Top Banner
Copyright 0 1989 by the Genetics Society of America Recombination between Small X Chromosome Duplications and the X Chromosome in Caenorhabditis elegans Robert K. Herman and Claire K. Kari Department of Genetics and Cell Biology, University of Minnesota, St. Paul, Minnesota 55108 Manuscript received August 26, 1988 Accepted for publication December 19, 1988 ABSTRACT Twelve new X chromosome duplications were identified and characterized. Eight are translocated to autosomal sites near four different telomeres, and four are free. Ten include unc-l(+), which in wild type is near the left end of the X chromosome, and two of these, mnDp72(X;ZV) and mnDp73(Xf), extend rightward past dpy-3. Both mnDp72 and mnDp73 recombined with the one X chromosome in males in the unc-1-dpy-3 interval at a frequency 15- to 30-fold higher than was observed for X-X recombination in hermaphrodites in the same interval. Recombinant duplications and recombinant X chromosomes were both recovered. Recombination with the X chromosome in the unc-1-dpy-3 interval was also detected for five other unc-l(+) duplications, even though their right breakpoints lie within the interval. Inhermaphrodites, mnDp72 and mnDp73 promoted meiotic X nondisjunction and recombined with an X chromosome in the unc-1-dpy-3 interval at frequencies comparable to that found for X-X recombination; mnDp72(X;ZV) also promoted trisomy for chromosome ZV. A mutation in him-8 ZV was identified that severely reduced recombination between the two X chromosomes in hermaphrodites and between mnDp73 and the X chromosome in males. Recombination between the X chromosome and duplications of either the right end of the X or a region near but not including the left end was rare. We suggest that the X chromosome has one or more elements near its left end that promote meiotic chromosome pairing. N UMEROUS translocated and free duplications of chromosomal segmentshave been generated and characterized in the nematode Caenorhabditis ele- guns (HERMAN, ALBERTSON and BRENNER 1976; HER- MAN, MADL and KARI 1979; HODGKIN 1980; HERMAN, KARI and HARTMAN 1982; ROSE, BAILLIE and CUR- RAN 1984; ANDERSON and BRENNER 1984; ROSEN- BLUTH, CUDDEFORD and BAILLIE 1985; DELONG, CAS- SON and MEYER 1987; ROGALSKI and RIDDLE 1988). Such duplications have been used to vary the dosage of genes (JOHNSON et al. 198 1 ; GREENWALD, STERN- BERG and HORVITZ 1983; MEYER and CASSON 1986; DONAHUE, QUARANTILLO and WOOD 1987), to bal- ance recessive lethal and sterile mutations (MENEELY and HERMAN 1979,1981; HOWELL et al. 1987), to facilitate manipulation of X-linked markers (HERMAN, ALBERTSON and BRENNER 1976), to vary the X chro- mosome-to-autosome ratio (MADL and HERMAN 1979; MENEELY and WOOD 1984, 1987; WOOD et al. 1985; MENEELY and NORDSTROM 1988), to mark particular chromosomes cytologically (ALBERTSON 1984), and to generate genetic mosaics (for review, see HERMAN 1989). Some of the X chromosome duplications that have been characterized are translocated to autosomes. The other X duplications and all autosomal duplica- tions that have beenstudied are free chromosome fragments, as shown by genetic and cytological crite- ria. The relatively high frequency at which free du- Genetics 121: 723-737 (April, 1989) plications have been recovered is probably at least in part attributable to the holokinetic (also referred to as diffuse centromeric) nature of C. elegans chromo- somes (ALBERTSON and THOMSON 1982). Nearly all duplications that have been studied previously, whetherfree or translocated,appear to recombine very infrequently with the homologous segments of the normal chromosomes. Those rare duplications that have been shown to recombine with their normal homologues are large; sDpl(Z;f), which recombines with chromosome I over the dpy-5-unc-54 interval at moderate frequency, includes more than half of the genes on chromosome Z (ROSE, BAILLIE and CURRAN 1984), and mDpl, which recombines over the unc-17- dpy-I3 interval of chromosome ZV at a frequency of 20-fold less than that found for chromosome-chro- mosome recombination, includes approximately half the genes on chromosome ZV (ROGALSKI and RIDDLE 1988). Examples of smaller duplications, for which recombination has been shown to be very rare, are duplications of the right end of the X chromosome. The rare apparently recombinant duplications of that region that were identified were at least in some cases, perhaps most, the result of mutation rather than re- combination (HERMAN 1984). We do not understand why recombination between a small duplication and its homologous region of chromosome is generally rare. From a practical point of view, the low recombination frequency is a conven-
15

Chromosome in Caenorhabditis elegans

Jan 12, 2017

Download

Documents

tranhanh
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Chromosome in Caenorhabditis elegans

Copyright 0 1989 by the Genetics Society of America

Recombination between Small X Chromosome Duplications and the X Chromosome in Caenorhabditis elegans

Robert K. Herman and Claire K. Kari

Department of Genetics and Cell Biology, University of Minnesota, St. Paul, Minnesota 55108 Manuscript received August 26, 1988

Accepted for publication December 19, 1988

ABSTRACT Twelve new X chromosome duplications were identified and characterized. Eight are translocated

to autosomal sites near four different telomeres, and four are free. Ten include unc-l(+), which in wild type is near the left end of the X chromosome, and two of these, mnDp72(X;ZV) and m n D p 7 3 ( X f ) , extend rightward past dpy-3. Both mnDp72 and mnDp73 recombined with the one X chromosome in males in the unc-1-dpy-3 interval at a frequency 15- to 30-fold higher than was observed for X-X recombination in hermaphrodites in the same interval. Recombinant duplications and recombinant X chromosomes were both recovered. Recombination with the X chromosome in the unc-1-dpy-3 interval was also detected for five other unc-l(+) duplications, even though their right breakpoints lie within the interval. In hermaphrodites, mnDp72 and mnDp73 promoted meiotic X nondisjunction and recombined with an X chromosome in the unc-1-dpy-3 interval at frequencies comparable to that found for X-X recombination; mnDp72(X;ZV) also promoted trisomy for chromosome ZV. A mutation in him-8 ZV was identified that severely reduced recombination between the two X chromosomes in hermaphrodites and between mnDp73 and the X chromosome in males. Recombination between the X chromosome and duplications of either the right end of the X or a region near but not including the left end was rare. We suggest that the X chromosome has one or more elements near its left end that promote meiotic chromosome pairing.

N UMEROUS translocated and free duplications of chromosomal segments have been generated

and characterized in the nematode Caenorhabditis ele- guns (HERMAN, ALBERTSON and BRENNER 1976; HER- MAN, MADL and KARI 1979; HODGKIN 1980; HERMAN, KARI and HARTMAN 1982; ROSE, BAILLIE and CUR- RAN 1984; ANDERSON and BRENNER 1984; ROSEN- BLUTH, CUDDEFORD and BAILLIE 1985; DELONG, CAS- SON and MEYER 1987; ROGALSKI and RIDDLE 1988). Such duplications have been used to vary the dosage of genes (JOHNSON et al. 198 1 ; GREENWALD, STERN- BERG and HORVITZ 1983; MEYER and CASSON 1986; DONAHUE, QUARANTILLO and WOOD 1987), to bal- ance recessive lethal and sterile mutations (MENEELY and HERMAN 1979, 1981; HOWELL et al. 1987), to facilitate manipulation of X-linked markers (HERMAN, ALBERTSON and BRENNER 1976), to vary the X chro- mosome-to-autosome ratio (MADL and HERMAN 1979; MENEELY and WOOD 1984, 1987; WOOD et al. 1985; MENEELY and NORDSTROM 1988), to mark particular chromosomes cytologically (ALBERTSON 1984), and to generate genetic mosaics (for review, see HERMAN 1989).

Some of the X chromosome duplications that have been characterized are translocated to autosomes. The other X duplications and all autosomal duplica- tions that have been studied are free chromosome fragments, as shown by genetic and cytological crite- ria. The relatively high frequency at which free du-

Genetics 121: 723-737 (April, 1989)

plications have been recovered is probably at least in part attributable to the holokinetic (also referred to as diffuse centromeric) nature of C. elegans chromo- somes (ALBERTSON and THOMSON 1982). Nearly all duplications that have been studied previously, whether free or translocated, appear to recombine very infrequently with the homologous segments of the normal chromosomes. Those rare duplications that have been shown to recombine with their normal homologues are large; sDpl(Z;f) , which recombines with chromosome I over the dpy-5-unc-54 interval at moderate frequency, includes more than half of the genes on chromosome Z (ROSE, BAILLIE and CURRAN 1984), and m D p l , which recombines over the unc-17- dpy-I3 interval of chromosome ZV at a frequency of 20-fold less than that found for chromosome-chro- mosome recombination, includes approximately half the genes on chromosome ZV (ROGALSKI and RIDDLE 1988). Examples of smaller duplications, for which recombination has been shown to be very rare, are duplications of the right end of the X chromosome. The rare apparently recombinant duplications of that region that were identified were at least in some cases, perhaps most, the result of mutation rather than re- combination (HERMAN 1984).

We do not understand why recombination between a small duplication and its homologous region of chromosome is generally rare. From a practical point of view, the low recombination frequency is a conven-

Page 2: Chromosome in Caenorhabditis elegans

724 R. K. Herman and C;. I(. Kari

ient property for most genetic applications of dupli- cations. On the other hand, it is conceivable that a better understanding of the factors affecting duplica- tion-chromosome recombination could help lead to a method for promoting gene disruption by homolo- gous recombination of injected DNA fragments-a powerful tool in yeast genetics (ROTHSTEIN 1983) not yet available in C. eleguns although methods for pro- moting integrative transformation have been worked out (FIRE 1985).

In this paper we show that certain small duplica- tions, both translocated and free, of the left end. of the X chromosome recombine at surprisingly high frequencies with the homologous region of the normal X chromosome in both males (which carry a single X chromosome in addition to the duplication) and her- maphrodites (which have two X chromosomes). This is not a general property of X chromosome duplica- tions because the frequencies of recombination be- tween other small X duplications and the X chromo- some appear to be very low in both males and her- maphrodites. We suggest that homologous pairing of the X chromosome of C. eleguns at meiosis is promoted by one or more sites near the left end of the X chromosome.

MATERIALS AND METHODS

Strains, genes and general procedures: C . elegans var. Bristol strain N 2 was the wild-type parent for all strains used in this work. Media, culture and mating techniques were as described by BRENNER (1974) and HERMAN (1978). The procedure used for staining chromosomes with DAPI (diam- idinophenolindole) was identical to the previously-described procedure for staining with Hoechst 33258 (HERMAN, MADL and KARI 1979) except for the exchange of dyes. General gene names of genes used were: che, chemotaxis defective; dpy , dumpy: egl, egg-laying defective: j l u , abnormal gut autofluorescence; him, high incidence male self progeny; lin, cell lineage defective; Lon, long; mab, male abnormal; osm, defective in avoidance of high osmolarity; sup, suppressor: unc, uncoordinated. Names of genes and alleles used were: linkage group (LC) I : egl-30(n686), dpy-5(ehl), unc-54- (e190). LGZIl: d p y - l ( e l ) , unc-Y?(e1500n224) = unc-Y3(0) in the text, unc-Y3(e1500), unc-?6(e251). LGZV: him-8(e1489 or mn253), dpy-4(e1166). LGV: unc-60(e667), dpy-1 l(e224). IGX: mab-7(e1599). che-2(e1033), unc-l(e538), dpy-?(e27 or e182), lin-32(~282), unc-2(e55), osm-5(p81?), unc-20(e112), unc-78(e1217), lon-Z(e678), dpy-8(e130), jlu-2(e1003), unc- h(e78), dpy-7(e88), unc-Y(elOl), unc-84(e1410), unc-?(e151), unc-7(e5), osm-l(p808), sup-l0(n 183). A genetic map show- ing the relative positions of these genes is given as Figure 1 . For ;t more complete map and references to origins o f mutations, see WOOD (1 988). The position of sup-10 to the right o f osm-1 (A. VILLENEUVE and B. MEYER, personal cornmunication) is a correction on our earlier placement (MENEELY and HERMAN 198 1). The position of che-2 about 0.4 map uni t left of unc-1 is based on the following data: three of 340 Unc-l progeny of che-2 unc-lldpy-3 her- maphrodites were non-Che [as determined by staining with fluorescein isothiocyanate (FITC): see below], and none of the three segregated Dpy (Unc) animals. The position of mab-7 slightly left of unc-1 (and not resolved from the position of che-2) is based on the following data: among 424

male progeny generated by a cross between mab-7 dpy-3/ unc-1 hermaphrodites and N 2 males, one was Dpy non-Unc non-Mab and one was Mab Unc non-Dpy. The position of flu-2 (BABU 1974) between dpy-8 and unc-6 is based on the following data: among the progeny of dpy-8 unc-h/’u-2 hermaphrodites, 26 of 50 Dpy non-Unc and 16 of 50 Unc non-Dpy animals were Flu. The derivation of the du- plication mnDp3 was described previously (HERMAN, MADL and KARI 1979). Genetic nomenclature follows the guide- lines described by HoRvrrz et al. (1979), with the fol- lowing addition: the genotype of a duplication is some- times stated explicitly in brackets along with the duplication name, as, for example, mnDp62(X,f)[unc-?(+) osm-l(+) sup- lO(nl83)l. The che-2, osm-1 and osm-5 mutations abolish the uptake of FITC by amphid and phasmid sensory neu- rons (HEDGECOCK et al. 1985; PERKINS et al. 1986), and i t was this trait that we scored for these genes. The FITC stain- ing protocol of HEDGECOCK et al. ( 1 985) was used. Some of our counts of Che-2 animals may be underestimates, by as much as about 20-30%, because of the suicidal tendency of these animals, particularly the males (HODGKIN 1983), to crawl off the agar plate. The mab-7 and lin-32 markers were scored, in males, by looking at the rays o f the copu- latory bursa by Nomarski microscopy. I n mab-7 animals, the bllrsdl fan is reduced and the bursal rays are swollen (HODG- KIN 1983); the lin-32 bursa is essentially devoid of rays (E. HEDGECOCK, personal communication).

Induction, identification and characterization of unc- I ( + ) duplications: The general strategies for generating and characterizingxduplications have been described (HER- MAN, ALBERTSON and BRENNER 1976; HERMAN, MADI. and KARI 1979). N2 males were exposed to y-rays and mated with unc-1 hermaphrodites. The screen was greatly aided by the use of the recessive unc-1 allele e538 (PARK and HORVITZ 1986) in preference to the previous reference allele eY4, which is semidominant to unc-I(+). ?-Rays were supplied by “’Ck in a Shepherd irradiator (model 143-45). Doses of 4200 roentgens (r) were used at a dose rate of 530 r/min. Progeny were screened for exceptional wild-type males, which were mated with unc-1 hermaphrodites. When roughly half the male progeny of the latter cross were wild type, it w a s concluded that they carried a copy of unc-l(+) that was not X-linked. The frequency of recovery of unc- I ( + ) duplications was about per male progeny. We learned from our first recovered unc-l(+) duplication, mnDp63, that many of the wild-type hermaphrodite progeny of a cross of unc-l/0:mnDp63 males and unc-I hermaphro- dites did not carry mnDp63 but carried a recombinant unc- I ( + ) X chromosome (see below). Therefore, i n order to guard against the possibility of losing duplication stocks, we maintained each duplication by continual backcrossing of wild-type males to unc-1 hermaphrodites. Each duplication was outcrossed dozens of times during the course of these experiments.

T o test whether or not a given duplication carried the X- linked gene g, wild-type males of genotype unc-l/O;Dp were crossed with g hermaphrodites. If roughly half (as opposed to virtually none) of the male progeny were non-G, it was concluded that the duplication carried g(+).

The possible autosomal map location of each duplication was sought i n the following way: unc-I/O:Dp males were mated with m;unc-1 hermaphrodites, where m is an autoso- mal marker. Wild-type hermaphrodite progeny were picked. A potential difficulty at this stage was in identifying those wild-type hermaphrodites that carried the duplication (gen- otype m/I)p;unc-l/unc-I or m/+:Dp/+:unc-l/unc-1) rather than a recombinant X chromoson~e (genotype m/+;unc-l/ +). Different criteria for the presence o f different duplica- tions were used. FOI- mnnp63, mnDp70 and mnDp72, the

Page 3: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C. elegans 725

X CHROMOSOME :

fin 5 MAP UNITS

. mnDp3(X;fl . mnDp62 (X;f)

AUTOSOMES : h 5 MAP UNITS

FIGURE 1 .-Linkage map showing genes used in this work. The extents of the X chromosome duplications used are indicated below the X map. Note that the map scale for the X chromosome is different from the scale for the autosomes.

relatively high incidence of male self progeny (see RESULTS) was sometimes used. Hermaphrodites carrying the duplica- tions mnDp68, mnDp71, mnDp72 and mnDp7? generally segregated more than 40% Unc-1 self-progeny rather than the 25% segregated by unc-llunc-l(+) hermaphrodites (see RESULTS). Finally, when the self-progeny ratios indicated pseudolinkage of m and unc-1, it was concluded that the hermaphrodite parent must have had the genotype m/ Dp;unc-llunc-1. (Conversely, in the absence of other evi- dence, the independent assortment of m and unc-l did not distinguish between the following two possible parental gen- otypes: m/+;unc-l/+ and m/+;Dp/+;unc-llunc-1). When pseudolinkage was found, the recombination frequency, calculated from the frequency of Unc-1 non-M or M non- Unc-1 recombinants, provided an estimate of the recombi- nation distance from the duplication to m. In each estimate, we ignored recombination events occurring within the du- plications in hermaphrodites as well as the possibility of mitotic loss of the translocated duplication; if such events contribute significantly to the incidence of recombinant types, our estimated recombination distance is an over- estimate. In previous work, two X duplications translocated to the right end of LGl were shown to be subject to mitotic loss ( H E R M A N , MADL and KARI 1979). We did not system- atically investigate mitotic loss of the duplications studied here, but we did observe that mnDp69(X;Z) homozygotes gave rare heterozygous progeny.

TO test whether or not hermaphrodites could become homozygous for a duplication, up to 50 wild-type her- maphrodite progeny of Dp/+;unc-llunc-1 hermaphrodites were picked individually, and their progeny were inspected for the absence of Unc-1 animals. When such true-breeding non-Unc-1 animals were found, proof that they were homo- zygous for the duplication, rather than homozygous for an

unc-l(+) recombinant chromosome, was obtained as follows: putative duplication homozygotes were mated to N2 males, and wild-type male progeny were picked and mated with unc-1 hermaphrodites to see whether the expected wild-type male progeny were produced.

For experiments involving him-8;unc-l/O;mnDp73 males, the males were generated as follows. Males of genotype unc- l/U;mnDp73 were mated with him-8;unc-1 hermaphrodites; wild-type male progeny were picked and backcrossed to him- 8;unc-1 hermaphrodites. This cycle was repeated several times. A single wild-type male was then mated to him-8;unc- 1 hermaphrodites; 15 wild-type hermaphrodite progeny were picked and all were Him, which indicated that their male parent and male sibs were homozygous for him-8. In the experiments that used him-8;unc-l/0;mnDp73 males (crossed to unc-1 dpy-? hermaphrodites), only one male was used per mating plate. The resultant progeny ratios were then used to confirm that a given male was of the desired genotype and not him-8;unc-l(+)/0. Once formed, males of the latter genotype might be expected to give progeny of the same genotype (by patroclinous inheritance) in the him- 8 stock because of the high frequency of nullo-X ova pro- duced by him-8 hermaphrodites (HODGKIN, HORVITZ and BRENNER 1979).

Identification and characterization of mnDp57 and mnDp62: These two duplications were identified and char- acterized by procedures identical to those described for the duplications of unc-l(+) with the following exceptions. Doses of 6600 r were used. mnDp57 was recovered in a screen for duplications of unc-20(+), and mnDp62 was recovered in a screen for duplications of unc-3(+). In the latter case, sup- 10 males were irradiated instead of N2 males. On the basis of complementation tests, we conclude that mnDp57 extends over unc-2, osm-5, unc-20, unc-78 and lon-2, but not dpy-3,

Page 4: Chromosome in Caenorhabditis elegans

726 R. K. Herman and C. K. Kari

FIGI:RE ‘L.-l;luorescence microscopy of oocytes at di;lkincsis stained with DAPI showing free duplications (a) mnDp62, (h) mn- 4 7 1 and (c) mnDp73. Magnifications: X1.500.

lin-32, dpy-8,Ju-2, unc-6 or dpy-7 (Figure 1); it is tightly linked to unc-54 I and is homozygous viable. mnDp62 ex- tends over unc-3, sup-IO and osm-1 but not unc-9 or unc-84; it is a free duplication, as judged by the high frequency of nullo-Dp self-progeny produced (60%) and the appearance of a small stained fragment in addition to the normal six bivalents in squashed oocytes treated with DAPI (Figure 2a).

Identification of unc-I(+) recombinant X chromosomes generated by unc-IIO;Dp males: We used three procedures to identify unc-I(+) recombinant X chromosomes generated by unc-I/O;Dp males. The first procedure involved mating the unc-l/O;Dp males whose sperm were to be tested with unc-1 dpy-3 hermaphrodites. Wild-type larval (unmated) her- maphrodite progeny were picked. Progeny counts allowed us to distinguish between hermaphrodites that carried an unc-I(+) X chromosome from those that did not (the latter being non-Unc by virtue of carrying an unc-I(+) duplica- tion). The expected self-progeny ratios of W T (wild type):Unc:Dpy:Dpy Unc for animals of genotype unc-1 dpy- 3/++ (class A) were 3:O:O:l (crossovers between unc-1 and dpy-3 account for only about 1 % of the self progeny and are ignored here). The expected hermaphrodite self-progeny ratios for unc-1 dpy-3/unc-l+;Dp[unc-1(+)] (class B) and unc- I dpy-3/++; Dp[unc-I(+)] (class C) animals depend on the nature of the duplication. For duplications that are translo- cated, homozygous viable, and do not carry dpy-3(+) (i.e., mnDp63, mnDp65, mnDp66, mnDp67, mnDp69, and mn- Dp70) the expected self-progeny ratios are 9:3:3:1 for class B and 12:0:3: 1 for class C. Every hermaphrodite tested agreed well with the predictions for one of these three classes (there was only one example of class C).

The expected results for classes B and C involving mn- Dp68 and mnDp71 were different because these free dupli- cations are less frequently transmitted to progeny. Animals we put in class B for mnDp68 gave average self-progeny ratios of 0.45:0.29:0.14:0.12. Class B mnDp71 gave nearly identical ratios. N o class C animals, for which expected self- progeny ratios were 0.74:0.00:0.14:0.12, were found for either of these duplications.

The expected results for classes B and C involving mn- Dp72 and mnDp73 are altered by the fact that these dupli- cations carry dpy-3(+). They also gave relatively high fre- quencies of nullo-duplication bearing self progeny, mnDp72 because animals carrying it were trisomic for chromosome IV (see RFSULTS) and mnDp73 because it is free. Class B animals for mnDp73 gave the following hermaphrodite self- progeny ratios: 0.57:0.32:0.00:0.11. Class B mnDp72 ani- mals gave similar ratios. For both of these duplications, the expected ratios for class C were about 0.89:0.00:0.00:0.11. For several animals involving mnDp72 or mnDp73 we counted enough progeny per brood to be sure that the parent carried an unc-I(+) recombinant X chromosome but we did not count enough to distinguish between classes A and C.

The second procedure was used only for some of the

mnDp63 experiments. In this case mnDp63/+;unc-l/0 males were mated with unc-1 dpy-3 hermaphrodites, and larval wild-type hermaphrodite progeny were picked and mated one per plate with N 2 males. Only male progeny were counted. Class A animals (of which there were 11) gave about equal numbers of W T and Dpy Unc males and very few Unc or Dpy males, whereas class B animals (of which there were 24) gave WT, Dpy, Unc, and Dpy Unc males in roughly equal numbers. Class C animals (of which there were none) were expected to give roughly equal numbers of WT, Dpy, and Dpy Unc males with very few Unc males. Some assignments to both classes A and B were tested further by mating wild-type male progeny with unc-1 dpy-3 hermaphrodites; as expected, only the wild-type male prog- eny of class B animals carried duplications of unc-I(+).

The third procedure was used only for some of the mnDp72 experiments. In this case mnDp72(X;IV)/+/+;unc- I/O males (trisomic for IV; see RESULTS) were mated with dpy-4 1V;unc-I X hermaphrodites. The dpy-4/+;unc-l/+ an- imals (class A) gave the expected 9:3:3:1 dihybrid ratios. The class B genotype was mnDp72[unc-l(+)] dpy-4(+)/dpy- 4(+)/dpy-4;unc-l/unc-I, for which the self-progeny ratios were 0.51 WT:0.36 Unc:O.OO Dpy:0.13 Dpy Unc; see RE- SULTS for further details. The expected hermaphrodite self- progeny ratios for class C, genotype mnDp72[unc-l(+)] dpy- 4(+)/dpy-4(+)/dpy-4;unc-l/unc-l(+), were 0.79 WT:0.09 Unc:0.09 Dpy:O.OJ Dpy Unc. With this procedure, how- ever, we generally, did not count enough self progeny per brood to be sure of the distinction between classes A and C. Seventy-six animals were assigned to classes A or C and 19 were assigned to class B by this procedure.

Identification of hirn-8(mn253) mutation: We used a strain of genotype dpy-8 unc-3llon-2 unc-7 X to screen for hermaphrodites homozygous for autosomal recessive muta- tions causing greatly reduced X chromosome recombina- tion. dpy-8 and lon-2 are closely linked (0.6 map unit), as are unc-3 and unc-7 (1.7 map units); but the dpy-8 to unc-3 distance is 25 map units. Over 98% of the wild-type self- progeny of this strain are heterozygous for all four loci because the two members of each closely linked pair tend to balance each other. About 1 1% of the wild-type progeny are dpy-8 unc-7/lon-2 unc-3 instead of the original genotype, owing to recombination in the large interval separating the pairs of loci. Either quadruple heterozygote, however, gen- erates the following four easily distinguished recombinant types, each at a frequency of about 10%: Dpy non-Unc, Lon non-Unc, Unc-3 non-Dpy non-Lon, Unc-7 non-Dpy non- Lon. We mutagenized with EMS (BRENNER 1974), picked first and second generation wild-type progeny, and then screened the broods of FP animals for a great reduction in the usual 40% recombinant types. Among about 4400 fertile F P broods, we found one mutant, which was outcrossed three times. Its recombination frequency for the dpy-8 to unc-3 interval was reduced (see RESULTS), it gave 39% male self progeny (among 3344 progeny), and it failed to comple- ment him-8(e1489) IV for the Him trait. We therefore assign the mutation, mn253, to him-8.

RESULTS

Duplications of unc-I(+): Ten independently de- rived unc-l(+) duplications are listed in Table l . All were identified following y-ray treatment, according to the procedures described in MATERIALS AND METH- ODS. The extent of each duplication was assessed by ascertaining whether or not it suppressed other mu- tations mapping near unc-l (see MATERIALS AND METH- ODS). By this criterion, all ten duplications carry the

Page 5: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C. elegans

TABLE 1

Properties of duplications of unc-I(+)

727

Duplication Other mutations Mutations not

suppressed suppressed

mnDp63

mnDp65

mnDph6

mnDp67

mnDp68 mnDp69

mnDp70

mnDp71 mnDp72

mnDp73

che-2, mab-7

che-2, mab-7

che-2, mab-7

che-2, mab-7

che-2, mab-7 che-2, mab-7

che-2, mab-7

che-2, mab-7 che-2, mab-7, dpy-3

che-2, mab-7, dpy-3 lin-32

dPY-3

dPY-3

dPY-3

dPY-3

dPY-3 dPY-3

dPY-3

dPY-3 lin-32, unc-2

unc-2

Location"

Percent Unc-1 True-breeding hermaphrodite

Percent male

homozygotes? self-progeny' self-progeny'

2.7% from unc-60 V (12 Unc-1 non-Unc-60/879) 0.1 % from egl-30 I (1 Unc-1 non-Eg1/2473) 0.1 % from unc-54 I (4 Unc-1 non-Unc-54/976) 0.6% from unc-54 I (4 Unc-1 non-Unc-54/1242) Probably free 0.1 % from unc-54 I ( 1 Unc-1 non-Unc-54/2873) 2 .8% from unc-60 V (7 Unc-1 non-Unc-60/505) Free 0.6% from dpy-4 IV (8 Dpy non-Unc/2507) Free

Yes

Yes

Yes

Yes

N o Yes

Yes

No ?

N o

30 (446)

27 (1840)

26 (2741)

25 (243 1 )

42 (5330) 29 (2479)

27 (946)

43 (3574) 48 (4792)d

43 (43 14)

9.9 (1471)

0.2 (1333)

N DC

ND

ND ND

5.5 (5310)

0.4 (1819) 3.4 (4955)

0.9 (291 1 )

In parentheses are given the number of recombinants/total progeny counted from parents of genotype Dp/m; unc-llunc-1 (except for

Hermaphrodite parents carried a single copy of the duplication. The total number of progeny scored is given in parentheses.

The hermaphrodite parents in this case were trisomic for chromosome IV and had the genotype mnDp72/+/dpy-4; unc-llunc-I.

rnnDp72/+/dpy-4; unc-llunc-1), where m is the autosomal marker listed. Only one of the recombinant classes could be scored in each case.

' ND = not accurately determined, but the frequencies were clearly low in these cases.

closely linked genes che-2(+) and mab-7(+) in addition to unc-I(+). Only two, mnDp72 and mnDp73, extend as far to the right of unc-1 as dpy-3, and only mnDp73 carries lin-32(+), just to the right of dpy-3, but mnDp73 does not extend as far as unc-2 (Figure 1).

Seven of the ten duplications are translocated to autosomal sites. In each of the seven cases, unc-I(+)- duplication-bearing hermaphrodites were constructed that were homozygous on the X chromosome for unc- I and heterozygous for an autosomal genetic marker (see MATERIALS AND METHODS). Progeny counts were then used to look for possible pseudolinkage of the autosomal marker and unc-I. In this way it was dem- onstrated that mnDp66, mnDp67 and mnDp69 map near unc-54, which is situated near the right end of linkage group (LG) I ; mnDp63 and mnDp70 map near unc-60, near the left end of LGV; mnDp65 maps near egl-30, near the left end of LGZ; and mnDp72 maps near dpy-4, near the right end of LGZV (Table 1). At least six translocated duplications (all but mnDp72, a special case discussed in more detail below) are homo- zygous viable, which indicates that none of the auto- somal breakpoints for these six translocated duplica- tions has inactivated an essential gene and raises the possibility that they are fusions to autosomal telo- meres. The seven translocated duplications are nec- essarily half-translocations because the scheme used to recover them precluded the recovery of the poten- tial complementary half-translocations, which would have included the bulk of the X chromosome.

Hermaphrodites heterozygous for any of the trans-

located duplications except mnDp72 gave approxi- mately the expected frequency of nullo-duplication (homozygous unc-I) self-progeny of 25% (Table 1). The high incidence of nullo-duplication self-progeny for mnDp72 is at least partly a consequence of the fact that the hermaphrodites in this case generally carried two normal linkage group ZV chromosomes in addition to the mnDp72(X;ZV)-bearing chromosome. The evi- dence for this claim will be presented in a later section.

mnDp71 and mnDp73 are free duplications. A small chromosome fragment was observed cytologically in addition to the normal six bivalents in oocytes of hermaphrodites carrying either duplication (Figure 2), and the frequency of nullo-duplication self progeny from duplication-bearing hermaphrodites was in each case greater than 40% (Table l) , which is character- istic of many free duplications (HERMAN, MADL and KARI 1979). The final unc-I(+) duplication, mnDp68, we tentatively judge to be free because of the high frequency at which mnDp68-bearing hermaphrodites segregated nullo-mnDp68 self-progeny (Table 1). We have been unable to confirm this assignment cytolog- ically, however. (For unknown reasons, mnDp68-bear- ing hermaphrodites are relatively slow growing, and their cytology has been difficult). None of the free duplications gave rise to true-breeding non-Unc-1 stocks. This does not mean that animals carrying two copies of free duplications are inviable; it is likely that they are viable but simply give a significant incidence of nullo-Dp progeny (HERMAN, MADL and KARI 1979).

Four of the duplications-mnDp63(X;V), mn-

Page 6: Chromosome in Caenorhabditis elegans

728 R. K . Herman and C. K. Kari

Dp7O(X;V), mnDp72(X;lV) and, to a weak degree, mnDp73(X,f)-are responsible for (or are tightly cou- pled to a genetic element responsible for) a dominant Him (high incidence of male self progeny) trait (Table 1). The wild-type incidence of males is 0.2% (HODG- KIN, HORVITZ and BRENNER 1979).

We note that the frequency at which mnDp65 was transmitted to progeny clearly changed during the course of our experiments. Our “early” measure- ments, all made prior to October 1986, disagree with the ‘‘late’’ measurements, all made after January 1987. In one early set of measurements, the average fre- quency of nullo-mnDp65 self progeny for nine broods was 0.50, with none of the nine approaching 0.25. Five more broods sampled on a different early day yielded an average frequency of 0.49. In addition, an early measurement of the frequencies of sperm gen- otypes produced by mnDp65;unc-l/0 males (discussed below) indicated that about 75% of the sperm were nullo-mnDp65. All late measurements of hermaphro- dite self-progeny ratios (some approximate but made on nine different days) agree with the value given in Table 1. And the male sperm ratios given below, which indicated a frequency of about 50% nullo- mnDp65 sperm, were late measurements. We suggest that in the early experiments mnDp65 was either highly subject to loss-some translocated duplications were previously shown to be unstable (HERMAN, MADL and KARI 1979)“or mnDp65 was originally a free duplication whose transmission properties were al- tered as a consequence of becoming attached to chro- mosome I . It is the first example of a duplication attached near the left end of LGZ.

Recombination in males between unc-Z(+) dupli- cations and the X chromosome: For seven duplica- tions, our first step in analyzing sperm genotypes was to mate unc-I/O;Dp males with dpy-1 I V;che-2 unc-1 X hermaphrodites and to score the non-Dpy progeny (Table 2). For the two duplications that carry dpy- 3(+)-mnDp72 and mnDp73“the unc-l/O;Dp males were mated with che-2 unc-1 dpy-3 (or unc-1 dpy-3) hermaphrodites. The results of both types of crosses are given in Table 2. We presume that the wild-type male progeny in each cross were generated by sperm carrying an unc-I(+) duplication and no X chromo- some. Many of the Unc male progeny in the crosses were scored with respect to their Che phenotype (by FITC staining; see MATERIALS AND METHODS) to detect crossover duplications (che-2(+) unc-I). A surprisingly large frequency of Unc non-Dpy males were found in the mnDp72 and mnDp73 crosses. All that were tested were non-Che; we conclude that the Unc males were generated by sperm carrying che-2(+) unc-I dPy-3(+) recombinant duplications (and no X chromosome) and not by duplications that had simply lost the unc-I(+) locus by deletion. Non-Che Unc males were also found for some of the other duplications. We conclude that

these males were generated by sperm carrying che- 2(+) unc-1 recombinant duplications; high incidences of such recombinant duplications were found for mnDp63 and mnDp71 (Table 2). Che Unc (or Che Dpy Unc for mnDp72 and mnDp73) male progeny we presume were generated by nullo-Dp nullo-X sperm. The Unc hermaphrodite progeny in both sets of crosses we presume were generated by sperm carrying an unc-1 X chromosome (the possible presence of a recombinant unc-1 duplication in addition we ig- nored). Finally, the wild-type hermaphrodite progeny in these crosses could have been generated either by sperm carrying recombinant unc-I(+) X chromosomes or by unc-I X;Dp[unc-I(+)] sperm. When the presence of a recombinant unc-I(+) chromosome was indicated, we often checked to see whether or not an unc-I(+)- bearing duplication was also present (the possibility of recombinant unc-I duplications was again ignored). The procedures used to measure the relative frequen- cies of these sperm genotypes generated by unc-I/ 0;Dp males are described in MATERIALS AND METHODS; the results are given in Table 3.

Recombinant X chromosomes were detected for seven of the ten duplication-bearing males. Among the rather limited numbers of X chromosomes ana- lyzed, at least two unc-l(+) recombinants were found for five of the ten duplication stocks. In only one of 69 cases examined was a recombinant X chromosome found accompanied by a duplication; this pattern was expected because the rare class requires nondisjunc- tion from an X-Dp chiasma. The data in Tables 2 and 3 were used to calculate frequencies of recombinant X chromosomes and recombinant duplications for the unc-l-dpy-3 interval transmitted by unc-l/O;Dp[unc- I(+)] males; see Table 4. Three duplications yielded extremely high frequencies of recombinant X chro- mosomes and recombinant duplications: mnDp63, mnDp72 and mnDp73. For these duplications in males, the frequency of recombination between the duplica- tion and the X chromosome is a factor of ten or more higher than the 1.3% recombination frequency be- tween X chromosomes in hermaphrodites (see below), for the same unc-l-dpy-3 interval. One other duplica- tion, mnDp71, also led to fairly high frequencies of recombinant chromosomes and recombinant duplica- tions from unc-I/O;Dp males. It should be emphasized that only mnDp72 and mnDp73 carry the full unc-I dpy-3 interval. Recombination involving any other duplication must occur between unc-I and the dupli- cation breakpoint. Perhaps the extent of this segment is greater €or mnDp63, and to a lesser degree mnDp71, than for the other six duplications.

The data in Tables 2 and 3 were also used to calculate frequencies of sperm genotypes produced by unc-I/O;Dp males; see Table 5. A notable feature of Table 5 is that the frequencies of X ; D p and nullo-X; nullo-Dp sperm were reduced for those duplications

Page 7: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C . eleguns

TABLE 2

Frequencies of progeny sired by che-Z(+) unc-l/0; Dp males

729

Progeny

No. of Fraction of progeny Wild-type Unc Wild-type Unc Unc males Dpy Unc

Duplication scored hermaphrodites hermaphrodites males males non-Che males

A. X dpy-I1 V; che-2 unc-1 X hermaphrodites: mnDp63(X;V) 4236 0.16 0.27 0.36 0.2 1 25/60 mnDp65(X;I) 1579 0.21 0.29 0.29 0.2 1 111'71 mnDp66(X,I) 8446 0.24 0.26 0.24 0.26 0129 1

mnDp67(X;I) 1823 0.24 0.27 0.25 0.25 11141 mnDp68(Xj ) 4950 0.19 0.34 0.20 0.27 01296 mnDp70(X;V) 1575 0.20 0.27 0.34 0.19 0140 m n D p 7 I ( X f ) 370 0.15 0.39 0.26 0.20 9/49

mnDp72(X;IV)" 462 0.32 0.33 0.27 0.06 919 0.0 1 mnDp73(Xf ) 1367 0.24 0.33 0.34 0.08 75/75 0.02

mnDp73(Xf)b 940 0.20 0.28 0.49 0.00 0.04

B. X rhe-2 m e - 1 dpy-3 X hermaphrodites:

C . X unc-1 dpy-3 X hermaphrodites:

These males were trisomic for chromosome IV. The males i n this case were homozygous for him-8(mn253).

TABLE 3

Classification of fertilizing sperm that carry unc-I(+) and an X chromosome, generated by unc-l/0; Dp males

Sperm

unc-I(+) X or unc-I(+) X that FraCtiOl1 with

Ihplication scored" Dp[unc-I(+)] Dp[unc-l(+)] Dp[unc-I(+)] Total no. unc-1 X: unc-I(+) X, also had

mnDp63(X;V) 53 34 mnDp65(X;I) 18 17 mnDp66(X;I) 24 24 mnDp67(X;I) 24 23 m n D p 6 8 ( X f ) 28 28 rnnDpfjY(X;I) 30 30 mnDp70(X;V) 20 18 mnDp71(Xf ) 19 16 mnDp72(X; 1 V ) 148 27 mnDp73(Xf ) 47 22 mnDp73(Xf)b 34 34

19 011 9 1 01 1

I 111 0

0 0 2 012 3 013

121 0/3 3 25 0110

0

Animals scored were wild-type hermaphrodites issuing from the cross of unc-I/0;Dp nlales and dpy-11; che-2 unc-1 hermaphrod- ites or (for mnDp72 and mnDp73) che-2 unc-1 dpy-3 hermaphrodites. ' The males in this case were homozygous for him-B(mn253).

showing recombination. Indeed, the extent of this effect correlated very well with the extent of X-Dp recombination, i . e . , in increasing order: mnDp65, mnDp70, mnDp71, mnDp63, mnDp73 and mnDp72. We attribute this effect to the disjunction of X and Dp following pairing and exchange. Two peculiarities of Table 5 deserve comment. First, mnDp72-bearing males gave an unexpectedly low frequency of nullo- X;Dp sperm (giving a low transmission of the dupli- cation and a deficiency of male progeny). As will be discussed below, mnDp72-bearing animals were tri- somic for chromosome ZV; it is possible that the tri- somic males had reduced viability. Second, the nullo- X;nullo-Dp class seems aberrantly small for both mnDp72 and mnDp73; possibly the Che Dpy Unc

TABLE 4

Recombination frequencies for the unc-I-dpy-3 interval in unc-1 X/@ Dp[unc-I(+)] males

Percent unc-I(+) Duplication

Percent unc-1 X chromosomes" duplications*

mnDp63(XV) 13 (9-19) 16 (1 1-21) mnDp65(XI) 2 (0-1 1) 0 (0-1) mnDp66(X;I) 0 (0-7) 0 (0-1) mnDp6 7(X;I) 2 (0-10) 0 (0-2) mnDp68(Xf ) 0 (0-4) 0 (0-1) mnDp6Y(X;I) 0 (0-6) N D'

mnDp70(X;V) 4 (1-14) 0 (0-3) mnDp71(X;f) 4 (1-1 1) 9 (4- 15) mnDp72(X;lV) 40 (37-44) I7 (11-23) mnDp73(Xf ) 22 (1 6-29) 15 (12-18) m n ~ p 7 ? ( ~ f ) ~ 0 (0-4) 0 (0-1)

" Fraction of unc-I(+) X sperm in Table 3 X ratio of wild-type hermaphrodites to total hermaphrodites in Table 2. (For mn- Dp69(XI) it was assumed that the ratio of wild-type hermaphrodites to total hermaphrodites would have been the same as for mn- Dp66(X;I); Table 2.) The extreme values given (in parentheses) are estimated 95% confidence limits, calculated from the 95% confi- dence limits for the frequencies of unc-l(+) X sperm in Table 3, assuming binomial distributions (MAINLAND, HERRARA and SUT- CLIFFE 1956). The frequency of unc-I(+) X sperm was by far the major source of error in estimating percent unc-I(+) X chrorno- somes. Values are given to nearest 1 %.

Frequency of Unc males in Table 2 X fraction of Unc males that are non-Che (Table 2) divided by {frequency of wild-type males in Table 2 + (frequency of Unc males in Table 2 X fraction of Unc males that are non-Che) + (frequency of wild-type hermaphrodites in Table 2 X fraction of Dp[unc-l(+)]-bearing hermaphrodites in Table 3)j. For all duplications except mnDp72 and mnDp73, the extreme values given are based on the 95% confidence limits (assuming binomial distributions) for the fraction of Unc males that are non-Che (lable 2); these values were by far the major source of error in these determinations. For mnDp72 and mnDp73, the major source of error was the frequency of Unc males (Table 2). and the confidence limits for these frequencies were used to calcu- late the extrenles given for percent unc-1 duplications. Values are given to nearest 1%.

' ND = not determined. The males i n this case were homo7ygous for him-R(mn253).

Page 8: Chromosome in Caenorhabditis elegans

730 R. K. Herman and C. K . Kari

TABLE 5

Frequencies of sperm genotypes produced by unc-l/O; Dp[unc-I(+)] males"

Duplication -

mnDp63(X;V) mnDp65(XI) mnDp66(X;I) mnDp67(X;I) mnDp68(XJ) mnDp7U(X;V) mnDP7I(Xf ) mnDP72(X;IV) mnDp73(Xj) mnDf173(X$)~

NCO X nullo-Dpg

0.27 0.29 0.26 0.27 0.34 0.27 0.39 0.33 0.33 0.28

IIullo-Dp'

0.06 0.01 0.00 0.00 0.00 0.02 0.02 0.26 0.13 0.00

CO X NCO u p '

0.10 0.20 0.24 0.23 0.19 0.18 0.13 0.06 0 . 1 1 0.20

NCO X NCO Dp'

co X

0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

N CO Dp nullo-X

0.36 0.29 0.24 0.25 0.20 0.34 0.26 0.27 0.34 0.49

co up' nullo-x

0.09 0.00 0.00 0.00 0.00 0.00 0.04 0.06 0.08 0.00

nullo-np nullo-X/

0.12 0.21 0.26 0.25 0.27 0.19 0.16 0.01 0.02 0.04

" N<:O (noncrossover) X is unc-I, and CO (crossover) X is unc-I(+); N C O Dp is unc-I(+), and CO Dp is unc-1.

' Frequency of wild-type hermaphrodites in Table 2 X proportion of wild-type hermaphrodites generated by indicated sperm genotype

" Frequency of wild-type males in Table 2. ' Frequency of non-Che Unc males in Table 2. 'Frequency of' Che Unc (or Dpy Unc) males in Table 2.

Frequency of Unc hermaphrodites in Table 2.

(Table 3).

The males in this case were homozygous for him-8(mn253).

males representing this class had reduced viability. Three additional crosses were performed to inves-

tigate further the nature of the recombination be- tween m n D p 7 3 and the X chromosome in males. In the first cross, che-2 unc-1 dpy-3/0;mnDp73 males were crossed to che-2 unc-1 dpy-3 hermaphrodites, and male progeny were scored. The results were as follows: 296 non-Unc non-Dpy (among which 21202 were Che), 57 Che Unc non-Dpy, 0 Dpy non-Unc and 43 Che Dpy Unc. Thus, two classes of recombinant duplica- tion were identified: che-2 unc-I(+) dpy-3(+), which were rare (1 %), and the more common che-2 unc-1 dpy-3(+), which comprised 16% of the duplications scored. Each recombinant class identified is most sim- ply explained as resulting from a single crossover between a linear duplication and linear chromosome (Figure 3A). Other possible duplication recombinant types, such as Dpy non-Unc, which were not found, would have required more complicated interpreta- tions, such as multiple crossovers or gene conversion. In a second cross, unc-1 dpy -3 /0 ;mnDp73 males were crossed to che-2 unc-1 dpy-3 hermaphrodites, and the male progeny were scored, with the following results: 222 non-Unc non-Dpy, 69 non-Che Unc, 0 Dpy non- Unc and 23 Unc Dpy (among which 3/22 were non-

A B

Che). Here too, all of the identified recombinant duplications can be explained by single crossovers between duplication and X chromosomes (Figure 3B). In this case, however, it was possible to detect cross- overs to the right of dpy-3, which generated non-Che Unc Dpy male progeny. They proved to be only about 5% as frequent as the crossovers occurring in the unc- I-dpy-3 interval, which in this case comprised 23% of the duplications scored; thus, whatever the nature of the putative right end of m n D p 7 3 , as diagrammed in Figure 3, it does not seem to be particularly recom- binogenic. In the third cross, che-2 dpy-3/0;mnDp73 males were crossed to che-2 dpy-3 unc-2 hermaphro- dites, and non-Unc-2 hermaphrodites were scored, with the following results: 199 non-Dpy and 367 Dpy; among the Dpy animals, 651256 were non-Che. This result shows that recombinant X chromosomes need not pick up all the genes carried by the duplication (Figure 3C), since about 25% of the X chromosomes inherited by the Dpy hermaphrodite progeny picked up che-a(+) but not dpy-3(+) from m n D p 7 3 by recom- bination in the male parent. Thus this result does not support a model in which the recombinant X chro- mosome is generated by the attachment o r integration of a complete mnDp73.

C

FIGURE :I.--Diagranls of identified recombination events occurring in males between mnDp73 and the X chrornosotne. The males in A and I3 Mere mated wi th che-2 unc-l dpy-3 hermaphrodites; the males i n C were mated with che-2 dpy-3 unc-2 hermaphrodites (see text).

Page 9: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C. ekgans 731

mnDp72(X;N) and mnDp63(X;V) promote tri- somy for chromosomes N and V, respectively: The wild-type hermaphrodite progeny of crosses involving unc-l/O;mnDp72 males and dpy-4 IV;unc-I X her- maphrodites fell into two classes. As expected from results already discussed (Table 3), the majority car- ried an unc-I(+) recombinant chromosome. Among the hermaphrodites that did not carry an unc-I(+) chromosome, 19 of 19 gave similar hermaphrodite self progeny ratios, which overall were as follows: 0.5 1 WT, 0.36 Unc, 0.00 Dpy, and 0.13 Dpy Unc (2507 total hermaphrodite progeny scored). Because vir- tually all Dpy animals were Unc, we conclude that mnDp72 is closely linked to dpy-4. But the ratios show three peculiarities: first, many of the Unc animals were not Dpy; second, the overall proportion of Dpy animals (Dpy plus Dpy Unc) was considerably less than one quarter; and, third, the overall proportion of Unc (Unc plus Dpy Unc) animals was nearly half. Our interpretation of these results is that the wild-type hermaphrodite parents in these cases had the geno- type mnDp72[unc-l(+)] dpy-4(+)/dpy-4(+)/dpy-4;unc- Ilunc-I. According to this interpretation, these ani- mals were essentially trisomic for LGZV. The viability and wild-type morphology of LGZV trisomics was dem- onstrated previously (SICURDSON et al. 1986). We progeny-tested a total of 14 wild-type hermaphrodite progeny of three of these wild-type hermaphrodites; all segregated Unc (plus Dpy Unc) progeny at the high frequency characteristic of their parents, as if they all, like their parents, carried a single copy of the mnDp72 chromosome in addition to two normal LGZV chro- mosomes. Among the 14, some gave progeny ratios like those of their parents; others either gave only WT and Dpy Unc animals or only WT and Unc animals. We conclude that the genotypes responsible for the latter broods were mnDp72 dpy-4(+)/dpy-4/ dpy-4;unc-l/unc-l and mnDp72 dpy-4(+)/dpy-4(+)/ dpy-4(+);unc-l/unc-l, respectively. The apparent ab- sence of mnDp72/IV;X/X animals among the 14 mnDp72-bearing hermaphrodites suggests that the two normal LGIV chromosomes tend to disjoin from each other during meiosis. This interpretation is also supported by the 3:l (rather than 7:l) ratio of Unc:Dpy Unc progeny and the relatively low propor- tion of WT progeny. If we assume IV/ZV disjunction and if we further assume that the mnDp72 chromo- some is transmitted to 35% of gametes (both sperm and ova) and that tetrasomics are inviable, we calculate the following expected progeny ratios for mnDp72 dpy-4(+)/dpy-4(+)/dpy-4;unc-l/unc-l hermaphrodites: 0.52 W T , 0.36 Unc, 0.00 Dpy and 0.12 Dpy Unc. This agrees well with the results obtained. (A higher frequency of transmission of mnDp72 with a reduced viability of trisomics would also be compatible with the observed results). Note that according to our interpretation, we would not expect homozygous dip-

loid mnDp72 animals to be generated as progeny of the trisomic hermaphrodites; thus we cannot find out whether or not such animals are viable by this route.

The trisomic animals must have been generated by duplication-bearing sperm disomic for LGZV, i e . , of genotype mnDp72[unc-I(+)] dpy-4(+)/dpy-4(+) ZV;unc- 1 X. This suggests that the duplication-bearing males, which were continually regenerated by crossing to homozygous unc-1 hermaphrodites, were themselves trisomic for LGZV;mnDp72/+/+;unc-l/0. During spermatogenesis the mnDp72 and X chromosomes would then be frequent pairing partners, as would the two normal LGZVchromosomes. Wild-type male prog- eny of matings with unc-1 hermaphrodites would then be generated by ZV/mnDp72[unc-l(+)] sperm and would therefore also be trisomic for ZV.

We also found hermaphrodites that we concluded carried two LGV chromosomes in addition to mn- Dp63(X;V). In this case, self-progeny ratios of 12 wild- type hermaphrodite progeny of crosses involving unc- I/O;mnDp63 males and unc-60 V;unc-1 X her- maphrodites revealed three genotypes: four were unc- 60/+;unc-l/+, segregating close to the expected 9:3:4 for WT:Unc-1:Unc-60 (which also includes Unc-60 Unc-1 animals because unc-60 is epistatic to unc-1); six were mnDp63 unc-60(+)/unc-60;unc-I/unc-I, seg- regating 3 WT: 1 Unc-60 (with 0.02 Unc-1 recombi- nants); and two we concluded were mnDp63 unc- 60(+)/unc-60(+)/unc-60;unc-l/unc-l; they segregated 0.47 WT, 0.39 Unc-1, and 0.14 Unc-60 (208 animals total), which are progeny ratios very close to those seen for trisomic mnDp72 hermaphrodites. As ex- pected, we found, among the wild-type progeny of the latter two animals, hermaphrodites that gave ap- proximately equal numbers of WT and Unc-60 prog- eny and almost no Unc-1 progeny; we concluded that these hermaphrodites were mnDp63 unc-60(+)/unc- 60/unc-60;unc-I/unc-1. The only difference between the results on mnDp63 and trisomy for LGV and the results on mnDp72 and trisomy for LGZV is that a minority of the mnDp63-bearing hermaphrodites ap- peared to be trisomic, whereas all of the mnDp72- bearing hermaphrodites appeared to be trisomic. We do not know whether the mnDp63/+;unc-l sperm that gave rise to the trisomic hermaphrodites were gener- ated by a minority of trisomic males in the mnDp63 stock or whether they were generated by nondisjunc- tion in (mnDp63-bearing) diploid males.

him-8(mn253) severely reduces X-X recombination in hermaphrodites and X-mnDp73 recombination in males: The mutation mn253 was identified in a screen for EMS-induced mutations (see MATERIALS AND

METHODS) that markedly reduce meiotic recombina- tion in the hermaphrodite (in both sperm and oocyte lines) between the two X chromosomes (Table 6). Animals homozygous for mn253 showed less than 10% as much recombination over the dpy-8-unc-3 interval

Page 10: Chromosome in Caenorhabditis elegans

732 R. K. Herman and C. K. Kari

TABLE 6

him-8(rnn253) reduces recombination on the X chromosome but not on autosomes

Genotype Recombination frequency"

dpy-8 unc-3/+ -+ X 0.28 (0.24-0.32) him4 IV; dpy-8 unc-3/+ + 0.02 (0.01-0.04) dpy-5 unc-54/+ + I 0.29 (0.25-0.33) dpy-5 unc-54/+ +; him-8 0.41 (0.37-0.45) dpy-1 unr-36/+ + 111 0.18 (0.16-0.20) dpy-1 unc-36/+ +; him-8 0.28 (0.24-0.33)

a Estimated errors are 95% confidence limits.

of the X chromosome as did wild-type animals. The mutation also resulted in a high incidence (39%) of male self progeny, probably as a consequence of the deficiency in X recombination (see DISCUSSION), and is an allele (see MATERIALS AND METHODS) of the previ- ously studied gene him-8 (HODGKIN, HORVITZ and BRENNER 1979). We measured autosomal recombi- nation frequencies for the intervals dpy-5 unc-54 I and dpy-I-unc-36 I I I and found about 40-50% higher recombination in him-8(mn253) animals (Table 6).

We have investigated the effect of him-8(mn253) on recombination between mnDp73 and the X chromo- some in males. In crosses between him-8(mn253);unc- 110 mnDp73 males and unc-1 dpy-3 hermaphrodites, no Unc non-Dpy male progeny were found (Table 2, last row). About 70 Unc males would have been expected if the yield had been equal to that obtained with him-8(+);unc-l/O;mnDp73 males (Table 2, next to last row). As already noted, the Unc males contain unc-1 dpy-3(+) recombinant duplications. Recombi- nant unc-I(+) X chromosomes were also sought in these same crosses (see MATERIALS AND METHODS). Among 34 unc-I(+)-bearing sperm generated by him- 8(mn253);unc-I/O;mnDp73 males, none carried a re- combinant unc-l(+)X chromosome (Table 3, last line). By contrast, 25 of 47 unc-I(+)-bearing sperm from him-8(+);unc-l/O;mnDp73 males carried recombinant unc-I(+) X chromosomes (Table 3, next to last line). We conclude that him-tI(mn253) severely reduces re- combination between mnDp73 and the X chromosome in males. We do not understand why the frequency of mnDp73-bearing sperm was aberrantly high (69%; Table 5, last line) in the crosses involving him-8;unc- l/O;mnDp73 males; possibly these males carried two copies of mnDp73.

Recombination in hermaphrodites between the X chromosome and mnDp72 or mnDp73: We detected recombination between either mnDp72 or mnDp73 and an X chromosome in unc-I dpy-3/unc-l dpy- 3;Dp[unc-I(+) dpy-3(+)] hermaphrodites: both classes of recombinant hermaphrodites were found at fre- quencies comparable to those found in unc-1 dPy-31 ++ hermaphrodites (Table 7). We followed the seg- regation of the dominant Him trait conferred by mnDp72 among some of the hermaphrodite recombi-

nants: 10/10 Unc recombinants were Him (all segre- gated >2% male self-progeny) and 1811 8 Dpy recom- binants were non-Him. These results are consistent with the view that the Unc recombinants carried a recombinant duplication of genotype unc-I dpy-3(+) and the Dpy recombinants carried a recombinant X chromosome of genotype unc-Z(+) dpy-3 (Figure 4A).

In the absence of detailed knowledge about either the segregation of recombinant and nonrecombinant duplications and X chromosomes or the relative via- bilities of all possible zygote genotypes, we have taken the following simplified approach to estimating re- combination frequencies in these hermaphrodites. Re- combinant X chromosomes were only detectable in progeny lacking the duplication. For mnDp72, we detected 26 recombinant X chromosomes in a total of 1084 nullo-mnDp72 hermaphrodites, each of which had two X chromosomes; this gives an X chromosome recombination frequency of 1.2%. We detected 15 recombinant duplications in 120 1 mnDp72-bearing hermaphrodites, a frequency of 1.2%. For mnDp73, the frequencies of recombinant X chromosomes and duplications were 0.6% (612 X 533) and 0.3% (2/ 648), respectively. The Dp-X recombination frequen- cies were clearly much lower in hermaphrodites than in males (Table 4).

Recombination in males between the X chromo- some and duplications of other parts of the X chro- mosome: We mated mnDp57(X;I)/+; unc-2 osm-5 lon- 210 males with unc-2 osm-5 Lon-2 hermaphrodites and screened 1502 hermaphrodite cross progeny for Unc non-Lon or Lon non-Unc recombinants. None was found. The unc-2-Lon-2 distance is about 7 map units.

We mated unc-9 unc-3 osm-l/O;mnDp3(X;f) males with dpy-1 I ;unc-9 unc-3 osm-1 hermaphrodites and screened non-Dpy hermaphrodite progeny for Unc-9 non-Unc-3 recombinants (Unc-3 non-Unc-9 recombi- nants cannot be readily recognized because unc-3 is epistatic to unc-9). Among 525 screened chromo- somes (carried by hermaphrodite progeny not carry- ing mnDp3), four recombinant chromosomes were found; all were unc-9 unc-3(+) osm-I(+), and all were homozygous viable (indicating that mnDp3 extends through the right-most vital locus). We presume that X;Dp sperm, which were approximately as frequent as X;nullo-Dp sperm, carried only nonrecombinant X chromosomes, in which case the overall frequency of recombinant X chromosomes was about 0.4% or ' / Z O

the normal 8% X-X recombination for the unc-9-unc- 3 interval (Figure 1).

Recombination in hermaphrodites between the X chromosome and duplications of the right end of the X chromosome: We screened self progeny of unc- 9 unc-3 osm-1 ;mnDp3(Xf) hermaphrodites for Unc-9 non-Unc-3 recombinants. Such recombinants would be expected to carry an unc-9 unc-3(+) osm-I(+) re- combinant chromosome. Among 4350 progeny lack-

Page 11: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C. elegans

TABLE 7

Recombination in the unc-I-dp.v-3 interval in hermaphrodites

733

Hermaphrodite frequencies Total progeny Male

Genotype scored W T Unc DPY Dpy Unc frequency

+ +/uric-1 dpy-3 4290 0.73 0.007 0.006 0.26 0.002" mnDp72/+/+; unc-1 dpy-3/unc-l dpy-3' 2355 0.52 0.007 0.01 1 0.46 0.03 1 unc-1 dpy-3junc-1 dpy-3; mnDp73 1190 0.55 0.002 0.005 0.45 0.008

a From HODGKIN, HORVITZ and BRENNER (1979). ' These animals are trisomic for chromosome IV, carrying one mnDp72-bearing chromosome and two wild-type (WT) IV chromosomes.

A B d / sup- 10 FIGURE 4,"Diagrams showing identified

mnDp62fx , f j recombination events between (A) mnDp72 and the X chromosome and (B) mnDp62 and the X chromosome. - " X

unc- 1 dpy-3 unc-3 osm-1 + ing the duplication, we found no recombinant X chro- mosomes, a frequency of 0/8700.

Finally, we screened the self-progeny of unc- 93(e15OO);unc-3 osm-I sup-lO(+);mnDp62(X;f)[unc- 3(+) osm-I(+) sup-IO] hermaphrodites for recombi- nant X chromosomes carrying sup-IO. Because the recombinant chromosome would be expected to ap- pear first in an animal whose other X chromosome would be sup-IO(+), the animal would be Unc-93, like its non-recombinant parent; one quarter of its prog- eny, however, would be non-Unc-93. All of the non- Unc-93 animals would also be non-Unc-3 if the cross- over occurred to the left of unc-3; otherwise the only progeny that would be non-Unc-3 would be those carrying mnDp62. We screened the broods of 18,135 mnDp62-bearing self progeny of unc-93;unc-3 osm-I supIO(+);mnDp62 animals for non-Unc-93 non-Unc-3 hermaphrodites and found five independent candi- dates for carrying recombinant sup-IO X chromo- somes. Two of the candidates no longer carried mnDp62 and were unc-3(+) and osm-I(+); we conclude that they carried a recombinant unc-3(+) osm-I(+) sup- IO X chromosome, produced by an event as dia- grammed in Figure 4B. Our estimated frequency of such events is 2/36,270. Two other candidates carried mnDp62 and X chromosomes that were unc-3 and osm- I(+); we conclude that they carried a recombinant unc-3 osm-I(+) sup-IO X chromosome, generated as diagrammed in Figure 4B, at a frequency of 2/36,270. The fifth candidate carried mnDp62 and X chromo- somes that were unc-3 and osm-I, which could have arisen by a recombination event as diagrammed in Figure 4B. Alternatively, the fifth candidate might have been formed by spontaneous mutation, in either sup-IO or another extragenic suppressor of unc- 93(e1500) (GREENWALD and HORVITZ 1980) (we did show in crosses to N2 males that the fifth candidate was not an unc-93 intragenic revertant). The sponta- neous frequency of mutation to extragenic suppres-

sion of unc-93(e1500) is about 3 x 1O"j (GREENWALD and HORVITZ 1980).

In summary, the frequency of recombination in hermaphrodites between mnDp62 and the X chromo- some in the unc-3-sup-IO interval is roughly 1/100 that found for mnDp73 (and 1/200 that found for mnDp72) in the smaller unc-I-dpy-3 interval on the other end of the X chromosome.

DISCUSSION

For each of ten duplications carrying unc-I(+), a gene near the left end of the X chromosome, we screened for products of recombination between the duplication and the X chromosome in unc-I/O;Dp[unc- I ( + ) ] males. Our screen involved assaying both for X chromosomes that had picked up unc-I(+) from the duplication and for duplications that had picked up unc-1 from the X; the crossovers in these cases must have occurred to the right of unc-I. It is perhaps not surprising that the highest frequencies of recombina- tion were found for mnDp72 and mnDp73, the only two duplications that extend far enough to the right of unc-1 to include dpy-3(+). We found a somewhat higher frequency of recombinant X chromosomes for mnDp72 than for mnDp73 despite the fact that the latter duplication extends farther to the right than the former; this difference might be related to the fact that mnDp73 is a free duplication whereas mnDp72 is translocated. For both of these duplications the fre- quencies of recovery of both recombinant X chromo- somes and recombinant duplications in X/Dp males was a factor of ten or more higher for the unc-l-dpy- 3 interval than the frequency of X-X or X-Dp recom- bination in hermaphrodites in the same interval. Al- together, we found one or more products of X-Dp recombination between duplication and X chromo- some for seven of the ten unc-I(+) duplications; mnDp63(X;V) exhibited recombination frequencies nearly as high as those found for mnDp72 and mn-

Page 12: Chromosome in Caenorhabditis elegans

734 R. K. Herman and C. K. Kari

Dp73. Two other duplications, mnDp70(X;V) and mnDp72(Xf ) ) , showed lower but still appreciable re- combination frequencies. We do not know the relative rightward extents of the unc-I(+) duplications other than mnDp72 and mnDp73.

We looked at recombination between m n D p 7 3 ( X f ) and the X chromosome in males in more detail. The classes of recombinant duplications that we recovered were all most simply explained as the consequence of single crossovers between the X chromosome and a linear duplication. Crossovers in each of the following three intervals were represented: che-2-unc-1, unc-l- dpy-3 and dpy-?-right breakpoint; crossovers in the middle segment were much the most frequent.

The frequency of recombinant X chromosomes in mnDp72-bearing males for the unc-I-dpy-3 interval, which we measured as 1.3 map units in hermaphro- dites, was about 40%. Why is this frequency of recom- bination so high? We suggest that it may be a direct consequence of, first, very efficient pairing between mnDp72 and X , and second, the way in which ex- changes are controlled. It seems likely that in wild- type animals bivalents completely lacking a crossover are rare. This suggestion in the case of the X chro- mosome can be based on the following argument. First, the total extent of the X map is about 50 map units. Assuming that crossing over occurs at the four- chromatid stage, a single exchange generates two recombinant and two parental chromosomes. There- fore, the average number of X chromosome exchanges per meiosis is one. Moreover, double crossover prod- ucts, at least i n the ovum line, appear to be rare or nonexistent for the X chromosome (HODGKIN, HORV- ITZ and BRENNER 1979). The average of one exchange per X per meiosis can therefore only be achieved by the occurrence of precisely one exchange for one meiosis. The idea that every bivalent normally has at least one exchange is also consistent with the view, discussed below, that bivalents lacking an exchange are likely to lead to meiotic nondisjunction. We sug- gest from these considerations that there must be a mechanism for insuring that each X bivalent has one exchange. We suggest further that the efficient pair- ing of mnDp72 and the one X in males brings this mechanism into play, generating one crossover per mnDp72-X pair. The bulk of the crossovers seem to occur in the unc-I-dpy-3 interval, since about 40% of X chromosomes recovered are recombinant in this interval and the maximum expected recombination frequency over the full extent of the duplication is 50%. (The frequency of recovery of recombinant duplications was lower; we do not know why.) Similar nlechanisnls for insuring at least one exchange per bivalent may be operative for autosomes, although autosomal double crossovers have been detected (HODGKIN, HORVITZ and BRENNER 1979). We have found, for example, that the chromosome that carries

mnC1, a dominant crossover suppressor for much of chromosome ZZ (HERMAN 1978), exhibits enhanced recombination outside the region in which crossing over is suppressed (R. K. HERMAN, unpublished data). Similar intrachromosomal effects have been observed in Drosophila (LUCCHESI 1976).

In view of the very efficient pairing between mn- Dp72 and the X chromosome in males, it is easy to see how our mnDp72 male stock maintains its trisomy for chromosome ZV, the chromosome to which mnDp72 is translocated. We imagine that mnDp72 and the single X chromosome in males constitute one set of meiotic pairing partners, and the two normal chro- mosome ZVs constitute another. Sperm carrying mnDp72 then necessarily also carry a normal chro- mosome ZV, and they generate trisomic progeny when they fertilize haploid ova. Pairing between mnDp73 and the X in males appears also to be very efficient, because high recombination frequencies were found, but mnDp73 is a free duplication, so no autosomal effects are expected. The duplication showing the next highest recombination frequency in males was mnDp63, and it was also found in trisomic animals, in this case trisomic for chromosome V , to which mnDp63 is translocated. Although it is easy to imagine how trisomy for these duplications can be maintained, a separate question is whether the duplications enhance the incidence of trisomy in diploid animals. We sug- gest that they do on the basis of our earlier work with the translocation mnTI2 , which is a fusion of chro- mosomes ZV and X . Hermaphrodites of genotype ZV/ mnT12/X show high frequency meiotic nondisjunction both between m n T I 2 and X and between m n T I 2 and ZV (SIGURDSON et al. 1986). We suggest that ZV/mn- Dp72/X animals (males in this case), for example, would behave similarly and hence generate the chro- mosome ZV trisomics we detected.

We screened for EMS-induced autosomal mutations that markedly reduce X chromosome recombination in hermaphrodites and identified one mutation, mn253. The mutation reduced X recombination in the dpy-8-unc-3 interval to about 10% the wild-type value of 28%, and it proved to be an allele of the previously- identified gene him-8 (HODGKIN, HORVITZ and BREN- NER 1979), which promotes nondisjunction of the X chromosome in hermaphrodites to give a high inci- dence of male self progeny. Among the recessive him mutants described by HODGKIN, HORVITZ and BREN- NER (1979), him-8(e1489) showed the most severe reduction of X chromosome recombination, giving about '/R the wild-type value in the Lon-2-unc-7 interval; autosomal recombination did not appear to be af- fected. GOLDSTEIN (1982) has shown that him- 8(e1489) hermaphrodites have six synaptonemal com- plexes; this result suggests that the mutant gene does not affect X chromosome pairing. We therefore sug- gest that the primary defect in him-8 mutants is in the

Page 13: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C. elegans 735

process of exchange itself and that the defect in dis- junction is a consequence of the defect in exchange. In many species, exchange between homologs is essen- tial for orderly chromosome segregation at meiosis (BAKER et al. 1976). We suggest that this rule also applies to C. elegans. If so, any mutation that drasti- cally reduced recombination of all chromosomes would be expected to lead to a large proportion of inviable zygotes, owing to aneuploid chromosome compositions. Mutants that appear to be of this type have been found by KEMPHUES, KUSCH and WOLF (1 988) among the class of maternal effect lethals.

A curious property of the him-8 mutations is their specificity for X chromosome recombination. We measured recombination frequencies for two autoso- mal intervals; in both cases we found recombination was not reduced by him-8(mn253) but was in fact increased, by about 40-50%. We suggest that the increases are a consequence of the reduced X recom- bination. Similar interchromosomal effects on recom- bination frequencies are well documented in Drosoph- ila, although there is apparently no wholly satisfactory model to account for them (LUCCHESI 1976). No increase in recombination on LGV was apparent for him-8(eI489) animals, however (HODGKIN, HORVITZ and BRENNER 1979), and we cannot exclude the pos- sibility that the increases we have observed are due to a secondary mutation.

We found that him-8 severely reduced recombina- tion between mnDp73 and the X chromosome in males, eliminating the recovery of both recombinant duplications and recombinant chromosomes. Thus, whatever function him-8(+) provides for X-X recom- bination in hermaphrodites appears also to be needed for mnDp73-X recombination in males.

Several of the unc-I(+) duplications showed a tend- ency to disjoin from the single X chromosome during meiosis in the male, i.e., there was generated a pre- ponderance of sperm containing only an X chromo- some or only a duplication at the expense of X;Dp and nullo-X;nullo-Dp sperm (Table 5). The strength of this effect correlated well with the X-Dp recombina- tion frequency (Table 4). The disjunction effect may be at least partly a consequence of the pairing of duplication and chromosome, rather than of recom- bination per se. We previously saw a tendency for free autosomal duplications to disjoin from the X chro- mosome in males, even though there was presumably no homology or recombination between the elements (HERMAN, MADL and KARI 1979); it was suggested that there is a weak pairing, analogous to the distrib- utive pairing that occurs in Drosophila females (GRELL 1976), between the nonhomologous univalents. Con- sistent with this interpretation is our finding in the present work that when him-8 prevented recombina- tion between mnDp73 and the X (but presumably not pairing; see above), the two elements still tended to

disjoin from each other. These considerations lead to the suggestion that those unc-I(+) duplications-such as mnDp66(X;Z) and mnDp67(X;I)-that did not show a tendency to disjoin from the X chromosome in males may not have paired efficiently with the X chromo- some and therefore would also have shown very little recombination in the region to the left of unc-I, had we been able to measure recombination there.

We measured X-Dp recombination frequencies for mnDp72 and mnDp73 in hermaphrodites in the unc-1- dpy-3 interval and found values comparable to the hermaphrodite X-X recombination frequencies for the same interval. We interpret these results as follows. Pairing between mnDp72 and the one X chromosome in males, we have already argued, may be nearly 100% efficient and be responsible for the generation of 40% recombinant X chromosomes for the unc-I-dpy-3 in- terval. mnDp72-bearing hermaphrodites give rise to about 3.4% male self-progeny, which corresponds to an average of 1.7% nullo-X gametes per germ line. We suggest that the nullo-X gametes are generated from meioses in which there is good pairing between mnDp72 and one of the X chromosomes; the other X chromosome in these meioses behaves as a univalent. An unpaired X chromosome, in X 0 hermaphrodites transformed by a her-I mutation, is not transmitted to about 70% of the gametes (HODGKIN 1980). Hence, in those meioses in which mnDp72 and one X are well paired, we would expect disjunction of these two elements, with a nullo-X;mnDp72 genotype generated in about 35% (0.7 X 0.5) of the gametes. Thus, a frequency of 5% meioses of this type could account for the 1.7% nullo-X gametes. Two predictions about the male self progeny follow from this picture: first, we would expect nearly all the males to carry mnDp72; among the 73 males counted during the scoring of hermaphrodites for Table 7, 68 (93%) were either wild-type or Unc non-Dpy males, all of which must have carried mnDp72. A second prediction is that the proportion of recombinant duplications carried by these males should have approximated the high X-Dp recombination frequency found in males: indeed, the observed frequency was 13/68 (19%). According to this picture, only about 5% of the meioses in her- maphrodites are capable of generating recombinant duplications and recombinant X chromosomes, which explains why the X-Dp recombination frequencies are much lower in hermaphrodites than in males. To summarize our picture, in hermaphrodites there is competition for pairing; in about 5% of the meioses there is pairing between mnDp72 and an X chromo- some, and in the other 95% of the meioses, pairing occurs between the two X chromosomes. Trivalents consisting of mnDp72 and two X chromosomes either are not formed or do not prevent nondisjunction of the two X chromosomes. The incidence of male self- progeny is thus a measure of the efficiency of X-Dp

Page 14: Chromosome in Caenorhabditis elegans

736 R. K. Herman and C. K. Kari

pairing in hermaphrodites. We therefore suggest that mnDp63 and mnDp70 are better competitors for pair- ing with an X in hermaphrodites than is mnDp72 (Table I); by comparison with mnDp72, we estimate that mnDp63 and mnDp70 pair with an X in about 15% and 8% of the meioses, respectively. The rela- tively high efficiency of pairing by these duplications in hermaphrodites may be related to the fact that both are situated near the left end of chromosome V. We could imagine, for example, that there is a tendency for the left ends of chromosomes V and X to be in proximity prior to meiotic pairing. Both mnDp73 and mnDp71 seem to fare rather badly in their ability to pair with an X in hermaphrodites (Table 1) compared with their ability to recombine with the one X in males (Table 4); this result could be related to the fact that both are free duplications.

Recombination between duplications of other re- gions of the X chronlosome and the X seems to be rare, in both males and hermaphrodites. In earlier work (HERMAN, MADL and KARI 1979) we detected very little if any recombination between the X and duplications of the right end of the X , in either males or hermaphrodites. We also were unable to detect recombination in hermaphrodites between the X and mnDp30, a free duplication that covers lon-2, dpy-8 and dpy-7 (Figure 1). In the present work we were unable to detect recombination in males between the X and mnDp57, a fairly large duplication (Figure 1) of the region near but not including the left end of the X . We also conducted additional screens for recom- bination between the X and duplications of the right end of the X . We detected less than one percent recombination between mnDp3 and the X chromo- some in males in an 8 map-unit interval; mnDp3 is large, about 30% of the X map. We were unable to detect recombination between mnDp3 and an X chro- mosome in hermaphrodites. By using a specially con- structed duplication, mnDp62, which carries a sup-IO mutation, we were able to recover recombinant chro- mosomes in hermaphrodites at a frequency of about

which is about two orders of magnitude less frequent than recombinant chromosomes were found in hermaphrodites carrying mnDp72. We suggest that the very low recombination frequencies found for many duplications are due to the rare pairing of these duplications with an X chromosome. Consistent with this view is the fact that none of these duplications promotes meiotic nondisjunction of the X chromo- somes in hermaphrodites.

The efficient pairing of several unc-I(+) duplica- tions with the X, particularly in males but also to a lesser degree in hermaphrodites, where the duplica- tion must compete with another X for pairing, leads to the suggestion that one or more sites that promote synapsis (HAWLEY 1980) are preferentially situated near unc-I(+). We do not know why there was consid-

erable variation in the frequencies of recombination among the different unc-I(+) duplications, although, as we have already suggested, we suspect that the efficiencies of pairing differed. Possible factors affect- ing pairing include the number of pairing sites carried by each duplication, the relative extent of each dupli- cation, whether or not the duplication carries an in- ternal aberration that could disrupt pairing, and whether the duplication is free or translocated, and if translocated, where.

Finally, we note that all eight of our new translo- cated duplications seem to be situated near autosomal ends. Several previously-identified duplications are situated near the right end of chromosome I , which is where three of the unc-I(+) duplications and mnDp57 map. Perhaps the ribosomal gene cluster, which maps to the right end of Z (ALBERTSON 1984) is a favored site for radiation-induced breakage. All of the dupli- cations mapping near the right end of Z are homozy- gous viable. Other unc-I(+) duplications mapped near the left end of I (mnDp65) , the right end of IV (mn- Dp72) and the left end of V (mnDp63 and mnDp70). All of these duplications, with the possible exception of mnDp72, were also homozygous viable. The trans- location m n T l 2 (SIGURDSON et al. 1986) is a homozy- gous viable fusion of the left end of ZV with the right end of X. Perhaps a common feature of the re- arrangement .junctions near chromosome ends will become apparent when the DNA sequences of the junctions are known.

We thank STEVE P R A n and MARCIA SCHUYLER for technical help. This work was supported by National Institutes of Health (NIH) grant GM22387. Some strains were obtained from the Cae- norhabditis Genetics Center, which is supported b y Contract NO1 KR-4-21 1 1 between the NIH Division of Research Resources and Cul-;ators of the Llniversity of Missouri.

LITERATURE CITED

ALBERTSON, D. G. , I984 Localiration of the ribosomal genes in Caenorhabditis elegans chronlosornes by in situ hybridization using biotin-labeled probes. EMBOJ. 3: 1227-1234.

ALBERTSON, D. G., and J. N. T H o M s O N , 1982 The kinetochores of Caenorhabditis elegans. Chromosoma 86: 409-428.

ANDERSON, I>. , and S. BRENNER, 1984 A selection for rnyosin heavy chain mutants in the nematode Carnorhabditis elegans. Proc. Mall. Acad. Sri. 1JSA 81: 4470-4474.

BABU, I] , , 1974 Biochenlical genetics of (hrnorhabditis elegans.

BAKKR, B. S . , A . T. <:. CARPENI'ER, M . S. Es~osrro, K. E. Esposrro ;and L. SANDIXR, 1976 Thr genetic control ofmciosis. Annu. R e v . Genet. 10: 53-134.

BRE;NNER, S.. 1974 The genetics of Caenorhabditis elegans. Ge- netics 77: 7 1-94,

~ I , o N c ; , I.., I.. 1'. CASSON and E. J. MZ:YKR. I987 Assessment of chromosome dosage conllmlsation i n Carnorhabditis elegans by phenotypic ;rnalysis oflin-14. Genetics 117: 657-670.

I)ONAHUE. L . M . . B. A . QUARANI'ILLO and W . B. WOOD, 1987 Molecul;ll- analysis of X c h r o ~ ~ m s o ~ ~ a e dosage ro111pt.11- satioll i n Caenorhabditis elegans. l'roc. Natl. Acad. Sci. [JSA 84: 7600-7604.

FIRE, :I., 1986 Integrative transformation of O'aenorhabditis d e - g a m . EMB0.J . 5: 2673-2680.

Mol. (;en. Genet. 135: 351-44.

Page 15: Chromosome in Caenorhabditis elegans

Dp-X Recombination in C. elegans 737

( ;OI,D~I 'EIN, P., 1982 The synaptonemal complexes of Caenorhab- diti.? elegans: pachytene karyotype analysis of male and hrr- tnaphrodite wild-type and him mutants. Chromosoma 86: 577- 593.

<;REENWAI.I), I . S., and H. R. HORVITZ, 1980 unc-Y?(e1500): a behavioral mutant of Caenorhabditis elegans that defines a gene with a wild-type null phenotype. Genetics 96: 147-164.

(;REENWALD, 1. S., P. W. STERNBERG and H. R. HORVITZ, 1983 The l in -12 locus specifies cell fates in Caenorhabditis elegans. Cell 34: 435-444.

(;REI,I., R. F . , 1976 Distributive pairing, pp. 436-486 in The Genetics and Biology of Drosophila, Vol. la , edited by M. ASH-

FK~RNER and E. NOVITSKI. Academic Press, New York. HAWLEY, R. S. , 1!)80 Chrotnosomal bites necessary for nortnal

levels of nleiotic recombination in Drosophila melanogaster. I . Evidence for and mapping of the sites. Genetics 94: 625-646.

HEDGECOCK, E. M., J. G. CuLon.1, J. N. THOMSON and L. A. PERKINS, 1985 Axonal guidance mutants of Caenorhabditis elegans identified by filling sensory neurons with fluorescein dyrs. DKV. Biol. 111: 158-170.

HERMAN, R. K.. 1978 Crossover suppressors and balanced reces- sive lethals in Caenorhabdztis elegans. Genetics 88: 49-65.

HERMAN, R. K., 1989 Mosaic analysis i n the nematode Caenorhab- ditis elegans. J. Neurogenet. (in press).

HERMAN, R. K., D. G. ALBERTSON and S. BRENNER, 1976 Chromosome rearrangements in Caenorhabditis elegans. Gelletics 83: 91-105.

HERMAN, K. K., C. K. KARI and P. S. HARTMAN, 1982 Donlinant X chromosome nondisjunction nlutants of Caenorhabditis ele- gans. Genetics 102: 379-400.

HERMAN, R. K., J. E. MADL and C. K. KARI, 1979 Duplications in (henorhabditis elegans. Genetics 92: 4 19-435.

HODGKIN, J., 1980 More sex-determination mutants of Caenor- habditis elegans. Genetics 9 6 649-664.

HODCKIN, J., 1983 Male phenotypes and mating efficiency in Caenorhabditis elegans. Genetics 103: 43-64.

HODGKIN. J . , H. K. HORVITZ and S. BRENNER, 1979 Nondisjunction mutants of the nematode Caenorhabditis elegans. Genetics 91: 67-94.

HORVITZ. H. R., S. BRENNER, J. HODCKIN and R. K. HERMAN, 1979 A uniform genetic nomenclature for the nematode Caenorhabditis elegans. Mol. Gen. Genet. 175: 129-133.

HOWF.I.I., A. M., S. G. GILMOUR, R. A. MANCEBO and A. M. ROSE, 1987 Genetic analysis of a large autosomal region by the use of :I free duplication. Genet. Res. 49: 207-2 13.

JOHNSON, C. D.,J. G. DUCKETT, J. G. CULOTTI, R. K. HERMAN, P. M. MENEELY and R. I.. RUSSELL, 1981 An acetylcholinester- ase-deficient mutant of the nematode Caenorhabditis elegans. Genetics 97: 261-279.

KEMPHUES, K. J., M. KUSCH and N. WOLF, 1988 Maternal-effect lethal mutations on linkage group I 1 of Caenorhabditis elegans. Genetics 120: 977-986.

LUCCHESI, J . C., 1976 Interchron~osonlal effects, pp. 315-329 in The Genetics and Biology of Drosophila, Vol. la , edited by M.

ASHBURNER and E. NOVITSKI. Academic Press, New York. MADL, J. E., and R. K . HERMAN, 1979 Polyploids and sex deter-

mination i n Caenorhabditis elegans. Genetics 93: 393-402. MAINLAND, D. L., L. HERRERA and M . I . SUTCLIFFE,

1956 Statistical Tables for Use with Binomial Samples, Contin- gency Tests, Confidence Limats and Sample Size Estimates. Depart- merlt of Medical Statistics, New York University School of Medicine, New York.

MENEELY, 1'. M.. and R. K. HERMAN, 1979 Lethals, steriles and deficiencies in a region of the X chromosome of Caenorhabditis elegans. Genetics 92: 99-1 15.

MENEELY, P. M., and R. K. HERMAN, 1981 Suppression and function of X-linked lethals and steriles i n Caenorhabditis ele- guns. Genetics 97: 65-84.

MENEELY, 1'. M., and K. D. NORDSTROM, 1988 X chromosonle duplications affect a region of the cl1rotnosotne they do not duplicate in Caenorhabditis elegans. Genetics 119: 365-375.

MENEELY, P. M., and W. B. WOOD, 1984 An autosomal gene that afftcts X chromosome expression and determination in Caenor- habditis elegans. Genetics 106: 29-44.

MENEEI-Y, P. M., and W. B. WOOD, 1987 Genetic analysis of X chro1nosome dosage compensation in Caenorhabditis elegans. Genetics 117: 25-41.

MEYER, B. J., AND L. P. CASSON, 1986 Caenorhabditis elegans compensates for the differences in X chron,osome dosage be- tween the sexes by regulating transcript levels. Cell 47: 871- 881.

PARK, E.-[;., and H. R. HORVIIZ, 1986 Mutations with dominant effects on the behavior and morphology of the nematode Caenorhabditis elegans. Genetics 113: 82 1-852.

PERKINS, L. A,, E. M. HEDGECOCK, J. N. THOMSON and J. G. CULOTTI, 1986 Mutant sensory cilia in the nematode Caenor- habditis elegans. Dev. Biol. 117: 456-487.

ROGALSKI, T. M., and D. R. RIDDLE, 1988 A Caenorhabditis ele- gans RNA polymerase 11 gene, ama-1 I V , and nearby essential genes. Genetics 118: 6 1-74.

ROSE, A. M., D. L. BAILLIE and J . CURRAN, 1984 Meiotic pairing behavior of two free duplications of linkage group I in Caenor- habditis elegans. Mol. Gen. Genet. 195: 52-56.

ROSENBLUTH, R. E., C. CUDDEFORD and D. L. BAILLIE, 1985 Mutagenesis i n Caenorhabditis elegans. 11. A spectrum of mutational events induced with 1500 R gamma radiation. Genetics 109: 493-5 1 1.

ROTHSTEIN, R. J . , 1983 One-step gene disruption in yeast. Merh- ods Enzytnol. 101: 202-2 11.

SIGURDSON, D. C., R. K. HERMAN, C. A. HORTON, C. K. KARI and S. E. PRAIT, 1986 An X-autosome fusion chromosome of Caenorhabditis elegans. Mol. Gen. Genet. 202: 21 2-218.

WOOD, W. B. (Editor), 1988 The Nematode Caenorhabditis elegans. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

WOOD, W. B., P. MENEELY, P. SCHEDIN and L. DONAHUE, 1985 Aspects of dosage compensation and sex determination i n Caenorhabditis elegans. Cold Spring Harbor Symp. Quant. Bid. 50: 575-583.

<:ommunicating editor: R. L. METZENBERC