0.02 8 0.03 Expected 7.5 0.02 1.2 0.004 CSICC 0.8 6.7

* These are directly comparable as the expected values (expected in the absence of corresponding-site interference) allow for the non-detection of a proportion of 6:2(2) and 2:6(2) events in accordance with.the above formulae.

/ CSICC Corresponding-site interference coincidence coefficient, observed/expected values. 92

In both IICF and LCF crosses, the numbers of 6:2(2) and 2:6(2) octads observed were higher than expected in the absence of corresponding-site interference by a factor of about 3 to 10. Still it is difficult to say whether this is an accurate assessment or not, because of the small numbers of octads detected for these classes, though the trend to negative corresponding-site interference was consistent for these classes.

From the wider ratio classes 0:8s may not be completely reliable as they have been shown by Lamb and Wickramaratne

(1973), to arise often by mutation, and they have not been checked here. By the same author's it was also shown that

8:0 and 7:1 segregations in Ascobolus immersus with w-78, were nearly all genuine.

The detection of other unique narrower ratios, whose occurrence has been predicted by Lamb & Wickramaratne (1973), namely of 5:3(3), 3:5(3) and 4:4(4), was not possible with the detection system used here. A system with very low T frequency would greatly assist the detection of these classes. 93

DISCUSSION

A) Aberrant 4:4 Octads

The detection of aberrant 4:4's was undertaken partly because they were needed for the calculation of the corresponding-site interference coincidence coefficients and partly because they would help determine the mechanisms by which high and low conversion crosses differ.

Aberrant 4:4's provide a measurement of the degree of non-correction of mispairs. The frequency of aberrant 4:4's in the present work was scored in two kinds of cross: high conversion frequency cross,

+ s (1)w-78 4.(HCF), pfrR,+,X w-78 (HCF),LI-3 ,pfr PP- and low conversion frequency cross,

(2)w-78 4-(HCF), it-3, pfrR,+,X w-78(LCF),1I-3+,pfrS,-, and it was found to be the same in both of them (0.15%).

High and low conversion crosses may differ in either or both the frequency of hybrid DNA formation, and the efficiency of correction (with high conversion crosses showing higher frequency of hybrid DNA formation or higher efficiency of correction than low conversion crosses respectively). They may also differ in relative frequencies of dual and single hybrid chromatid DNA.

There are different interpretations on the aberrant 4:4's frequencies in the two types of crosses - (1) and (2) - depending on whether there is only dual hybrid DNA occurring, or only single hybrid DNA or both. 94

1. If only dual hybrid DNA occurs, then (a) if the efficiency of correction in both crosses was the same and their

difference was due to higher frequency of hybrid DNA formation in high conversion frequency crosses than in

low conversion frequency ones, the frequency of aberrant

4:4's would be expected to be higher in high conversion frequency crosses than in low conversion frequency ones. And (b) if the frequency of hybrid DNA formation was the

same in both types of cross and their difference was due

to a difference in the efficiency of correction, with high conversion crosses representing a more efficient system than low conversion ones, then the frequency of aberrant 4:4's is expected to be higher in low conversion frequency crosses than in high conversion frequency ones.

The present results of no difference in the frequencies of aberrant 4:4's between high and low conversion frequency crosses, and in the case of only dual hybrid DNA occurring-

indicate that the two types of cross differ in both factors,

namely hybrid DNA formation (more frequent in high

conversion frequency crosses) and efficiency of correction (higher in high conversion frequency crosses), with the

raising and lowering effects of the two factors on aberrant 4:4 frequencies balancing out.

2. If only single hybrid DNA occurs, no aberrant 4:4's are expected, apart from cases of what might be corresponding

site events.

3. If dual and single hybrid DNA occur , then, since the 95

detection of aberrant 4:4's is generally accepted as

evidence for the occurrence of dual hybrid chromatid mechanism of recombination, the same frequency of aberrant 4:4's in both types of crosses could come from both types of crosses having the same frequency of dual

hybrid DNA formation and the same frequency of correction

but high conversion frequency cross having additional single hybrid DNA, increasing the total conversion frequency, but not the frequency of the aberrant 4:4's.

Aberrant 4:4 octads are usually considered to be due to a 2:2 (2) event in one pair of non-sister chromatids of a bivalent and a normal 2:2 segregation in the other pair, but could also arise from a 3:1 event in one pair of non- sister chromatids and a 1:3 event in the other, at exactly corresponding sites. In the second case they may arise by the single hybrid chromatid hypothesis as well as by the dual.

This possibility is not usually considered because corresponding- site events are usually overlooked, but we know that they occur at the w-78 site of Ascobolus immersus (Lamb & Wickramaratne,

1973). It is possible therefore that a part or even the whole of the aberrant 4:4's scored here may arise according to the single hybrid chromatid hypothesis, depending on the extent that this hypothesis actually works here.

There is good reason to believe that both dual and single hybrid chromatid hypothesis may be responsible for gene conversion at the w-78 locus of Ascobolus immersus (PART A, SECTION II). So it is possible that part of the aberrant 4:4's scored here may have arisen by dual and part of them by single hybrid chromatid hypothesis. 96

The frequency of aberrant 4:4's that are expected to occur by corresponding-site events, in the absence of corresponding-site interference, is given by Wickramaratne (1974) as follows

4:4(2) = 2 x 1 5:3 x 1 3:5 = 1 (5:3 x 3:5), derived from Lamb and Wickramaratne (1973), by ignoring negligibly rare events.

According to this formula and to the frequencies of

5:3 and 3:5 octads in ECF and LCF crosses in Table 5 we find:

expected frequency of 4:4(2) for HCF cross = 0.004% expected frequency of 4:4(2) for LCF cross = 0.001% compared with the observed frequency of aberrant 4:4's for either cross = 0.15%

The expected values are significantly lower than the actual frequencies of 4:4 (2) scored for either cross. This evidence suggests that nearly all aberrant 4:4 octads came from dual hybrid DNA in two chromatids, rather than from corresponding-site events involving two lots of single chromatid hybrid DNA.

B) The detection of 6+:2m (2) and 2+:6m (2). The results here (Table 5) show that it is possible to detect, in unordered octads, postmeiotic segregation among the 6:2 and 2:6 classes, provided a system of suitable markers especially visual ones, is available. Genetic analysis of these octads reveals that even where there is no certainty as 40 wh/c4 rore_s exac-A, form, a mars-idealicaL sisZer---wore pair, 97 they belong to the unique classes (: classes only obtained from hybrid DNA formation at corresponding sites in both pairs of non-sister chromatids of a single bivalent).

The observed and expected values, in the absence of corresponding-site interference, for classes 6:2 (2) and

2:6 (2) are given in Table 6. Among these two classes in low conversion frequency and high conversion frequency crosses there were in each of the four cases more octads of these types observed than would have been expected in the absence of corresponding•-site interference. Observed values were about 3 - to 10 - fold higher than the expected ones, but this assessment may not he completely accurate, due to the small numbers involved.

Formulae and modifying factors used for the calculation of the expected values of these classes are given in the Introduction of this SECTION (I), paragraphs (ii) and (iii).

C) Wider Ratio Octads The wider ratio octads which were detected here are given in Table 5. The work of Lamb (1972), Leblon (1972) and Lamb & Wickramaratne (1973) with Ascobolus immersus, Kitani & Whitehouse (1974), with Sordaria fimicola, Yu-Sun et al (1977), with Sordaria brevicollis has shown beyond doubt the occurrence of genuine wider ratio octads, that is 8:0, 0:8, 7:1 and 1:7 segregations in + X m crosses. 98

Observed and expected values, without corresponding- site interference, are given in Table 6, with fairly good agreement between them except in two cases: the IICF 8:0's and the 7:1's, where fewer were observed than expected (X2=19.9, P = 1% and X2 = 4.9, P = 2.5% respectively).

It is possible that the data on 0:8 may not be completely reliable, since such octads have been shown (Lamb & Wickramaratne, 1973) to arise often by mutation and not by conversion, and they were not checked by isolation and crossing here. On the contrary, they found 8:0's and 7:1's to be nearly all genuine.

D) Corresponding-site events

'Corresponding-site interference' is defined (Lamb

& Wickramaratne, 1973) as interference between the two pairs of non-sister chromatids of a bivalent in hybrid DNA formation at exactly corresponding-sites.

The observed/expected values (corresponding-site interference coincidence coefficient-CSICC) for 6:2 (2), 2:6 (2) and wider ratio octads are given in Table 6. According to the present data these values range from 0.3 to 10 showing that corresponding-site interference can be negative or positive, with much variation between the different segregation classes and between high and low conversion frequency crosses. These results are in fair agreement with the ones reported earlier by Lamb & Wickramaratne (1973) for wider ratio classes at the w-78 and w10 loci although no explanation for the observed variation is available. 99

The Sordaria brevicollis results of Yu-Sun et al (1977) also varied widely between different wider ratio classes for the same mutant, between different mutants of the same locus and between different loci, but corresponding-site interference varied from non-interference (CSICC 1.1) to strong negative interference (CSICC 28.8).

In the present data the 6:2 (2) and 2:6 (2) segregations gave moderate but consistent negative interference in both high conversion frequency and low conversion frequency crosses. But in general the present Ascobolus data confirm the previous ones that corresponding-site interference may be negative or positive.

4. CONCLUSIONS The present work leads to the following conclusions:

1. Aberrant 4:4's at the w-78 locus of Ascobolus immersus in both HCF and LCF crosses largely arise by dual hybrid DNA

in two chromatids rather than by corresponding-site events

(which could be due to two lots of single hybrid chromatids).

2. The frequency of aberrant 4:4's in HCF and LCF crosses was found to be the same. This could be explained in two ways: Either a) HCF and LCF crosses form only or mainly dual hybrid DNA at w-78 locus, with HCF crosses having

more hybrid DNA and higher efficiency of

correction, the two factors giving balance and

producing equal aberrant 4:4's in both types of

crosses. 100

Or b) HCF and LCF crosses have the same frequency

of dual hybrid DNA, (resulting in the same frequency of aberrant 4:4's in both) but HCF

crosses have additional single hybrid DNA, which

does not give aberrant 4:4's, but raises its total conversion frequency.

3. Unique narrower classes (6:2 and 2:6 with postmeiotic

segregation) do occur at the w-78 locus of Ascobolus

immersus, strongly suggesting hybrid DNA formation in both pairs of chromatids, in line with the detection here of the

wider ratio octads, which occurred with frequencies similar to those of previous studies.

4. Corresponding-site interference in the Pasadena strains of Ascobolus immersus, with findings up to now, varies

from moderately negative to moderately positive, with CSICC's 5.1 - 0.3 for the larger samples, both of these being significantly different from no interference at

P = 1%. 101

SECTION II Analysis of the 5:3 and 3:5 classes in IICF x LCF

monohybrid crosses

1. INTRODUCTION

The detection of aberrant 4:4s (4:4(2)), at the w-78 locus - SECTION I of this PART - has been discussed as indicating the occurrence of symmetrical hybrid DNA in this region. It is of interest to investigate whether hybrid DNA occurs predominantly symmetrically or asymmetrically, and if so, to what extent. Stadler and Towe (1971) provided evidence in Ascobolus for meiotic recombination involving only one member of a tetrad, using a system of three markers, namely the w-17 locus and two flanking ones, and analysing the 5:3 and 3:5 segregations in crosses of w-17 x wt. Each of these two segregations may appear in the cross in two types in relation to the flanking markers, type 1 resulting from either symmetrical or asymmetrical hybrid DNA, and type 2 resulting exclusively from symmetrical hybrid DNA. Assuming that symmetrical hybrid DNA produces with equal frequency type 1 and type 2 of either 5:3 or 3:5 segregations, a comparison of the frequencies of these two types of each class in a cross may yield valuable information as to which kind of hybrid DNA occurs predominantly.

This principle was applied in this work, with the w-78 white spore locus and its closely linked K/P gene conversion controlling factors, with which it rarely co-converts (Emerson

& Yu-Sun 1967, Nickramaratne 1974).

Two kinds of cross were used in this investigation

a: LCF wild-type x HCF w-78

b: MCF wild-type x LCF w-78 102

Theoretical analysis of the 5:3 and 3:5 ratios for w-78 in the present system reveals that the expected octad genotypes resulting from symmetrical hybrid DNA are of two types mth with one forcone ratio type being identical to the octad genotype resulting from asymmetrical hybrid DNA. The analysis of 5:3s and 3:5s may therefore be used to investigate the relative frequencies of the two kinds of hybrid DNA, at the w-78 site.

2. THEORETICAL CONSIDERATIONS a) Cross: LCF wild-type x HCF w-78

(i): The analysis of 5:3 ratio octads A 5:3 octad, when symmetrical hybrid DNA occurs, will be the

result of repair to the wild-type genotype in one of the two mispairs present in the bivalent. When asymmetrical hybrid DNA occurs a 5:3 octad will be the result of non-

repair of the single mispair present in the octad (Figure 1)

after invasion of a m chromatid by a + one

Symmetrical "S/A" Type Asymmetrical hybrid DNA hybrid DNA + K + K + K + K + p m P m P m K m P "S" Type + K + K + K m K + p + p m P m P FIGURE 1 The "S" (symmetrical only) and "S/A" (symmetrical or asymmetrical hybrid DNA) types of 5:3s at the mutant site (w-78) in the cross wt(K)xw-78(P). The segregation at linked K/P factor is 4: 103

Two kinds of 5:3s are expected with symmetrical hybrid DNA:

Case 1. If mispair (a) is corrected to wild-type and

mispair (b) is not corrected, the result is: a 5+:3111 with 4 spores : +,K (wild-type, LCF) 1 spore : +,P (wild-type, HCF) "S/A" type 3 spores : m,P (mutant, IICF)

Case 2. If mispair (a) is not corrected and mispair (b) is corrected to wild-type, the result is:

a 5:3 with 3 spores : (wild-type, LCF) 2 spores : +,P (wild-type, HCF) usu type 1 spore : m,K (mutant, LCF) 2 spores : m,P (mutant, HCF)

In the case of asymmetrical hybrid DNA occurring, the result is:

a 5:3 with 4 spores : +,K (wild-type, LCF) 1 spore : +,P (wild-type, IICF) "S/A" type 3 spores : m,P (mutant, HCF)

It is apparent that the genotype of 5:3s resulting from asymmetrical hybrid DNA is identical to the one of case 1 from

symmetrical hybrid DNA. This genotype will be referred to as "S/A" type as it can be produced by either symmetrical or

asymmetrical hybrid DNA. The 5:3 octad with the genotype of case 2, comes exclusively from symmetrical hybrid DNA and it

will be referred to as "S" type.

The "S" and "S/A" types are termed "tetratype" and "tritype" respectively by Whitehouse (1974), in similar studies with

a three marker system in Sordaria fimicola.

104

(ii): The analysis of the 3:5 ratio octads

A 3:5 octad when symmetrical hybrid DNA occurs, will be the

result of repair to the mutant genotype in one of the two mispairs present in the octad. When asymmetrical hybrid DNA

occurs a 3:5 octad will be the re:.;ult of non-repair of the

single mispair present in the octad. (Figure 2).

Symmetrical "S" Type Asymmetrical hybrid DNA + K hybrid DNA + K m K m- K

m P m P + K m P m K m K "S/A" Type m P + K m P • K

m K m P m P m P m P

FIGURE 2. The "S" (symmetrical only) and "S/A" (symmetrical or asymmetrical hybrid DNA) types of 3:5s at the mutant site (w-78) in the cross wt(K)xw-78(P).

When symmetrical hybrid DNA occurs, two kinds of 3:5s are

expected: Case 1. If mispair (a) is corrected to the mutant and mispair (b) is not corrected, the result is:

a 3:5 with 2 spores : +K (wild-type, LCF) 1 spore : +P (wild-type, IICF) "S" type 2 spores : mK (mutant, LCF) 3 spores : mP (mutant, IICF) 105

Case 2. If mispair (a) is not corrected and

mispair (b) is corrected to mutant, the result is: a 3:5 with 3 spores : +K (wild-type, LCF) 1 spore : mK (mutant, LCF) "S/A" type 4 spores : mP (mutant, IICF)

In the case of asymmetrical hybrid DNA occurring, the result is:

a 3:5 with 3 spores : +K (wild-type, LCF)

1 spore : mK (mutant, LCF "S/A" type 4 spores : mP (mutant, IICF)

It is apparent again that the genotype of 3:5s resulting from asymmetrical hybrid DNA is identical to the one of case 2 from the symmetrical hybrid DNA: "S/A" genotype for 3:5s, in

this cross. Case 1 from symmetrical hybrid DNA is the "S" genotype for 3:5s in this cross (coming only from this type of hybrid DNA).

b) Cross: HCF wild-type x LCF w-78

(i): The analysis of 5:3 ratio octads The distribution of the markers in a 5:3 octad for w-78 locus, expected from symmetrical and asymmetrical hybrid DNA in the

cross wt(P)xw-78(K) is shown in Figure 3.

106

"S/A" Type + p P + p + p + p + p + P + K + p m K + P + P (a) m K m P m K + P "S" Type (b) m K m K m K m K m P

m K m K FIGURE 3. The "S" (symmetrical only) and "S/A" (symmetrical or asymmetrical hybrid DNA) types of 5:3s at the mutant site (w-78) in the cross wt(P)xw-78(K)

When symmetrical hybrid DNA occurs two kinds of 5:3s are expected: Case 1. If mispair (a) is corrected to wild-type and mispair (b) is not corrected, the result is: a 5:3 with 4 spores +P (wild-type, UCF) 1 spore : +K (wild-type, LCF) "S/A" type 3 spores : mK (mutant, LCF)

Case 2. If mispair (a) is not corrected and mispair (b) is corrected to wild-type, the result is a 5:3 with 2 spores : +K (wild-type, LCF) 3 spores : +P (wild-type, IICF) usu type 1 spore : mP (mutant, IICF) 2 spores : mK (mutant, LCF)

In the case of asymmetrical hybrid DNA occurring, a 5:3 octad 107

is produced from non-correction of a single mispair : type (b) from the symmetrical hybrid DNA.

a 5:3 with 4 spores : +P (wild-type, HCF) 1 spore : +K (wild-type, LCF) "S/A" type 3 spores : mK (mutant, LCF)

: The analysis of 3:5 ratio octads The distribution of the markers in a 3:5 octad for w-78 locus, expected from symmetrical and asymmetrical hybrid DNA in the cross wt(P)xw-78(K) is shown in Figure 4. Symmetrical "S" Type Asymmetrical hybrid DNA + p hybrid DNA • + p m P m P + P m K + P + K + P + a) m K P p m K m m P "S/A" Type K b) m K m K m K P m K m K m K

FIGURE 4. The "S" (symmetrical only) and "S/A" (symmetrical and asymmetrical hybrid DNA) types of 3:5s at the mutant site (w-78) in the cross wt(P)xw-78(K) Case 1. If mispair (a) is corrected to the mutant and mispair (b) is not corrected, the result is: a 3:5 with 2 spores : +P (wild-type, HCF) 1 spore : +K (wild-type, LCF) ”s” type 2 spores : mP (wild-type, HCF) 3 spores : mK (mutant, LCF) 108

Case 2. If mispair (a) is not corrected and

mispair (b) is corrected to the mutant the result is: a 3:5 with 3 spores : +P (wild-type, HCF) 1 spore : mP (mutant, HCF) "S/A" type 4 spores : mK (mutant, LCF)

From the asymmetrical hybrid DNA - non correction of the single mispair, type (a) the result is: a 3:5 with 3 spores : +P (wild-type, HC'F)

1 spore : mP (mutant, HCF) "S/A".type 4 spores : mK (mutant, LCF)

For this analysis it will be assumed that the mispairs (a) and (b), Figs 1,2,3,4, have equal chances of correction: this assumption is not necessarily correct. Emerson (1966), for example, gave equations for quantitative analysis of correction with different correction properties in the two mispairs at a site, but other authors doing the present kind of study have, for simplicity, assumed equal correction, e.g. Stadler and Towe (1971), Whitehouse (1974).

This implies that types "S" and "S/A" within each class - 5:3 and 3:5 - due to occurrence of symmetrical hybrid DNA, are expected to be of equal frequency, in a given cross. A significant excess of the "S/A" types therefore would indicate the occurrence of asymmetrical hybrid DNA, while complete absence of the "S" type would indicate that symmetrical hybrid

DNA probably does not occur at all in the specific system. If hybrid DNA occurs asymmetrically more often than symmetrically one expects the "S/A" type to occur more than 3 times more 109

often than the "S" type. From these theoretical considerations, it was decided that the practical analysis of the 5:3 and 3:5 classes in crosses wt(K)xw-78(P) and wt(P)xw-78(K) would help reveal the type of hybrid DNA operating at the w-78 locus, at which aberrant 4:4s had already been detected.

3. RESULTS The 5:3 and 3:5 segregations at the w-78 locus analysed here, were from the crosses:

i)9K5 +(wt,K) x EC11w-78.7(m,P) and ii) 61(wt,P) x 5w5KEe(m,K)

The frequency of normal (4:4) and aberrant (narrower ratio) classes in these crosses is given in Table<7.

TABLE 7 Frequencies (%) of several classes in crosses (i) and (ii) Total Cross 4:4 6:2 2:6 5:3 3:5 NRCF* octads

( i) % 96.34 2.62 0.25 0.60 0.17 3.65 No. 28,968 788 76 182 30,075 97.59 0.60 0.68 0.19 0.88 2.37 No. 23,084 144 162 46 210 23,654

*NRCF: Narrower Ratio Conversion Frequencies

All detected 5:3 and 3:5 octads were used for germination, isolated in single spore cultures and backcrossed, in order to determine their genotypes in respect to K/P factors: white spore cultures were backcrossed to wild-type (P), + and - mating types, 110 and red. spore cultures to white (P) +, and -.

The attempt was to classify 5:3s and 3:5s as either of the "S" or the "S/A" type, according to the spore genotypes in each octad. The analysis was much hindered by poor germination of the spores and high infertility among the crosses. As a result in not one octad of the 239 totally isolated, out of the 492 detected ones, were all 8 spores analysed.

In nearly all cases, only one kind of spore (red or white) provided the crucial information as to the type of octad ("S" or "S/A").

The least number of spores - red or white - that it is necessary to have fully analysed in order for the octad to be classifiable as either "S" type or "S/A" type is as follows:

If we define as: class of : the 5:3 ratio from wild-type (K)xw-78(P) class p : the 3:5 ratio and class Joie' : the 5:3 ratio from wild-type (P)xw-78(K) class p` : the 3:5 ratio from Figures 1, 2, 3, and 4 we may conclude that:

to De "S" 2 red spores HCF or 1 white spore LCF are needed for 04 to be "S/A" 4 red spores LCF or 3 white spores HCF are needed

to be "S" 1 red spore HCF or 2 white spores LCF are needed for p to be "S/A" 3 red.spores LCF or 4 white spores HCF. are needed 111

to be "S" 2 red spores LCF or 1 white spore IICF are needed for .1Y to be "S/A" 4 red spores IICF or 3 white spores LCF are needed

to be "S" 1 red spore LCF or 2 white spores IICF are needed for r to be "S/A" 3 red spores IICF or 4 white spores LCF are needed

According to the above it is possible to find out the type of an octad, "S" or "S/A", with less than 8 spores analysed.

There is even a chance, however small, for an octad with only one spore analysed to be conclusive, if the spore was the critical one, e.g. 1 white spore LCF in class c)( .

The number of octads in each of the o' , r , 0, and classes, which were isolated, analysed and classified as either

"S" or "S/A", is given in Table 8, along with the chances that "S" and "S/A" types had of being detected in the present data from red or white spores in the classes oe , p c)( and f). TABLE 8 The "S" and "S/A" types of octads in each of the , and classes.

"S" type "S/A" type Conclusive Analysed Isolate. Class *R only W only R+W Total R only W only R+W Total octads octads octads No. 3 1 0 4 2 1 0 3 7 21 68 **Chances 5.00 3.67 8.70 1.60 2.00 3.60 No. 3 0 0 3 0 0 0 0 3 14 37 0 1 Chances 5.00 0.70 5.70 1.00 0.20 1.20 No. 3 1 0 4 1 2 0 3(+1)+ 8 11 38 Chances 3.90 4.66 7.90 1.00 2.00 2.60

n ' No. 20 4 1 23 4 8 0 12 35 60 96 Chances 27.65 14.80 42.30 8.00 5.60 13.60 Total No. 29 6 1 34 7 11 0 19 53 106 239 Total chances 40.95 23.83 64.60 11.60 9.80 21.00

R: conclusion on the type of octads derived from red spores only; w: from white spores only; R+W : from both kinds of spore. ** Chances that each type of octad had of being detected from the respective kind of spore and class from both conclusive and inconclusive octads One octad in class 04' was deduced indirectly to be "S/A" type: this octad,c 1 32, is discussed in the text. 113

The number of octads of "S" and "S/A" type in each of

the o(, A and r> classes in which either all red spores F or all white ones were fully analysed, is given in Table 9.

TABLE 9

Octads with either all red spores or all white spores analysed

Class No. of octads with

All red spores analysed All white spores analysed

"S" type "S/A" type "S" type "S/A" type

1 0 1 1 1 0 0 0 0 0 1 1 4 4 0 3

Total 6 4 2 5

The chances were calculated as the sum of the chances of all analysed octads to be "S" or "S/A" type according to the amount and kind of spores - red, white or both - contained in

each one of them. For example in class f, (3:5's in wt(P)xw-78(K)) the chances of an octad to be "S" or "S/A" (Table 10) were

calculated as follows:

When all 3 red spores or all 5 white spores are

analysable, the octad is 100% classifiable. When less than 3 red spores are analysable, "S/A" type is not detectable while "S" type has 67% and 33% chances to be detected with 2 and 1

spores analysable respectively.

With 4 white spores analysable, "S" type will be detected 3 if the missing white spore is LCF: chance = 3 = 60%. "S/A" will 114

be detected if the missing white spore is HCF: chance = 3 = 20%. With 3 white spores analysable, "S" type will be detected if 1 3 both the missing white spores are LCF: chance = 4 x 7 = = 30% (.1 2 : the chance of the second missing spore being LCF, since, after the loss of the first LCF white spore, there are two LCF left among four spores). "S/A" type will not be detected with less than 4 white spores analysable. With 2 white spores analysable, "S" type will be detected if all 3 missing white 1 spores are LCF: chance = -- x 7 x 3 = 10% 1 : the chance of the third missing spore to be LCF, after the two other missing white spores were LCF). With only one white spore analysable neither'of the two types of octad can be detected.

The corresponding chances of detection of "S" and "S/A" types in class p are calculated in a similar way while those in classes 0( and 0(1 , for the red spores are like the ones for the white spores in tv, and for the white spores are like the ones for the red spores in p

TABLE 10 Chances for "S" and "S/A" types of octad in class p , (calculation method suggested by Dr. Lamb).

Chances of detecting

Number and kind of spores available "S" type "S/A" type

3 red spores 1.00 1.00 2 " It .67 I " It. .33 5 white spores 1.00 1.00 4 " It .60 .20 3 It 11 .30 2 " It .10 1 tt 115

For example, according to Table 11, octad p 85, described later on in this section as having 2 red and 4 white spores analysed, had (.67) from the red spores and (.60) from the white ones, total (1.27) chances to he "S" type, while it had (.20) chances to be of the "S/A" type. This particular octad however, was not conclusive.

The difference between the detected number of octads in each class (Table 7) and the isolated ones (Table 8) is due to bad germination, while the difference between the number isolated and the analysed ones is due to infertility of the single spore cultures. This infertility is also the main cause for the difference in number between the analysed octads and the conclusive ones, although some of the analysed octads are not conclusive because they are missing the critical spores. One such example is octad 85 (Table 11) in which two red spores and four white ones were fully analysed and yet no conclusion could be reached for the type of this octad.

TABLE 11 NRCFs of the analysable spores in the octad 85

Sample Spore Crossed to NRCF Size r85w1 9r+ 1.45 LCF 207 (3'85w2 9r+ 1.84 LCF 217 p5w3 9r+ 3.45 LCF 318 r85w5 92- 13.08 IICF 382 185R1 9-4w78- 13.10 IICF 321 V85R 2 9-4w78- 9.45 IICF 381

continue or page 117 116

These were three red spores and 1 white one and should be the members of two pairs of sister-spores. Of the four spores, a(' 32R3- and <"32R4- clearly form a pair of identical sister- spores, since they are both red, of - mating type and HCF. As a result 32w/- and a 32R2- should be the pair of non- identical sister-spores contained in the octad, (as all 5:3s contain at least one such pair), on the basis that theY are identical in the mating type (-) and the gene conversion type, LCF, differing only at the w-78 locus. In the analysis of class ce (Figure 3) it was shown that the pair of non-identical sister-spores was accompanied by the K factor in "S/A" type of octads, the spores showing low conversion frequency, and by the P factor in "S" type of octads, the spores showing high conversion frequency. Accordingly, octad a' '32 was deduced to be of "S/A" type.•

continue on page lig 117

This octad lacks the critical spores that could identify it either as "S" (namely 1 red spore LCF or 2 white spores IICF) or as "S/A" (namely 3 red spores IICF or 4 white ones LCF).

There was only one octad of the "S" type which was concluded to be such on the basis of both red and white spores.

This was octad (31 65 and contained 1 red spore LCF and 2 white ones IICF.

In another octad classified as "S/A" the conclusion was reached indirectly, using both red and white spores. This was octad oe 32 (5 red: 3 white) and contained four analysable red and 1 white spores (Table 12).

TABLE 12

NRCFs of the analysable spores in the octad Ge 32.

Sample Spore Crossed to NRCF Size 04'32w1 9r+ 2.10 LCF 190 c"32R2 1w3KEC+ 4.26 LCF 469 o<132R3 1w3KEC+ 10.60 IICF 264 c(' 32R4 1w3KEC+ 10.23 IICF 391 41 32R5 9-4w-78- 9.05 IICF 265

This octad contained neither 4 red spores IICF nor 3 white ones LCF as needed to be of the "S/A" type.

It was noted however that among the analysed spores of

this octad were all the four ones of minus (-) mating type.

continue on page tts 118

4. DISCUSSION

The analysis of the 5:3 and 3:5 ratio octads was carried out according to the theoretical considerations given in this section, paragraph 2, with the aim of finding out whether these octads result from hybrid DNA in two chromatids, (symmetrically) or in one chromatid only (asymmetrically). These octads have been considered to occur in two types: "S" type, coming exclusively from symmetrical hybrid DNA and "S/A" type coming from either symmetrical or asymmetrical hybrid DNA. On the assumption that symmetrical hybrid DNA yields "S" and "S/A" types of octads with equal frequency, one may have the following possibilities:

1. If "S" and "S/A" types of asci appear with equal frequency, one may conclude that they are nearly all due to symmetrical hybrid DNA.

2. If "S" type of asci appears with less frequency than "S/A" type, one may conclude that both symmetrical and asymmetrical hybrid DNA occurs and may calculate the relative frequencies of

occurrence. And

3. If "S/A" type of asci occur only, one may conclude that they are all due to asymmetrical hybrid DNA.

Before discussing the data in the present work, I would like to discuss the basis for justification of the above assumption.

Let us consider the occurrence of the "S" and "S/A" types in 5:3s after dual hybrid DNA formation according to: 119

1. The Holliday (1964) model, and

2. The Whitehouse and Hastings (1965) model.

Let us also suppose that a GC base pair represents the wild-type allele and AT base pair represents the mutant allele.

Figure 5 shows that, on the Holliday dual hybrid DNA, correction of GT mispair to + can give either "S/A" or "S" type of 5:3 octad, depending on which chromatid it occurs in.

Similarly, correction of AC to + gives "S/A" or "S" type of

5:3 octad. Nispairs GT and AC may have different correction frequencies, but this is not expected to affect the frequencies of "S" and "S/A" types in the whole crop of spores produced in cross, since any such difference in one meiosis would be balanced by a difference in the opposite direction in another meiosis. However, if G and A strands are exchanged with a different frequency from C and T strands - if for example there is preference in strand breakage - then different correction frequencies between GT and AC mispairs would affect the frequencies of "S" and "S/A" types of 5:3 octads. This may provide any explanation for a possible excess of "S" type of asci over "S/A"

type which otherwise is unexpected.

120

A + B

A + B chromatids in Holliday a m A b dual hybrid DNA

a m A b

this gives

either or

A +G B A +G B C C

A + G B A + A B m T m C

a +m CA b a +m T b A A a m T b a m T b

after correction to + (production of 5:3s) GT + AC-4+ GT-4+

A+B A+B A+B A+B A+B A+B A+B A+B A+B A+B AmB MB A+B AmB A+B A+B amb a+b a+b a+b atb a+b a+b amb amb amb amb amb amb amb . amb amb

"S/A" type "S" type "S" type "S/A" type ,

FIGURE 5 "S" and "S/A" types of 5:3 octads expected from

correction of the two different mispairs resulting from hybrid DNA formation according to Holliday (1964) 121

On the Whitehouse & Hastings (1965) model, things are somewhat different, (Figure 6). Here the two mispairs occurring in a single meiosis are the same, though there are two different possibilities in the the kind of mispair that may occur. If symmetrical hybrid DNA results in the same mispairs in both chromatids, types "S" and "S/A" may be expected to appear with the same frequency, especially in the absence of cryptic heterozygosity. The frequencies with which the two kinds of mispair may be formed in a number of meioses do not affect the frequencies of "S" and "S/A" types of either 5:3 or 3:5 octads, since these types represent two alternative events in a single meiosis.

122

A C B

A C B chromatids in Whitehouse & A Hastings dual hybrid DNA a m T b A am T b this gives

either or G A + B A +G B C C G 11 cA (1) A +m T B A B (1) (2) a + G b m1' a T Ac b (2) b am A TA T am b

correction to + (production of 5:3) AC—t+

(1) (2) (1) (2) A+B A+B A+B A+B A+B AfB A+B A+B A+B A+B A+B AmB A+B AmB A+B A+B a+b a+b amb a+b amb a+b a+b a+b amb amb amb amb amb amb amb amb

"S/A" type' "S" type "S/A" type" "S" type

FIGURE 6 "S" and "S/A" types of 5:3 octads expected from

correction of mispairs resulting from hybrid DNA formation according to Whitehouse & Hastings (1965) 123

It seems therefore that the assumption of equal frequencies of the "S" and "S/A" types of 5:3. and 3:5 octads resulting from symmetrical hybrid DNA, may be justified when hybrid DNA results in two different mispairs in the two chromatids involved, (Holliday model), if the two alternative events producing them have equal chances to occur in any one cross; it may be on stronger basis when the same mispairs are formed in both chromatids (Whitehouse & Hastings model). In both cases there is a reservation that there may still be some influence from possible heterozygous cryptic mutations nearby.

There have been a few reports of results that were explained on the basis of different behaviour of the mismatches in the two chromatids. For example Emerson (1966) found a difference in the direction of correction between the odd - and even - ratio conversion classes (6:2s were more than 2:6s but 5:3s were less than 3:5s), with the w-62 mutant in Ascobolus immersus. He attributed this to a different mismatch in the two chromatids (which Holliday's model predicts) and to a differential response by the correction enzyme.

Differences between odd- and even- ratio conversion at the g locus of Sordaria fimicola were explained by Whitehouse (1974) on the basis of a two-enzyme hypothesis of conversion, involving a correction and a degradation enzyme. According to his suggestion the odd-ratio asci could be generated by either enzyme alone, while the even-ratio asci would require the correction enzyme acting in one mispair and either the degradation or the correction enzyme in the other. Thus there is a possibility in a Sordaria fimicola system, which is defective 124 in the correction enzyme, that all 5:3 and 3:5 octads result from the action of the degradation enzyme, which produces only tritypes, that correspond to "S/A" types here (Figure 3 in Whitehouse 1974). This would falsely suggest, especially if aberrant 4:4s are aot deteCtable, that odd-ratio octads originate from hybrid DNA formation in one chromatid rather than in two. with subsequent enzyme action.

Whitehouse (1974) also comments on the findings by Spatz

& Trautner (1970) that mismatch correction—as revealed by single burst experiments in transfection studies using phage SPP1 of

Bacillus subtilis, may favour wild-type in one heterduplex and mutant in the reciprocal one for the same mutant..

It seems therefore advisable to proceed in the discussion of the data with the appropriate reservation on the justification of the above stated assumption.

Table 8 gives the numbers of the "S" and "S/A" types of octads in each of the of , , 0( and p class6s, which refer to the following octads:

: 5:3s from wild-type (K) x w-78(P)

'' '' It f : 3:5s ,fir : 5:3s from wild-type (P) x w-78(K) and ( 5 3:5s If

It is unfortunate that the number of octads in each class that were conclusively analysed for reasons explained in the

RESULTS of this SECTION, is too small to support a credible theoretical analysis. 125

However, "S" type of octads, which can only be attributed

to hybrid DNA occurrence in both chromatids at the w-78 locus, were detected in both kinds of cross. A comparison between the detected numbers of "S" and "S/A" types of octads would be

meaningful only if we consider the chances that these types of octad had of being detected in the total number of octads analysed (Table 8).

On the assumption that the number of detected octads of

either type ("S" or "S/A") is equally proportional to the chances that they had of being detected in the total number of octads analysed, then

No of "S" octads detected Total chances of detection of "S" octad_ No of "S/A" octads detectedTotal chances of detection 771-77777

From Table 8 we find that

64.60 Total chances of detection of "S" octads = 3.07 (1)Total chances of detection of "S/A" octads 21.00

and (2 No. of "S" octads detected _ 34 = 1 79 'No. of "S/A" octads detected - 19

It is evident that, if the above assumption holds true,

there is a discrepancy in the present data between the numbers of octads detected in each type and their total chances of

detection as such.

According to the value of ratio (1) there should be three

times as many "S" type of octads as "S/A" ones. This is not

shown by the value of ratio (2), and it may be that the number

of "S" octads detected is too small or the number of "S/A" octads

detected is too high, compared with the expected one from their

total chances of detection. 126

The total number of conclusive octads was 53 (Table 8) and according to ratio (1) there should be

Expected No. of "S" octads detected = 40 and

Expected No. of "S/A" octads detected = 13

So if these numbers had actually been detected here, we could conclude that the results were compatible with nearly exclusive occurrence of hybrid DNA in two chromatids, since then the numbers of "S" and "S/A" octads could be expected to be equal, if their chances of detection were equal.

The present results, though not completely satisfactory,. do nevertheless show the occurrence of dual hybrid DNA and they also set a trend on the value of ratio (2). The number of "S" octads is almost twice that of the "S/A" octads. This is evidence against the asymmetrical hybrid DNA being formed with equal (or higher) frequency than symmetrical one.

If the two kinds of hybrid DNA were formed with equal frequency, then, with also equal chances of detection, "S/A" type of octads would be three times more frequent than "S" type ones.

Considering however that in these data "S" octads have three times higher chances of detection than "S/A" ones, the two types would appear with equal numbers here and this would suggest equal frequency of symmetrical and asymmetrical hybrid DNA formation.

This is not occurring with the present data where the values of "S", (34), and "S/A", (19), octads are significantly different. 127

For a predominent formation of asymmetrical hybrid DNA, the number for "S/A" octads would be expected here to be significantly higher than that of the "S" ones, which in no case occurs here.

So the present data support rather wide occurrence of symmetrical hybrid DNA and do not exclude the possibility of asymmetrical hybrid DNA occurring sometimes, but not as frequently as the symmetrical one.

Let us now consider those octads with either all red spores or all white ones fully analysed, (Table 9). Each of these octads had chances of detection 1, so their numbers are directly comparable. Though numbers are small we may say that the ratio of

No. of "S" octads = 8 No. of "S/A" octads 9 is consistent with all hybrid DNA'at this site being symmetrical

(giving No of "S" octads = No of "S/A" octads) or with hybrid

DNA being mainly symmetrical but sometimes asymmetrical. It is not consistent with hybrid DNA being mainly or entirely . asymmetrical.

It is most probable that asymmetrical DNA is actually formed, even to a comparatively small extent, as it is also suggested by Wickramaratne (1974), who found that meiotic segregation classes at w-78 and w-10 and postmeiotic segregation ones change in opposite directions with changes in temperature and this is best explained with single hybrid chromatid hypothesis. 128

Finally, it is noticeable that the "S" type was detected in both 5:3 and 3:5 classes and in both (reciprocal) crosses used here. This indicates that symmetrical hybrid DNA is a regular event in this site and it is not a special feature associated with a specific cross or a specific class.

In the analysis of 5 : 3's and 3 : 5's the possible effect of crossovers associated with conversion must be considered (Whitehouse, personal communication). These crossovers may be up to 50% of the conversion events (Whitehouse, 1974) and if they occur in the interval between w-78 and P/K markers 'they may change an "S" type of octad to one of "S/A" type and vice-versa, thus influencing the relative numbers of detected "S" and "S/A" types, if the two types are formed with unequal frequency in the progeny. With the present results and in the case of nearly exclu- sive occurrence of symmetrical hybrid DNA, the discrepancy between the expected numbers of "S" and "S/A" types and the detected ones may be due to such crossovers occurring between w-78 and P/K markers. "S" type is more frequent here than "S/A" type and a figure of 10% to 15% of conversion-associated crossovers would explain the size of the present discrepancy. This would not occur with the general system proposed on page 229 , whether the auxiliary marker is linked or unlinked to the marker under study, since then PDs, NPDs and Ts inter- change after such crossovers occur. The frequencies of these classes however had been considered in the formulation of the system. 129

CONCLUSIONS 5. CONCLUSIONS

1. There is evidence that in LCF crosses of the w-•78 locus of Ascobolus immersus, the hybrid DNA that gives rise to 5:3 and

3:5 conversion classes occurs mainly symmetrically in both chromatids of a- pair.

2. If "odd" narrower ratio classes arise mainly by symmetrical

hybrid DNA, we may conclude that the same is true for the "even"

narrower ratio classes (6:2 and 2:6), since they only differ

from the first ones in the correction of a single mispair, unless the correction system distinguishes between mispairs produced by dual or single hybrid DNA. This case is rather improbable because a 5:3, for example, may contain one of two

different kinds of mispairs which are the same two kinds, Whether they are produced by dual or single hybrid DNA. We

may therefore conclude that gene conversion in LCF crosses of the w-78 locus arises mainly from symmetrical rather than asymmetrical hybrid DNA.

3. Symmetrical hybrid DNA occurs at the w-78 locus, in LCF

crosses, irrespective of whether the mutant site is linked to the (P) factor and crossed to wild-type (K), or it is linked to the (K) factor and crossed to wild-type (P).

4. Asymmetrical hybrid DNA may be formed occasionally but most

probably not as frequently as symmetrical hybrid DNA. 130

SECTION III

Background Studies

1. The nature of w-78 phenotype a) The development of the "collapsing" character

The vast majority of the white ascospore mutants studied here (PART B) produce white spores that remain unchanged on the collecting medium, turgid and smooth, as they were the day

of collection and as long as the medium does not dry out,

(maximum period tested: 4 months). After that they become sickle-shaped, just as the wild-type ones do without changing size or texture.

The w-78 white ascospores look normal and smooth for about

a week after dehiscence. After this time, and as if they

absorb water, they start swelling and become bigger than wild- type spores. At the same time they lose their smooth surface

and look rather rough, while they often become fragile upon

transferring to germination medium. It is noticeable that from this stage on, no w-78 ascospore has germinated.

With drying out of the spore collecting medium, the w-78 ascospores shrink to a size almost half the wild-type ones, presumably through loss of water. As they shrink they

acquire a. light pinkish colour and do not regain their original shape, as other white as well as red ascospores do, after the addition of fresh water to the dried spore

collecting medium. In cases when the spore collecting medium

was deep enough to last for two or so months before starting

drying out, the w-78 ascospores had almost all collapsed by 131

PLATE 15 A 0:8(PD)octad from cross of w-78 x a non-collapsing

white spore mutant, one week after dehiscence, (see text)

0

PLATE 16 A 0:8 (PD) octad from cross w-78 x a non-collapsing

white spore mutant, 3 months after dehiscence 132

+ PLATE 17 A 4:4 (NPD) octad from cross w-78 x w-78 , one week

after dehiscence, when white spores start collapsing

+ PLATE 18 A 4:4 (NPD) octad from cross w-78 x w-78 , 3 months

after dehiscence, with white spores collapsed and

smaller than red spores 133

PLATE 19 A 2:6 octad of a tetratype from cross w-78 x ainon-

collapsing'white spore mutant, containing 2 red spores

2 white 'non-collapsing' ones and 4 w-78 'collapsed'ones

PLATE 20 A 2:6 octad from cross w-78 x a 'non-collapsing' white

spore mutant, resulting from conversion at the w- 78 site

(see text) 134

the end of the third week.

The "collapsing" feature of the w-78 phenotype is very

characteristic and has been found in another three mutants described in PART B. First in the UVKw8 mutant, 'which gives no recombination at all with w-78. Second in NGw11, which

shows the same course of events as w-78, but is unlinked and not synnemal to it. And third in BW4.6 mutant, whose white

ascospores collapse as soon as two (2) days after collection.

The linkage relations of the last mutant to the other three are unknown since it was completely infertile in crosses to other white spore mutants.

b) w-78 in crosses to other white mutants The w-78'mutant in crosses with other "non-collapsing" white

spore mutants, gives parental, non-parental and tetratype

segregations which, after dile time, appear as following: 0:8'

(genotypically PD) contained 4 collapsed (w-78) and 4 non-

collapsed (other mutant) spores, Plates Nos. 15 & 16. Non- parental ditypes contained 4 red (wild-type) and 4 white collapsed spores. This indicates that the "collapsing" character of w-78 is epistatic to the "non-collapsing"

character of the other mutant, Plates No. 17 & 18. The tetratype segregation contains 2 red spores, 2 white non- collapsed spores and 4 white collapsed spores, in accordance with the epistatic nature of w-78, Plate No.19. The 2:6

tetratype segregation in this cross is distinguishable from 2:6's resulting from gene conversion to wild-type at the w-78

locus, due to the "collapsing" and epistatic nature of this

mutant. These latter 2:6s contain 4 white non-collapsed spores, 135

2 white collapsed ones an6 2 red ones, Plate No. 20. In this

way one can assign gene-conversion to w-78 in crosses to other white spore mutants as long as there is no need for germination of these ascospores, because by the time the 2:6s can be visually classified as either tetratype segregations

or gene conversion at the w-78 locus, the germinability of the w-78 spores has been lost.

2. The re-isolation of "K" derived wild-types a) Method of isolation The w-78 and w-10 mutants had originally been derived from crosses of the two original wild-type strains P5- x K5+

(Emerson & Yu-Sun, 1967). Both these mutants in crosses to P5- gave low conversion frequencies (about 3%) while from such crosses the w-78(P) and w-10(P) (presumably the mutants in P

background for the neighbouring region to the w-78 and w-10 sites) were reported to have been isolated. The w-78(P) x wt(P) and w-10(P) x wt(P) crosses gave high conversion frequencies

,(about 12% to 18%). The K5+ wild-type (which does not give fertile crosses any more) gave only poor crosses with w-10(P),

with a predominant 3:5 class among the aberrant ones, while the

previous two crosses were quite similar in the pattern of conversion with 6:2 as the predominant aberrant class, (Emerson & Yu-Sun, 1967).

It was considered here that the original w-78 strains probably had the mutation in "K" background, much as the w-78(P) was the mutation in "P" background. According to this reasoning one

could re-isolate the "K" wild-type in 6:2 conversion octads from

crosses w-78 x wt(P), where 4 wild-type spores were expected to 136 be of "P" type, while the other 2 wild-type spores coming from conversion to wild-type at the w-78 locus would be of the "K" background. There is evidence (Emerson & Yu-Sun, 1967; Wickramaratne, 1974) that the P/K factors do riot usually co- convert with w-10 or w-78, although being located close to them.

From the six wild-type spores of a 6:2 octad from 9r+xw-78-, only four were fully testable and three of them gave high conversion frequencies while the fourth gave low conversion frequencies (Table 13). The same result was found with + another 6:2 octad from 9r x22-7w-78 (Table 13). These two low converting wild-types were deduced to have arisen from conversion at the w-78 locus and arc believed to be of the "k" type.

A parallel attempt was made to isolate the one wild-type spore of the 1:7 class in y-78LCFxw-10LCF. This cross is usually infertile but eventually one such wild-type was isolated and is also considered to be "K" type.

These new "K"s were crossed with several new and old w-78(P) and w-10(P) strains and the result was invariably low conversion frequency (about 2% to 5%), with 3:5s being the least represented narrower ratio aberrant class (Table 13).

The original K5+ was however described as high converting by Emerson and Yu-Sun, who found a total conversion frequency of 17% in w-10(P) x K, and in their results 3:5s were the predominant conversion class, unlike any other w-78 or w-10 crosses. The above authors state that crosses of w-10(P) to. 137

K were poor, with few counts made, so their results may be

unrepresentative, especially in view of the consistency of the data in the present work.

TABLE 13

Analysis of 6:2s from wt(P) x w-78(K), for re-isolation of the "K" wild-types

Octad ratios 7, Spore Total Cross isolate 4:4 6:2 2:6 5:3 3:5 8:0 0:8 7:1 1:7 NRCF Octads A-1R- 82.19 13.243.19 1.37 17.80 219 A-2R+ 87.25 8.753.25 0.50 0.25 12.75 400 *A-4R+ 95.70 3.10.37 0.56 0.18 4.30 535 sl A-5R- 89.05 8.901.85 0.18 10.94 539 8 A-3R: infertile A-6R, A-lw, A-2w: did not germinate 9K1- 84.1810.201.53 15.81 392 co 9K2- 84.6713.311.79 0.10 0.10 15.22 946 1 :.--.= I 9K4- 83.62 11.6130.87 0.86 16.37 232 (-4 (-4 9K5+ 96.63 2.4 .37 0.49 0.06 3.36 1604 9K3: infertile 9K6, 9Kwl, 9Kw2: did not germinate

*"K" type re-isolates in crosses: A-4R+x9-4w-78-, 9K5+xEC11w-78.7-

138

m K

1 m ) K ) w-78 m K ) (LCF) 2 m K ) 3 + P + 9r P ) (wt HCF) 4

Formation of hybrid DNA symmetrically asymmetrically

m K m K m K' m K ;_— + K K m P + P + P + P + P + P + P + p Correction of mispairs to wild type

1 m K 3 w-78(LCF)

2 - K wt "K" (LCF) + P 3 ) + P ) wt "P" (HCF) + P ) 4 ) + P

FIGURE 7 The re-isolation of "K" derived wild-type strains from a w-78(LCF) x wt(HCF) cross, using a 6:2 octad, where conversion at the w-78 locus is not accompanied by conversion at the K/P site(s). The factor K in this work corresponds to factor "0" in Wickramaratne, 1974. 139 b) Tests of the "K" derived wild-types

The "K" strains isolated were subsequently crossed to w-78LCF

strains and from these crosses the six white spores of the

2:6 octads were isolated, germinated and backcrossed to "P" wild-type to check their conversion frequency. The idea was that if the "K"s are low converting, there should be not

one high converting spore among the six white spores, because crosses to w-78LCF would be homozygous for "K", Figure 6. Due to bad germination, only in one case could five spores out

of the six be fully tested (octad Kw7, Table 14). Not one

gave high conversion frequency. Moreover, with eight other white spores tested from five different 2:6s, all gave low

conversion frequency (Table 15).

A second test was done with crosses of "K" wild-type to w-78(P) strains from which the 2:6s were isolated. In one such 2:6 five white spores were fully analysed (none of them

was fully fertile) and there was evidence that the four white spores were giving high conversion frequencies while the fifth

one - 1Kw78.3- in the Pedigree chart - gave low conversion

frequency (2.55%). It is considered that this mutant spore was produced by conversion to mutant at the w-78+ site, and was carrying the "K" controlling factor. Helmi has also found some white spores with LCF in five octads he analysed

of 2:6 ratio in the cross (K)wt x w-78(P). One of these

w-78(K)s is 5w5KEC+ used in this work.

It is finally concluded that the re-isolated "K" wild-types

are low converting strains, carrying factor "K" closely

linked to the w-78 and w-10 loci.

140 m K 1 m ) K ) w-78 m K ) TLCF) 2 in K )

3 ) ) derived "K" wild-type ) 4 ) Formation of hybrid UNA symmetrically asymmetrically 111 K m K In K m K m •K K K K Correction of mispairs to mutant

1 ) m K ) w-78 ) TLCF) 2 m K ) ) w-78(LCF) from 3 conversion m K K 4 Iderived "K" wild-type

FIGURE 8

Test of the derived "K" wild-type: a2:6 class from "K"xw-78LCF contained six LCF white spores, of which four originated from strain w-78LCF and two from the "K" wild-type, through gene conversion at the W-78 site. TABLE 14

Conversion frequencies of "K" wild-type re-isolates in different crosses

Total Cross Octad ratios % NRCF Octads

4:4 6:2 2:6 5:3 3:5 8:0 0:8 7:1 1:7

A-4R+xECllw-78.7- 95.81 3.46 0.21 0.38 0.07 0.02 0.02 4.11 4130 A-4R+xS1-u-10- 95.20 2.99 1.19 0.59 - - 4.79 167 91:5+xEC11w-78.7- 96.31 2.62 0.25 0.60 0.17 0.01 - 0.006 0.006 3.65 30068 9K5+x421w-10- 97.26 2.28 0.45 - - 2.73 219 TABLE 15

Conversion frequencies of white spores, from 2:6s of "K" wt re-isolates x w-78LCF, in crosses to "P" wild-types (61- or 9r+)

Octad ratios (%) NRCF Total Spore Octads isolate 4:4 6:2 2:6 5:3 3:5 8:0 0:8 7:1 1:7 Kw7.1+ 97.60 0.91 0.91 0.22 0.34 2.40 875 Kw7. 2+ 97.71 0.96 0.43 0.52 0.26 2.25 1140 Kw7.3- 98.43 0..62 0.94 1.57 1595 Kw7.4+ 94.68 2.5 1.25 1.25 0.31 5.31 320 Kw7.5+ 96.50 1.78 0.89 0.56 0.16 0.08 3.47 1232 Kw2.3+ 98.09 0.70 0.65 0.05 0.10 1.50 1996 Kw2.4- 99.56 0.87 0.87 1.74 459 Kw2.5+ 96.94 1.10 0.76 0.50 0.67 3.03 1180 Kw6. 4+ 97.84 1.07 0.74 0.20 0.13 2.14 1483 Kw6. 5+ 94.49 1.38 0.42 0.31 0.31 0.05 2.47 1879 Kw3. 4+ 97.79 1.28 0.36 0.36 0.18 2.18 545 Kw8.2- 97.75 0.84 0.70 0.14 1.68 1427 Kw9.2+ 98.09 0.70 0.65 0.05 0.10 1.50 1996

Broken line separates spores coming from different octads 143

3. Germination and Fertility tests a) Germination Ascospore germination in Ascobolus immersus has many problems

. (Van Brummelen, 1968). Several workers failed to obtain germination of these ascospores, even when they were passed through the alimentrary canal of a rabbit, by feeding to the

animal with bread, (Zukal, 1889, quoted by Van Brummelen,

1968).

There were two reports of good germination of Ascobolus immersus ascospores, one by Yu-Sun, 1964, on the Pasadena

strains. However, when her germination medium was used for

most of the present work, ascospore germination of the original and other derived strains rarely exceeded 30%, being usually

under 20%.

This extremely poor germination greatly impaired the progress of this work, where analysis of full octads was very often needed. There was great variability in the germination

frequencies among crosses of different strains, among repeated crosses of the same strains, even among groups of

spores of the same cross, and with factors such as spore age, treatment or non-treatment with.pronase and kind of spores (wild-type or mutant). Repeatability was poor so no standard

optimum conditions for germination could be established.

An experiment designed to relate germinability to the age of the spores,their time of dehiscence and the presence or

absence of pronase treatment reflected this variability

(TABLE 16). 144

TABLE 16 Ascospore germination, %, in the cross Bal-x91-2R+

Day of collection of the spores during dehiscence of perioderiod Spores: With pronase Without pronase days after collection 1st 2nd 4th 6th 1st 2nd 4th 6th 1 15.3 15.22 0.0 15.74 26.77 37.17 7.55 0.97 2 13.67 0.64 0.0 11.68 2.77 8.36 3 2.34 4.15 11.4 11.62 4 5.52 24.18

The data in Table 16 represent one single experiment, with samples of over a hundred spores for each value. A repeat experiment gave significantly different values for each particular case and a different overall pattern.

Tests were also made using germination medium supplemented with thiamine and biotin, two vitamins that were added to minimal medium as Pasadena strains of A. immersus are deficient in their synthesis (YU-Sun, 1964); furfural, reported to activate dormant ascospores of Neurospora crassa (Emerson 1954), Sordaria brevicollis germination medium (YU-Sun et a1,1977). No pronase treatment was included, (TABLE 17). 145 TABLE 17 Germination of wild-type spores from cross BIH-xpfr-1+, in

different media

Media and No. of spores germinated Time at 30°C (Hours) SG(80)* G(64) G+F(60) G+T(40) G+B+T(40)

19 0 0 3 0 0 34 0 1 5 1 0 47 0 6 9 0 0 62 0 1 3 0 0 77 0 2 4 3 0 2nd heat shock; all were then covered with G+F+T

105 0 0 0 3 5 Total Germination 0.0% 15.62% 37.5% 17.5% 12.5%

SG: Sordaria brevicollis germination medium G: Ascobolus immersus germination medium (Yu-Sun, 1964)

F: furfural: 1ppm B,T: Biotin, thiamine, as in minimal medium

Parentheses: number of spores tested

Furfural seems to increase the germination frequency but, in

absolute terms, the values obtained are far from satisfactory.

The germination reached high levels - up to 90% - after the introduction of the horse dung germination medium by flelmi.

b) Fertility A second big problem apart from bad germination has been a high

degree of infertility among crosses. During most of this work 1 only 3 of the crosses were fertile, and only one in eight to 146 ten times were the fertile crosses fully fertile, with half or-,-, of the petri dish surface covered with apothecia.

As with germination, fertility varied unpredictably with different strains, and repeated crosses of the same strains.

There were strains, such as K5+ which, for reasons unknown, stopped producing any fertile crosses with any other strain of opposite mating type, or give in general very poor crosses like the w-10 LCF strains.

Fertility, defined as the ability of a cross to produce viable and dispensable ascospores, was found to be impaired at any stage between protoapothecium formation and octad dehiscence. This agreed with reports by Van Brummelen (1968) of sudden inhibition of fruit-body development in species of Ascobolus, at different stages of the development, even when asci and ascospores have ripened; these ascospores are set free by decay only.

Van Brummelen (1968) also reported that on most media used,

(before Yu-Sun's 1964, defined media) for fructification of Ascobolus immersus, the number of fruit-bodies formed in each generation gradually diminished after several isolations. The infertility of the crosses in the present work may indicate a similar situation with Yu-Sun's (1964) defined media.

It was thought that the strains used may have been adapted to continuous subculturing in rich medium, and a poorer than usual crossing medium may help strains complete their 147 reproductive cycle. Two kinds of cross were made: wild-type x wild-type, (1311R1-x91-21::,) and wild-type x mutant

(1311R1-xEC11w78.5+)' on three media: minimal, minimal + 1.5g/1 yeast extract and minimal + 3g/l.yeast extract (as in usual crossing medium). No casamino acids were added. The results were as following: In minimal + 3g/1 yeast extract, seven of the twenty replicas of each kind of cross were fully fertile.

This is similar to the result with usual crossing medium. In minimal + 1.5g/1 yeast extract, six of the twenty replicas of each kind of cross were partially fertile (small number of apothecia and spores produced). In minimal medium all 40 • crossing plates of both crosses were sterile.

It is concluded that poorer than usual crossing medium does not help the present strains to form fully fertile crosses.

Prakash (1963) reported that the introduction of non-absorbent cotton wool into crossing medium increased fertility in Neurospora crassa, so this was tried here. The outer surface of a 5cm diameter petri dish lid was covered thinly with cotton wool and was placed1 wool surface up/ in a 9cm petri dish, which was then autoclaved. Several such sterilized dishes were prepared and sterile crossing medium was poured into them, with the cotton wool covered small petri dish in the middle of the larger one. These dishes were then inoculated with two wild-types or wild-type and mutant strains for crossing. The results were that the fungus hardly ever grew over the cotton wool, with production of apothecia and ripe ascospores even poorer than in usual crosses. 148

Another attempt to improve fertility of the crosses was by

using the germination horse dung medium for crosses used plain or supplemented with all the contents of the usual crossing medium. Unfortunately, no promising results appeared

in this case either.

The fertility problem of the Ascobolus immersus strains still

awaits solution.

4. Search for outside markers Appropriate markers flanking the w-78 and w-10 sites would greatly assist studies of the control of gene conversion and of the wider ratio segregations (Lamb 1972).

The search for outside markers started with mutagenesis of crossing cultures by UV irradiation, or NG treatment (PART B).

The ascospores produced were collected, germinated and kept as single spore cultures in small test tubes. These single spore cultures were left to grow for a week and were then tested on minimal medium for auxotrophy and on media supplemented with antifungal agents for resistance. The following antifungal agents were used: p-fluorophcnyalanine, 100 pg/ml, griseofulvin 10 pg/ml,

Chloramphenicol 100 pg/m1 and sodium desoxycholate lmg/m1 (Ghikas 1973).

Twelve auxotrophic mutants were isolated but were not fully investigated, because none grew well enough on crossing medium to produce even partially fertile crosses. In most of the cases growth lacked vigour and stopped after 4 or 5 days; the

mycelium was thinner than usual. 149

These mutants were possibly carrying many harmful mutants, which were impairing normal growth and crossing even on rich medium.

From the antifungal agents tests two p-fluorophenylalanine and one priseofulvin resistant strains were isolated. Of the former ones the pfr-1 strain has been used here in the detection of aberrant 4:4s, 6:2(2)s and 2:6(2)s. The other two mutants proved to be rather leaky: after 2 or 3 days on the druz medium their growth drastically changed in pattern and rate and eventually stopped, about 5 days after inoculation.

Some of the auxotrophs mentioned before were also leaky: after 2 or 3 days on the minimal medium they would start growing in a pattern similar to the one of the wild-type on the same medium, though with a somewhat slower rate.

Some of the single spore cultures showed a mutated type of growth, restricted or colonial. Five such mutants were found, which in crosses to w-78 all proved unlinked to it.

There was a continuous search for mutations affecting ascospore colour and shape: four granular and one pink mutant were isolated, all unlinked to w-78. Several other types of spore e.g. "elongated", "banded" (with the pigment as a band round the middle of the spore, the rest being white), "restricted" (with a restriction round the middle) and "half red" (with the other half along the longer axis of the spore, white). All these types of spore were phenocopies and not mutants in test crosses. 150

A round spore mutant of Ascobolus immersus has been reported by Hamelin & Gousineau (1974). Round spores were sometimes spotted during this work, but these were usually of an odd number in the octad (more than four spores), and so they were considered malformations rather than mutations.

Use was also made of a method whereby one isolates

Morphological mutants showing parental ditype segregation with a spore colour marker (Wickramaratne 1976, here w-78). Such mutants have a higher probability of being linked to w-78 than the ones isolated in tetratype segregations, in crosses where only one parent was mutagenized and PDs are identifiable. The. difficulty with this method is that red spores, when germinated, shed their red spore wall looking white and so making the recognition of the ditypes uncertain, and germlings of the wild- type spores grew very vigorously, compared with the ones of the potential mutant, and covered the whole area, making recognition of the mutant's growth virtually impossible.

Another disadvantage of this method with Ascobolus immersus ascospores is that the wild-type spores hardly ever germinated on minimal medium, so that the method could not be used for isolation of auxotrophs.

There haS been a constant seareh..for outside markers, especially among ascospore mutants which do not need backcrossing to reveal their genotype.

5. Attempts at higher temperature crosses

Crosses in this work were routinely carried out at 17.5°C. 151 Lamb & Wickramatne (1975) slif.mved that there is a variation in gene conversion parameters with temperature, and so it was decided to investigate the detection of 4:4(2)s, 6:2(2)s and 2:6(2)s in

crosses at 22.5°C. The Pasadena strains of Ascobolus immersus,

according to the above report, grow and cross relatively well at

22.5°C, which was used routinely by Yu-Sun (1964).

In spite of that, all attempts to repeat the crosses described in PART A, SECTION II at 22.5°C failed. Trials were

continued for six months and many strains, in very many combinations were used. Crosses were impaired at any stage between apothecium formation and ascus dehiscence. In most cases

,crossing plates after about 18 days at 22.5°C contained many

small immature apothecia and cultures were turning brown.

A few crosses produced asci and ascospores but rarely would the asci project above the surface of the apothecia, and

even then nearly all collapsed on the apothecia: extremely small

numbers of asci were dehisced.

At Dr. Lamb's suggestion some crossing plates were placed

in a 14.5cm Petri dishes containing sterile water, to provide

a high humidity. No positive results were obtained.

It is not known why the strains used in this work were

unable to cross at higher than usual temperature, while similar

strains did in the very near past. 152

PLATE 21 Electron microscope photograph of a presumably white spore (by Sharon Roberts, BSc Project, Imperial College 1977).

PLATE 22 Electron microscope photograph of a presumably red spore

(by Sharon Roberts, BSc Project, Imperial College 1977) 153

-211/111‘.

PLATE 23 Electron microscope photograph of a presumably red granular spore, (by Sharon Roberts, BSc Project,

Imperial College 1977) 154

PART B

The study of white-spored mutants

of different mutagenic origin 155 I INTRODUCTION

1. Problems investigated

This is a study of the conversion behaviour, that is of the frequency, direction and kind (meiotic versus postmeiotic) of conversion, of several white-spored mutants of Ascobolus immersus in relation to their origin. These mutants were either spontaneous or induced and the purpose of this study was to investigate:

a) The relation between the conversion spectrum of each mutant and its origin, in crosses to a wild-type strain, and b) the consistency of this relation in crosses of each mutant to different derived wild-types. These wild-types were the (P), (K) and (91) ones (General Methods, The Stock's ), defined as such in relation to their behaviour in crosses

to w-78. This is of interest for mutants linked or unlinked

to w-78.

2. Isolation of mutants

Ascospore colour mutants were detected by scanning collecting lids from wild-type x wild-type crosses for octads containing four wild-type and four colourless spores. Single white spores from such asci were isolated, germinated and crossed to wild-type to check for one gene segregation, namely the production of 4+:4m octads as the main pattern. Pairwise repulsion crosses between the isolated white spored mutants were used to test for linkage. Unfortunately, not all such crosses were fertile and so many mutants could not be assigned to any particular linkage-group. Mutants therefore are 156 referred to with their isolation number. The recombination frequency between any two mutants was calculated as NPD+1T PD+NPD+T (Strickberger, 1976). Mutants were provisionally assumed to represent the same gene if a cross between them gave more than 99% parental ditype asci (0+:8m): direct allelism tests were not possible for spore colour mutants. All new mutants were crossed to w-78 to test for linkage.

3. Induction of mutations

Induction was by one of three mutagens: UV irradiation.

N-methyl-N'-nitro-N-nitrosognanidine (NC), and acridine mustard ICR170 (ICR). The methods of induction applied were modifications of those used by Stadler et al (1970). a) UV induced mutations i) Induction with UV

For UV treatment two wild-type parents were inoculated at opposite edges of a 9cm petri dish of crossing

medium and allowed to grow at 17.5°C until they had nearly reached the centre of the plate (three or four

days). The treatment was given by placing this culture for 60 seconds under a CAMAG universal UV lamp, with emission centered around 254nm and with dose rate at 2 the position of the exposed dish of 300 ergs/sec/cm . After treatment the plates were incubated in the dark

for 24 hours and then continuously in the light until all asci had dehisced their spores onto the collecting

lids.

157

ii) Frequency of UV induc::d- mutations In order to find out the optimal conditions, duplicate cultures of the cross 9r+x61- were treated under UV

light for various lengths of time,

TABLE 18

Changes in mutation measure (% of 4:4's) with time of exposure to UV light.

Exposure time Total octads No. of 4:4's 'Frequency of 4:4's (seconds) detected 0(Control) 20,868 44 0.21 30 7,649 35 0.46 60 3,761 21 0.56 90 4,372 10 0.27 120 6,882 15 0.23 150 4,394 11 0.25 180 5,626 12 0.21

The exposure time of 60 seconds gave the highest frequency of 4:4's and was used throughout this work. UV irradiation had a drastic effect on the number of total octads produced in a cross (Table 18),which appeared 3 to 5.5 times lower than in the control crosses. For exposure times of 90 seconds up to 180 seconds the frequency of 4:4's almost declined to the level of the control value (Figure 10):there was no case of a lower value or of lethality to the cross (no completion of

the sexual cycle) due to the mutagenic treatment. It is possible that the mycelium on the surface of the

petri dish did not survive the 90 seconds or more UV Frequency o f4: 4s FIGURE 10Partial measureofmutationfrequency. Seetext:InductionwithUV 30

TIME (sec) 60

90

120

150

180 159 irradiation, and that completion of the sexual cycle was by the mass of undamaged mycelium grown inside the crossing medium and hence protected from the UV. In this case the mutation measure in the cross would be unrelated to the time of surface exposure to UV irradiation, and is expected to be comparable to the spontaneous mutation measure, as found here for long UV exposures (Figure 10).

Concerning the frequency of spontaneous mutations one has to consider that both parents are subjected to mutagenesis and that each 4:4 event represents a mutation present in one of the two nuclei which, through fusion, produced the ascus. The frequency of spontaneous mutations therefore, defined as the proportion of mutated nuclei among a population of nuclei subjected to mutagenesis, is equal to 1 of the frequency of 4:4's observed in a cross, provided there is no selection for or against the mutated nuclei and even though each 4:4 on a plate may not be an independent event, because of nuclear multiplication since mutagenesis.

This is also true for induced frequency of mutation, whenever both parents are undergoing the mutagenic treatment. However, in the present experiment of UV induction in both parents, an unknown fraction of the nuclei participating in ascus formation was coming from 160

hyphae below the culture medium surface and therefore

protected from irradiation and showing only the

spontaneous mutation frequency. This fraction of nuclei is expected to have lowered the induced mutation

frequency proportionally. For this reason the induced

mutation frequency, which actually is not needed for the purpose of the experiment, is not calculated here.

b) NG induced mutations

i) Induction with NG Petri dishes of 4.0cm diameter containing crossing

medium covered with a single layer of cellophane (British Sidac Ltd) were inoculated with one mating type. Simultaneously petri dishes of 9cm diameter containing crossing medium were inoculated off centre,

with the opposite mating type. All cultures were left

to grow for 3 days at 17.5°C. At the third day (about 3 to 4cm growth) the cellophane of the small petri dishes

was peeled off and floated (mycelium up) in a dish of

NG solution - 50pg/m1 in - pH 7.0 phosphate buffer (41.3% (v/v) of 1/15 Molar KH2 PO4, and 58.7% (v/v) of 1/15 0) - for 10 minutes. After this Molar Na HP04 • 2H2 time the cellophane with the mycelium was rinsed three times in minimal liquid medium and finally transferred inverted (mycelium on the crossing medium) on a 9cm petri

dish containing the opposite mating type, and untreated culture. It was left like this for 5 to 6 hours, when the cellophane layer was easily removed, leaving behind

the treated culture. The petri dishes containing one 161

treated and one untreated culture of the two mating

types were incubated and harvested as usual.

Mycelia for the control crosses were exposed to the

same procedure, with buffer substituting for mutagen solution.

ii) Frequency of NG induced mutations NG treated crosses had a 10 to 14 fold higher mutation

frequency than the control ones (Table 19) spontaneous mutation :Frequency of the control crosses is calculated

as 1 the :Frequency of 4:4's. (3,a, (ii) - Frequency of

UV mutations - of this PART). The mutation frequency of

the treated crosses is equal to the frequency of 4:4's since only one parent is treated here.

TABLE 19

Frequency of NG induced mutations

Mutation Frequency % Cross Total No. of Frequency 9 r+x61- oc.:.ads- 4:4's of 4:4's(%) Spontaneous Induced Control 27,556 55 0.20 0.10

Treated 13,786 170 1.23 0.10* 1.03 (one parent)

*presumed c) ICR induced mutations

i) Induction with ICR Induction with ICR was the same as with NG but using 120

minutes exposure in darkness. The ICR solution consisted 162

of 10µg/ml of the chemical in pH 7.0 phosphate buffer. Control crosses Were the same as with NG induction.

ii) Frequency of ICR induced mutations

ICR treated crosses had a 55 to 60 fold higher frequency of mutations than the control ones, calculated the same way as in the experiments with NG, (table 20).

TABLE 20

Frequency of ICR induced mutations

Total No. of octads 4:4's % of 4:4's Mutation frequency % Spontaneous Induced Cross B: 25-5R+xBBR1-

Control 29,153 52 0.18 0.09 Treated 15,984 777 4.86 0.09* 4.68 (one parent) Cross C: 9r+x92-

Control 23,010 31 , 0.13 0.07 Treated 13,894 556 4.00 0.07* 3.87 (one parent)

*presumed

4. Spontaneous mutations All spontaneous mutations studied here were isolated from crosses 9r+x3r- and 9r+xBIIR2-, which showed respectively

0.21% and 0.25% frequency of 4:4's. Since a spontaneous mutation may be present in either of the two nuclei fusing to form the ascus, the mutation frequency is equal to 1 the frequency of 4:4's. The frequencies of spontaneous mutations 163 in the above crosses were 0.11% and 0.13% respectively.

The frequency of spontaneous mutations was usually around 0.12%. Nevertheless, significantly lower (0.07%) or higher (0.19%) values appeared occasionally. 164 II RESULTS

1. The mutants

A total of 40 white spore colour mutants were studied

(including stock mutant w-78), and their names, parental crosses and origins are given in Table 21.

TABLE 21

The mutants, their origins and parental crosses

Origin Mutant Parental cross

Sw6 3r- x 9r+ Sw8 3r- x 9r+

Sw9 BHR2- x 9r+ us

o Sw20 BHR2- x 9r+ SW21 BHR2- x 9r+

tane Sw24 BHR2- x 9r+ on Sw25 BHR2- x 9r+ Sp Sw26 BIIR2- x 9r+ w-78 P5- x K5+

BHo BUR1- x 91-2R+ BBm BHR1- x 91-2R+ BHj BHR1- x 91-2R+ d UVkw8 KIII- x KIV+ BB9w8 BBR1- x 9r+ duce

in BB9w9 BBR1- x 9r+

V BBw11 BBR1- x 9r+ U BBw21 BBR1- x 9r+ BBSw BBR1- x 25-5R+ 165

TABLE 21 Cont....

Origin Mutant Parental Cross

NGwl 61- x 9r+ NGw5 61- x 9r+ NGw11 61- x 9r+ 0 NGw12 61- x 9r+ 0 -73 NGw14 61- x 9r+ NGw16 61- x 9r+ NGw17 61- x 9r+ 72, NGw18 61- x 9r+ NGw21 61- x 9r+

Bw3.10 BBR1- x 25-5R+ Bw4.2 BBR1- x 25-5R+ Bw4.3 BBR1- x 25-5R+ Bw4.6 BBR1- x 25-5R+ d BwS.3 BBR1- x 25-5R+ Bw6.1 BBR1- x 25-5R+ duce

in Bw6.2 BBR1- x 25-5R+ 0

7 Cwl.1 92- x 9r+

R1 Cw1.4 92- x 9r+

IC Cw1.7 92- x 9r+ CwS.2 92- x 9r+ 3C1* 92- x 9r+

*induced by myself and isolated by Shaker Helmi

2. Linkage relations

All mutants were crossed to w-78 as well as among themselves in an attempt to establish linkage relations.

Unfortunately, many of these crosses were sterile in spite of repeated efforts, sometimes using different isolates

(e.g. the high converting UV induced mutants were crossed among 166

themselves as many as 15 times at different intervals, with no fertility occurring. Tables 22a, 22b and 22c give the recombination frequency between any two mutants which succeeded in producing a fertile cross. The empty spaces in the tables as well as the missing pairwise crosses of mutants correspond to infertile crosses. The number of octads scored was usually well above a hundred while in cases of close linkage it was several thousands. For example from the cross w-78 x UVKw8,

13,094 octads were scored (with no recombination at all); in

Sw26 x NGw18, 5,850 octads were scored (again with no recombination), while in the cross w-78 x NGwS, 5,088 octads were scored (recombination frequency 0.33%).

With the sample sizes usually used, recombination frequencies below about 40% indicate clear linkage. Above about 40% non-linkage or very loose linkage are indicated.

Applying the rather arbitrary special criterion for allelism given in Introduction, paragraph 2 (isolation of mutants), according to which alleles are considered to be the mutants showing 99% of their asci in the 0+:8 mutant phenotypic class in repulsion crosses, four different "loci" were recormised among the mutants. These are as follows, with numbers in parentheses giving the recombination frequency between the respective mutants. 167

TABLE 22a

Recombination frequencies (%) between different mutants from pairwise crosses

NGwl NGw5 NGw11 NGw12 NGw14 NGw16 NGw18 NGw21 BB5w 42.1 NGw12 48.0 42.9 0.20 NGw14 50.8 40.8 50.0 49.2 NGw18 49.9 50.3 48.2 NGw21 50.0 50.2 49.4 63.1 48.5 Bw3.10 50.0 35.5 34.1 Bw4.2 24.8 26.8 35.4 34.6 45.0 17.4 59.5 Bw4.3 17.9 18.8 32.5 28.6 41.1 47.2 Bw4.6 48.9 59.2 39.3 Bw6.1 51.5 30.4 46.3 44.4 49.3 Cw1.4 45.0 28.92 38.2 41.5 36.6 35.0 10.8 32.4 Cwl.7 37.9 40.60 44.0 51.1 40.1 48.2 Cw5.2 45.28 37.12 38.5 36.5

*Empty spaces as well as unreported crosses correspond to infertile crosses 168

TABLE 22b

Recombination frequencies (%) between different mutants from pairwise crosses

w-78 Sw20 Sw21 Sw24 Sw25 Sw26 UVKw8 Cw1.4 NGw14 48.1 45.8 46.8 52.3 47.4 NGw16 42.7 2.50 NGw18 48.3 0.00 NGw21 52.3 47.2 50.0 Bw3.10 40.8 0.07 Bw4.2 31.4 17.4 0.14 Bw4.3 50.0 24.9 23.9 33.3 37.7 Bw6.1 43.7 41.9 Cw1.4 51.0 32.3 45.22 41.6 49.0 Cw1.7 53.5 45.6 45.08 49.2 46.7 Cw5.2 62.6 47.5 55.77 48.3 5.20* 0.28 3C1 0.46 Bw4.6 47.2

*High conversion frequency (IICF) w-78x(HCF) 3C1 gave 5.20% Low conversion frequency (LCF) w-78x(IICF) 3C1 gave 0.46% 169

TABLE 22c

Recombination frequencies (%) between different mutants from pairwise crosses

w-78 Sw20 Sw21 Sw24 Sw25 Sw26 BHo BIIj Cw1.4 BBm 46.8 BHj 47.1 40.5 52.3 50.0 UVKw8 0.00 BB9w8 48.0 BB9w9 45.3 BBw11 41.5 45.5 47.5 BBw21 41.4 BB5w 48.9 45.4 36.5 NGw1 48.0 47.5 43.9 28.3 NGw5 0.3 40.5 43.8 35.3 NGw11 50.8 34.6 50.7 NGw12 45.9 41.5 53.1 51.0 31.5 52.2

170 a) Locus I includes the mutants

NGwS w-78 UVKw8 3C1 (NG induced) (Spontaneous) (UV indliced) (ICR170 induced)

(0.33) (0.00) (0.28)

(0.46) (5.20)

The position of the mutant NGw5 is not known since its crosses to UVKw8 and 3C1 were not fertile.

Mutants w-78 and 3C1 have two values for recombination

frequency, corresponding to the following crosses:

1)5.20 in high conversion frequency (P)w-78 x (P)3C1 and 2) 0.46 in low conversion frequency (K)w-78 x (P)3C1.

All other crosses for this locus were of the (P) x (P) type.

It is possible that the w-78 and 3C1 mutants do not show a high frequency of co-conversion and that in a repulsion cross involving high conversion frequency strains, independent conversion in either site will increase the percentage of recombination frequency between them to a value disproportional to the physical distance between them (cross (1)). b) Locus II includes the mutants: 12. 51

Sw26 and NGw18 NGw16 (Spontaneous) (NG induced) (NG induced)

(0.00) 171

Dotted line mutant NGw16 is linked to locus II, but because cross Sw26 x NGw16 was not fertile, its relevant position is not known.

c) Locus III includes the mutants:

Bw4.2 Gw1.4 Bw3.10 (all three induced by ICR)

0.14) (0.07)

Cross Bw4.2 x Bw3.10 was not fertile.

Loci II and III are linked, though not very closely, with

recombination frequency ranging from about 10.80% (in Cwl.4 x NGw18) to 17.4% (in Bw4.2 x NGw18). d) Locus IV includes the mutants:

NGwll and NGw12 (both induced by NG)

(0.2)

The relation between the loci described here and the ones found in the linkage groups of Ascobolus immersus given by

Yu-Sun (1966), is not generally known. Nevertheless, loci II and III here strongly resemble the loosely linked clusters of closely linked markers (w-32, w-44, w-79, w-80) and (w-4, w-8, w-72, w-87) in linkage group II in Yu-Sun (1966). Also marker w-75 reported by Yu-Sun (1966) to be closely linked to 172 w-78 and w-10 could be simflqr in position to either of the two mutants (NGw5 and 3C1) found here closely linked to w-78. There is less probability that it may correspond to the onsition of mutant UVKw8 since the latter gives no recombination at all with w-78 and no such case was reported by Yu-Sun.

3. Conversion spectra All mutants were crossed to (P), (K) and (91) wild- type strains (PART A, General materials and methods). In each cross the frequency of 4:4's and all aberrant classes was determined by scoring under the microscope the octads discharged on consecutive collecting lids. Controls for repeatability, phenocopies, new mutations, and false spore-clusters were given by Emerson and Yu-Sun (1967) and Wickramaratne (1974) for mutants w-78 and w-10 only. Octads with aborted spores or groups with other than 8 spores were ignored.

All 4 types of conversion spectra (A,B,C, and D), which were defined by Leblon (1972a) by two criteria, namely the frequency of postmeiotic segregation, and the relative frequencies of conversion to the mutant allele or to wild-type, were recognised among the mutants studied here. In addition intermediate types (AB and CD) also occurred. The characterisation of the conversion spectra by the different types is as follows:

A type: postmeiotic segretation rare, excess of conversion, to

the wild-type.

B type: postmeiotic segregation rare, excess of conversion to

the mutant 173

AB type: postmeiotic segregation rare, conversion to the

wild-type not markedly different from conversion to the mutant.

C type: postmeiotic segregation frequent, excess of conversion to wild-type.

D type: postmeiotic segregation frequent, excess of conversion to the mutant. CD type: postmeiotic segregation frequent, conversion to wild-

type not markedly different from conversion to the

mutant.

Mutants showing the C, D and CD type of conversion spectra have been treated by Yu-Sun et al (1977) as a single category called class C. Evidence that these classes intergrade is revealed by the work of Leblon (1972a) who reports 17 mutants of C type, 2 of D type and 16 of CD type in Ascobolus immersus, as well as of that of Lamb and Wickramaratne (1975), where two spore colour mutants of the same organism, presumed to have arisen by base-pair substitution, were of C type at 22.5°C but in some cases were of D type at 17.5°C. However, in the present work it is considered desirable to keep C, D and CD classes separate, much as A, B and AB classes are kept separate though it is found in this work that these too sometimes intergrade .

Leblon (1972a) found that 075% of the aberrant asci showed postmeiotic segregation in A, B or AB mutants of Ascobolus immersus and a much higher proportion for C mutants (45-88% in 174 gene b1, 27-64% in gene b2), In the present work with different strains of the same organism the difference in the frequency of postmeiotic segregation between the C and non C mutants is smaller: relative frequency of postmeiotic segregation ranges from 0-10% for A, B or AB mutants and from

15-100% for C, D or CD mutants. Very similar differencesin the values of postmeiotic segregation between the two groups of mutants have been reported for Sordaria brevicollis by Yu-Sun et al (1977).

The conversion spectra obtained from the different crosses of each mutant are given in Tables: 23, 24, 25, 26, 28,

30, 32 and 34.

The allocation of the mutants into the different classes (A, B, AB, C, D and CD) is given in Tables 27, 29, 31, 33 and 35. This was carried out according to the relative frequencies (as a % of total detected aberrant segregation octads) of meiotic segregation (6:2 + 2:6), postmeiotic segregation (5:3 + 3:5), conversion to the wild-type (6:2 + 5:3) and conversion to the mutant (2:6 + 3:5), that each mutant showed in each cross.

The formulae used here for the type (meiotic or postmeiotic) and the direction of conversion(to the wild-type or to the mutant), do not allow for undetected octads like aberrant 4:4's resulting from either dual or single hybrid DNA occurrence. However, they still give a strong indication of the type and direction of conversion while providing measurements directly comparable with the ones reported by other workers,

(Leblon 1972a, Yu-Sun et al 1977). TABLE 23

Numbers of aberrant asci from crosses between mutants of locus I and wild-types

Asci (wild-type: mutant spores) Aberrant asci Crossed Origin Mutant to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total 0

(P)+,61- 10773 792 348 106 102 2 3 9 3 12144 1371 11.21 t tr) w-78* (K),KIII- 12043 411 47 85 36 1 1 2 0 12626 583 4.62 g 0 (EC2w78.5+) 0 0 /D Q$ (91),91- 6031 131 39 26 25 0 1 2 2 6257 225 3.6(

(P),61- 6328 0 7 1 6 0 0 0 0 6342 14 0.22

UV UVKw8 (K),KIII- 7513 3 11 6 8 0 0 0 0 7541 28 0.37

(91),91- 5902 0 15 0 5 0 0 0 0 5922 20 0.34

(P),61- 2947 51 122 7 69 0 1 0 0 3197 250 7.81

NG NGw5 (K),KIII- 3412 29 32 4 23 0 0 . 0 1 3501 89 2.54

(91),91- 1696 . 24 8 2 0 0 0 0 0 1730 34 1.97

(P),61- 3087 141 221 0 1 0 0 1 0 3452 364 10.5/

ICR170 3C1 (K),KIII- 1180 17 19 0 2 0 0 0. 0 1218 38 3.11

(91),91- 1707 20 19 1 1 0 0 1 0 1749 42 2.40 *Mutant w-78 is one of the stock strains and is used here as a mutant belonging to locus I +Wild—type strains (P), (K) and (91), (General Materials and Methods: 2 the Stocks) give high conversion frequency (if P), or low conversion frequency, (if K or 91) in crosses to (P)w-78. TABLE 24

Numbers of aberrant asci from crosses between mutants of locus II and wild-types

Asci (wild-type: mutant spores) Aberrant asci Crossed Origin Mutant to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total

(P),92- 5976 12 4 7 1 0 0 0 0 6000 24 0.40 o ° Sw26 (91) ,91- 8028 1 3 16 12 0 0 0 0 8060 32 0.30

(P),92- 5508 136 41 44 22 1 1 0 0 5753 245 4.25

NG NGw18 (K),KIII- 10242 12 0 4 2 0 0 0 0 10260 18 0.18

(911,91- 8257 180 30 SO 22 1 0 1 0 8541 284 3.32

0 5

F CO 0 S -a 177

Wider ratio octads are dealt with separately and they are not included in the calculations of the above relative frequencies. This was done so again for the purpose of obtaining values directly comparable with the ones of other workers, and because they were usually of small numbers not greatly affecting the overall result.

covvtifnue oM racde 182 TABLE 25

Numbers of aberrant asci from crosses between mutants of locus III and wild-types

Asci (wild-type: mutant spores) Aberrant asd Crossed Origin Mutant to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total

(P),9r+ 10916 8 2 6 2 0 0 0 0 10934 18 0.16

ICR170 Bw4.2 (K),KIV+ 4227 6 3 6 1 0 0 0 4 4247 20 0.47

(91),91- 4392 25 4 32 0 3 0 2 0 4458 66 1.48

(P),9r+ 7500 4 6 14 0 0. 0 0 0 7524 24 0.32

ICR170 Bw3.10 (K),KIV+ 12176 8 16 4 2 2 0 0 0 12206 32 0.26

(91),91- 4828 2 0 9 0 1 0 1 0 4841 13 0.27

(P),9r+ 6817 10 24 11 0 0 0 0 0 6862 45 0.66

ICR170 Cw1.4 (K),KIV+ 5460 7 28 3 0 0 0 0 0 5490 38 0.69

(91),91- 4312 8 20 10 0 2 0 0 0 4352 40 1.37 TABLE 26

Numbers of aberrant asci from crosses between mutants of locus IV and wild-types

Asci (wild-type: mutant spores) Aberrant asc Crossed Origin Mutant to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total 0

(P),92- 5062 4 8 4 14 0 0 0 0 5092 30 0.59

NG NGw11 (K),KIII- 5498 2 7 6 9 0 0 0 0 5522 24 0.4

(91),91- 5925 6 10 2 9 0 0 0 0 5952

(P),92- 9981 6 15 3 5 0 0 0 0 10010 29 0.29 NG NGw12 (91),91- 8325 3 3 6 4 0 0 0 0 8347 16 0.19 180 TABLE 27

Classification of the-mutants of loci I, II, III and IV into

classes A, B, AB, C, D and CD, according to their relative

frequencies (as a % of detected aberrant segregation octads) of

meiotic segregation (6:2 + 2:6), postmeiotic segregation

(5:3 + 3:5), conversion to the wild-type (6:2 + 5:3) and conversion

to the mutant (2:6 + 3:5)

Mutant and Crossed Origin to Class 6:2+2:6 5:3+3:5 6:2+5:3 2:6+3:5

Locus I

(P),61- C 83.20 15.2 65.5 32.8 w-78 (K),KIII- C 78.9 20.8 85.1 14.2 (Spontaneous) (91),91- C 75.6 22.7 67.8 28.4

(P),61- D 50.0 50.0 9.09 90.9 UVKw8 (K),KIII- D 51.4 48.6 32.4 70.2 (UV) (91),91- D 73.5 26.5 0.00 100.0

(P),61- D 69.4 30.5 23.3 76.6 N Gw 5 (K),KIII- D 68.5 30.3 37.0 61.8 (NT) (91),91- A 93.9 6.09 76.7 23.4

(P),61- B 99.4 0.28 38.6 61.1 3C1 (K),KIII- AB 94.7 5.26 44.7 55.3 (ICR170) (91),91- AB 92.9 5.00 50.0 47.9

Locus II

(P),92- C 66.7 33.3 79.2 20.8 Sw26 (Spontaneous) (91),91- CD 12.5 87.5 53.1 46.9

(P),92- C 72.7 28.0 73.9 25.8 NGw18 (K),KIII- C 66.6 33.3 88.9 11.1 (NG) (91),91- C 74.1 25.6 81.3 17.7 181.

TABLE 27 Cont

Mutant and Crossed Origin to Class 6:2+2:6 5:3+3:5 6:2+5:3 2:6+3:5

Locus III

(P),9r+ C 56.3 43.5 75.0 25.0 Bw4.2 34.0 29.0 (IUR170) (K),KIV+ C 44.7 56.6 (91),91- C 43.9 48.7 86.5 6.08

(P),9r+ C 40.6 59.4 75.0 25.0 Bw3. 10 (K),KIV+ D 75.0 18.8 37.5 56.3 (IUR170) (91),91- C 14.8 70.4 85.2 0.00

(P),9r+ CD 75.0 24.2 46.9 53.0 Cwl.4 (K),KIV+ B 92.1 7.9 26.3 73.7 (IUR170) (91),91- CD 75.0 25.0 45.0 50.0

Locus IV

(P),92- D 40.0 60.0 26.7 73.3 NGw11 (K),KIII- D 37.5 62.5 33.3 66.7 (NG) (91),91- D 59.2 40.7 29.6 70.4

(P),92- D 72.4 27.6 27.6 69.0 NGw12 (NG) (91),91- CD 37.5 62.5 56.2 43.8

covitivwe oln ?coie TABLE 28

Numbers of aberrant asci from crosses between UV induced mutants and wild-types

Asci (wild-type: mutant spores) Aberrant asci

Mutant Crossed to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total

(P),92- 7659 81 46 8 2 0 0 0 0 7796 137 1.76

BB9w9 (P),BBR1- 700 54 4 42 6 2 0 0 0 808 108 13.36

(91),91- 7069 67 25 7 1 0 0 0 0 7169 100 1.39

(P),25-5R+ 2138 151 25 90 17 0 0 1 0. 2422 284 11.72

BBwll (P),9r+ 434 32 .8 18 4 1 0 0 0 497 63 12.67

(91),352+ 2797 29 16 2 0 0 0 0 0 2844 47 1.65

(P),BBR1- 1516 141 21 96 16 0 0 0 0 1790 274 15.30

BBw21 (K),KIII- 6226 39 51 6 2 0 0 0 0 6324 98 1.55

(91),91- 2371 15 20 4 3 0 0 0 0 2313 42 1.74

(P),9r+ 8222 382 104 71 76 4 4 0 4 8867 645 7.27

BHj (K),KIII- 986 40 6 0 14 0 0 0 0 1046 60 5.74

(91),91- 6682 688 146 166 ill 4 0 6 2 7805 1123 14.39 TABLE 28 Cont

Asci (wild-type: mutant spores) Aberrant asci

Mutant Crossed to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total

(P),3r- 12707 3 29 0 28 0 9 0 14 12790 83 0.65 BHo (91),91- 5754 2 2 14 0 0 0 0 0 5772 18 0.31

(P),9r+ 9678 0 0 8 8 0 0 0 0 9694 16 0.17 BBm (K),KIV+ 5390 8 0 8 0 0 0 0 0 5406 16 0.30

(P),61- 4152 206 22 147 18 3 1 0 0 4549 39 7 8.72 BB9INE, (91),91- 7408 61 88 8 2 0 0 0 0 7567 159 2.10

BB5w (P),25-5R+ 13580 14 0 0 0 0 0 0 0 13594 14 0.10 TABLE 29 184

Classification of UV induced mutants into classes A, B, AB, C,

D and CD, according to their relative frequencies (as a % of detected aberrant segregation octads) of meiotic segregation (6:2 + 2:6), postmeiotic segregation (5:3 + 3:5), conversion to the wild-type (6:2 + 5:3) and conversion to the mutant (2:6 + 3:5)

Mutant Crossed to Class 6:2+2:6 5:3+3:5 6:2+5:3 2:6+3:5 (P),92- A 92.7 7.29 64.9 35.0 BB9w9 (P),BBR1- C 53.7 44.5 88.9 9.3 (91),91- A 92.0 8.00 74.0 26.0 (P),25-5R+ C 61.9 37.7 84.9 14.8 BBw11 (P),9r+ C 63.5 34.9 79.4 19.0

(91),352+ A 95.8 4.24 66.1 33.9 (P),BBR1- C 59.2 40.9 86.5 13.4 BBw21 (K),KIII- AB 92.3 7.75 45.8 54.2

(91),91- CD 83.3 16.7 45.4 54.6 (P),9r+ C 75.4 22.9 69.3 27.9

BIij (K),KIII- C 76.5 23.3 66.6 33.3 (91),91- C 74.0 24.7 76.0 22.9

(P),3r- D 33.9 3.07 68.7 Blio (91),91- C 22.2 77.7 88.8 11.1 (P),9r+ CD 0.00 100.0 50.0 50.0 BBm (K),KIV+ C 50.0 50.0 100.0 0.00 (P),61- C 57.5 41.6 88.9 10.09 BB9w8 (91),91- AB 93.8 6.67 43.8 56.6 BB5w (P),25-5R+ A 100.0 0.00 100.0 0.00 TABLE 30

Numbers of aberrant asci from crosses between NC induced mutants and wild-types

Asci (wild-type: mutant spores) Aberrant asci

Mutant Crossed to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total

(P),92- 7425 13 12 17 2 0 0 0 0 7469 44 0.59

NGwl (K),KIII- 8452 17 3 26 5 1 0 0 0 8504 52 0.61

(91),9l- 10439 22 4 17 9 0 0 0 0 10491 52 0.50

NGw14 (91),91- 6469 5 4 6 3 0 0 0 0 6487 18 0.28

(P),92- 2433 17 5 61 19 0 0 0 0 2535 1G r.00

NGw16 (K),KIII- 3813 54 7 106 10 7 0 2 0 3999 186 4.65

(91),91- 4967 79 16 166 41 1 0 1 0 5270 304 5.76

(K),KIII- 3870 14 12 3 3 0 0 0 0 3902 32 0.82 NGw17_ (91),91- 3759 3 6 9 3 0 0 0 0 3780 21 0.56

(P),61- 2386 10 1 2 2 0 0 0 0 2401 15 0.62 NGw21 (91),91- 3061 3 0 3 6 0 0 0 0 3075 12 0.39 186

TABLE 31

Classification of NG induced mutants in classes A, B, AB, C, D and CD according to their relative frequencies (as a % of detected aberrant segregation octads) of meiotic segregation (6:2+2:6), postmeiotic segregation (5:3+3:5), conversion to

wild-type (6:2+5:3) and conversion to the mutant (2:6+3:5).

Mutant Crossed to Class 6:2+2:6 5:3+3:5 6:2+5:3 2:6+3:5 (P),92- C 55.9 44.1 67.8 32.2 NGwl (K),KIII- C 39.4 60.7 83.6 16.4 (91),91- C 48.0 52.0 74.0 26.0 NGw14 (91),91- C 50.0 50.0 60.0 40.0 (P),92- C 21.8 78.3 77.0 23.0 NGw16 (K),KIII- C 32.5 62.4 86.0 8.82 (91),91- C 31.3 68.2 80.7 18.8 (K),KIII- CD 81.3 18.7 53.1 46.9 NGw17 (91),91- CD 42.9 57.0 57.0 42.9 (P),61- C 74.2 25.8 80.6 19.4 NGw21 (91),91- CD 25.7 74.4 51.3 48.8

• TABLE 32

Numbers of aberrant asci from crosses between ICR170 induced mutants and wild-types

Asci (wild-type: mutant spores) Aberrant asci

Mutant Crossed to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total

(P),9r+ 7315 9 1 31 0 6 0 5 0 7369 52 0.71

Bw4.3 (K),KIV+ 5225 12 0 14 0 3 0 3 0 5237 32 0.61

(91),91- 3282 10 0 9 0 2 0 2 0 3305 23 0.70

Bw4.6 (K),KIV+ 1387 20 9 3 3 0 0 0 0 1422 35 2.46

(P),9r+ 1292 0 30 2 0 0 0 0 0 1324 2.41

Bw6.1 (K),KIV+ 2446 8 25 3 0 2 0 0 0 2484 38 1.52

(91),352+ 2364 6 35 3 0 2 0 0 1 2411 47 1.95

Bw6.2 (K),KIV+ 3855 9 61 0 1 0 0 0 0 3926 71 1.81

(P),9r+ 2893 40 35 1 1 0 0 0 0 2970 77 2.59

Cw1.1 (K),KIV+ . 2701. 59 33 2 0 0 0 0 1 2796 95 3.40

(91),352+ 1991 17 29 0 3 0 0 0 0 2042 49 2.40

(P) ,9r+ 1082 2 24 2 0 0 0 0 0 1110 28 2.31 Cw1.7 (K),KIV+ 3351 7 29 3 0 0 0 0 0 3390 39 1.15

(91),352+ 842 2 14 2 1 0 0 0 0 861 19 2.21

(P),9r+ 2438 2 18 1 1 0 0 0 0 2460 22 0.89 Cw5.2_ !-- (91),352+ 7886 6 25 2 1 0 0 0 0 7920 34 0.43t 188

TABLE 33 Classification of ICR170 induced mutants into classes A, B, AB, C, D and CD, according to their relative frequencies (as a % of detected aberrant segregation octads) of meiotic segregation (6:2+2:6), postmeiotic segregation (5:3+3:5), conversion to the wild-type (6:2+5:3) and conversion to the mutant (2:6+3:5)

Mutant Crossed to Class 6:2+2:6 5:3+3:5 6:2+5:3 2:6+3:5 (P),9r+ C 17.3 59.2 76.1 1.41 Bw4.3 (K),KIV+ C 37.5 43.8 81.3 0.00 (91),91- C 43.5 39.1 81.6 0.00 Bw4.6 (K),KIV+ C 82.9 17.3 65.7 34.3 (P),9r+ B 93.8 6.25 6.25 93.8 Bw6.1 (K),KIV+ B 86.9 7.89 27.9 65.8 (91),352+ B 87.1 8.61 21.3 74.4 Bw6.2 (K),KIV+ B 98.3 1.66 14.4 85.6

(P),9r+ AB 98.1 2.34 53.3 47.1 Cwl.1 (K),KIV+ A 97.8 2.06 64.1 34.7 (91),352+ B 93.7 6.25 34.6 65.4

(P),9r+ B 92.9 7.14 17.3 85.7 Cw1.7 (K),KIV+ B 92.3 7.69 25.6 74.4 (91),352+ D 84.2 15.8 21.1 78.9 (P),9r+ B 90.9 9.09 13.6 86.4 Cw5.2 (91),352+ B 91.2 8.82 23.5 76.5 TABLE 34

Numbers of aberrant asci from crosses between spontaneous mutants and wild-types

Asci (wild-type: mutant spores) Aberrant asci

Mutant Crossed to 4:4 6:2 2:6 5:3 3:5 7:1 1:7 8:0 0:8 Total Total %

Sw6. (P),3r- 1639 3 2 2 10 0 0 0 0 1656 17 1.02

Sw8 (P),BHR2- 3145 10 1 6 2 0 0 0 0 3164 19 0.60

Sw9 (P ) ,9r+ 2376 6 10 10 1 0 0 0 0 2403 27 1.08

(P),9r+ 16190 10 8 1 1 0 0 0 0 16210 20 0.12 Sw20 (K),KIII- 14020 18 3 0 0 0 0 0 0 14041 7:1 0.15

(91),91- 13510 15 4 0 0 0 0 0 0 13529 19 0.14

(P),9r+ 10552 6 6 7 0 0 0 1 0 10571 20 0.18 Sw21 (K),KIII- 8552 7 3 4 8 0 1 0 - 0 8574 23 0.27

Sw24 (P),61- 8167 4 0 16 0 0 0 0 0 8187 20 0.17

Sw25 (K),KIII- 15005 5 2 13 12 1 0 0 0 15038 33 0.22 190

TABLE 35

Classification of spontaneous mutants into classes A, B, AB, C, D and CD, according to their relative frequencies (as a % of detected aberrant segregation octads) of meiotic segregation (6:2+2:6), postemiotic segregation (5:3+3:5), conversion to the wild-type.(6:2+5:3) and conversion to the mutant (2:6+3:5)

Mutant Crossed to Class 6:2+2:6 5:3+3:5 6:2+5:3 2:6+3:5 Sw6 (P),3r- D 29.4 70.6 29.4 70.6 Sw8 (P),BIIR2- C 57.9 42.1 84.2 15.8 Sw9 (P),9r+ CD 59.3 40.7 59.3 40.7 (P),9r+ AB 90.0 10.0 55.0 45.0 Sw20 (K),KIII- A 100.0 0.0 85.7 14.3 (91),91- A 100.0 0.0 78.9 21.1 (P),9r+ C 60.0 40.0 65.0 35.0 Sw21 (K),KIII- CD 43.5 52.2 47.8 47.8 Sw24 (P),61- C 20.0 80.0 100.0 0.0 Sw25 (K),KIII- CD 21.2 75.8 54.5 42.4

continue on page 193 TABLE 36

Observed and expected numbers of wider ratio asci in crosses of UV induced mutants to wild-types

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected+ Observed Expected Observed Expected Observed Expected

(P),61- D 0 0 0 0.00 0 0 0 0.00

UVKw8 (K),KIII- D 0 0.00 0 0.01 0 0.00 0 0.00

(91),91- D 0 0.00 0 0.01 0 0 0 0.01

(P),9r+ C 4 1.63 4 0.49 0 4.2 4 0.38

(K),KIII- C 0 0 0 0.00 0 0 0 0,00

(91),91- C 4 7.67 0 1.07 6 15.87 2 0.72

(P),3r- D 0 0 9* 0.09 0 0.00 14* 0.10 BHo (K),KIV+ C 0 0.00 0 0.00 0 0.00 0 0.00

(P),9r+ CD 0 0.00 0 0.00 0 0.00 0 0.00 BBm (K),KIV+ C 0 0.01 0 0.00 0 0.00 0 0.00

BB5w (P),25-5R+ A 0 0.00 0 0.00 0 0.00 0 0.00

+ In the absence of corresponding-site interference. IC

* Significant difference between observed and expected values at P=1% TABLE 36 Cont

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected Observed Expected Observed Expected Observed Expected (P),61- C 3 3.45 1 0.05 0 2.4 0 0.03 BB9w8 (91),91- AB 0 0.03 0 0.01 0 0.12 0 0.26 (P),92- A 0 0.04 0 0.01 0 0.21 0 0.07 BB9w9 (P),BBR1- C 2 1.52 0 0.01 0 0.97 0 0.00 (91),91- A 0 0.03 0 0.00 0 0.17 0 0.02 (P),25-5R+ C 0 2.84 0 0.09 1 2.4 0 0.06 BBw11 (P),9r+ C 1 0.62 0 0.03 0 0.55 0 0.03 (91),352+ A 0 0.01 0 0.00 0 0.07 0 0.02 (P),BBR1- C 0 3.78 0 0.09 0 2.78 0 0.62 BBw21 (K),KIII- AB 0 0.02 0 0.01 0 0.06 0 0.10 (91),91- CD 0 0.01 0 0.01 0 0.02 0 0.04 D a U nuq a J no 3 5Y a. l frb 193

4. Aberrant asci with wider Fatios

Aberrant asci with wider ratios occurred in different crosses of several mutants. Their observed and expected numbers are given in Tables 36, 37, 38 and 39. The expected frequencies of wider ratios in the absence of corresponding-site interference are given by the following partly corrected formulae which are based on observed octad frequencies (not %'s) and were.derived by Lamb and Wickramaratne (1973).

Wider ratio class partly corrected formulae 8:0 (16:2 + 8:0 + 17:1)2

0:8 (12:6 + 0:8 + 11:7)2

7:1 2(16:2 +8:0 + 17:1)(15:3 + 17:1) 1:7 2(12:6 + 0:8 + 11:7)(13:5 + 11:7)

Where in the formulae 6:2 etc., stand for the frequencies of that segregation class as a fraction of all octads.

In most of the crosses reported no wider ratio octads were deteCted at all. In these cases the expected values of the wider ratio octads were calculated for the purpose of detecting any strong negative corresponding site interference occurring.

coil -hi/we on page Iq I TABLE 37

Observed and expected numbers of wider ratio asci in crosses of NG induced mutants to wild-types

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected Observed Expected Observed Expected Observed Expected

(P),92- C 0 0.01 0 0.00 0 0.01 0 0.00 NGwl (K),KIII- 1 0.03 0 0.00 0 0.01 0 0.00 (91),91- C 0 0.02 0 0.00 0 0.01 0 0.00

(P),61- D 0 0.06 1 1.35 0 0.20 0 1.18

NGwS (K),KIII- D 0 0.02 0 0.11 0 0.06 1 0.01

(91),91- A 0 0.01 0 0.00 0 0.08 0 0.01

(P),92- D 0 0.00 0 0.00 0 0.00 0 0.00

NGw11 (K),KIII- D 0 0.00 0 0.00 0 0.00 0 0.00

(91),91- D 0 0.00 0 0.00 0 0.00 0 0.00

(P),92- D 0 0.00 0 0.00 0 0.00 0 0.00 NGw12 (91),91- CD 0 0.00 0 0.00 0 0.00 0 0.00

NGw14 (91),91- C 0 0.00 0 0.00 0 0.00 0 0.00

(P),92- C 0 0.20 0 0.02 0 0.03 0 0.00

NGw16 (K),KIII- C 7 0.86 0 0.01 2 0.23 0 0.00

(91),91- C 1 1.30 0 0.06 1 0.32 0 0.01 TABLE 37 Cont....

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected Observed Expected Observed Expected Observed Expected

(K),KIII- CD 0 0.01 0 0.00 0 0.01 0 0.01 NGw17 (91),91- CD 0 0.00 0 0.00 0 0.00 0 0.00

(P),61- C 0 0.00 0 0.00 0 0.01 0 0.00 NGw21 (91),91- CD 0 0.00 0 0.00 0 0.00 0 0.00

(P),92- C 1 0.54 1 0.08 0 0.82 0 0.08

NGw18 (K),KIII- C 0 0.00 0 0.00 0 0.00 0 0.00

(91),91- C 1 0.55 0 0.04 1 0.98 0 0.03 TABLE 38

Observed and expected numbers of wider ratio asci in crosses of 1CR170 induced mutants to wild-types

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected Observed Expected Observed Expected Observed Expected

(P),61- B 0 0.00 0 0.03 1 1.48 0 3.54

3C1 (K),KIII- AB 0 0.00 0 0.02 0 0.06 0 0.07

(91),91- AL 0 0.01 0 0.01 1 0.07 0 0.05

(P),9r+ C 0 0.01 0 0.00 0 0.00 0 0_90

Bw3.10 (K),KIV+ B 2 0.00 0 0.00 0 0.00 0 0.01

(91),91- C 1 0.01 0 0.00 1 0.00 0 0.00

(P),9r+ C 0 0.00 0 0.00 0 0.01 0 0.00

Bw4.2 (K),KIV+ C 0 0.00 0 0.00 0 0.00 4 0.01

(91),91- C 3 0.13 0 0.00 2 0.06 0 0.00

(P),9r+ C 6 0.01 0 0.00 5 0.01 0 0.00 Bw4.3 (K),KIV+ C 3 0.03 0 0.00 3 0.02 0 0.00

(91),91- C 2 0.00 0 0.00 2 0.00 0 0.00

Bw4.6 (K),KIV+ C 0 0.02 0 0.01 0 0.07 0 0.01 TABLE 38 Cont

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected Observed Expected Observed Expected Observed Expecte'd

(P),9r+ B 0 0.00 0 0.00 0 0.00 0 0.17

Bw6.1 (K),KIV+ B 0 0.01 0 0.00 0 0.01 0 0.06

(91),352+ B 2 0.01 0 0.00 0 0.01 1 0.13

Bw6.2 (K),KIV+ B 0 0.00 0 0.01 0 0.01 0 0.24

(P),9r+ AB 0 0.01 0 0.01 0 0.13 0 0.10

Cwl.1 (K),KIV+ A 0 0.02 0 0.00 0 0.31 0 0.10

(91),352+ B 0 0.00 0 0.02 0 0.04 1 0.10

(P),9r+ CD 0 0.01 0 0.00 0 0.00 0 0.02

Cw1.4 (K),KIV+ B 0 0.00 0 0.00 0 0.00 0 0.04

(91),91- CD 2 0.01 0 0.00 0 0.01 0 0.02

(P),9r+ B 0 0.00 0 0.00 0 0.00 0 0.13

Cwl.7 (K),kIV+ B 0 0.00 0 0.00 0 0.00 0 0.06

(91),352+ D 0 0.00 0 0.01 0 0.00 0 0.06

(P),9r+ B 0 0.00 0 0.00 0 0.00 0 0.03 Cw5.2 (91),352+ B 0 0.00 0 0.00 0 0.00 0 0.02 TABLE 39

Observed and expected numbers of wider ratio asci in crosses of.spontaneous mutants to wild-types

7:1 1:7 8:0 0:8

Mutant Crossed to Class Observed Expected Observed Expected Observed Expected Observed Expected

Sw6 (P),3r+ D 0 0.00 0 0.01 0 0.00 0 0.00 Sw8 (P),BHR2- C 0 0.01 0 0.00 0 0.01 0 0.00

Sw9 (P),9r+ CD 0 0.01 0 0.00 0 0.00 0 0.01

(P) ,9r+ AB 0 0.00 0 0.00 0 0.00 0 0,00 Sw20 (K),KIII- • A 0 0.00 0 0.00 0 0.01 0 0.00 (91),91- A 0 0.00 0 0.00 0 0.00 0 0.00

(P) ,9r+ C 0 0.00 0 0.00 1 0.00 0 0.00 Sw21 (K),KIII- CD 0 0.00 1 0.00 0 0.00 0 0.00

Sw24 (P),61- C 0 0.00 0 0.00 0 0.00 0 0.00

Sw25 CD 1 0.00 0 0.00 0 0.00 0 0.00

(P),92- C 0 0.01 0 0.00 0 0.01 0 0.00 Sw26 (91),91- CD 0 0.00 0 0.00 0 0.00 0 0.00 199

III DISCUSSION

LOCUS I

The four mutants of locus I are all of different origin: , one is spontaneous (w-78), one is UV induced (UVKw8), one is NG induced (NGw5) and the last one is ICR 170 induced (3C1),

Table 23.

Crosses of these mutants to the same (P), (K) and (91)

wild-type strains show that the frequency of gene conversion in the three of them (w-78, NGw5 and 3C1) is controlled by the same factors, namely 'P', 'K' and '91', (see GENERAL MATERIALS AND METHODS,2. The 'Stocks; also Lamb & Helmi, 1978). Mutant

w-78 used in these crosses carried the 'P' factor, and so presumably did the other two - NGw5 and 3C1 - since they were induced in crosses +(P) x +(P), (Table 21, and Pedigree chart).

The frequency of gene conversion in UVKw8 mutant is

comparatively very low (0.22% to 0.37%) and doesn't seem to be under the influence of the 'P', 'K' and '91' control factors. This mutant was induced in a +(K) x +(K) cross, (Table 21 and

Pedigree chart) but in crosses to (P), (K) and (91) wild-types

it does not behave like w-78(K).

Crosses of w-78 x UVKw8, with over 13,000 of asci scored,

gave no wild-type recombin'ant spores (Table 22c). This implies

that, if the two mutants do not actually occupy the same position in the map, they are extremely closely linked, so that they always co-convert. Similar cases were reported by Kuszewska and

Gajewski (1967) with the 77 and 73 mutants of the Y locus of

Ascobolus immerses (which do not recombine and have completely 200 different conversion frequer.:ies) and by Ros'signol (1969) with the 1472, 1848 and 1983 sites of gene 75 in Ascobolus immersus -4 which give very low frequency of recombination (about 1x10 ) and yet they are grouped in different classes (1472 and 1983 in class

of and 1848 in classy ), because they have different dissymmetry coefficients (3:1/1:3 segregation frequencies). With this criterion the total of 15 mutants of the central region of gene 75 fall into five distinct classes. According to

Rossignol (1969) the probability of correction in each direction, which on a hybrid DNA hypothesis results in the dissymmetry pattern, is conceivably determined either by the identity of mispaired bases at the mutant site or by interaction between the neighbouring and the mispaired site. lie considers that findings where mutants of a single gene fall into a small number of homogeneous and distinct categories seem to favour the idea that the chemical specificity of the mispaired site itself has at least a predominant influence.

The different crosses of the mutants of locus I to the three derived wild-types (Table 27) show that in general their conversion spectra are in accord with the expectations from the mutants' nature, as presumed from their origin. The different conversion spectra are allele-specific and not gene-specific and the type of conversion spectrum shown in each case is independent of the frequency of conversion shown.

Thus mutant w-78 gives only C type of conversion in all three crosses (Table 27) and mutant UVKw8 gives only D type of conversion. 201

Mutant 3C1 gives a B Lype in crosses to (P) wild-type and an AB type in crosses to either (K) or (91) wild-types. AB type of spectrum was reported by Leblon (1972a) for 4 of the 31 presumably frameshift mutations, but it was not found in any of the 21 ICRI70 induced mutants in Sordaria brevicollis by Yu-Sun et al (1977); the nature of the mutants giving AB conversion spectrum is not clear.

Unexpected results are given by NGw5 mutant: it shows D type of conversion spectrum in crosses to either (P) or (K) wild-types, but an A type in crosses to (91) wild-type. This shows that a single mutant may give rare postmeiotic segregation in a certain cross (about 6% of the total aberrant asci, in this case) and significantly more frequent postmeiotic segregation in another one (about 30% of the total aberrant asci). According to the relevant assumptions by Leblon (1972b), mutant NGw5 could be classified as a base substitution on one set of data and as a frameshift mutation on another set of data and this implies that most probably the frequency of postmeiotic segregation is not predominantly decisive in helping to conclude the nature of the mutation NGw5.

1,011A Mutants NGw5 and 3C1,0,Tere induced in (P) x (P) wild-types crosses, behave similarly to w-78(P) in crosses to (P), (K) and (91) wild-types, as said earlier. Accordingly, mutant UVKw8, which was induced in a (K) x (K) background, would be expected to behave like w-78 in the above crosses, which it does not

(Table 23). The lack of any effect of the 'K' factor as it is known from w-78 crosses on the gene conversion of UVKw8 mutant, 202 may be due to one or more of the following possibilities:

a. The 'K' factor may not affect the whole of the locus I. The relevant sequence of the mutants in locus I in the map is not known, due to infertility of the appropriate crosses, so it is not known whether UVKw8 is located in between the rest of the markers or outside them. However, the problem does

not seem to be one of site specificity of the 'K' factor; first because UVKw8 is extremely close to, if not occupying

the same position as w-78, which is under the control of 'K'

factor, and second because the 'K' factor is shown to control the gene conversion frequencie's of two other sites further apart in the locus I (NGw5 and 3C1, Table 23). b) The chemical nature of the UVKw8 mutant is such that it is not

affected by the 'K' factor. Mutant w-78 was presumed by Lamb

& wickramaratne (1975) to be due to a base substitution, based on the Leblon (1972a,b) criteria for classification of a mutant's nature according to its conversion spectrum. With the same criteria, mutation NGw5 may well be a base substitution

since it is giving conversion spectrum type D in crosses to (P)

and (K) wild-types (Table 27) and it was induced by NG. Similarly mutation 3C1 may be considered as a frameshift one since it shows rare postmeiotic segregation and it was induced

by ICR170. Gene conversion in these three mutants (w-78, NGw5

and 3C1). of locus I, is apparently under the control of the 'P', 'K' and '91' factors (Table 23), which means that factor 'K' may affect base substitutions and frameshift mutations

alike and so it would be expected to affect gene conversion of

UVKw8 as well. Since this does not actually happen it is 203

either that the assumptions on the nature of the mutations

involved are not completely correct or that the lack of interaction between 'K' and UVKw8 is due to a situation other than mispair specificity of the 'K' factor.

c) The 'K' factor is altered in strain UVKw8 or absent.

This could happen (i) if UVKw8 strain contains two mutations, one at the 'K'

factor and one at the UVKw8 site or

(ii)if the UVKw8 mutation involves a deletion or other alteration of a considerable length, covering both sites. We can distinguish between the two possibilities if we cross UVKw8 (altered K) x wild-type (K): recombinants will be

produced only if UVKw8 carries an "altered K" factor. This

has not been done here. If however mutant UVKw8 was due to a deletion, it would be expected, according to Leblon (1972a,b),

to show a conversion spectrum of type A (postmeiotic

segregation rare, conversion mainly to wild-type). In fact this mutant shows constantly conversion spectrum D (postmeiotic

segregation not rare, conversion predominantly to the mutant,

Table 27), indicating that it may be a base substitution

mutant.

d) Finally it is possible that UVKw8 differs from w-78(K) in the efficiency of correction (lower in UVKw8) rather than in the

frequency of hybrid DNA formation.

We already know that hybrid DNA in w-78(K) x +(P) crosses is

most probably formed predominantly in two chromatids (PART A,

SECTION II). If this is also true for UVKw8 x +(P) cross, 204

then it would be expected to produce more aberrant 4:4s than

cross w-78(K) x +(P), (at least 10 times more). In crossing

strain UVKw8 to the(wild-type,)granular used for detection of aberrant 4:4s at w-78, (PART A, SECTION I), it was found that

aberrant 4:4s were of about the same frequency as w-78

(0.1%, in 4,322 octads scored). This means that the two strains differ in the frequency of dual hybrid DNA formed, with w-78(K) forming more than UVKw8 strain. (Alternatively it

could mean that they have the same frequency of dual but w-78

has additional single hybrid DNA, which would account for its

higher total conversion frequency. This however would imply that hybrid DNA at w-78 in LCF cross is predominantly formed according to the single hybrid chromatid hypothesis, but this

is not supported by other data in this work (PART A, SECTION II)

In summary gene conversion in locus I is as follows:

- The frequency of gene conversion of w-78, NGw5 and 3C1 mutants is controlled by 'P', 'K' and 'L;1' factors, being higher in crosses to (P.) wild-type, and lower in crosses to (K) and (91)

wild-types. - The frequency of gene conversion of UVKw8 mutant does not show any clear effect of the 'P', 'K' and '91' factors.

- From Leblon's (1972a,b) classification: w-78 may well be a base substitution giving a C type of conversio

spectrum and 3C1: a frameshift mutation, probably an addition. NGw5: may be a base substitution, whose decrease in postmeiotic segregation in crosses to (91) wild-type may be due to

influence by the neighbouring sites, or a frame shift mutation

whose increase in postmeiotic segregation to (P) and (K)

types may be due to neighbouring sites. 205

UVKw8: may be a base .subs“tution, but it is not known why its

frequency of gene conversion is so different than that of w-78, with which it does not recombine. Probably the best

explanation for the gene conversion at UVKw8 would be that

UVKw8 strain contains two mutations, namely a base substitution

at UVKw8 site and a deletion at the 'K' site: this is consistent with no interaction between 'K' and UVKw8 sites, with UVKw8 showing U type of conversion spectrum, and the UKVw8 site

giving aberrant 4:4 segregations, which may not be expected

from long deletions (Meselson & Radding, 1975).

LOCUS II

Locus II contains two mutants; a spontaneous one (Sw26) and an NG induced one (NGw18), which did not recombine with each other (Table 22) in crosses where over 5,000 asci were scored, and which showed a big difference in the frequency of total gene conversion (Table 24). From the types of conversion spectrum that they gave in crosses to different wild-types (Table 27), they may be concluded to be base substitutions. Their differences in conversion frequencies in crosses to the same wild-type strains may be due to differences in the nature of the mispairs involved or in nearby cryptic heterozygosity (Lamb 1975). Gene conversion in locus II does not seem to be controlled by factors 'P', 'K' and '91' of locus I, to which locus II is unlinked.

Mutant NGw16 does not.belong,on the rather arbitrary allelism criterion used here, to locus II but shows 2.5% recombination frequency in crosses to NGw18. These two mutants show about the same frequency of gene conversion in crosses to 206 the (P) and (91) wild-types (about 4.5%) but not in crosses to

(K) wild-type, where NGw18 give a much lower value than NGw16

(0.18% and 4.64% respectively). These along with' NGw5, are also the only mutants of the NG induced ones that give relatively high frequency of conversion.

LOCUS III

Three mutants are involved here (Table 25), all ICR170 induced, showing similar frequency of gene conversion in the different crosses to the same wild-types and similar conversion spectra, namely of C, D or CD type, with one exception: mutant Cwl.4, crossed to (K) wild-type gave a B type of conversion spectrum (Table 27). This is not compatible with data of Leblon (1972a) where all ICR170 induced spore colour mutants in

Ascobolus immersus showed B type conversion spectrum. It is also not compatible with data of Yu-Sun et al (1977) in Sordaria brevicollis, where most of the mutants induced by ICR170 gave B type of conversion spectrum and the rest of them A type.

The frequency of mutations produced after ICR170 treatMent of the crosses used here was 50 to 60 times higher than that of the corresponding control crosses, (Table 20), so it is rather unlikely that the mutations isolated as induced by ICR170 here are actually spontaneous ones.

Leblon (1972b) inferred that the class C mutants, which he obtained with Ascobolus immersus, originated from base-pair substitution, and the report by Yu-Sun et al (1977) supports this. 207

Leblon (1972b) also concluded that the acridine mustard

ICR170 induces chiefly additions, since Imada et al (1970) showed that the acridine proflavin preferentially induces addition frameshifts in phage T4 of Escherichia coli.

There is therefore a discrepancy in the present data concerning the nature of the mutants in locus III as implied by their origin and by the types of their conversion spectra.

ICR170 compound however contains two parts:

1) an acridine ring, which is required to produce base-pair insertions and/or deletions, as proflavin does, and

2) a nitrogen mustard side chain - referred to as the alkylating part of the ICR170 molecule,- which is essential for mutagenicity of ICR170, and to which the induction of base-pair

substitutions is attributed (Mailing 1967).

According to this ICR170 possesses two different mutagenic activities, namely induction of base-pair insertions or deletions and induction of base-pair substitutions. Mailing (1967) studied the relative specificity of ICR170 for inducing both kinds of mutations, using inversions in Neurospora, and he concluded that at low doses mutagenesis of frameshifts may predominate, but he stresses that the ratio between the reversion frequency induced by ICR170 in base-pair substitution mutants and in a base-pair insertion or deletion mutant can be varied to some extent by the mutants used. 208

Later, in 1970, Munz & Leupold reported the characterization of 67 ad-6 and 50 ad-7 mutants induced by

ICR170 in Schizosaccharomyces pombe. From these only 19% of the ad-6, and 32% of the ad-7 mutants were of the frameshift type, the rest being base substitutions or other alterations.

In 1976 Auerbach concluded that the diagnostic value of the ICR compounds for frameshift mutations is diminished by the fact that they may also act as alkylating agents, and that this duality of action creates an ambiguity that has to be taken into account when mutations in cellular organisms are classified by means of their response to ICR compounds.

It is possible therefore that the three ICR170 induced mutations of locus III may not be frameshifts,'but base substitutions, showing a conversion spectrum expected by such a mutation', that is a C, U or CD type, indicating at the same time that there may be a site specificity of ICR170 for base substitutions' induction at the locus III.

Factors 'P', 'K' and '91', which control gene conversion frequencies in locus I, do not seem to affect gene conversion frequencies in locus III, at least not in a parallel way.

LOCUS IV

The two mutants belonging to locus IV are both NG induced, showing low frequency of gene conversion (from 0.19% to 0.59%,

Table 26) and mainly U type of conversion spectrum. (Table 27).

This is compatible with expectations from Leblon (1972a,b) and Yu-Sun et al (1977),. according to which NG induced mutations, 209. presumed to originate from 1:se substitutions, show relatively high postmeiotic segregation, with conversion spectra being of type C, D or CD.

Here again the frequencies of gene conversion shown by the two mutants in the different crosses seem to be independent of the influence of the factors 'P', 'K' and '91' of locus I.

UV induced mutations

Table 29 contains the conversion spectra and relative frequencies of conversion classes for all the UV induced mutants that it was not possible to assign to any particular locus, except mutant UVKw8, which belongs to locus I.

Most of the UV induced mutants give at least one cross with high frequency of gene conversion (from nearly 9% to 15.3%, Table 28) which is much higher than their frequencies in other crosses or the average gene conversion frequency given by mutants of other origins (under 2%), Tables 30, 32 and 34.

Four of these mutants show their high frequencies of conversion in crosses to (P) wild-type specifically: mutants

BB9w9, BBwll, BBw21 and BB9w8. This could,indicate that factor 'P' may have a wider range of action than just on locus I. However mutant BB9w9, although giving 13.36% of gene conversion in crosses to the (p) wild-type BBR1-, it gives only 1.76% in crosses to another (P) wild-type, namely 92-. This may indicate that the frequency of gene conversion in this mutant is independent of the control of factor 'P', but it may also be that other factors apart from 'P' are influencing the conversion 210 frequency of this mutant's cross to 92 - wild-type, which are not involved in the mutant' cross to BBR1-. It is unfortunate that the 92- wild-type gave no fertile crosses with the other three of these mutants, and that crosses among the 4 of them were also infertile.

Mutant Bilj gave a frequency of gene conversion over 5% in all three crosses (Table 28), with its highest value in its cross to (91) wild-type, which is usually giving the lowest values with the rest of these mutants.

The conversion spectra of the UV induced mutations were not investigated by Leblon (1972a), but they were studied by Yu-Sun et al (1977), and they were found to be one of class

B and 10 of class C.

Results from phage T4 of Escherichria coli and Neurospora crassa have indicated that both base substitutions and frameshifts occur after UV induction (references in Auerbach, 1976), and Yu-Sun et al (1977) consider that the diversity of molecular structure implied by their results is expected with UV, which is believed to derive its mutagenicity from error-prone repair of damage to the DNA.

The present results with UV induced mutations show even greater diversity, not only between mutants but also between crosses of the same mutant to different wild-types (Table 29).

Only mutant Blij gives the , same spectrum (C) in all three crosses used, while the conversion spectra of mutants Mk) (C and D) and

BBm (CD and C) show clearly the intergradation of the C, D and 211

CI) classes, as was shown by Lamb & Wickramaratne (1975) and discussed by Yu-Sun et al (1977). These mutants could presumably be base substitutions.

Mutants BB9w9, BBw11, BBW21 and BB9wY however are giving conversion spectra of both frameshifts and base substitutions, depending on the wild-type they are crossed to. This is similar to the situation found in mutant NGw5 of locus I. There is an interesting correlation in these five mutants between the total frequency of gene conversion and the conversion spectrum shown: crosses of high conversion frequency give all C type of conversion spectrum, and crosses of low conversion frequency give either A or AB type of conversion (Tables 23, 27, 28 and 29). That is to say that in low conversion frequencies these mutants behave like frameshifts, while in high conversion frequencies they behave as base substitutions, with one exception; mutant BBw21 crossed to (91) wild-type gives a CD type of spectrum and low conversion frequency.

In other words, these mutants, whose direction of correction is almost always to the wild-type, show rare postmeiotic segregation in low conversion frequency crosses and relatively high one in high conversion frequency crosses.

It is also noticeable that crosses of these mutants that gave high conversion frequency and C type of conversion are crosses to the (P) wild-type. And most of the crosses that gave low conversion frequency and A or AB type of conversion are crosses to (91) wild-type (5 out of 8), the rest being 2 crosses to (K) and 1 cross to (P) wild-types, (Table 29 and for NGw5 Table 27). 212

This situation with these five mutants, where changes in frequency of gene conversion are accompanied by changes in the relative frequencies of the aberrant classes, is similar to that presented by Lamb &. Helmi (1978) for the closely linked mutants w-78, w-10, 3d1 and NG1, where gene conversion is under the control of factors 'P', 'K' and '91'.

There is a possibility therefore that gene conversion in these UV induced mutants, which are unlinked to the w-78 locus

(Table 22c) is also affected by factors 'P', 'K' and '91'.

Lamb & Helmi (1978), inferred for the w-78 system that on the dual models alone, the controlling factors would have to affect the frequency of hybrid DNA formation and directly affect correction too, since a change in the frequency of hybrid DNA formation only, would bring about parallel change in the absolute . frequencies of all narrower ratio conversion classes but no change in their relative frequencies.

On single hybrid chromatid models (as in Lamb & Helmi 1978) there are two separate kinds of hybrid DNA formation: invasion of the + bearing chromatid by part of the mutant-bearing chromatid (in + x m crosses), or invasion of m by +. If control factors equally (proportionately) increased both types of invasion, there would again he no change in relative frequencies of conversion classes; this could only result from different frequencies between the two kinds of invasion.

Finally, mutant BBm presents another interesting feature: in crosses to (P) wild-type.it gave only postmeiotic segregation, 213 though in very low frequency and equally to the wild-type and to the mutant (Table 29). In crosses to the (K) wild-type KIV + however it gave equal frequency of meiotic and postmciotic segregation, all to the.wild-type, while the total frequency of conversion was the same as before. Though the numbers of aberrant asci in these crosses are very small - and therefore bigger samples are probably required to confirm the present data - it is possible that the cross BBm x 9r+ is deficient in correction while cross BBm x KIV+ is not.

In conclusion UV mutations here show a great variety of gene conversion spectra, which is unexpected from other relevant works (Leblon (1972a,b) and Yu-Sun et al (1977)). 214 NG induced mutations

The mutations induced by NG in general give low frequencies of gene conversion. Table 30 contains all NG induced mutants which were not assigned to any particular locus. These mutants in general gave low frequencies of gene conversion, (under 1%), except mutant NGw16 (4% to 6%), which is linked to locus II.

NG induced mutants are expected to be base substitutions

(review in Auerbach, 1976) and as such are expected to give conversion spectra of C, U or CD type (Leblon 1972a,b, and Yu-Sun et al, 1977). The present data of NG induced mutations in

Ascobolus immersus confirm these expectations and are therefore comparable with the ones reported by other people.

ICR170 induced mutations

Tables 32 and 33 contain all the ICR170 induced mutations which were not assigned to any particular locus, with their conversion spectra and total frequencies of gene conversion.

These mutants showed in general a low frequency of gene conversion (highest value: 3.40%), and a variety of conversion spectra (Table 33).

Some of these show a C type of conversion spectrum and some show B type .(sometimes A or AB) conversion spectrum.

Mutant Cw1.7 which twice shows spectrum B and once spectrum D, shows the lowest frequency of postmeiotic segregation

for D spectra (15.8%). 215

Mutant Cwll shows all three types, A, B and AB conversion spectra, presenting difficulties when it comes to deciding whether it is a deletion or an addition.

According to Leblon (1972a,b) and Yu-Sun et al (1977) the other three mutations showing B type of conversion spectrum

(Bw6.1, Bw6.2 and Cw5.2), Table 33, may be concluded to arise from frameshifts of the addition kind.

Much of the discussion concerning the nature of the mutants induced by ICR170 was done in paragraph on locus III.

The A and B classes of mutants reported by Leblon (1972b) were of complementary molecular structure, with ICR170 favouring the induction of B and the reversion of A, and EMS often inducing

A and reverting B. Leblon concluded that ICR170 induces chiefly additions, just as the acridine proflavin does (Imadaaa41970). EMS is believed often to induce' deletions since it reverts T4 mutants that have been induced with proflavin (Drake 1970, cited by Leblon 1972b). In this way, the 13 type of spectrum was deduced to result from addition, and A type from deletion, frameshifts.

This assignment however is the opposite of what would be expected on another set of data, reported by Benz & Berger

(1973). Here E. coli was mixedly infected with T4D wild-type and rII deletion mutants or with the same wild-type and addition lysozyme mutants. The result was that in both cases the excess

DNA - the wild-type allele in the first case and the mutant

allele in the second one - was lost. They concluded that their 216 results demonstrate a rep air system which removes "loops" in heteroduplex DNA molecules.

According to this, a deletion would result in gene conversion directed mainly to the mutant - giving B type of conversion - and an addition would result in gene conversion directed mainly to the wild—type - giving A type of conversion -. This means that frameshifts from ICR170 would be of class A, and not B, and from EMS of class B, and not A, which is the opposite of Leblon's conclusions.

In the present work it is found that, on the basis of their conversion spectra only, six mutants of the eleven ICR170 induced ones could be presumed to be frameshift mutations and the remaining five could be base substitutions, with three of

them constituting locus III.

It is noticeable however that the crosses which gave B type of conversion spectrum in this work were all from mutants induced by ICR170 and not by other mutagens, or spontaneous.

The present data of ICR170 induced mutations indiCate that their gene conversion spectrum is a better guide than their origin in concluding the nature of the mutants.

Spontaneous mutations

Table 34 contains all spontaneous mutants which were not classifiable to any particular locus, with their frequencies of total gene conversion, and Table 35 gives the conversion

spectra obtained in their different crosses. 217

The conversion spectra of spontaneous mutations in relation to their nature were studied only by Leblon (1972a), where they were reported to give 2 A type of spectrum, 3 B type,

1 AB type, 2 CD and 2 1) type. This diversity among spontaneous mutants is in accord with Leblon's postulation that the nature of the mutants is responsible for the difference observed in the conversion spectra, because then one would expect mutants induced by very specific mutagens to have homogenous conversion spectra, whereas mutants induced by a mutagen of low specificity or arising spontaneously, would give various types of conversion spectra.

The spontaneous mutants isolated here show low conversion frequencies (ranging from 0.12% to 1.08%) and according to their conversion spectra they may be presumed to be base substitutions, except mutant Sw20 which is probably due to a frameshift mutation.

Mutagen specificity and interchangeability of conversion spectra

It has been shown in this work that the conversion spectra of a mutant are often interchangeable, that is a mutant may give more than one type of conversion spectrum, depending on genetic factors in the strains used. In order to define the relationship between mutagen on one side and conversion spectra on the other we have to consider the number of crosses that give each class of conversion spectrum among the mutants of a specific origin rather than the number of mutants that do so, (Table •40), each 218 mutant usually being used in three different crosses, that is to wild-types carrying 'P', 'K' or '91' control factors.

TABLE 40

Relationships between mutagens and conversion spectra in crosses of various types of mutant

Origin of Total mutants Classes of conversion spectra crosses No. of Total mutants A B AB C D Cl) mutants/ studied UV *4/3 - 2/2 10/6 4/2 2/2 22/15 9 NG - - - 12/5 4/2 5/3 22/10 9 ICR17O 1/1 11/7 3/2 9/4 1/1 2/1 28/16 11 Spontaneous 2/1 - 1/1 7/5 1/1 4/4 15/12 9

* Numerator: Number of crosses giving that class of conversion spectrum

Denominator: Number of mutants from which the crosses in the numerator arose.

/ Total mutants: the sum of the denominators, giving the number

of mutants that would be required to give the respective total crosses, if each mutant gave the same class in all its

crosses.

According to Table 40 conversion class C is the most common for spontaneous as well as for UV- and NG- induced mutants, while class B is only given by ICR170 induced mutants, for which mutants it is the commonest class. 219

The interchangeability of the conversion classes for each origin of mutants increases with increasing difference between the number of mutants studied in each case and the denominator of the fraction: total crosses/total mutants, (Table 40). This difference is 6 for UV mutants, 5 for ICR mutants, 3 for spontaneous and 1 for NG mutants.

With this criterion we may say that UV induced mutations show the highest degree of interchangeability, with ICR170 induced mutations. coming second, spontaneous ones third and NG induced

mutations fourth.

NG is shown therefore here to be the most specific mutagen, inducing mutants showing only the spectra of base substitutions (C, D, CD) and with the least interchangeability of spectra. UV induced Mutants show the highest degree of interchangeability while ICR170 is the only mutagen here giving mutants with crosses of all classes of conversion spectra, and

also the only mutagen giving mutants of B class spectrum.

Intergradation of conversion spectra

Leblon (1972a) reported that in Ascobolus the difference

in the frequency of postmeiotic segregation between classes A, B and AB (0-5%) on one side and• C, D and CD on the other side

(45-88% for gene bl and 27-64% for gene b2) was definite and unambiguous. In Sordaria (Yu-Sun et al, 1977) this difference was much smaller (postmeiotic segregation was 0-8% for A, B

and 14-54% for C classes). 220

Figure 12 gives the dLstribution of the mutants of genes b and b 1 2 studied by Leblon (1972a), according to their relative postmeiotic segregation. It is shown that while the separation

of the two groups is clear for gene b1, it is not so clear for gene b2, where more than one relatively big gaps appear.

In the present work there is no really clear distinction between rare (0-10%) and frequent (15-100%) postmeiotic

segregation classes. This is so whether we regard the different classes given by the mutants in each of recognised loci (Fig 11) or the ones given by these pluS the rest of the mutants (Figs 13, 14).

For example in locus I the highest value of postmeiotic segregation between the "rare postmeiotic segregation" classes

is 6.09%, and the lowest value for the "frequent postmeiotic

segregation" classes is 15.2%. In the same locus the biggest gap in a numerical sequence of values for % postmeiotic

segregation is between.the D type crosses of mutant NGw5 and

the also D type crosses of mutant UVKw8, which involve the same wild-types (Fig 11).

Locus III, which contains only ICR170 induced mutants, shows crosses with relative postmeiotic segregation from 7.9% to 70.4%, with seven more values in between these two, evenly arranged all along. The biggest gap between two successive points appears between two C types of cross, namely Bw4.2 x 91-

.and Bw3.10 x 9r+ (Fig 11). 221

Concluding Remarks

The data here do not show a clear distinction between mutants of the groups (A, B, AB) and (C, D, CD), on the basis of their relative frequency of postmeiotic segregation and they do show that a mutant may give very different conversion spectra, depending on the wild-type used in the cross.

This could be interpreted as lack of a clear correlation between the nature of a mutant and the conversion spectrum of a particular class. Such a lack of correlation is not unexpected on theoretical grounds if a correlation between a mutant's origin and its self triggered correction pattern is obscured in its conversion pattern in either or both these two cases: a) When there is a "single hybrid chromatid" model operating,

the conversion spectrum will depend in part on the frequencies

with which an m chromatid is invaded by a strand of a +

chromatid (producing 6:2, 5:3 and 4:4 classes) or a + chromatid is invaded by a strand of an m chromatid (producing 2:6, 3:5 and 4:4 classes). These two frequencies of chromatid invasion may be unequal and probably are independent of the

nature of the mutant involved.

b) Cryptic heterozygosity may well influence the conversion spectrum of a mutant in a manner independent of the nature

of the mutant itself and depending on the nature and position

of the cryptic heterozygous site.

corqillue Cr) page 223 222 cross Bib x 3r-, which showed negative corresponding-site interference for classes 1:7 and 0:8.

The lack of strong corresponding-site interference in

Ascobolus immersus has been demonstrated before (Lamb & Wickramaratne, 1973; Ghikas & Lamb, 1977), for the w-78 and w-10 sites.

The present data from mutants from different parts of the genome of the same organism do not contradict previous findings, although the numbers scored are rather small.

Mutant Bw4.3 (Table 38) gives consistently, in all three crosses, higher numbers of observed than expected asci in classes 7:1 and 8:0. Although the difference is not significant and the scored wider ratio octads were not tested by isolation and back-crossing, it would probably be worthwhile to see in bigger samples if this mutant shows consistently negative corresponding- site interference in wider ratio asci that show conversion to the wild-type.

coniin ue on pag e 224 223

The fact of many examples here of one mutant giving

different conversion spectra in different crosses implies one or more of the following conditions occur: (1) lack of a good correlation between a mutant's origin and its conversion spectrum; (2) single hybrid chromatid DNA is formed at appreciable frequencies, with unequal frequencies of invasion by the two kinds of chromatid, at a high proportion of the. sites investigated; (3) most sites here have in their vicinity sites

of cryptic mutant heterozygosity, correction at which has major effects (by co-conversion) on conversion results at the known site.

Supporting these data are several reports (Fink & Styles, 1974; Fink, 1974; Lawrence, Sherman, Jackson & Gilmore,

1975), which have shown that there are many similarities between the conversion of point mutations and the conversion of deletions

and that "parity" (equality of conversion types 3+:1m and 3m:1+

in heterozygous crosses) can be a property of gene conversion

of both types of mutations. However, postmeiotic segregation was not detectable in these works and therefore no comparison between the yeast deletions and point mutations was possible on

this basis.

Wider ratio asci

The observed and gxpected numbers of wider ratio asci, in crosses to the different wild-types are given for the UV

induced mutants (Table 36), the NG induced mutants (Table 37)

the ICR170 induced mutants (Table 38) and the spontaneous ones

(Table 39), without significant difference between them except

C0v7i I. 41 We oviPale 222 224 IV CONCLUSIONS

1. The conversion spectrum shown by a mutant in a particular cross very often depended on the wild-type strain participating

in the cross, and this either does not allow for a strong

correlation between the chemical nature of the mutant and the type of gene conversion spectrum it may show, or it implies that other genetic factors apart from the nature of the mutant May play a major role in the gene conversion process of this mutant.

A third possibility is that both these may occur at the same time.

2. There is not a really clear distinction of the mutants

in two groups - (A, B, AB) and (C, D, CD) - on the basis of their relative frequencies of postmeiotic segregation.

3. According to the present data NG was specifically inducing mutants that give C class of conversion spectra mainly, while B class of conversion spectra were specifically given by mutants induced by ICR170.

4. There was usually no correlation in these data between the frequency of gene conversion shown by a mutant in a specific cross and the type of conversion spectrum shown in this cross. A probable exception is the case of the four UV induced mutants, BB9w9, BBw11, BBw21 and BB9w8, and the NG induced NGw5 (of locus

I), which are of C class in high conversion frequency crosses (to (P) wild-type) and of A..class in low conversion frequency crosses (mainly to (91) wild-type). - postmtio-eic segregation k%) •(11 • C/1 t•;' . ta-1 C •• C 0 - c U o 7-78+ x 61 • i-78 x

O w-78 x 91- N UVKw8+ x 61- O UVKw8+ x KII O

=UVKw8+ x 91- oE

eT .c.:NIGw5+ x 01- T \1Gw5+ x Kill

l aAT Gw5+ x 91- soc

ta tzi3C1+ x 61- C-) an >13C1+ x Kill- oT C

on. e 3C1+ x 91-

•co", 'riSw26+ x 92- O o (i)• nSw26+ x KIII -,NCw18+ x 92-

r:NGw18+ x KII G7 - r:NGw18+ x 91-

Bw4.2- x 9r+ r-i r. Bw4.2 x KIV+

1w4.2 x 91-

Bw3,10 x 9r+

Bw3.10 x KIV

c-) Bw3.10 x 91-

Cw1.4-x 9r+

rz Cw1.4- x KIV

Cw1.4+ x 91- •

1 1+ x)2 -

72n' N Cw 11 x I I C- NCw11+ x 91-

=NGw12+ x 92- SZZ r\iCw12+ x 91- .41 No. of crosses 22

20

18

10 Leblon (1972a) Leblon (1972a) 14 NG 10 NG 3 ICR17O 5 ICR170 18 Locus b EMS 5 1 EMS 11 Locus b 2 Spont 5 Spont 5 10 31 mutants 31 mutants

6

4 2 n rr -1nm nmn Ii 0 4 16 24 32 40 48 56 64 72 80 88 3 16 24 32 40 48 56 64 72 SO

Relative postmeiotic segregation (o)

FIGURE 12 Distribution of the mutants studied by Leblon (1972a) according to their relative postmeiotic segregation No. of crosses 12

11

10

9

8

7

1r n 6 12 18 24 30 36 42 48 54 60 66 72 78 S4 90 96 100

Relative postmeiotic segregation (%) FIGURE 13 Distribution of all crosses obtained from all mutants according to their relative postmeiotic segregation frequencies No. of crosses

b

UV NG

4

3

ti

Spontaneous ICR170 5

2

••••••■ 1 1 1 [1 L6 24 32 40 48 56 64 72 80 88 )6 100 8 16 24 32 40 48 56 64 72 80 8 Relative postmeiotic segregation (a) FIGURE 14 Distribution of the crosses obtained by all the mutants spontaneous or induced (UV, NG and ICR170) accordinz to their relative nostmeiotic segregation 229

OPEN PROPLEMS

AND

SUGGESTIONS FOR FURTHER WORK

The present work was an attempt to provide answers to the questions stated in the "GENERAL INTRODUCTION". However, new questions and possibilities arose, which are of interest for further work.

1. The detection of 5:3(3) and 3:5(3)

After the detection of 6:2(2) and 2:6(2) here only the 5:3(3) and 3:5(3) of the classes predicted by Lamb &

Wickramaratne (1973) remain undetected. These two last classes could not be detected with the system used for the detection of the 6:2(2) and 2:6(2), but in theory should be detectable in suitably multiply-marked crosses, where the T segregations of various pairs of markers define the spore pairs. This however would be extremely laborious unless it provides a specific selection system. Such a system would be for example one with three visual markers which would provide the basis for the initial selection, with further tests for other markers to confirm the aberrant segregation for the marker under study. Another possibility would be a special system, which would involve two visual markers like the ones used for the detection of 6:2(2) and 2:b(2) here, but would show 'very low or no T' as suggested by Lamb, (Ghikas Lamb 1977).

2. A new method for assessing the relative frequencies of symmetrical and asymmetrical hybrid DNA at a mutant site. - ▪

230

An attempt was made t',1 assess the relative frequencies of occurrence of dual and single hybrid DNA in site w-78 (PART A, SECTION II), through the analysis of the 5:3 and 3:5 classes, in crosses heterozygous for the control factors 'P' and 'K'.

This analysis even when it is carried out fully and with satisfactory results can only be applied in one sort of cross here, namely the LCF ones (low conversion frequency), and it does not permit the study of crosses involving the same factor in both the crossing strains (+PXmP or +KXmK). The study of these last crosses would be very useful in determining how factors 'P' and 'K' actually affect gene conversion at the w-78 site, in terms of hybrid DNA formation and/or correction.

This sort of study is possible with the systems described below:

Consider the system of visual markers that were used in this work for the detection of 6:2(2), 2:6(2) and 4:4(2) classes (PART A, SECTION I), and let us take the cross:

w-78 (P), gr-3+, x w-78 (P), gr-.) , + and consider conversion at the w-78 site. A S+:3w octad in this cross may come in one of four different phenotypes

1) 4 red: 1 granular: 3 white spores, (4:1:3) 2) 1 red: 4 granular: 3 white spores, (1:4:3) 3) 3 red: 2 granular: 3 white spores, (3:2:3)

4) 2 red: 3 granular: 3 white spores, (2:3:3) where red spores are the genotypically w-781- , gr-3+ ones

granular spores are the genotypically w-78+, gr-3 ones, and white spores are the genotypically w-78 , gr-3 or

w-78- , gr-3 ones 231

Consider a 5:3 formation in octad with parental ditype

(PD) and non-parental ditype (NPD) segregations for.w-78 and gr-3

I PD II NPD w-78 R1-3 w-78 gr-3

M 2 m2 m a 2 a M M • 1 2 • 1 ' + m2 b m M 1 • In1 2 m M M 1 1 2 m m m 1 1 2 I In PDs

1. If symmetrical hybrid DNA occurs and mispairs a and b are

formed, then

i) correction of a to + and non-correction of b gives: 4 spores: red granular

1 spore: red non-granular

3 spores: white non-granular that is a (1:4:3) phenotype of a 5:3

ii) correction of b to + and non-correction of a gives

3 spores: red granular 2 spores: red non-granular

1 spore: white granular 2 spores: white non-granular that is a (2:3:3) phenotype of a 5:3

2. If asymmetrical hybrid DNA occurs only mispair b is formed,

which is not corrected, and the result is the same as in case (i 232

with the symmetrical hybrid DNA, that is a (1:4:3) phenotype

of 5:3.

II In NPD

1. If symmetrical hybrid DNA occurs and mispairs a and b are

formed, then

i) correction of a to + and non-correction of b gives:

4 spores: red, non-granular 1 spore: red, granular

3 spores: white, granular that is a (4:1:3) phenotype of a 5:3

ii) correction of b to + and non-correction of a gives:

3 spores: red, non-granular 2 spores: red, granular 1 spore: white, non-granular

2 spores: white, granular That is a (3:2:3) phenotype of a 5:3

2. If asymmetrical hybrid DNA occurs only mispair b is formed,

which is not corrected, and the result is the same, as in case (i) with the symmetrical hybrid DNA, that is a (4:1:3)

phenotype of 5:3.

In Tetratype (T) segregations there are four different

arrangements, in which mispairs a and b may occur.

An analysis of the tetratype segregation (T) along the

same lines reveals that the different phenotypes of 5:3 class • are produced from either symmetrical or asymmetrical hybrid DNA

as following: 233

1. If symmetrical hybrid DNA occurs, the four different

arrangements of the mispairs in Ts produce eight possibilities for the 5:3 phenotypes, according to whether it is mispair a or b corrected in each arrangement. These eight possibilities are represented in the 4 phenotypes of 5:3 as

follows:

a) 1 x (4:1:3)

b) 1 x (1:4:3)

c) 3 x (3:2:3) d) 3 x (2:3:3)

2. If asymmetrical hybrid DNA occurs, the four different arrangements of the single occurring mispair in Ts produce four possibilities for the 5:3 phenotypes, with the following

outcome.

a) 2 x (3:2:3) and

b) 2 x (2:3:3)

This analysis shows that there are different expectations

for the possible phenotypes of the 5:3 octads resulting from symmetrical or asymmetrical hybrid DNA formation at the locus

under study, in Plls, NPDs or Ts of the system.

On the assumption that the correction to + of the mispairs a and b in the symmetrical hybrid DNA occurs with equal frequency,

we may consider the frequencies of the different phenotypes of

5:3s as follows:

▪• -

234

If the frequency of 5:3s re'llting from PDs = A

the frequency of 5: 3s resulting from NPDs = B and

the frequency of 5:3s resulting from Ts = C

then 1) in the case of symmetrical hybrid DNA formation:

the frequency of (4:1:3)s = +

the frequency of (1:4:3)s = +

(3:2:3)s and the frequency of = 2 + -=-3C A 3C the frequency of (2:3:3)s = +

2) in the case of asymmetrical hybrid DNA formation: the frequency of (4:1:3)s = B

the frequency of (1:4:3)s = A the frequency of (3:2:3)s = T2- and

the frequency of (2:3:3)s = 2C 4

(4:1:3) + (1:4:3) W e may now consider the fraction (3:2:3) + (2:3:3) an d see how it is evaluated, using the frequencies of the different 5:3

phenotypes given above for each type of hybrid DNA separately.

1. In the case of symmetrical hybrid DNA

B C (4:1:3) + (1:4:3) _ 7 8- 7 -g 2(A4B)+C (3:2:3) + (2:3:3) B 3C 4. A 3C 2(A+B)+3C 2 8 2 + 8

2) In the case of asymmetrical hybrid DNA

(4:1:3) + (1:4:3) B+A _ B+A (3:2:3) + (2:3:3) 2C + 2C C 4 4 235

If we now assume that. 5:3 octads occur in equal proportions in the groups PD, NPD or T, then the relation among the variables A, B and C will be the same as among PDs, NPDs and Ts in a cross. For example: if PDs = NPDs = Ts = X then A = B = C = Y, and then the fraction

with symmetrical = 2(Y+Y)+Y, _ 0.71 hybrid DNA 2(Y+Y)+5Y

(4:1:3)+(1:4:3) - (3:2:3)+(2:3:3) with asymmetrical = Y+Y 2 hybrid DNA

This shows that for P=NPD=T the ratio of the frequencies

of (4:1:3)+(l:4:3) over the frequencies of (3:2:3)+(2:3:3) is

0.71 if hybrid DNA occurs mainly symmetrically, while it is equal to 2 if hybrid DNA occurs mainly asymmetrically.

This method for assessing the relative frequencies of occurrence for symmetrical and asymmetrical hybrid DNA is most

helpful with a system of markers where the frequencies of PDs, NPDs and Ts allow the greatest possible difference between the two (4:1:3)+( 4: 3) according to each type values of the fraction (3:2:3)+( 2:3:3)' of hybrid DNA. As pointed out by Dr. Lamb (Personal communication)

the best such system would be that of low T frequency, compared to PD and NPD, but high enough in order to permit accurate estimation of the frequencies of the (3:2:3) and (2:3:3) phenotypes

which in the case of the asymmetrical hybrid DNA occur only in Ts.

The second marker in the "low T" system could be unlinked, or

linked but non-co-converting with the main marker. 236

However, this system could not be used here with the system used for the detection of 6:2(2), 2:6(2) and 4:4(2), from which we take the particular cross considered at the beginning of this analysis. This is because in this system PDs = NPDs = X

and Ts = 2(PD+NPD) = 4X therefore A = B = Y and C = 4Y and therefore

with symmetrical hybrid DNA = 2(Y+Y)+4Y = 8 = 0.5 2(Y+Y)+12Y 16 (4:1:3)+(1:4:3) (3:2:3)+(2:3:3) = Y+Y - O.5 with asymmetrical 41 hybrid DNA

In this system, as we see from the two values given for the above fraction, the frequencies of the phenotypes in the numerator and the denominator are expected to be the same with either symmetrical or asymmetrical hybrid DNA, and therefore

it can not help in evaluating the relative frequencies with which

the two kinds of hybrid DNA may occur at the w-78 site.

3. Studies with the mutant UVKw8 The possible nature of this mutant is discussed elsewhere

(DISCUSSION of PART B), as probably involving two mutations: one

at the UVKw8 (white spore) site and one at the 'K' site giving a 'altered K' character. The case of an 'altered K' mutation would be very interesting, since it could give us the possibility of

studying its effects on the rest of the mutants that belong to the same locus (locus I), whose conversion is controlled by the

factors 'P', 'K' and '91'. 237

We'could also see whit would be the frequency and spectrum of gene conversion at the UVKs8 site, when the 'altered K' factor is absent or replaced by any of the 'P'. 'K' and '91' factors.

It is also probable that there is only one mutation involved in the UVKw8 strain, and in this case it would be of interest to investigate what would be the effect, if any, of this

mutant on the gene conversion of the other mutants in the locus

in interallelic (coupling and repulsion) crosses, where the other mutant may carry the 'P', 'K' or 'DJ.' control factor.

4. Studies with some UV induced mutants UV induced mutants BB9w9, BBw11, BBw21 and BB9w8 (as well

as the NGw5 mutant of locus I), show a correlation between conversion frequency and conversion spectrum and a probable correlation between conversion frequency and kind of wild-type strain ((P), (K), or (91)) used in the cross. They have been discussed as possibly affected by the control factors 'P', 'K'

and '91' of locus I and an investigation towards this direction

would be of interest since it may reveal that these factors have a wider effect in the genome of the Ascobolus immersus strains

used here, and do not affect only locus I.

It would also be interesting to know (through new mutant isolates that may give fertile crosses), whether they are linked or not, and whether the correlation between their conversion frequency and conversion spectrum is real or accidental. 238

It would also be of pinch interest to know what lies behind such a correlation, which shows these mutants to give the base substitution spectrum in one cross and the frameshift spectrum in another one. The use of a new method for assessing the relative occurrence of symmetrical and asymmetrical hybrid DNA, proposed here (paragraph 2), could probably assist in resolving this question, along with other methods (e.g. detection of aberrant

4:4s).

5. Studies with the mutants of locus II

The mutants Sw26 and NGw18, which do not recombine, show very different frequencies of gene conversion. It would be interesting to know what would be the conversion frequencies of other mutations induced in this locus, and what factors are responsible for the observed difference in the conversion frequencies of the Sw26 and NGw18 mutants, which give (both of them) C conversion spectra. There is a possibility that they differ only in the frequency of hybrid DNA formed or only in the efficiency of correction of mispairs, and in either case it would be interesting to localize the factors involved. The fact

that the two mutants do not recombine makes their investigation somewhat difficult (for example they are not distinguishable from

the double mutant) but closely linked to the locus II is mutant

NGw16, which shows 2.5% frequency of recombination in crosses to

NGw18. It is possible that the NGw1.0 mutant may belong to locus

II in spite of not fitting the present arbitrary criterion of

allelism, but does not co-convert with NGw18 with a very high.

frequency, showing as only closely linked to it. :Mutant NGw16

gives similar conversion frequencies and spectra as NCw18 and

therefore it can be taken with the mutants of locus II for studies 239 concerning the difference of conversion frequencies between

Sw26 on one hand and NGw18 and NGw16 on the other.

b. Studies with mutant BBm

This is a UV induced mutant which in crosses to 9r+ wild- type gave only postmeiotic segregation, while it did not do so in crosses to KIV + wild-type (Table 28). It would be of interest to investigate the relative frequencies of symmetrical

and asymmetrical hybrid DNA and the occurrence of aberrant 4:4s, in the two crosses, as it may be that BBm mutant is correction-

deficient in one cross and not so in another. 240

REFERENCES IN ALPHABETICAL ORDER

AUMAD, A., W.K. IIOLLOMAN and R. HOLLIDAY, (1975). Nuclease that preferentially inactivates DNA containing mismatched bases. Nature. 258 AHMAD, A., and U. LEUPOLD, (1973). On a possible correlation between fine-structure map expansion and reciprocal recombination based on crossing-over. Mol. Gen. Genet. 123: 143-158

AINSWORTII, and BISBY, (1971) . DICTIONARY OF THE FUNGI, ed. COMMONWEALTH MYCOLOGICAL INSTITUTE ANGEL, T. , B. AUSTIN and D.C. CATUESIDE, (1970). Regulation of recombination at the his-3 locus in Neurospora crassa Aust. J. biol. Sci, 23: 1229-1240

AUERBACH, C. (1976). In MUTATION RESEARCH: Problems, Results and Perspectives. 504pp. London: Chapman & Hall BADMAN, R., (1972). Deoxyribonuclease-deficient mutants of Ustilago maydis with altered recombination frequencies. Genet. Res. 20: 213-229 BAKER, B.S., A.T.C. CARPENTER, M.S. ESPOSITO, R-E. ESPOSITO, and L. SANDLER (1976). The genetic control of meiosis. Ann. Rev. Genet. 10: 53-134. BALLANTYNE, G.H., and A.. CHOVNICK, (1971). Gene conversion in higher organisms: non-reciprocal recombination events at the rosy cistron in Drosophila melanogaster. Genet Res. 17: 139-149. BENBOW R.M., A.J. ZUCCARELLI, R.L. SINSIIEIMER. (1974b). An aggressive role for single strand breaks in bacteriophage 0 x 174 genetic recombination J. Mol. Biol. g2K629-651.- BOND, D.J. (1973). Polarity of gene conversion and postmeiotic segregation at the buff locus in Sordaria brevicollis. Genet. Res. 22: 279-289. BOON, T., and N.D. ZINDER (1969). Proc. nat. Acad. Sci. U.S.A. 64 573-577. ,9 m8cAarmsn7 /or genetic reco/nbinallior7 Bener-Exion one parenf and one reconthinome. CARLSON, P.S., (1971). A genetic analysis of the rudimentary locus of Drosophila melanogaster. Genet. Res. 17: 53-81

CARPENTER, A.T.C., (1975). Electron microscopy in Drosophila melanogaster females. II. The recombination nodule - a recombination associated structure at pachytene? Proc. Natn. Acad. Sci. U.S.A. 72: 3186-3189.

CASE, M.E., and N.H. GILES, (1964). Allelic recombination in Neurospora: tetrad analysis of a three-point cross within the pan-2 locus. Genetics 49: 529-540. 241

CATCHESIDE, D.G., (1966). second gene controlling allelic recombination in :.eurospora crassa Aust. J. Biol. Sci. 19: 1039-1046 CATCHESIDE, D.C., (1968). The control of genetic recombination in Neurospora crassa. In Replication and Recombination of Genetic Material, ed by W.J. Peackock and R.D. Brock, pp. 216-226. Proc. Aust. Acad. Sci. Canberra

CATC1IESIDE, D.G. , (1974). Fungal genetics Ann. Rev. Genet. 8: 279-300 CATCHESIDE, D.G., (1977). The genetics of recombination.LONOON, Arnold. CATCHESIDE, D.G. , and B. AUSTIN (1969). The control of allelic recombination at histidine loci in Neurospora crassa. Am. J. Bot. 56: 685-690. CHANDLEY, A.G., (1968). The effect of X-rays on female germ cells of Drosophila melanogaster. III. A comparison with heat-treatment on crossing-over in the X- chromosome. Mutation Res. 5: --;3-107 CHOVNICK, A., (1973). Gene conversion and transfer of genetic information within the inverted regions of inversion heterozygotes. Genetics 75: 123-131. DAVERN, C.I., (1971). Progress in Nucleic Acid Research and Molecular Biology, eds J.N. Davidson, W.E. Cohn, New York: Academic Press, EMERSON, M., (1954). Some physiological characteristics of ascospore activation in Neurospora crassa. Plant Physiology, 29, 418-428

EMERSON, S. (1966). Quantitative implications of the DNA-repair model of gene conversion. Genetics 33: 475-485. EMERSON, S. (1969). Linkage and recombination at the chromosome level. In Genetic Organization (ed. E.W. Caspari and A.W. Ravin). pp 267-360 . New York and London, Academic Press. EMERSON, S., and C.C.C. Yu-Sun, (1967). Gene conversion in the Pasadena strain of Ascobolus immersus. Genetics 55: 39-47. ESSER, K., and R. KUENEN, (1967). , Genetics of Fungi. Translated by E. Steiner, Berlin, Heidelberg, N.Y. Springer- Verlag. FIESER, L.F. , and M. FIESER (1956). Organic Chemistry, 3rd ed. Reinhold

FINCHAM, J.R.S., and R.R. DAY, (1911). Fungal Genetics. ed. Oxford. Blackwell 242

J.R.S. , and R. IIOLLtUAY. (1970). An explanation of fine structure map-expansion in terms of excision repair. Mol. Gen. Genet. 109: 309-322.

FINK, G.R., (1974). Properties of gene conversion of deletions in Saccharomyces cerevisiae. In: Mechanisms in Recombination (ed. R.F. Grell), pp. 287-293. New York and London, Plenum Press.

FINK, G.R. , and C.A. STYLES., (1974). Gene conversion of deletions in the his-4 region of yeast Genetics 77, 231-244. FOGEL, S., and R.K. MORTIMER, (1969). Informational transfer in meiotic gene conversion. Proc. nat. Acad. Sci.(Wash) 62: 96-103

FOGEL, S., and R.K. MORTIMER, (1970). Fidelity of gene conversion in yeast. Malec. Gen. Genet. 109 177-185 FOX, M.S., and M.K. ALLEN (1964). On the mechanism of deoxyribonucleate integration in pneumococcal transformation. Proc. Nate. Acad. Sci. U.S.A. 52: 412-419

FROST, L.C. (1961). Heterogeneity in recombination frequencies in Neurospora crassa. Genet. Res.2: 43-62 GHIKAS, A., (1973). 'Selection of thermal mutants in respect to deficiency to conidiation and/or to growth in the fungus Trichoderma viride'. M.Sc. Thesis, Tel-Aviv University, ISRAEL.

GIIIKAS, A., and B.C. LAMB, (1977). The detection in unordered octads, of 6+:2m and 2+:6m ratios with postmeiotic segregation and of aberrant 4:4s, and their use in corresponding-site interference studies. Genet. Res. Camb. 29, 267-278.

GIRARD, J., and J.L. ROSSIGNOL, (1974). The suppression of gene conversion and intragenic crossing-over in Ascobolus immersus: evidence for modifiers acting in the heterozygous state. Genetics 76: 221-243 GRELL, R.F., .(1967). Pairing at the chromosome level. J. Cell. Physiol. Suppl. 1.70: 119-146 GUTZ, H., (1971a). Gene conversion: remarks on the quantitative implications of hybrid DNA models. Genet. Res. 17, 45-52 GLITZ, H., (1971b). Site specific induction of gene conversion in Schizosaccharomyces_pombe. Genetics 69: 317-362.

HAMELIN, C., and G.H. COUSINEAU, '(1974). A comparative study of the production of RNA by round spore mutations, in Ascobolus immersus. Canadian J. Genetics Cytology 16: 39-48 243

HASTINGS, P.J., (1972). A pc)ssible source of primary nucleotide chain breaks in If,:combinant structures in eucaryotes. Genet. Res. 20: '253-256

HASTINGS, P.J., (1975). Some aspects of recombination in eucaryotic organisms. Ann. Rev. Genet. 9: 129-144 HASTINGS, P.J., and H.L.K. WHITEHOUSE (1964). A polaron model of genetic recombination by the formation of hybrid DNA. Nature London 201:1052-1054.

HOLLIDAY, R., (1964). A mechanism for gene conversion in fungi. Genet. Res. 5: 282-304

HOLLIDAY, R., (1967), Altered recombination frequencies in radiation sensitive strains of Ustila.go. Mutation Res. 4: 275-288.

HOLLIDAY, R., (1968). Genetic recombination in fungi, 157-174. In Replication and Recombination of Genetic Material, Ed. by W.J. Peackock and R.D. Brock. Austr. Acad. Sci. Canberra

HOLLIDAY, R., (1974). Molecular aspects of genetic exchange and gene conversion. Genetics 78: 273-287 HOLLIDAY, R., (1977). Recombination and meiosis. Phil. Trans. R. Soc. Lond. B. 277, 359-370

HOLLIDAY, R. , and H.L.k. WHITEHOUSE, (1970). The wrong way to think about gene conversion? Mol. Gen. Genet. 107: 85-93 HOTCHKISS, R.D., (1971). Toward a general theory of genetic recombination in DNA, pp 325-348. In: Advances in Genetics. Vol. 16. Academic Press, Inc., N.Y. and London. HOTCHKISS, R.D., (1974a). The evolution of recombination models. In: Mechanisms in Recombination, uu. R.F. Grell, Plenum Press (pp.439-443). HOTCHKISS, R.D., (1974b).. Molecular basis for genetic recombination. Genetics 78: 247-257. HURST, D.D., S. FOGEL, and R.K. MORTIMER (1972). Conversion associated recombination in yeast. Proc. nat. Acad. Sci. U.S.A. 69, 101-105 IMADA, M., M. INOUYE, M. EDA, and A.T. SUGITA, (1970). Frameshift mutations in the lysozyme gene of bacteriophage T4: demonstration of the. insertion of four bases and the preferential occurrence of base addition, in acridine mutagenesis. Journal of Molecular Biology, 54: 199-217 244

JESSOP, A.P.,,and D.G. CATCHcSIDE (1965). Interallelig recombination at the his-1 loch:, in Neurospora crassa and its genetic control. Herediiy 20, 237-256

KITANI, Y., L.S. OLIVE,(1967). Genetics of Sordaria fimicola. VI. Gene conversion at the g locus in mutant X wild- type crosses. Genetics 57, 767-782.

KITANI, Y., and L.S. OLIVE, (1969). The Genetics of Sordaria fimicola- VII Gene conversion at the g locus in interallelic crosses. Genetics 62: 33-66

KITANI, Y. , L.S. OLIVE, and A.S. EL-ANI. (1962). Genetics of Sordaria fimicola V. Aberrant segregation at the g locus. Am. J. Botany, 49: 697-706 KITANI, Y., and H.L.K. WHITEHOUSE, (1974). Aberrant ascus genotype:- from crosses involving mutants at the .g locus in Sordaria fimicola. Genet. Res. 24, 229-250

KRUZEWSKA, A., and W. GAJEIVSKI (1967). Recombination within the Y locus in Ascobolus immersus. Genet. Res. 9: 159-177

KUSNEV, V.V., (1974). Mechanisms of Genetic Recombination. Studies in Soviet Science. Translated by B. Haigh ed. Consultants Bureau New York & London LAMB, B.C., (1967). The differential maturation of asci and its relevance to recombination studies of Ne,Arospora, Sordaria and similar Ascomycetes. Genet. Res. 10: 1-12 LAMB, B.C., (1968). Gene conversion: temperature data from Sordaria fimicola on the correction of mis-paired bases. Nature 217, 353-354. LAMB, B.C., (1969a) Related and unrelated changes in conversion and recombination frequencies with temperature in Sordaria fimicola, and their relevance to hybrid-DNA models of recombination. Genetics 62,,67-78

LAMB, B.C., (1969b) Evidence from Sordaria that recombination and conversion frequencies are partly determined before meiosis, and for a general model of the control of recombination frequencies. Genetics 63, 807-820. LAMB, B.C., (1971). Some details and effects of the premeiotic controls of recombination frequencies in Neurospora crassa. Genetical Research 18, 255-264.

LAMB, B.C., (1972). 8:0, 0:8, 7:1 and 1:7 conversion ratios in octads from wild-type X mutant crosses of Ascobolus immersus. Heredity 29, 397. LAMB, B.C., (1975). Cryptic mutations: their predicted biochemical basis, frequencies and effects on gene conversion. Mol. Gen. Genet. 137, 305-314 245

LAMB, B.C., (1977a). The ey.l.ent and frequency of intimate pairing in meiosis. Clcr evidence from gene conversion for corresponding-site pairing. (Two discussion con- tributions). Phil. Trans. Royal Soc. B, 277, 199 and 357-358. LAMB, B.C., (1977b). The use of gene conversion to study synaptinemal complex structure and molecular details of chromatid pairing in meiosis. Mol. Gen. Genet. 157, 31-37 LAMB, B.C., and S. IIELIII (1978). The P, K, and 91 genetic controls of gene conversion in Ascobolus immersus. Heredity, in press LAMB, B.C., and M.R.T. WICKRAMARATNE (1973). Corresponding-site interference, synaptinemal complex structure, and 8+:0m and 7+:1m octads from wild-type X mutant crosses of Ascobolus immersus. Genet. Res. 22, 113-124 LAMB, B.C., and M.R.T. WICKRAMARATNE (1975). Induced changes in gene conversion spectra. Nature, London 258, 645-646.

LANUNER, L., (1970). Variation of recombination frequency in Neurospora crassa following temperature changes prior to and during meiosis and evidence for a pre-meiotic sensitive stage. Molec. Gen. Genet. 109, 219-252 LAWRENCE, C.W., (1965)- Influence of non-lethal doses of radiation on recombination in Chlamydomonas reinhardii. Nature, 206, 789-791 LAWRENCE, C.W., (1967). Influences of non-lethal doses of radiation on allelic recombination in Chalydomonas reinhardii. Genet. Res. 9, 123-127

LAWRENCE, C.W. , F. SHERMAN, M. JACKSON and R.A. GILMORE (1975). Mapping and gene conversion studies with the structural gene for iso-l-cytochrome c in yeast. Genetics 81, 615-629 LEBLON, G., (1972a). Mechanism of gene conversion in Ascobolus immersus. I. Existence of a correlation between the origin of mutants induced by different mutagens and their conversion spectrum. Molec. Gen. Gen. 115, 36-48.

LEBLON, G., (1972b). Mechanism of gene conversion in Ascobolus immersus. II. The relationship between the genetic alterations in-b or b 2mutants and their conversion spectrum. Molec.1 Gen. Genet. 116, 322-335 LEBLON, G., and J-L. ROSSIGNOL, (1973). Mechanism ofggene conversion in Ascobolus immersus. III. The interaction of heteroalleles in the conversion process. Molec. Gen. Genet. 122, 165-182 246 LINDAHL, T. , (1974)., An N-1-' ytosidase from Escherichia coli that releases fren uracil from .DNA containing deaminated cytosine residues. Proc. nat. Acad. Sci. (Wash). 71, 3649-3653 LINDEGREN, C.C., (1953). Gene conversion in Saccharomyces. J. Genet., 51, 625-637 LISSOUBA, P.J. MOUSSEAU, G. RIZET, J-L, ROSSIGNOL, (1902). Fine structure of genes in the ascomycete Ascobolus immersus, Adv. Genet, 11, 343-380. MAGUIRE, M.P. (1977). Homologous chromosome pairing. Phil. Trans. R. Soc. Lond. B277, 245-258 DIALLING, H.V., (1967). The mutagenicity of the acridine mustard (ICR17O) and the structurally related compounds in Neurospora. Mutation Res. 4, 265-274 McCARRON, M., W. GELBART and A. CHOVNICK (1974). Intracistronic mapping of electrophoretic sites in Drosophila melanogaster: Fidelity of information transfer by gene conversion. Genetics 76, 289-299, McNELLY-ANGLE, C.A., B.C. LAMB. and L.C. FROST, (1966). The effect of temperature on recombination frequency in Neurospora crassa. Genet. Res. 7, 169-183 MESELSON, M.S., and C.M. RADDING (1975). A general model for genetic recombination. Proc. nat. Acad. Sci. U.S.A. 72(1), 358-361. MESELSON, M., and J.J. WEIGLE, (1961). ChroMosome breakage accompanying genetic recombination in bacteriophage. Proc. nat. Acad. Sci. U.S.A. 47, 857-868 MITCHELL, M.B., (1955). Aberrant recombination of pyridoxine mutants of Neurospora. Proc. nat. Acad. Sci., U.S.A. 41, 215-220 MOORE, D., (1974). Dynamic unwinding of DNA helices: a mechanism for genetic recombination. J. Theor. Biol. 43 167-186. MOSES, M.J. (1968) Synaptinemal complex. Ann. Rev. Genet. 2, 363-412 MUNZ, P., and U. LEUPOLD (1970). Characterization of ICR170 induced mutations in Schizosaccharornyces pombe, Mutation Res. 9, 199-212 MURRAY, N.E., (1963). Polarized recombination and.fine structure within the me-2 gene of Neurospora crassa. Genetics 48,' 1163-1183 -- PARAG, Y., and G. PARAG (1975). Mutations affecting mitotic recombination frequency in haploids and diploids of the filamentous fungus Aspergillus nidulans. Molec. gen. Genet. 137, 109-123 247

PASZEWSKI, A. , (1970). Ger... conversion: observations on the DNA hybrid models. Cnet. Res. 15, 55-64 PASZEWSKI, A., and W. PRAZMO (1969). The bearing of mutant and cross specificity on the pattern of intragenic recombination. Genet. Res. Camb. 14, 33-43 PRAKASH, V., (1963). Perithecial production; use of non- absorbent cotton. Neurospora Newsl. 4:25

PUKKILA, P.J., (1977). Biochemical analysis of genetic recombination in Eukaryotes. Heredity 39(2), 193-217

RIZET, G., J-L, ROSSIGNOL, and C. LEFORT (1969). Sur la variete et la specificity des spectres d'anomalies de segregation chez Ascobolus immersus. Compt. Rend. Acad. Sci. Paris 269: 1427-1430.

ROSSIGNOL, J-L, (1969). Existence of homogenous categories of mutants exhibiting various conversion patterns in gene 75 of Ascobolus immersus. Genetics 63, 795-805. SCHROEDER, A.L., (1970). UV-sensitive mutants of Neurospora, crassa I. Genetic basis and effect on recombination. Molec. Gen. Genet. 107, 291-304

SCHROEDER, A.L., (1970). UV-sensitive mutants of Neurospora crassa II. Radiation studies. Molec. Gen. Genet. 107, 305-320, SIGAL, N., and B. ALBERTS (1972). Genetic recombination: the nature of a crossed strand-exchange between two homologous DNA molecules. J. Mol. Biol. 71, 789-793.

SIMCNEN, C. , and V. CONOLLY (1968). 'Changes in recombination following inbreeding in Schizophyllum. Genetics 58, 319-326

SIMCIiEN, G. , and J. STAMBERG (1969). Fine and coarse controls of genetic recombination Nature 222, 329-332 SIMONET, J.M. , . (1973) Mutations affecting meiosis in Podospora anserina. II. Effect of mei 2 mutants on recombination. Mol. Gen. Genet. 123, 263-281 SIMONET, J.M. , and D. ZICKLER (1972) Mutations affecting meiosis in Podospora anserina. I. Cytological studies. Chromosoma 37, 327-351 SMITH, B.R., (1966). Genetic controls of recombination. I. The recombination-2 gene of Neurospora crassa. Heredity 21, 481-498 SMITH, P.D., V.G. FINNERTY and A. CHOVNICK, (1970). Gene conversio in Drosphila: Non-reciprocal events at the maroon-like cistron. Nature 228, 441-444 248

SMYTH, D.R. , (1971). Effect of rec-3 on polarity of recombination in :he amination-1 locus of Neurospora crassa. Austr. J. Biol. Sci. 24, 97-106 SPATZ, B.C., and T.A. TRAUTNER (1970). One way to do experiments on gene conversion? Transfection with heteroduplex SPP1 DNA. Molec. Gen. Genet. 109, 84-106 STADLER, D.R., and A.M. TOME, (1962). Genetic factors influencing crossing-over frequency in Neurospora. Genetics 47, 839-846 STADLER, D.R., A.M. TOViE and J-L ROSSIGNOL, (1970). Intragenic recombination of ascospore color mutants in Ascobolus and its relationship to the segregation of outside markers. Genetics 66, 429-447 STAHL, F., (1969). One way to think about gene conversion. Genetics 61: Suppl. 1-13 STAMBERG, J., (1968). Two independent gene systems controlling recombination in Schizophyllum commune. Molec. Gen. Genet. 102, 221-228 STAMBERG, J. , (1969a) Genetic control of recombination in Schizophyllum commune: separation of the controlled and controlling loci. Heredity 24, 306-309 STAMBERG, J. , (1969b). Genetic control of recombination in Schizophyllum commune: the occurrence and significance of natural variation. Heredity 24, 361-368

STAMBERG, J., and Y. COLTIN (1971). Selectively combining B incompatibility factors of Schizophyllum commune. Molec. Gen. Genet. 113, 157-165 SOBELL, H.M., (1972). Molecular mechanism of genetic recombination. Proc. nat. Acad. Sci. U.S.A. 69, 2483- 2487. STADLER, D.R., (1973). The mechanism of intragenic recombination Ann. Rev. Genet, 7, 113-127 STADLER, D.R., and B. KARIYA (1969). Intragenic recombination of the mtr locus of Neurospora with segregation at an unselected site. Genetics 63, 291-316 STADLER, D.R., and A.M. TORE, (1971). Evidence for meiotic recombination in Ascobolus involving only one member of a tetrad. Genetics 68, 401-413

STRICKBERGER, M.1., (1976). Genetics, 2nd ed. MacMillan, New York; Collier MacMillan, London TAYLOR, J.H., (1965). Distribution of tritium-labelled DNA among chromosomes during meiosis. I. Spermatogenesis in the grasshopper. J. Cell. Biol. 25(2) 57-67 249

Van BLUMELLEN (1968). A world monograph of the genera Ascobolus and Saccobolus. (Persoonia, 1, suppi.); Eckblad (1968): 27

VOGEL, H.J., (1957). A convenient growth medium for Neurospora Microbial. Genet. Bull,. 13, 42-43

Von WETTSTEIN, D. (1977). The assembly of the synaptinemal complex. Phil. Trans. R. Soc. Lon B, 277, 235-243 WAGNER, R. , and M. MESELSON (1976). Repair tracks in mismatched DNA heteroduplexes. Proc. nat. Acad. Sci. (Wash). 73, 4135-4139 •WESTERGAARD, M.D., Von WETTSTEIN, (1972). The synaptinemal complex. Ann. Rev. Genet. 6, 71-110

WHITEHOUSE, fl.L.K. (1957). Mapping chromosome centromeres from tetratype frequencies. J. Genet. 55, 348-360 WHITEHOUSE, H.L.K. (1963). A theory of crossing-over by means of hybrid deoxyribonucleic acid. Nature 199, 1034-1040 WHITEHOUSE, H.L.K. (1967). Secondary crossing-over. Nature 215, 1352-1359 WHITEHOUSE, H.L.K. (1973). Towards an Understanding of the Mechanism of Heredity 3rd ed. London. Arnold WHITEHOUSE, H.L.K. (1974a). Advances in recombination research. In Symposium No.5: Genetic exchange. Genetics 78, 237-245

WHITEHOUSE, H.L.K. (1974b). Genetic analysis of recombination at the locus in Sordaria fimicola. Genet. Res. Camb. 24, 251-279 WHITEHOUSE, 11.L.K., and P.J. HASTINGS (1965). The analysis of genetic recombination on the polaron hybrid DNA model. Genet. Res. 6, 27-92 WICKRAMARATNE, M.R.T. , (1974). Gene conversion in Ascobolus immersus and the control of recombination in Nuerospora crassa. PhD Thesis, University of LONDON WICKRAMARATNE, M.R.T. (1976). A method for selecting linked mutants in Sordaria fimicola. Neurospora Newsl. No.23, 27. W C KRAMARATNE , M. R. T._ and B . C . LAMB (1978) . The estimation of conversion parameters and the control of conversion in Ascoholus immersus.- Molec. Gen. Genet. in press

WINKLER, cited by H.L.K. Whitehouse (1973) YU-SUN, C.C.C. (1964a) Nutritional studies of Ascobolus immersus. Amer. Jour. Bot. S1(3), 231-237. 250 YU-SUN, C.C.C. (1964b). Biochemical and morphological mutants of Ascobolus immevsus, Genetics 50, 987-998,

YU-SUN, C.C.C. (1966). Linkage groups in Ascobolus immersus. Genetica, 37, 569-580 YU-SUN, C.C.C. (1969) cited by Emerson, 1969 YU-SUN, C.C.C., M.R.T. WICKRAMARATNE and H.L.K. WHITEHOUSE (1977). Mutagen specificity in conversion pattern on Sordaria brevicollis. Genet. Res. Camb. 29, 65-81 ZICKLER, D., (1973). Fine structure of chromosome pairing in ten Ascomycetes: Meiotic and Pre-meiotic (mitotic) Synaptonemal Complexes. Chromosomes. (Berl.) 40, 401-416

(1968).Evidence on the stage of heat induced crossover effect in maize.Genetics 60,353-362.

MOORE,C.W. & F.SHERMAN (1977).Role of DNA sequences in genetic re- combination in the iso-I-cy- ochrome c gene of the yeast. II.Comparison of mutants altered at the same and nearby base pairs.Genetics

VirICKRAMAHATNE, R. T (1975) .The determinants of gene conversion properties in Ascobolus immersus.Iieredity, "2.2,428. 251

ACKNOWL;DGEMENTS

I am grateful to my supervisor Dr. B.C. Lamb for introducing me to this line of work, his generous help throughout this course and for corrections and valuable suggestions in the preparation of this thesis.

This work was mainly supported by.the Greek taxpayers' money, through a 3-year scholarship of the Greek National Scholarships Foundation, to which I address my thanks.

Thanks are also due to the Old Students of the Royal College of Science and to the University of London Convocation Trust Fund for grants provided, as well as to the Registry of the Imperial College for remitting my last year's College fees.

I also wish to thank Prof. Rutter, Head of the Botany Department, Mr. Whitworth, Chief Librarian of the Lyon Playfair Library, Mr. Ward, deputy Registrar, Mr. Adlington, senior welfare officer and the people at the Bursar's office, all at Imperial College, for the kindness and real help they gave to me during difficult times at the College.

I am thankful also to the staff of the Day Nursery of the Imperial College, which my son happily attended, and to the technical staff of the Botany Dept. for their excellent assistance.

Last, but not least, I wish to thank my husband, Dimitris Chikas, for sharing with me our family's responsibilities and for his compassion and the endless discussions, throughout these years, that helped me to pursue this work with joy.

Genet. Res., Camb. (1977), 29, pp. 267-278 267 Printed in Great Britain

The detection, in unordered octads, of 6+ : 2m and 2+ : 6m ratios with postmeiotic segregation, and of aberrant 4:4s, and their use in corresponding-site interference studies

BY AGLAIA GHLKAS AND B. C. LAMB Department of Botany, Imperial College, London SW7 2BB

(Received 31 January 1977)

SUMMARY Systems for the detection of postmeiotic segregation in unordered octads are described for 4:4(2), 6:2(2), 2:6(2), 5:3(3), 3:5(3) and 4:4(4) segregations, where the number in parentheses is the number of pairs of non-identical sister-spores. Linkage relations, centromere distances and phenotypic detectability of suitable combinations of markers for initial selective identification of such segregations, and subsequent confirma- tion tests, are discussed; correction formulae are given to allow for non- detection of certain classes in particular circumstances. Using one such system in Ascobolus immersus, we detected 4:4(2), 6:2(2) and 2:6(2) segregations at the w-78 site: 2: 6(2)s were previously unknown. Initial visual detection in this system was from octad phenotype ratios for w-78 (white ascospore) and gr-3 (granular ascospore pigmentation), with mating type and parafluorophenylalanine resistance as confirmatory markers after germination. Corresponding-site interference for wider ratio and — for the first time — narrower ratio unique classes was determined : it was moderately positive or moderately negative, depending on the cross and segregation class involved.

1. INTRODUCTION The study of aberrant segregation ratios in fungal tetrads and octads has been of great value in understanding recombination and gene conversion. In wild-type ( + ) x mutant (m) crosses, the ratios 6+ :2m, 2:6, 5:3 and 3:5 are well known from several fungi and have been termed 'narrower ratios' as opposed to the `wider ratios' 8:0, 0:8, 7:1 and 1 : '7 (Lamb, 1972). Lamb & Wickramaratne (1973) reported the occurrence of genuine wider ratio octads 8 + :Om and 7 + : lm in the Pasadena strains of Ascobolus immersus and pointed out that such segrega- tions required that both pairs of non-sister chromatids of a bivalent must be involved in hybrid DNA formation at exactly the same point, although narrower ratio segregations only require one pair of non-sister chromatids to pair at a point. As well as the wider ratio octads, they predicted the occurrence of other 'unique classes' (those requiring hybrid DNA formation in both pairs of chromatids at a particular site), namely 6:2(2), 2:6(2), 5:3(3), 3:5(3) and 4:4(4), where the number in parentheses is the number of pairs of non-identical sister-spores; that is, 16-2 268 A. GI LAS AND B. C. LAMB tetrad products showing post-meiotic segregation when the octad stage is reached. One aim of the present work was to devise methods for identifying as many unique classes with postmeiotic segregation as possible, and to identify aberrant 4:4(2) segregations, even in unordered octads where sister-spores are not identi- fiable just from their physical arrangement in an ascus as they are in ordered octads. Even with ordered octads, as in Neurospora crassa, Sordaria fimicola and Sordaria brevicollis, additional markers are advisable for checking that apparent postmeiotic segregation has not arisen by spindle overlap (partial or complete) at the third division in the ascus, or nuclear passing or spore slippage. The systems described here for unordered octads can also be used with ordered octads, es- pecially as a safeguard against spurious postmeiotic segregations. The second aim was to use segregation types such as 6:2(2) and 2:6(2) in Ascobolus immerses for studying corresponding-site interference, for comparison with the data on 8:0 and 7:1 classes obtained by Lamb & Wickramaratne (1973). They defined corresponding-site interference as : 'interference between the two pairs of non-sister chromatids of a bivalent in hybrid-DNA formation at exactly corresponding sites', and derived formulae for expected frequencies of all 'unique classes' in the absence of corresponding-site interference. In the present work, correction factors have been deduced to allow for the expected non-detection of certain proportions of postmeiotic segregation events in the special systems used here with unordered octads. Information on 4:4(2) frequencies was required for these corresponding-site interference calculations and also for the interpretation of conversion results (Wickramaratne & Lamb, 1975) for the w-10 and w-78 sites in Ascobolus immerses, where a factor closely linked to the two sites determines high or low conversion frequencies (HCF or LCF, respectively) at the two sites, which are themselves closely linked. The possibility of detecting 4:4(2) segregations in unordered octads was men- tioned by Emerson (1966), who used the method of two visual ascospore mutants plus other marker genes in Ascobolus (Emerson, 1969), but although he gave 4:4(2) data, the method itself was not described.

2. THE DETECTION SYSTEM IN UNORDERED OCTADS (i) Number and linkage relations of markers In theory, nearly all postmeiotic segregations should be detectable in suitably multiply-marked crosses, where various pairs of markers can show tetratype segregations that define the spore-pairs and thus enable postmeiotic segregation - giving non-identical sister-spores — to be recognized. Almost any kinds of marker gene could be used but in practice the detection of rare postmeiotic segregation classes can be greatly facilitated by using a system with visual markers, such as ones for ascospore colour production or distribution, or ascospore size or shape, which can be scored directly in the octad, without the need for germinating and testing each spore before even preliminary assessments can be made. If a suffi- Postmeiotic segregation in unordered octads 269 cient number of visual markers is not available in an organism, a compromise is to use perhaps two visual markers for an initial detection of octads possibly having the relevant postmeiotic segregations for one of the visual markers, then isolate that octad, germinate the spores and make confirmatory tests with other, non- visual markers. In the following discussion, it will always be aberrant segregation at marker 1 that is being studied, with markers 2, 3, 4 and so on as aids for its detection. It is essential that marker 1 should not regularly co-convert with other key markers being used. A minimum of two markers, 1 and 2, is required for the initial detec- tion of postmeiotic segregation but there should preferably be other markers present for confirmation. The situation where only two markers, 1 and 2, are required is where marker 2 has been shown in other crosses to have no conversion of its own, so that any unusual mutual segregation pattern for 1 and 2 can be attributed solely to aberrant ratios at 1. If, however, both markers show appre- ciable conversion (the term is used here to include allele-interaction events giving meiotic or postmeiotic segregation), at least one further marker, preferably more, is required to determine which site, 1 or 2, underwent aberrant segregation. Where groups of eight spores are scored after ascal dehiscence, as is usually the case with Ascobolus; the segregation patterns of the additional markers provide a useful check against false clusters; that is, octads formed of spores from more than one ascus and simulating the product of a single ascus. If two markers are used for the initial, visual detection, provided they do not co-convert, they can be non-synnemal (` non-syntenic ') or synnemal, unlinked or linked, though the linkage relations can affect the efficiency (in a calculable way) of detection of particular segregation patterns. For example, in detecting 4:4(2)s for site 1, all such segregations should be detected when 1 and 2 give parental ditype (PD) or non-parental ditype (NPD) segregations but only half are de- tectable in tetratype (T) octads. For 6: 2(2)s from a 3:1 event in both pairs of chromatids, such segregations are not detected in PD or NPD combinations of 1 and 2, but are in all T octads. The tetratype frequencies, which depend on centromere distances for non-synnemal loci and on recombination frequencies for synnemal loci, affect in a predictable way the detection efficiencies. Different suitable marker combinations could therefore be chosen for maximum detection of different segregations. When further markers are used, it is best if they show high tetratype frequencies with each other, to assist identification of spore-pairs, so some markers should be far from their centomeres. In the following account of individual segregations, genotypes will be given for repulsion-phase crosses, 1, + x +, 2, and when details are given for linked markers, only single crossovers will be considered, but the basic analysis can easily be applied to include coupling crosses and multiple crossovers. With linked markers, accurate assessments would require knowledge of any interference between crossovers and the possibility of conversion-associated crossovers should be considered. 270 A. GHIKAS AND B. C. LAMB

(ii) 4:4(2) The 'singly aberrant' 4:4, 4:4(2), is not one of the 'unique classes' mentioned earlier. It is detectable in octads PD and NPD for 1 and 2, giving in repulsion crosses genotype ratios of 1 + + , 3 1+ , 3 + 2, 1 12 and 3 + +, 1 I+, 1 +2, 3 12, respectively. In T octads, three arrangements are possible: 3 + +, 1 I+, 1 +2, 3 12; 2+ +, 2 1+, 2 +2, 2 12; 1 + +, 3 1+, 3 +2, 1 12. For non-synnemal markers, with independent assortment, these will occur in the ratio 1:2:.1 re- spectively, where the second element is not different in unordered octads from a normal tetratype. so those particular 4:4(2)s would go undetected. The other two kinds will be phenotypically distinct, so for non-synnemal markers, only half the 4:4(2)s in T octads will be detected, with the over-all ratio of 4:4(2) events detected to those occurring being (PD+NPD+2T)1(PD+NPD+T), with detection being from octad ratios 1 + + , 3 I+ , 3 +2, 1 12 and 3 + + , 1 I+ , 1 +2, 3 12. For synnemal markers, one has to consider various crossover and non-crossover possibilities between sites 1 and 2. In the repulsion cross, the absence of a crossover in this region gives 1 + + , 3 I+, 3 + 2, 1 12, as does a crossover between the two converting chromatids, while a crossover between the two non-converting chromatids gives 3 + +, 1 1+, 1 +2, 3 12: all these would be recognizable as a 4:4(2) segregation. A crossover between one converting and one non-converting chromatid would give 2 + + 2, 1+, 2 + 2, 2 12, which in unordered tetrads is the same as from an ordinary T octad, so that particular proportion of 4:4(2) segrega- tions would not be detected, but would need to be calculated and allowed for.

(iii) 6:2(2) These have two possible origins: either by a 3:1 ratio in each pair of non-sister chromatids, or by a 4 : 0 event in one pair of chromatids and 2:2(2) in the other pair. With the first of these origins, 6: 2(2)s are not usually detectable as such in unordered PD or NPD octads because they give genotype ratios identical to those from ordinary 6 : 2s (without postmeiotic segregation) - that is, 2 + + , 2 I+, 4 +2 and 4 + + 2 +2, 2 12, respectively. The 6:2(2)s of this origin are all detectable in T octads, giving 3 + +, 1 1+, 3 + 2, 1 12, unlike normal 6:2s which would give 4 + +, 2 +2, 2 12 or 2 + +, 2 1+, 4 +2. For 6:2(2) segrega- tions from two 3:1 events, the ratio of detected segregations to those occurring is therefore T l(PD +NPD +T) for non-synnemal genes. For synnemal genes, such 6:2(2) events with no crossover between sites 1 and 2 will be undetected as the 2 + +, 2 1+, 4 + 2 genotype could come from ordinary 6:2s. A crossover between the sites, in either pair of non-sister chromatids, gives 3 + +, 1 1+, 3 + 2, 1 12, which is recognizable as a 6 : 2(2). 6:2(2)s from a 4:0 event plus a 2:2(2) event should be detected in all PD and NPD octads, giving 3 + +, 1 1+, 3 + 2, 1 12. In T octads, there are three possible genotypes: 4 + +, 2 +2, 2 12; 3 + +, 1 / + , 3 +2,1 12 and 2 + +, 2 / + , 4 +2. For non-synnemal markers, these will occur in the ratio 1:2:1, with the first and Postmeiotic segregation in unordered octads 271 third elements not distinguishable from ordinary 6:2s. For 6:2(2)s from a 4:0 plus a 2:2(2) event, the ratio of detected segregations to those occurring is thus (PD+NPD+g)1(PD + NPD +T). For synnemal markers, 6:2(2)s of this origin with no crossovers will be detectable as 6:2(2)s, giving 3 + +, 1 1+, 3 + 2, 1 12, and of those with crossovers between 1 and 2, half (those with a crossover between the two chromatids involved in the conversion, or the two not involved) will be detected, giving 3 + + , 1 1+ , 1 +2, 3 12, and half (those with a crossover between an involved and non-involved chromatid) will be undetected, giving 2 ++,2/+,4 nor 4 /+,2 +2,212. The 0 : 2(2)s from the two origins (3 : 1 4- 3 : 1, or 4: 0 4- 2 : 2(2)) could not be distinguished from each other using genotype ratios for 1 and 2 alone, but they could be distinguished if outside markers were used for site 1. The same applies to 2: 6(2)s of two different origins.

(iv) 2:6(2) This class comes either from a 1: 3 ratio in each pair of chromatids or from a 0:4 in one pair and 2:2(2) in the other pair of chromatids. With the first of these origins, 2: 6(2)s are not detected as such in unordered PD or NPD octads because they give identical genotype ratios to those from ordinary 2: 6s, that is, 4 1+, 2 + 2, 2 12 and 2 + +, 2 1+, 4 12, respectively. 2: 6(2)s of this origin are all detectable in T octads, giving 1 + +, 3 1+ , 1 + 2, 3 12. For 2: 6(2)s from two 1:3 events, the ratio of detected segregations to those occurring is therefore T I(PD+NPD +T) for non-synnemal genes. For synnemal genes, such 2: 6(2)s with no crossover between sites 1 and 2 will go undetected as the 4 1+ , 2 + 2, 2 12 genotype could come from ordinary 2: 6s. A crossover between the sites gives 1 + +, 3 1+, 1 + 2, 3 12, which would be recognised as a 2: 6(2). 2:6(2)s from a 0:4 event plus a 2:2(2) event should be detected in all PD and NPD octads, giving 1 + +, 3 1+ , 1 + 2, 3 12. In T octads, there are three possible genotypes: 2 + +, 2 1+, 4 12; 1 + +, 3 1+,1+2, 3 12; 4 1+. 2 +2, 2 12. For non-synnemal markers, these will occur in the ratio 1:2:1, with the first and third elements not distinguishable from normal 2: 6s. For 2 : 6(2)s from a 0:4 plus a 2:2(2) event, the ratio of detected segregations to those occurring is thus (PD + NPD + iT)1(PD + NPD+T). With synnemal genes, 2: 6(2)s of this origin with no crossovers will be detectable, giving 1 + + , 3 1+ , 1 +2, 3 12, and of those with crossovers between 1 and 2, half will be detected, giving 1 + +, 3 1+, 1 +2, 3 12 and half will be undetected, giving 2 + +, 2 1+ , 4 12 or 4 1+ , 2 + 2, 2 12.

(v) 5 : 3(3) 5:3(3)s arise by a combination of a 3:1 event in one pair of chromatids with a 2:2(2) in the other pair. If the detection system only has markers 1 and 2, 5:3(3)s cannot usually be distinguished from ordinary 5:3(1)s in unordered tetrads as the same genotype ratios are produced: 3 + +, 1 1+, 2 +2, 212 and 272 A. GIIIICAS AND B. C. LAS 2 + +, 2 1+, 3 + 2, 1 12 (5:3(1)s also give other ratios distinct from those of 5: 3(3)s). There is one situation in which a system with just markers 1 and 2 could be used for initial detection of 5:3(3)s. This is where the two markers, if non- synnemal, are very close to their centromeres, showing almost entirely first division segregation at each locus and hence almost no tetratype segregations. PD and NPD for 1 and 2 give 2 + + , 2 1+ , 3 + 2, 1 12 and 3 + + , 1 1+ , 2 + 2, 2 12 with a 5:3(3) for 1, and T gives either of these, whereas with a 5:3(1), PD gives 1 + + , 3 1+ , 4 + 2, NPD gives 4 1+ , 1 +2, 3 12, and only the tetratypes for 5:3(1), giving 3 + +, 1 1+, 2 +2, 2 12 and 2 + +, 2 1+, 3 + 2, 112, could cause confusion with the 5:3(3)s in unordered octads. If markers 1 and 2 gave no Ts, then 5: 3(3)s would be the only class giving the above two genotypes. Such a system would be usable with 1 and 2 as visual markers for the initial detection, with other markers being used after isolation, germination and testing, to confirm that this was a 5:3(3), not a rare tetratype for 1 and 2 with a 5:3(1) segregation. It could also be used, with more labour, where 1 and 2 had a low frequency of T octads, with many octads of appropriate genotype ratio for 1 and 2 being isolated and tested to distinguish ordinary 5: 3(1)s from rare 5: 3(3)s. For linked markers, this system could also be used if the loci were close enough to give only a low frequency of crossovers and hence of tetratypes. If such special 'very low or no T' systems cannot be used, 5:3(3)s are not distinguishable from 5:3(1)s just by using markers 1 and 2 in unordered tetrads, though they could be in ordered tetrads. Further markers would have to be used, to give tetratypes defining the four spore-pairs, so the postmeiotic segregations of 1 in 5:3(3)s could be detected. Ideally, such a system would have three markers, including 1, scoreable visually for the initial detection, with further tests for other markers to confirm that it is marker 1, not 2 or 3, with the aberrant segregation. Without visual markers or alternative selection systems, the finding of such rare classes as 5:3(3) in unordered tetrads would be extremely laborious.

(vi) 3:5(3) These could be identified under the same circumstances as 5:3(3)s. The crucial genotypes are: 1 + + , 3 1+ , 2 + 2, 2 12 (from PD or T with 3:5(3)) and 2 + + , 2 1+ , 1 +2, 3 12 (from NPD or T with 3 :5(3)). (vii) 4:4(4) Irrespective of whether the octad type for 1 and 2 would have been PD, NPD or T, 4:4(4)s give genotype ratios of 2 + +, 2 1+, 2 + 2, 2 12, not usually dis- tinguishable from regular 4:4 segregation in T octads when unordered. The only reported 'doubly aberrant 4:4' is one detected visually from patterns seen in the ordered asci of Sordaria fimicola (Kitani & Whitehouse, 1974). As well as regular 4:4s from T octads, half the 4:4(2)s (see (ii) above) from T octads give the same ratio, 2 + +, 2 1+ , 2 +2, 2 12 as 4:4(4)s. If markers 1 and 2 were such as to give no or very few tetratypes (as discussed in v above) all or a large proportion' of octads with this genotype ratio would come from Postmeiotic segregation in unordered octads 273 originally PD or NPD 4:4()s, as there would be no or few Ts with 4:4 or 4:4(2) segregations to give that ratio. If 'very low or no T' systems cannot be used, further markers must be employed to provide identifiable Ts, so 4:4(4)s can then be recognized by their four pairs of non-identical sister-spores.

(viii) Wider ratios Wider ratio octads, with 8: 0, 0:8, 7:1 or 1:7 segregations, are detected in both unordered and ordered octads simply from the allele ratios observed.

3. MATERIALS AND METHODS General methods, materials and the Pasadena strains of Ascobolus immerses were usually as described by Lamb & Wickramaratne (1973) and Wickramaratne (1974). One difference was that in these experiments, 0.7 g/1 of methyl -p-hydroxy- benzoate was used as an in situ germination inhibitor in the spore collection agar. Crosses were incubated at 17.5 °C and length of heat-shock at 50 °C given to ascospores before germination was 30 min for white spores, 1-2 h for red, non- granular spores; granular spores were not heat-shocked. The markers used were: white(w)-78, a spontaneous white ascospore mutation (Emerson & Yu-Sun, 1967) which, depending on closely linked genetic controlling factors, gives high conversion frequencies (HCF) or low conversion frequencies (LCF); granular (gr)-3, a UV-induced mutation giving ascospore pigment in prominent granules outside the spore wall instead of the wild-type uniform distribution over the wall; pfr-1, an NMG-induced mutation, parafluorophenylalanine resistance, con- ferring the ability to grow in the presence of 100 mg/1 of this chemical. mating type, (mt), + or -. These markers are all unlinked to each other: the relations between our gr-3 and Emerson's (1969) gr-1, or Yu-Sun's (1966) gr-2, and between our pfr-1 and Stadler, Towe & Rossignol's (1970) fpr are unknown. From tetratype frequencies (method of Whitehouse, 1957), centromere distances have been calculated for w-78, gr-3 and pfr-1 as 30, 1.5 and 23.5 recombination units respectively, and Yu-Sun (1966) gave 2 as the value for mating-type. Dehisced octads were scored visually for segregation of w-78, corresponding to marker 1 in the previous section, and gr-3, acting as marker 2, with isolation, germination and subsequent testing for pfr and mt in octads of interest. w-78 is epistatic to gr-3, so w-78-, gr+ and w-78-, gr- have the same phenotype, white: these correspond to 1+ and 12. As can be seen from Table 1, this epistasis of 1 over 2 does not affect detection of the desired segregations for 1, although it would affect detection of certain aberrant ratios for 2 if these were being simul- taneously studied. With this kind of system, the marker whose segregation is being primarily studied may be epistatic but should not be hypostatic to any of the other markers being used. In control crosses of w+, gr+ x w+, gr—, the granular site showed extremely low 274 A. GRIT:CAS AND B. C. LAMB conversion, only 0.004 %. This is so much lower than the conversion frequency of w-78 (3-14 % in different crosses, LCF and HCF) that nearly all unusual geno- typic ratios for w-78, gr segregations are likely to be from aberrant ratios at the w-78 site: in the present experiments, all such unusual ratios were shown, with aid of pfr and mt where necessary, to be due to events at the w-78 site.

Table 1. Genotype and phenotype ratios and their origins for narrower-ratio unique and non-unique classes in a repulsion cross, 1+ x + 2 Phenotype ratios with 1 epistatic to 2+/2- Genotype ratios* ,--A-----, Origin:allele ratio at site 1 and + + +2 1+112 unordered octad type for 1, 2 + + 1+ +2 12 R G Wt (with 4:4 for site 2) 4 2 2 4 2 2 6:2(2), NPD, T; 6:2, NPD, T 4 1 3 4 1 3 5: 3, NPD 4 . . 4 4 4 4:4, NPD $3 1 3 1 3 3 2 6:2(2), PD, NPD, T 3 1 2 2 3 2 3 5:3(3), NPD, T; 5:3, T §3 1 1 3 3 1 4 4:4(2), NPD, T 3 1 . 4 3 . 5 3:5, NPD 2 2 4 2 4 2 6:2(2), PD, T; 6:2, PD, T 2 2 3 1 2 3 3 5:3(3), PD, T; 5:3, T 2 2 2 2 2 2 4 4:4(4), PD, NPD, T; 4:4(2), T; 4:4, T 2 2 1 3 2 1 5 3:5(3), NPD, T; 3:5, T 2 2 . 4 2 . 6 2:6(2), NPD, T; 2:6, NPD, T 1 3 4 . 1 4 3 5:3, PD §1 3 3 1 1 3 4 4:4(2), PD, T 1 3 2 2 1 2 5 3:5(3), PD, T; 3:5, T $1 3 1 3 1 1 6 2:6(2), PD, NPD, T 4 4 4 4 4:4, PD 4 3 1 3 5 3:5, PD 4 2 2 2 6 2:6(2), PD, T; 2:6, PD, T * With no epistasis, and if 1 and 2 give different phenotypes, these genotype ratios will also be the phenotype ratios. ¶ R, red, non-granular ascospores, w+, gr-F; G, red granular, w+, gr-; W, white ascospores, w-, gr+ and w-, gr-. t Narrower-ratio unique classes distinguishable as such. Unique classes are those re- quiring both pairs of non-sister chromatids of a bivalent to pair at the segregating site. § Aberrant 4:48, i.e. 4:4(2)s, are not a unique ratio class but their postmeiotic segregation can be detected here.

4. RESULTS Octads with phenotype ratios expected for 4:4(2), 6:2(2) and 2:6(2) segrega- tions at w-78 were detected visually; after germination and further tests, including crossing to check w-78 and gr genotypes, all proved genuine, with genotypes corresponding to visual phenotypes and gr, mt and pfr showing regular 4:4 segregation. Table 2 shows data for high and low conversion frequency crosses of w-78+, gr-3-, pfrR, mt+ x w-78-, gr-3±, pfrs, mt-. Postmeiotic segregation in unordered octads 275

Table 2. Segregation classes for w-78, and unordered tetrad types 4 : 4(2) 4:4(2) Cross 6+ :2w 5:3 3:5 2:6 3R:1G:4W 1R:3G:4W 6:2(2) 2:6(2) HCF* No. 2511 378 258 1016 28 26 7 4 % 7.04 1.06 0.72 2.56 0.08 0.07 0.02 0.01 LCF* No. 406 94 129 344 22 20 2 4 % 1.47 0.34 0.47 1.25 0.08 0.07 0.007 0.01 4:4 4:4 Cross 8:0 7:1 1:7 0:8 T PD +NPD Total HCF No. 12 4 5 6 21 040 10 373 35 668 % 0.03 0.01 0.01 0.02 58.99 29.08 LCF No. 2 1 1 8 17 818 8 750 27 596 % 0.007 0.004 0.004 0.03 64.55 31.71 * HCF, high conversion frequency crosses, total narrower ratio conversion frequency, 11.86%; LCF, low conversion frequency crosses, total narrower ratio conversion frequency 3.56%.

Table 3. Specimen 4:4(2), 6: 2(2) and 2 : 6(2) octads to show confirmed genotypes and spore pairs with postmeiotic segregation for w-78 No. of No. of spores w-78 gr pfr mt spores w-78 gr pfr mt 4:4(2), 3R: 1G: 4W 4:4(2), 1R: 3G: 4W 2 + 2 + + S + 2 — — R 2 — — R 1 + + S * 1 + + S 1 — + S 1 — + S -4-E } 1 + — R 1 — — R _1 1 — — R 1 + — R 6:2(2), 3R: 3G: 2W 2:6(2), 1R:1G:6W 2 + + R + 2 — + R + 2 + — R — 2 — — S - 1 + + S 1 + + R +l 1 — + S —1 1 — + R + 1 + — S + 1. 1 + _ S _ 1 — — S + 1 1 _ _ S 1 * The brackets show pairs of spores with postmeiotic segregation for w-78. In the 6:2(2), the segregations of gr, pfr and mt identify unambiguously which spores are sisters, especially as pfr and mt showed tetratype segregation. In the 4:4(2), 3R, 1G, 4W, one spore ( — , +, S, +, and +, —, R, —) of each pair of non-identical sister-spores is unique, clearly belonging to a pair of non-identical sister-spores, but the other spore in each case (+, +, S, +, and —, —, R, —) has two similar spores in the octad and it is not known which of the three belongs to the non-identical sister-spore pairs. 276 A. GHIKAS AND B. C. LAMB 4:4(2)s can give two phenotype ratios, 3R:1G:4W or 1R:3G:4W, depending on whether they arose in NPD or T, or PD or T octads (see Table 1) and, as expected, the two types were about equally frequent. Because of the high fre- quency of tetratypes, 5:3(3), 3:5(3) and 4:4(4) segregations could not be detected. Of ten 4:4(2) octads tested, all nine fully analysable octads proved genuine: in the tenth octad, not all spores germinated. All three 6:2(2) and 2:6(2) octads tested were fully analysable and genuine.

Table 4. Corresponding-site interference data Frequencies of events per pair of chromatids: a( = 4:0) = i6 :2m+8:0+ -17 1 ; b(= 0:4) = 12:6+0:8+1:7; c(= 3:1) = 15:3+17:1; d(= 1:3) = e(= 2:2(2)) = I-4:4(2). The above formulae have been adapted from the 'partly corrected' formulae of Lamb & Wickramaratne (1973). 6:2(2) formed = 2ae+c2 ; 2:6(2) formed = 2be + d2. PD+NP 6:2(2) detected here = 2ae ,pD+TD+1T+c2x xPD+N PD+NPD+T. PD+ NPD+1T 2:6(2) detected here = 2be x +d2x PD + NPD +T PD+NPD+T. Proportion of 4:4(2)s detected here PD+NDP PD +NPD T Cross ... HCF LCF HCF LCF

No. No. No. No. Class: 6:2(2) Class: 2:6(2) Observed* 7 0.02 2 0.007 4 0.01 4 0.01 Expected* 2.5 0.007 0.4 0.001 Fl 0.003 0.4 0.001

CSICCt 2.8 5.0 3.6 10 Class: 8:0 Class: 7:1 Observed 12 0.03 2 0.007 4 0.01 1 0.004 Expected 46 0.13 1.5 0.006 14 0.04 0.71 0.003

CSICC 0.3 1.3 0.3 1.4 Class: 1:7 Class: 0:8 Observed 5 0.01 1 0.004 6 0.02 8 0.03 Expected 3.8 0.01 0.9 0.003 7.5 0.02 1.2 0.004

CSICC F3 1.1 0.8 6.7 * These are directly comparable as the expected values (expected in the absence of corresponding-site interference) allow for the non-detection of a proportion of 6:2(2) and 2:6(2) events in accordance with the above formulae. t CSICC Corresponding-site interference coincidence coefficient, observed/expected values.

Typical tested octads, giving segregations for w-78, gr, pfr and mt, are given in Table 3 to show how postmeiotic segregation could be recognized in 4:4(2), 6:2(2) and 2:6(2) octads. Formulae (Lamb & Wickramaratne, 1973), modifying factors for the detection Postmeiotic segregation in unordered octads 277 system used here (see section 2), and observed and expected values for unique narrower ratio classes are given in Table 4 for corresponding-site interference estimation: wider-ratio classes are also considered. Amongst the 6:2(2) and 2:6(2) classes in LCF and HCF crosses, there were in each of the four cases more octads of these types observed than would have been expected in the absence of corre- sponding-site interference. The excess is about 3- to 10-fold, though the small numbers make an accurate assessment difficult. Observed and expected values agree fairly well for: 8:0s, LCF; 7:1, LCF; 1:7 HCF and LCF; and 0:8 HCF. The interesting wider ratio values are the HCF 8:0s and the HCF 7: ls, where fewer were observed than expected (x2 = 19.9, P = < 1 % and x2 = 4.9, P = 2-5 %, respectively). Lamb & Wickramaratne (1973) found that nearly all 8: 0s and 7: is in Ascobolus crosses (many of which used w-78, as here) were genuine but that 0: 8s often arose by mutation, so the 0:8 data here may not be entirely reliable as the 0: 8s were not isolated for checking.

5. DISCUSSION These results show that unordered octads can be used to detect events giving rise to certain postmeiotic segregation classes such as 6:2(2), 2:6(2) and 4:4(2), even when they are very rare, providing suitable marker combinations, especially of visual markers, are available. Genetic analysis for rare unique narrower ratio classes can therefore be carried out in fungi with unordered octads. Even where there was uncertainty as to which of three spores belonged to a pair of non-identical sister spores, the postmeiotic segregation was still recognizable (examples in Table 3). Leblon & Rossignol (1973), with the European strains of Ascobolus immersus, detected 4:4(2)s in unordered octads using two-point coupling crosses between heteroallelic intragenic suppressors at the b2 ascospore colour locus. A disadvantage of such closely linked and co-converting sites is that simultaneous 4:4(2)s, from hybrid DNA formation at both sites without correction, are not detected. Of all possible segregation patterns for a pair of alleles, the non-unique narrower ratio classes 4:4, 6:2, 2:6, 5:3, 3:5 and 4:4(2) are well known from a number of organisms. The work of Kitani (e.g. Kitani & Whitehouse, 1974) with Sordaria fimicola, Lamb (1972), Lamb & Wickramaratne (1973) and Leblon (1972) with Ascobolus immersus, and of Yu-Sun, Wickramaratne & Whitehouse (1977) with Sordaria brevicollis, has firmly established the existence of all wider ratios, 8:0, 0:8, 7:1 and 1:7. Of the unique narrower ratios, one 4:4(4) was reported by Kitani & Whitehouse (1974); they also found one 6:2(2) and we (Table 2, present work) found nine of these; we also found eight 2:6(2)s, the first such report. Of the remaining possible classes predicted by Lamb & Wickramaratne (1973) only the 5:3(3) and 3:5(3)s await discovery. The present corresponding-site interference data, although puzzling, are in keeping with the earlier findings for w-78 and w-10 (Lamb & Wickramaratne, 1973) that coincidence coefficients, and even whether corresponding-site inter- ference was positive or negative, differed for different segregation classes and between HCF and LCF crosses. Thus the previous results gave corresponding-site 278 A. GHIKAS AND B. C. LAMB interference coincidence coefficients (CSICCs) of 0.5 for 8:0, HCF, but 2.6 for 8:0, LCF, and of 1.5 for 7:1, HCF, but 5.1 for 7:1, LCF. The present values of 0.3, 1.3, 0.3 and 1.4 respectively give fair agreement with the previous ones for 8: 0s, taking into account sample sizes, but the 7:1 CSICCs were much lower than before, although the HCF/LCF difference was proportionately maintained. The Sordaria brevicollis results of Yu-Sun et al. (1977) are equally puzzling, with CSICCs often varying widely between different wider ratio classes for the same mutant, for the same classes between different mutants at the same locus (e.g. 3.9-17.5 for 8:0s at grey-5) and between different loci (e.g. 3.3-28.8 for 1:7s at grey-3 and grey-5). In the S. brevicollis data, most CSICCs were between 1.1 and 28.8, so corresponding-site interference varied from almost no interference to quite strong negative interference. By contrast, this interference in the Pasadena strains of Ascobolus immersus varied from moderately negative to moderately positive (CSICCs 5.1-0.3 for the larger samples, both of these being significantly different from no interference at P = 1 %). The present Ascobolus data thus confirm the previous findings that corre- sponding-site interference can be negative or positive. These first assessments of corresponding-site interference for the 6:2(2) and 2:6(2) segregations gave moderate but consistent negative interference in both HCF and LCF crosses.

REFERENCES EMERSON, S. (1966). Quantitative implications of the DNA-repair model of gene conversion. Genetics 53, 475-485. EMERSON, S. (1969). Linkage and recombination at the chromosomal level. In Genetic Organi- sation, vol. I (ed. by E. W. Caspari and A. W. Ravin), p. 289. New York: Academic Press. EMERSON, S. & Yu-Sun, C. C. C. (1967). Gene conversion in the Pasadena strain of Ascobolus immersus. Genetics 55, 39-47. KITANI, Y. & WHITEHOUSE, H. L. K. (1974). Aberrant ascus genotypes from crosses involving mutants at the g locus in Sordaria fimicola. Genetical Research 24, 229-250. LAMB, B. C. (1972). 8: 0, 0:8, 7:1 and 1:7 conversion ratios in octads from wild-type x mutant crosses of Ascobolus immersus. Heredity 29, 397. LAMB, B. C. & WICKRAMARATNE, M. R. T. (1973). Corresponding-site interference, synaptine- mal complex structure, and 8+ :Om and 7+ :lm octads from wild-type x mutant crosses of Ascobolus immersus. Genetical Research 22, 113-124. LEBLON, G. (1972). Mechanism of gene conversion in Ascobolus immersus. I. Existence of a correlation between the origin of mutants induced by different mutagens and their conver- sion spectrum. Molecular and General Genetics 115, 36-48. LEBLON, G. & ROSSIGNOL, J.-L. (1973). Mechanism of gene conversion in Ascobolus immersus. III. The interaction of heteroalleles in the conversion process. Molecular and General Genetics 122, 165-182. STADLER, D. R., Tows, A. M. & RossieNoL, J.-L. (1970). Intragenic recombination of asco spore color mutants in Ascobolus and its relationship to the segregation of outside marker Genetics 66, 429-447. WHITEHOUSE, H. L. K. (1957). Mapping chromosome centromeres from tetratype frequencies. Journal of Genetics 55, 348-360. WIORRAMARATNE, M. R. T. (1974). Gene conversion in Ascobolus immersus and the control of recombination in Neurospora crassa. Ph.D. thesis, Imperial College, London. WICERAMARA'TNE, M. R. T. & LAMB, B. C. (1975). The determinants of gene conversion pro- perties in Ascobolus immersus. Heredity 35, 428. Yu-SUN, C. C. C. (1966). Linkage groups in Ascobolus immersus. Genetica 37, 569-580. Yu-Sun, C. C. C., WICERAMARATNE, M. R. T. & WHITEHOUSE, H. L. K. (1977). Mutagen specificity in conversion pattern in Sordaria brevicollis. Genetical Research 29, 65-81. Heredity, 1977, :22, 274.

THE DETECTION OF 6+:2m AND 2+ :6m ASCI WITH POSTMEIOTIC SEGREGATION, AND OF ABERRANT 4:4s, IN UNORDERED OCTADS OF ASCOBOLUS

A. GHIKAS and B. C. LAMB Department of Botany, Imperial College, London

By using (-anthill:din:1s of visual markers, it is possible to identify visually cot taint tt pos of posuneiotie segregation which arc usually undetertahle in unodlered ase j• A whit' (re) ascospore mutant and a granular (gr) pigmentation marker in the Pasadena straili, of Atotbehtt immeoirts base heCII used 11c1C. 12elevant segregations, identitied \isu.illy Lunt phenotype ratios in ;triads, sere confirmed by germinating spores, backerossing genotypes. and testing for other segregating loci (mating type and paratluorophenylal.mine resistance). N1'ilil-type spores are red and lion-granular. and to is epistatic to gr: ti.erc are three spore phenotypes: rid nun-granular (RN), FM granular (1:(;) and white (11' , . At the white locu s (ti-716, aberrant •1..1 segregations usually gave oetad plienott tie ratios or 3RN, it(;, -1W or 112N, 312(1, Of much more interest \tote the 312N..112(1, 211' and 11{N, I12G. 611' ()clads, coming respectively from (1-+ :2:e and 2+ :Ott' [antis t,ith postmeimic segregation for te. Lamb and NVickritinaratue (Gencl. 113, 1fi,.3) medic-Led the occurrence or such 6:2 and 2:6 classes arising from hybrid 1)NA. furmaiton at corresponding sites in both pairs of nitit-sister chromatids of a bivalent. NVith just two visual markers in a cross, postmeiotic segregations of the three kinds (lest-tilted trill some- times go undetected out u: were derived to allow for these undetected ones, su the present data Can lc used in corresponding-sin: intnierence studies.