Copyright 0 1990 by the Genetics Society of America
Chromosome I Duplications in Caenorhabditis elegans Kim S. McKim and Ann M. Rose Department of Medical Genetics, University of British Columbia, Vancouver, BritishColumbia, Canada Manuscript received April 26, 1989 Accepted for publication September 29, 1989 ABSTRACT We have isolated and characterized 76 duplications of chromosome I in the genome of Caenorhabditis elegans. The region studied is the 20 map unit left half of the chromosome. Sixty-twoduplications were induced with gamma radiation and 14 arose spontaneously. The latter class was apparently the result of spontaneous breaks within the parental duplication. The majority of duplications behave as if they are free. Threeduplications are attached to identifiable sequences from other chromosomes. The duplication breakpoints have been mapped by complementation analysis relative to genes on chromosome I. Nineteen duplication breakpoints and seven deficiency breakpoints divide the left half of the chromosome into 24 regions. We have studied the relationship between duplication size and segregational stability. While size is an important determinant of mitotic stability, it is not the only one. We observed clear exceptions to a size-stability correlation. In additionto size, duplication stability may be influenced by specific sequences or chromosome structure. The majority of the duplications were stable enough to be powerful tools for gene mapping. Therefore the duplications described here will be useful in the genetic characterization of chromosome I and the techniques we have developed can be adapted to otherregions of the genome.
HARACTERIZATION of theCaenorhabditis eleguns genome is progressing at a rapid rate. The progress has been recently stimulated by the physical mapping of contigs which are tied to the genetic map (COULSON et al. 1986, 1988). Parallel to the physical mapping are the genetic mapping effortsof C. elegans researchers who are isolating mutations and finding their positions on both the genetic and physical maps. A few labs have undertaken the task of genetically characterizing large regions of the genome with regard to the organizationof essential genes (MENEELY and HERMAN 1981; SIGURDSON, SPANIERand HERMAN 1984;CLARK et al. 1988; ROSENBLUTH et al. 1988). The present study was undertaken to provide two contributions toward our understanding of chromosome I organization. The first was to make rearrangements to be used in the mapping of genes in a large region of chromosome I . Previously HOWELL et al. (1987) described a system for identifying complementation groups on the left half of chromosome I using the duplication sDp2. The large number of mutations that have been isolated requires the generation of rearrangements in order to increase the mapping resolution and reduce the number of complementation tests. The ultimate goal of these studies is to correlate the physical and genetic maps; that is to identify the mutant phenotype for coding regions in the cloned DNA.
C
T h e publication costs of this article were partly defrayedby the payment of page charges. This article nust therefore be hereby marked "advertisement" in accordance with I8 U.S.C. 91734 solely to indicate this fact. Genetics 124: 1 1.5- 132 (January, 1990)
The second contribution of this study was to use these rearrangements to study the physicalbasisof chromosomebehavior.Earlier studies with translocations (ROSENBLUTH and BAILLIE1981; MCKIM, HOWELL and ROSE1988) and duplications (HERMAN, MADLand KARI 1979; HERMAN and KARI 1989; ROSE, BAILLIEand CURRAN 1984)have demonstrated the utility of this approach. A large numberof rearrangement breakpointshave been isolated as gamma radiation-induced or spontaneous shorter derivatives of preexisting duplications. We used two duplications to isolate our derivatives. The first duplication, sDp2(Z;f), was initially characterized by ROSE, BAILLIEandCURRAN(1984)and et al. (1 987) to recoverlethal mutaused by HOWELL tions. The second duplication is a chromosome segregated from a reciprocal translocation strain. This chromosome, I L X L s z T l (McKIM, HOWELLand ROSE 1988), is one half of the s z T I ( I ; X ) translocation (FoDOR and DEAK1985) and is used here in addition to the euploid genetic background.I L X L s z T l is somewhat a largerthan sDp2 and is attachedatoneendto fragment of the X chromosome.Neither of these duplications recombine with the normal chromosome I. This paperdescribes the characterization of 76 new duplications with regardtothe genetic position of their breakpoints and their meiotic/mitotic properties. Most of these duplications are free; that is they are not translocated toanother chromosome. The duplications contribute at least 19 breakpoints, which
116
K. S. McKim and A. M. Rose
in combination with seven deficiency breakpoints, divide the left half of chromosome Z into at least 24 regions. Using these duplications we have studied the relationship between chromosome size and segregationalstability. In addition, we have observed that free duplicationsspontaneouslybreakwith a frequency that correlates with their segregational instability. MATERIALS AND METHODS
General: Nematodes were grown and maintained on nematode growth medium streaked with Escherichia coli strain OP50 (BRENNER 1974). Unless otherwise noted, all experiments were carried out at 20" (ROSEand BAILLIE1979). The standard wild-type strain, N2, and some mutant strains Simon derived from N2 were obtained from D. L. BAILLIE, Fraser University, Burnaby, Canada or from the Caenorhabditis Genetics Center at the University of Missouri, Columbia, Missouri. fog-1 (q253ts) was received from K. BARTON, University of Wisconsin. sup-I I (n403n682)and nDf25 were received from R. H. HORVITZat MIT. spe-ll(hc90) was received from D. SHAKES, Carnegie Institutionof Washington. The nomenclature used in describing these strains is the uniform system outlined in HORVITZet al. (1979). The wild-type phenotype is oftenabbreviated Wt. The following mutant genes and alleles were used: I dpy-5(e61); dpy-5(h660); bli-3(e767);b M ( e 9 3 7 ) ; dpy14(e188);egl-30(n686); egl-30(n715);fog-l (q253ts); speI I (hc90); lin-6(h92); lin-I 7(n671);sup-I I (n403n682); unc-1I(e47); unc-I3(e450); unc-I4(e57); unc-29(e403); unc-35(e259); unc-37(e262); unc-38(e264); unc-40(e271); unc-55(e402); unc-57(e406); unc-63(e384); unc73(e936); unc-74(~19); unc-89(e1460). 111 unc-36(e251). X dpy-3(e27);dpy-7(e88); dpy-7(sc27); dpy-8(e130); lonZ(e678); unc-3(e151); und"(e112). Most of the EMS induced lethal mutations on chromosome I were isolated and described by ROSE and BAILLIE (1980) and HOWELL et al. (1 987).Also included in this study are three mutations, let-373(h234), let-374(h25I) and let375(h259),mapped to the hDf6 region by HOWELL(1989). One additional mutation, let-400(h269),has been positioned between dpy-14 and unc-I3 by T. STARR(personal communication). The reciprocal translocation s z T l ( 1 ; X ) was originally described by FODORand DEAK(1985) and further characterized byMCKIM, HOWELLand ROSE(1988). The reciprocal translocation hT2(I;Ill) was isolated by K. PETERS. AS a heterozygote it suppresses recombination on chromosome I from the left end to the right of unc-29. The dpy-5(h660) mutation is linked to hT2(I;Zll) and was induced by J. BABITY. The deficiencies used in this study were hDf6 (McKIM, HOWELL and ROSE1988), nDf23, nDf24 and nDf25 (FERGUSON and HORVITZ 1985)andsDf4 (ROSE1980; HOWELLet al. 1987). Some characteristics of sDpZ(1;f) (ROSE, BAILLIEand CURRAN 1984) and ILXLszTI (McKIM, HOWELL and ROSE1988) aredescribed in RESULTS. Mutagenesis with gamma radiation:The screening protocols are described in RESULTS. Gravid worms were treated with either 1500 or 3000 rad of gamma radiation from a cobalt-60 source (as recommended by ROSENBLUTH, CUDDEFORD and BAILLIE1985). The dose rate varied between 9.0 and 7.5 rad/sec. Levamisole selection: Some shortened derivative duplications were selected on the basisof resistance to 1 mM levamisole usingplates prepared as described in LEWISet al.
(1980). In general, two or three unc-74 dpy-5; Dpx or unc63 dpy-5; Dpx hermaphrodites were placed on each small Petri plate. When the Fz progeny were larvae, they were washed off in M 9 buffer (BRENNER 1974) and seeded onto new large (100 X 15 mm) plates containing 1mM levamisole. Two or three days later the plates were screened for levamisole resistant non-Dpy worms. Survivors were placed onto normal plates to observe their progeny. Recombinationmapping: Recombination frequencies between pairs of markers were determined by scoring the self progeny of cis-heterozygous hermaphrodites under the conditions described by ROSEand BAILLIE (1 979).For standard crosses, the total number of progeny was calculated as 4/3 (the number of Wt plus one recombinant class). The recombination frequency, p , was calculated using the formula p = 1 - (1 - ZR)", where R is the fraction of marked (non-Wt) recombinant individuals over total progeny (BRENNER 1974). Map distances are reported as map units (m.u. = loop). Confidence limits of 95% were calculated using the statistics of CROWand GARDNER (1959). Special consideration was given to the calculation of recombination in crossesinvolving duplications. When scoring recombination from the progeny of dpy-x unc-y/+ Dp(1;f ) hermaphrodites, the calculation was modified for the fraction of gametes which carried the duplication. In most cases this was 41-42% (see RESULTS). The recombinant frequency was calculated as
+;
p
= [(0.17W - 0.170)
- ((0.170 - 0.17W)'
- 4(0.09W + 0.03D)(0.74D))"]/2(0.09W + 0.030)
+
where W was the number of wild types (or Wt Unc-13 with sDp2). Depending on the cross, D was the number of Dpy or Dpy Unc progeny. The frequency at which a duplication was found in the germ line was calculated by scoring the progeny of m/m/ Dpx hermaphrodites; where m was dpy-5, unc-I1 or unc-13. The fraction of duplication bearing gametes was x/(2 x); where x = Wt/mutant. In these calculations the duplication homozygotes were not considered as they were not scored. These worms were usually small and slow growing. Complementation tests: Visible mutation with duplication [Dp(Z;f)]:Mostof the mapping of the duplications with respect to visible markers was done using markers linked to either dpy-5, dpy-14 or unc-13. The exception to this was the mapping of Dp[unc-1I (+) dpy-5(-)] duplications. This mapping was done using visible mutations linked to unc-11. T o facilitate mapping, most duplications isolated with chromosomes other than dpy-5dpy-14 were resegregated into dpy-5 dpy-14;Dp strains. In order to map a mutation (unc-x),the progeny of uncx/+; Dp(Z;f)hermaphrodites were examined to see if their unc-x; Dp progeny were Unc (i.e., Dp is unc-x(-)) or wild type ( D p is unc-x(+)). Most parental hermaphrodites were constructed by crossing unc-x bearing males to a duplication stock. For example, unc-x dpy-5 hermaphrodites were crossed to unc-11 dpy-14; O / s z T I ( I ; X ) [ + ;lon-21 males. The only resulting male progeny (+ unc-xdpy-5 +/uric-I1 ++ dpy-14) were crossed to duplication strains. In crosses to dpy-5 dpy-14;h D p z ( 1 ; f )strains, most wild-typehermaphrodite progeny were unc-x dpy-5 +/+ dpy-5 dpy-14; hDpz. If the duplication was [dpy-14(+)], then + dpy-5 dpy-14 / unc-1I + dpy-14; hDpz hermaphrodites were also recovered. These possibilities were distinguished by examining the resulting self fertilization progeny. Upon examining the progeny of the desired (unc-x dpy-5 +/+ dpy-5 dpy-14; hDpz) hermaphrodites, Dpy-5 and Unc-x progeny were observed if the duplication did not carry unc-x(+)(in the ratio 4 Wt: 2 Dpy-
+
Duplications in C. elegans
117
5: 2 Dpy-14: 2 Unc-x: 1 Dpy-5 Dpy-14: 1 Dpy-5 Unc-x). No used to indicate themarkers which are deleted. Unc-X progenywereobserved if the duplication carried Dp[dpy-5(+) dpy-14(-)]refers to a duplication which unc-x(+). The only exceptions to the above procedure recarries the wild-type dpy-5 allele but is deleted in the sulted from recombinant chromosomes. dpy-14 region. This is not meant to indicate the duIn crosses of unc-x dpy-5 +/uric-1 1 dpy-14 males to unc-11 dpy-5; h D p z ( 1 ; f ) strains (where hDpz was u n c - l l ( - ) plication carries a dpy-14 mutation. dpy-5(+)),most wild-type hermaphrodite progenywere uncInduceddeletion of duplicationchromosomes: 1 1 dpy-5 dpy-5 dpy-14; hDpz(1; f ). For crosses tounc-11 Duplications of different sizes with one end in comdpy-5; hDpz(1;f ) strains (where hDpz was u n c - l l ( + ) dpymon would provide a linear array of breakpoints for 5 ( - ) ) , the mutation to be tested was linked to unc-11 and rapidmapping of mutations. T o generate such an brought in through the male of genotype unc-11 unc-x +/ ++ dpy-5. Usingsimilarmethodology, dpy-14 unc-13; array, we used gamma radiationto shortenpreexisting hDpz[dpy-5(-) unc-I3(+)] duplications were crossed touncduplications. Shortened duplications were recovered x dpy-14 +/++ unc-13 males when mapping a visible mutaby screening for the exposureof mutant phenotypes tion. previously rescued by the duplication. The protocol In some of the complementation experimentsunc-x dpyfor one such experiment is diagrammed in Figure 1. 5 / + dpy-5; hDpz hermaphroditeswere generated by crossing duplicationbearingmalestohomozygous mutanthersDp2 derivatives: In the first screen (Figure l), demaphrodites. For example,h D p l 8 males were generated by rivatives of sDp2 were recovered which had breaks in ; males todpy-5 dpy-14; crossing d p y - 5 ; O / s z T l ( I ; X ) [ +lon-21 the unc-ll-dpy-5 region. T h e Po had a wild-type pheh D p l 8 hermaphrodites.Wild-type malesfromthiscross notype and segregated Wts and Dpy Uncs. T h e F I (dpy-5 + /dpy-5 dpy-14; hDpl8)were crossed to unc-x dpy-5 progeny of gammaradiation treated unc-11dpy-5/ hermaphrodites to generate unc-x dpy-5/ + dpy-5; hDpl8 hermaphrodites. worms were screened for rare Uncunc-11 dpy-5; sDp2 sDp.2-balancedlethal with duplication [Dp(Z;f ) ] : Lethal 11 and Dpy-5 individuals. Because no recombination bearing strains (i.e., dpy-5 let-x unc-13; sDpZ(1; f )) were has been observed between the normal chromosomes crossed tounc-11 dpy-14;O/ szTl (I;X)[ lon-21 males. The and CURRAN 1984), the rare and sDp2 (ROSE,BAILLIE resulting male progeny (+ dpy-5 let-x unc-l3/unc-ll dpy-14 +) were then crossed to dpy-5 dpy-14; hDpz hermaUnc-1 1s and Dpy-5s could have resulted from: (1) phrodites. Maleprogenycarrying sDp.2 have a distinctive shorter duplications derived from a break between phenotype and do not mate. Wild-type hermaphrodite progu n c - l l ( + ) and dpy-5(+) on sDp2, (2) a point mutation eny from this cross were dpy-5 let-x unc-13/dpy-5 dpywithin unc-11 or dpy-5 on sDp2, (3) somatic loss of dpy-5 dpy-14 /unc-11 + dpy-141 hDpz if 14 +/ hDpz or sDp2 resulting in a mosaic worm, or (4) intragenic hDpz was dpy-14(+). These twoclasseswere distinguished by examining their progeny. If the duplication carried letrevertants.Thirteen exceptional events were rex(+), Dpy-5 and Unc-13 progeny were observed in the ratio covered. T h e complementation testing described be6 Wt: 2 Dpy-5: 1 Dpy-5Dpy-14: 2 Unc-13: 2 Dpy-14. low revealed 12 class (1) mutations and one class (2) hDp[dpy-5(-) unc-l3(+)] strains were tested by generating mutation (sDp2 [ dpy-5( h585)l). Four duplications with dpy-5 let-x unc-131 dpy-14 unc-13; hDpz hermaphrodites and looking for viable Dpy-5 progeny. These hermathe right end and eightduplications with the left end unc-13; phrodites were recovered from crossing dpy-14 of sDp2 deleted were recovered(Table1).Intwo let-x unc-13 hDpz hermaphrodites to dpy-14 ++/dpy-5 cases, two duplications were recovered fromthe same males. Po. hDp3 and hDp7 were isolated from the same Po Lethal with. lethal or deficiency: These complementation tests requiredthat the two lethal mutations be tightly linked and hDp5 and hDp21 were isolated from the same Po. to a common visible marker. Mostof the lethal mutations hDp3 and hDp7 have identical breakpoints while the or deficiencies isolatedin this lab were linked to dpy-5. Any latter two are different. Both of these situations could that were not had a dpy-5 marker put on the chromosome have beencaused by a premeiotic mutationwhich was, by recombination.Complementationtestsweredone by in the case of hDp21, accompanied by a spontaneous crossing one lethal heterozygote to another lethal or deficiency heterozygote and looking for Dpy-5 males and fertile postmutation shortening (see below). Dpy-5 hermaphrodites in the F1 (or Dpy-5 Unc-13 if both A similar approach was used to obtainbreaks in the lethals were also linkedto unc-13). dpy-5 dpy-14 region. In this screen, dpy-5 dpy-l4/dpy5 dpy-14; sDp2 worms were treated with gamma raRESULTS diation and their F1 progeny screened for rareDpy-5 and Dpy-14 worms. T h e same possibilities as outlined T h e primary goal of this research was to identify a above apply to this screen. Eight Dpy-14 worms were set of mutations which could be used in the physical has recovered(Table 2). Complementationtesting characterization of chromosome I in C. elegans. We describe here theisolation of duplications with breaks shown all of these to be shortened derivatives of sDp2. in the left half of chromosome I . These were mapped Inaddition,one unlinkeddominant Dpy mutation with respect to known visible and lethal mutations in ( h 6 3 0 ) was recovered. Unlike theprevious screen, no this region. Finally we have characterized these dupliduplications with the left end (dpy-5) deleted were cations with respect to segregational stability and recovered. structural integrity. When describing duplication ILXLszTl derivatives: A third screen using a different chromosomes, the following terminology has been duplication, ZLXLszTl, was used. This duplication is
+
++
+/+
+; +
+
+
+
+
++
+
+
+
+
+
K. S. McKim and A. M. Rose
118 sDpZ(Z;f)
unc;ll dpy-5 unc- 11 dp7-5 Treat with gamma r a d i a t i o n
I
FIGURE 1 .-Protocol for isolating shortened duplications. This figureillustrates the experiment summarized in Table 1.
1-
= =
" "
unc-11 dpy-5
11
unc-
dpy-5
"
"
unc-11 dpy-5
unc-11 dpy-5 "
"
Uric- 11
DPY - 5 TABLE 1
TABLE 2
Isolation of duplications breaking between unc-11 and dpy-5
Isolation of duplications breaking between dpy-5 and dpy-14
Dose
1500 R 3000 R
Dp[ull
+ d5-]
1 3
Dp[ull-d5+] Chromosomes
4 4
16,000 7,900
Rate
3.1 8.9
X X
one of the translocation chromosomes fromszTl (I;X) that has been recovered in a strain where it exists in addition to the euploid complement of chromosomes (McKIM, HOWELL andROSE 1988). The nomenclature describes thestructure of this chromosome. I"X"szT1 carries the portion of chromosome I left of unc-29 ( I L )joined to a fragment of the X-chromosome (X"). This X-chromosome portion carriesunc-I (+) and dpy-3(+) butnot unc-20(+) (McKIM, HOWELLand ROSE1988). dpy-5 unc-13; I"X"szT1 worms are wildtype and Him (High incidence of males) due to X nondisjunction. The dpy-5unc-13; I"X"szT1 worms were treated with gamma radiation and their F1 progeny were screened forrare Dpy-5 and Unc-13 individuals. Twenty-four Unc-13 and 18Dpy-5 hermaphrodites were recovered (Table 3). Complementation mapping (below) showed 23 of the Uncs and all of the Dpys were shorter derivatives of ILXLszTI.One Unc13 strain carried no dpy-5 marker. Two dpy-5(+) dpy1 4 ( - ) duplications, hDp40 and hDp63, caused the hermaphrodites to be sick and slow growing. dpy-5 dpy-14; Dp strains could not be constructed for these two duplications and therefore no extensive mapping of their breakpoints was done. Unlike the sDp2 screen of this region, we recovered duplications with the left end (dpy-5) deleted. Included in Table 3 as entries in columns two and three is a worm which segregated Unc-13, Dpy-5, Wt and Dpy-5 Unc-13progeny. The Wts were picked and these segregated more Wt, Unc-13, Dpy-5, and Dpy-5 Unc-13 worms. We were able to isolate two
Dose
1500 R 3000 R
D p [ d 5 + d l 4 - ] D p [ d Y d l 4 + ] Chromosomes
5 3
0 0
4500 2500
Rate
1.1 X lo-' 1.2 X 10"
duplications from the Wt strain, the Unc-13 worms carried an unc-13(-) duplication ( h D p 6 4 ) while the Dpy-5 worms carried a dpy-5(-) duplication ( h D p 7 3 ) . Since hDp64 and hDp73 have similar breakpoints, they could be products of the same mutational event. The recovery frequency for theILXLszTlderivative duplications (Table 3) is approximately the same as that for the sDp2 derivative duplications in the dpy-5 dpy-14 screen (Table 2). The duplications with breaks in the unc-I I dpy-5 region were recovered four times less frequently (Table 1) thanduplications with breaks in the dpy-5 unc-13 region. Spontaneous derivatives: A fourth source of duplications was the spontaneous shortening of various duplications. Their isolation is described laterunder "SegregationAnalysis." We have characterized 14 of these events and their analysis is reported with the analysis of the gamma ray induced duplications. Mapping duplication breakpoints: sDp2-balanced mutations: HOWELLet al. (1987) used sDp2 to isolate and characterize 58 EMS inducedmutations in 40 essential genes. These authors also reported the isolation of 29 gamma ray induced mutations of essential genes. None of these mutations were found to delete more than one locus. In the present study the latter mutations are indicated by allele numbers h 1 7 I and below. Genes identified by an EMS induced mutation are indicated by their name. All sDp2-rescued mutations were recovered and maintained by providing the wild-type gene copy in the form of the free duplication. Chromosome Z genes identified by HOWELLet
Duplications in C . eleguns TABLE 3 Isolation of duplications breaking betweendfiy-5 and unc-I3 Dose
1500 R a b 3000 R a b
Dp[dS+u13-] Dp[dS-u13+]
Chromosomes
Rate
46006 6
4 3
4600
2.2 X lo-’ 1.9 X lo-’
9 3
8 3
5000 4700
3.4 X lo-’ 1.3 X lo-’
al. (1 987),ROSEand BAILLIE(1 980) andEDGLEY and RIDDLE(1 987)were tested for complementationwith our duplications. Translocation(szTI(Z;X))balancedmutations: We previously described the properties of the translocation s z T l ( Z ; X ) (McKIM, HOWELLand ROSE 1988) a reciprocal interchange between chromosomes Z and X . Thatreport describedaset of chromosome I gamma ray induced lethal mutations balanced by s z T l ( Z ; X ) . Only one deficiency, hDf6, was found in that set. The rest were point mutations at the resolution of the mapping. Mutations designated h546 and above are from this set. Mapping strategy: Complementationtesting using autosomal duplications requires observation of the F2 generation to determine if the duplication carries the wild-type copy ofthe gene. The object of our protocol was to selectively recover duplication-bearing strains from a cross between a duplication strainand a strain carrying the mutation to be tested. The F2 progeny from these crosses were observed to see if the duplication complemented the tested mutation. The following protocol allowed for rapid mapping of large numbers of mutations. If the male was of the correct genotype, then in a cross to a duplication hermaphrodite,all the Wt progeny carried the duplication and at least half of these also carried the mutation to be tested. We found the pseudolinkage betweenchromosomes Z and X in s z T I ( Z ; X ) heterozygotes to be useful for producing therequired males. s z T l ( Z ; X ) heterozygous males have a Lon-2 phenotype (FODORand DEAK1985) and carry only therearranged X chromosome. These males cannot sire hermaphrodite progeny with a paternal normal chromosome Z. For the same reason these males sire only male progeny which carry the paternal normal chromosome I. The Lon males were used togenerate males truns-heterozygous forthe mutation a to be tested and anothermutation b in the duplicationstrain (MATERIALS AND METHODS). The mutation to betested was linked to a common marker c. The trans-heterozygous males ( a c b ) were then crossed to duplication strains ( c b; Dp).With the exception of rare recombinants, at least half the wildtype hermaphrodite progeny carried both the duplication and themutation to be tested ( a c +/+ cb; Dp).
+/++
119
If the duplication complementedthe mutation a , then a non-mutant a c; Dp strain could be established. Mappingresults: Duplications hDp8, hDp9, hDpl0 and h D p l 1 were mapped using mutations linked to unc-11 (e47).These four have a single breakpoint in the unc-73 dpy-5 region. Their left endpoint is probably the same as sDp2 because all four complement bZi-3. The rest of the duplications were mapped with mutations linked to dpy-5(e61), dpy-l4(e188)or unc13(e450). Considerable mapping against t h e essential genes of HOWELLet al. (1987) in this region was carried out. We also mapped 16 gamma ray induced mutations of essential genes [ 11 isolated with szTI (McKIM, HOWELLand ROSE 1988) and five isolated with sDp2 (HOWELL et aZ. 1987)Jwith the duplications; these are indicated by their allele designation ( h ) . The results of this mapping combinedwith the deficiencies in the region are shown in Figures 2, 3, 4 and 5 . Mapping of genes with respect to thedeficiencies sDf4 and hDf6 is published elsewhere (HOWELL et al. 1987; MCKIM,HOWELLand ROSE1988; HOWELL 1989). T h e duplications in combination with sDf4 and hDf6 have subdividedthe 4 m.u. unc-I1 unc-13region into 18 intervals. None of the duplications tested had a break in the unc-37-dpy-14 interval (Figures 4 and 5), even though this interval is recombinationally 1/5 of thedpy-5 unc-I3 region (EDGLEY and RIDDLE1987). Twelve of 41 ILXLszTl duplications broke in the dpyI 4 u n c - 1 3 region; an interval genetically 1/5 the size of the dpy-5 unc-13 region (ROSEand BAILLIE1979). We have mapped these duplication breakpoints with five mutations in the dpy-14 unc-13 region (Figures 4 and 5). Ofthe12 duplicationbreakpoints in this interval, seven had a breakpoint similar to sDp2 (between unc-14 and Zet-75) and the other five had different breakpoints. The breakpoint of hDp62 is now the closestphysical marker on the right of dpy-14. Two of the duplications breaking between let-75 and unc-13,hDp64 and hDp73, were isolated in the F, worm and, as mentionedearlier,might have been reciprocal products of the same mutational event. Spontaneous shortening of duplications in the region to the left of dpy-5 (see below) provided some new breakpoints. All duplications which break between unc-38lunc-63 and dpy-5 were of spontaneous origin. Most duplications appeared to result from a single break in the selected regionproducingaterminal deletion (Figures 2, 3, 4 and 5). At least five of the gamma ray induced duplications, however, contained morethanonebreakpoint(Figure 2). hDp69 and hDp71, two ZLXLszTl[dpy-5(-)unc-13(+)] duplications, carry an internal deletion around dpy-5. These two duplications are dpy-5(-) but complemented Zet362, unc-I1 and unc-74. hDpl8, in addition to being deleted for the dpy-I4 region, was also found to be
K. S. McKim and A. M. Rose
120
A
1 -
h61,h167,h17J,h655 blt-3 egi-30 unc-35
let-362 I
ltn-6
1%"-17
I
sup-ll
I
I
unc-11
f0q-J
$7,-5
I
=+ =+ hDpl0
-4
I
unc-40bit-4 l
l
h654 h601 dpy-14 unc-13 unc-29 I
1
I
1
-
hDp69 hDp71
B ILXLSZTI
==iFt unc-l d q - 3 unc-fO
ion-2
ldfy-8
dpr-7
- - Unf-3
X
H
2 I.U.
FIGURE2.-(A) Duplications and deficiencies on the left half of chromosome I . Chosen examples of duplications are shown here. All the markers on the map except bli-?, egl-30 and unc-35 were tested for complementation with each duplication. When it is not known if a duplication complements a marker, the duplication is drawn with a gapped double line. Gamma induced lethal mutations h61, h167, h171 to the left of hDp18.(B) Genetic map of the X chromosome. T h e X chromosome portion of ILXLszTland the and h655 were only. mapped .. positions of X-linked markers used in this study are shown.
deficient for the left end of the chromosome. hDpl8 complemented unc-11 and fog-I but not sup-11, linI7,lin-6 or let-362. Thus, atleast one unselected break occurred in theregion between sup-I I and fog-I. hDp62 and hDp72, two of the I L X L s z T l [dpy-5(+)unc13(-)] derivatives, failed to complement egl-30, let362 and lin-6. In addition, hDp72 complemented supI I but hDp62 did not. N o other distal markers have been tested but, like hDpl8, these duplications probably have a terminal deletioncaused by a second break to the left of unc-1I. Some of the gamma ray induced lethal mutations mapped in this study were not separable from dpy-5 or unc-13 in recombination experiments(R. MANCEBO and K. MCKIM, unpublished results). Duplication mapping of these mutations has positioned some of
them into unexpectedplaces on the map. Four gamma ray induced lethal mutations (h61, h167, h171, h655) recombinationally inseparable from dpy-5 have been mapped to the left end of the chromosome because they failed to complement hDpl8 while complementing hDp12. These four mutations fully complement each other. Two other mutations (h601, h654)recombinationally inseparable from dpy-5 mapped to the right of unc-13 based on complementation with I L X L s z T I . These two mutationswerecomplementationtested against the deficiencies in the unc-29 region. h654 maps inside nDf24 and nDf25 but outside nDf23. h601 is inside all three deficiencies. h654 and h601 fully complement each other and unc-29. Allsix of these gamma ray inducedmutations
h549 Eet-361
121
Duplications in C. elegans h550
Zet-359
unc-11 I
Zet-353 Zct-356 let-354 Eet-366 Zet-374 let-373
h565
let-351
unc-73 unc-89 unc-74 unc-57 let-375 I
1
I
I
/
let-364 unc-38 unc-63
let-363' let-371 spe-11 dpy-5
I
I
I
unc-40 I
b12-4 I
unc-37 dpy-14 UnC-13 !
I
I
sDf4 I
I
hD022*. ?5*
l"--1 m.u.
FIGURE3.-Position of breakpoints of sDp2 derivative duplications which break in the unc-11 dpy-5 region. Genes andgamma ray-induced mutations of essential genes (designated by allele number)are indicated above the line. Mapping of the essential genes with respect to sDf4, hDf6,hDp3, hDp2O and h D p 2 2 is the work of HOWELL (1989). When it is not known if a duplication complements a marker, the duplication is drawn with a gapped double line. All duplications were tested for complementation with all the other markers on the map. Duplications marked with an asterisk arose spontaneously in a duplication strain. The rest were gamma-ray induced.
should have been recombinationally separated from dpy-5.h61 and h655 suppressed recombination in trans to a bli-3 unc-11 chromosome while h60I and h654 suppressed recombination in trans to a unc-1 I dpy-14 chromosome. These results suggest these four mutations disrupt sequences required for normal recombination. Gamma ray induced lethal mutations were complementationtested against known EMS mutations of essential genes. h549 was shown by HOWELL (1989) to be an allele of let-354. HOWELL(1989) also tested h550 and h565 for allelism with known genes in the hDf6 to dpy-5 interval. No allelism was found indicating h550 and h565 define new loci. We have found that h590 is an allele of let-400 and h563 is an allele of let-378. h170 and h546 fail to complement each other but have not been tested against other lethal mutations in the region. Among the mutations mapping near the left end of the chromosome (Figure 2), h17I was found to be allelic to let-362. h D p l 4 ( I ; X ) :The duplication hDpl4 (Figure 4)was inserted into the X chromosome. T h e insertion event did not disrupt an essential gene on theX since hDpl4 homozygotes were viable. Hermaphrodites of the gen-
otype dpy-5/dpy-5;hDpl4/hDpl4 were veryslow growing; they had a generation timeof 6 instead of 4 days for hDpl4 heterozygotes. These homozygous duplicationstrains also had a Him phenotype, and were never observedto segregate Dpy-5 progeny. dpy5 / d p y - 5 ; h D p l 4 / 0males were fertileand contributed the duplication to all their hermaphrodite progeny. The insertion event was associated withloss of the sequences at the right end of sDp2 (bli-4-dpy-14) but no detectable loss occurred at the left end of sDp2. let-362,egl-30 and unc-35 were complemented by hDpl4 (Figure 2). T h e insertion site was positioned by three factor mapping the Him phenotype. T h e Him mutation must have been tightly linked to the duplication since the duplication and theHim phenotype always segregated together. The Him mutation was mapped to the right of dpy-3 and unc-20 but to the left of dpy-8 (Table 4). T h e single non-Him Dpy-3 recombinant could have been the result of adouble crossover event,one between dpy-3 and unc-20, and the second between unc-20 and the insertion. In the unc-20 dpy-8 experiment, more Unc-20 recombinants had the duplication than Dpy-8 recombinants. It may be in error to con-
122
K. S. McKim and A. M. Rose
dpy-5
h3 70 h54 6 h563
unc-40
h55G h564
I
I
I
I
let-86 bli-4
-.Y
dpy-14 unc-37
I
I
8-'\r: 7 2 2
I
I
I
let-75 unc-13 I
I
I
I L I L ~2 TI
I
hDp36:43,64
1
sDp2, hDp38,@,60 hDp58
I
hDp62
i
hDpl9,33,35,39,54,56 hDp14,16,18,31,32,34
2 hDp12,17,55,61, 72 hDp37
I hDp13,,15,41,57
FIGURE4,"Position of breakpoints of sDp2 and ILXLszT1derivative duplications which break in the dpy-5 unc-13 region. Most of the duplications extend to the left end of the chromosome (Figure 2). This was tested by complementation to let-362. With the exception of hDpl8, all of the gamma ray induced sDpZ[dpy-5(+) dpy-l4(-)] derivatives (hDps 12-19) carried this locus. Nine ILXLsrT1[dpy-5(+)unc13(-)] derivatives (hDp31, 32, 33, 34, 35, 36,37, 38,58) tested complemented let-362. hDf62 and hDp72 have left end deletions (Figure 2). See the legend of Figure 3 for more details.
unc-14 h590 let-400 dpy - 5
bli-4
unc-40
ZLXLszTl hDp48,49,52,53,7
I I==
I
hDp47,67,68,69,70
-
hDp44,51
t
I
hDp45,46,66 hDp50 65
=I+
'
hDp73
H
FIGURE5.-Position of breakpoints of dpy-5(-) derivatives of ILXLszT1. Most of these duplications do not carry any left end sequences; they are terminal deletions. Six ILXLszT1[dpy-5(-) unc-13(+)]duplications (hDp44, 45, 48, 49,53, failed 65) to complement let-362. As shown in Figure 2, hDp6Y and hDp71 are internal deletions of ILXLszTI.See the legend of Figure 3 for more details.
Duplications in C. eleguns
123
TABLE 4
TABLE 5
Mapping of hDpl4(I;X)
Duplication Stability:I. Hermaphrodite Gametes
Genotype"
Wtb
Recombi- Fraction nants' m.u.Him
(C.1.y
dpy-3 unc-2U/hDpl4
389 19 Dpy-3 18/19 6.3 (4.6-8.9) 12 Unc-2Od 0112 7 Dpy-5 017 10.6 (4.4-22.4)
unc-20 dpy-blhDp14
273 27 Dpy-8 3/27 16 Unc-20 13/16
dpy-7 unc-3/hDp14
248Unc-3 3 313 2 D ~ y - 7 ~ 012
dpy-7 unc-3/++;hDpl6
296 36 Unc-3
ND'
(0.9-3.9) 1.8
All strains were heterozygous for dpy-5(e61) on chromosome I except in the h D p l 6 experiment in which dpy-5 was homozygous. Male progeny not included. ' C.I. = 95% confidence interval. One-fourth of this classwas not observed because of dpy-5 segregating. ' m.u. were not calculated because these markers flank the insertion site.
clude from this that the insertion site maps closer to dpy-8 because, as described next,the insertion disrupts the normal distribution of recombination on the X chromosome. The insertion of chromosome I material disrupted the normal distribution of recombination events on the X chromosome (Table 4). T o the left of the insertion, recombinationwas normal or slightly higher than normal. The dpy-3 unc-20 interval was 6.3 m.u. in duplication heterozygotes compared to 5.0 m.u. on the normal map (McKIM, HOWELL and ROSE 1988). In the dpy-3 unc-20 experiment, the Dpy-5 recombinants were the result of recombination between unc20 and the insertion. In these experimentsdpy-5(e61) was heterozygous on chromosome I , therefore, in the calculation of this distance the number of Dpy-5 recombinants was multiplied by four. The 10.6m.u. observed in the u n c - 2 0 h D p l 4 interval is only slightly higher than the 7.2 m.u. normally observed in the unc-20dpy-8 region (McKIM, HOWELLand ROSE 1988). In contrast, to the right of the insertion recombination was greatly reduced. In h D p l 4 heterozygotes, recombination in the dpy-7 unc-3 interval was reduced tenfold from 21 m.u. (EDGLEYand RIDDLE1987) to 1.8 m.u. As control, a recombination in dpy-5; h D p 1 6 ( I ; f ) ; dpy-7unc-3/ hermaphrodites was scored (Table 4). h D p l 6 is of comparable size to h D p l 4 but is free. This result indicates the hyperploidy in h D p l 4 strains did not cause the reduction in X chromosome recombination frequency. Segregation analysis of the duplications: Duplication stability in the hermaphrodite: All of the duplications from the screens described in Tables 1 and 2, some of those from Table 3, andtwo of spontaneous origin, were tested for segregation stability. Strains of the genotype dpy-5/dpy-5/hDpx[dpy-5(+)unc-13(-)] or u n c - l l l u n c - 1 1 ; h D p x [ u n c - l l ( + ) d p y - 5 ( - ) ] or unc-
++
Duplication
Wt"
dpy-5; sDP2 563 1067 dpy-5; hDp2 374 dpy-5; hDp3 dpy-5; hDp4 339 351 dpy-5; hDp6 183 dpy-5; hDp20 dpy-5; hDp22 302 dpy-5; hDp23 45 1 613 dpy-5; hDpl.2 dpy-5; hDpl4 231 (27) dpy-5; hDpl5 834 464 dpy-5; h D p l 8 dpy-5; ILXLszTl 401 (30) dpy-5; hDp34 268 38 1 dpy-5; hDP36 dpy-5; hDp37 730 dpy-5; hDp54 751 dpy-5; hDp57 209 151 dpy-5; hDp62 828 unc-1 I ; sDp2 574 unc-11; hDplU unc-11; hDpll 525 (16) I unc-13; I ~ ~ ~ S Z T 438 unc-13; hDp49 597 unc-13; hDp50 500 348 unc-13; hDp66 277 (15) unc-13; hDp69 553 unc-13; hDp73 a
Dpy or Unc"
Frequency of Dp gametes
405 1782 328 529 693 122 520 415 434 82 (32) 59 1 352 273 (20) 206 283 668 488 128 814 570 628 934 305 (6) 45 1 480 590 276 (3) 248
0.4 1 0.23 0.36 0.24 0.20 0.43 0.22 0.35 0.41 0.58 0.41 0.40 0.42 0.39 0.40 0.35 0.43 0.45 0.08 0.42 0.31 0.22 0.42 0.40 0.34 0.23 0.33 0.53
Male progeny scored are in parentheses.
13/unc-13; hDpx[dpy-5(-) unc-13(+)] were constructed. The ratio of wild type to mutant progeny was used to calculate the gametic frequency of each duplication (see MATERIALS AND METHODS). These results are summarized in Table 5 and Figure 6. The results from dpy-5;sDp2,dpy-5;ILXLszT1,unc-11; sDp2 and unc-13; ILXLszTl were not significantly different and thus experiments with different markers and duplications are comparable. Every duplication tested is notreported because those with similar breakpoints usually gave similar results. T h e calculations for gamete frequency assume the contributions fromthespermand oocyte were equal. As shown below, this was not always the case. The data in Table 5 does not include duplication homozygotes, which were slow growing, small and clear. sDp2 homozygotes were smaller than other D p homozygotes and only rarelyproducedprogeny. Homozygotes of sDp2 derivatives deleted at the left end also had this phenotype. Dp[dpy-5(+) dpy-14(-)] strains, however, produced homozygotes which were fertile but still small and clear. While wild types made upalargerfraction of the progenyfrom D p ( I ; f ) homozygotes than from hemizygotes, on the orderof three to five fold, this corresponded to a D p gamete frequency of only 60-70%. Thus it can be concluded
K. S. McKim and A. M. Rose
124 blt-3 let-362 I
s u Ipm - 1-16 I
Itn-17 I
fog-!
I
tmc-!I I
dpy-5
unc-74 unc-57 I
1
I
unc-40 I
rDp2
biz-4 1
unc-13 dpy-14 I
1
[417.; 417.1
hop12 [417.; 427.; 437.1 hDpl8 [407.; 327.;]
I
hop2
[232; varmble; i7.1
hDp3
[362; 212;
hDp4
[24%; 1.27.; 521
327.1
-
hDp20 [437.; -; 547.1 hDp22 [ Z Z ; - ;
li7.1
hDp62 [SZ]
+"" &!O
[317.; 377.1
4
" " " "
h D p l l [227,; 2.421
"i
" "
&69
r337.1
k"---L 1 m.u.
FIGURE6.-The relationship between the extent of duplicated material and the stability of the chromosome. Within the brackets, the percentages indicate the fraction of gametes carrying a duplication. If only a single percentage is shown, this refers to the average recovery of hermaphrodite sperm and oocyte as measured from self fertilization experiments. When more than one percentage is shown, each number is derived from a different assay and presented in the form: [hermaphrodite self-fertilization;oocyte; male sperm]. See the text for details.
that theduplications do not segregate fromeach other in hermaphrodite homozygotes. Most duplications derivedfromeither sDp2 or ILXLszTI in which part of the dpy-5 unc-13 region was deletedhad similar recovery frequencies;approximately 40% of the gametes carried the duplication. This included hDpl8 which is deleted at both endsof sDp2. Significant differences in duplication stability sDp2 or ILXLszTl were found with comparedto hDpl4,hDp62 and hDp73. hDpl4 was expected to be very stable since it is inserted into the X chromosome. hDp73 was alsovery stable. Strainscontaining this duplication segregated worms which grew to adults but were sterile. Like hDpl4, hDp73 may be an insertion into another chromosome, but unlike hDpl4,the insertion disrupted an essential gene. hDp62 is very unstable. It is also missing the left end of the chromosome but this alone cannot explain its instability because other duplications missing the left end, such as hDpl8,are much more stable. While most duplications with breaks in the dpy-5 unc-13region had similar stabilities, duplications with breaks in the unc-11 dpy-5 region varied with respect to their recovery frequency in gametes. Most duplications with breaks in this region had reduced stabilities. In addition,duplications carrying the unc-75unc57 region were often more stable than those deleting it. Thus unlike the dpy-5 unc-I3 region where deletion of chromosomal material had only small effects on duplication stability, deletions in the unc-1I dpy-5 region had significant effects on stability. hDp20 was much more stable than expected for its
size; it breaks between unc-57 and unc-63 but was found in 43% of the hermaphrodite gametes. Other duplications of similar size (e.g., hDp6) were recovered in fewer (20%) gametes. The hDp20 chromosome also contains part of chromosome V (K. MCKIM,unpublished results). These added sequences may confer the added stability to hDp20. hDp2 includes unc-74 but was found to be unstable like unc74(-) duplications (Table 5 and below). This may result fromadifference in thestructure of hDp2 compared to hDp3,5,and hDp7. The four unc-1I (+) dpy-5(-) duplications were of two classes. hDp9 and hDpl0 were found in approximately 31% of the gametes while hDp8 and hDpl I were found in approximately 22% of the gametes. We do not know how the stabilities of these four duplications relates to their structure because we have not precisely defined their breakpoints. The stabilities of ILXLszTl derivatives which had of chromosome lost the left end [dpy-5(-) unc-13(+)] I were less than sDp2 but greater than or equal to those sDp2 derivatives with breaks in the unc-II dpy5 region. This level of stability could be explained if the X chromosome sequences are attached to the right end of ILXLszTl. We have no otherevidence indicating to which end of ILXLszTI the X chromosome material is attached. Assaying for the Him phenotype was not informative because, while ILXLszTl strains were Him, most of the derivatives were not. Only the two duplications which carried an internal deletion around dpy5 (hDp69 and hDp71) had a Him phenotype. Assuming the X nondisjunction phenotype of ILXLszTI and
Duplications in C.elegans TABLE 6 Duplication Stability:11. Oocytes Dpy-5 hermaWt hermaphrodite Wt male
112
phrodite
125
see if the duplication or a recombinant chromosome was present. When hT2(Z;ZZZ)[+ dpy-5 +;+]/uric-1 I dpy-14; males were crossed to dpy-5 dpy-14; sDp2 hermaphrodites, 4 % 1 of the cross progeny carried the duplication. Thus the contribution from sperm and oocyte in the sDp2 hermaphrodite was equal. The smaller hDpl5 was recovered in slightly fewer oocytes. Similar experiments with hDpl2 showed it was also recovered in similar frequencies in the two hermaphrodite germ lines. In a cross with hDpl8 hermaphrodites, 32% of the cross progeny carried the duplication. Thisreduced recovery frequency of hDp18 compared to hDpl2 (42%) could be attributed to the left end deletion of hDpl8. For hDp[unc-ll(-) dpy-5(+)] strains, unc-1I dpy14/dpy-5 unc-13 males were crossed to unc-11 dpy-5; hDpx hermaphrodites. hDp4 was found in only 1.2% of the oocytes (Table 6). When the same cross was done with hDp3,21% of the oocytes carried the duplication. In self fertilization experiments, hDp3 was found in 36% of the gametes (Table 5). Knowing the hermaphrodite oocyte and male spermduplication frequencies, we calculated that hDp3 was found in 41 % of the hermaphrodite sperm.hDp4 was found at a similar levelin hermaphrodite sperm. Therefore, the low frequency of oocytes carrying hDp3 or hDp4 accounted for the reduction from 4 1% in sDp2 self fertilization experiments to 36% in hDp3 and 24% in the hDp4 self-fertilization experiments. hDp2 was lost frequently during oogenesis, but the results were inconsistent. When +dpy-5+unc-l3/unc11 +dpy-14 males were crossed to seven dpy-5 dpy14; hDp2 hermaphrodites,thefrequency of hDp2bearing oocytes varied from1to25%(datanot shown). This is probably because hDp2 is significantly more unstable in the somatic cells than hDp3. The frequency of hDpl0 and hDpl1 in the oocytes hT2(Z;IZZ)[++ was tested by crossing bli-3 unc-11 dpy-5;+] males to unc-11dpy-5;hDpx hermaphrodites. hDpl0 was recovered in 32% of the oocytes (Table 6); a rate similar to the self-fertilization frequency (Table 5). This indicates that duplication loss, while greater than that for sDp2,occurred equally in both germ lines. In contrast, hDpll was recovered in only 2.4% of the oocytes; a rate tenfold lower thanthe selffertilization frequency. In the matingexperiments, we oftennoted selffertilization progeny amongst cross progeny. Our experimental procedure allowed us to observe with certainty that self-fertilization progeny were produced amongst cross progeny. Either the duplication hermaphrodites multiply mated or the male sperm was not always used preferentially tohermaphrodite sperm asis normally found (WARD and CARREL 1979). The instability of these duplications results in part from somatic loss.Mosaic worms were frequently
+
+
Frequency of Dpy-5 male Dp oocytes"
dpy-5 hT2/unc-l I dpy-14 X dpy-5 dpy-14; ~ D p 2 ~ 118 66 0.41 dpy-5 hT2/unc-ll dpy-14 X dpy-5 dpy-14; hDpl2 23 32 0.42 38 39 dpy-5 hTZ/unc-II dpy-14 X dpy-5 dpy-14; hDp15 58 92 51 97 0.37 dpy-5 hT2/unc-ll dpy-14 X dpy-5 dpy-14; hDpl8 62 56 132 117 0.32 dpy-5 unc-13/unc-ll dpy-I4 X unc-11 dpy-5; hDp3 53' N DC 261 248 0.21 dpy-5 unc-13/unc-l1 dpy-14 X unc-11 dpy-5; hDp4 3' N DL 243 0.012 258 bli-3 unc-llldpy-5 hT2 X unc-11 dpy-5; h D p l 0 52 16 17 50 0.32 bli-3 unc-1 lldpy-5 hT2 X unc-11 dpy-5; hDpl1 3 0 0.024 71 53 a Fraction of duplication carrying oocytes calculated on the basis of Wt and Dpy progeny.
Dp oocytes= Dp worms/Total progeny. Dp worms = 2 X Wt except in the sDp2 experiment
+
where Dp worms = Wt. Total progeny = Dp worms (2 X Dpy). The crosses are indicated above the data. The male genotype is on the left and the hermaphrodite genotype in on the right. ' These experiments required progeny testing o i the Wt progeny because no balancer was used in the male. Males were not progeny tested and thus not included in the data.
the stability of the Dp[dpy-5(-) unc-13(+)] chromoX chromosome sequences somes was duetothe (McKIM, HOWELL and ROSE1988), the failure of most Dp[dpy-5(-) unc-13(+)] strains to be Him may have resulted from their reduced size. With regard to comparing the effects of deleting DNA in the unc-11 dpy-5 region to the dpy-5 unc-13 regions, hDp69 is of interest because it is a deletion of ZLXLszTlin the unc-11 dpy-5 region but still carries the flanking regions intact. Its stability was similar to hDp3 and hDpl0 despite the fact it contains more DNA (Figure 2). This may be because of the deletion around dpy-5. Most of the reduction in stability of these duplications can be accounted for by loss during oogenesis. T h e duplications were usually recovered at a higher frequency in hermaphrodite sperm than oocyte. The fraction of oocytes carryingaduplicationchromosome was assayed by crossing heterozygous males to duplicationhermaphroditesandscoringthe cross progeny.In the crosses described below, thefrequency of Wt progeny (which carried the duplication) was equivalent tothe frequency of oocytes which carried the duplication (Table 6). In Table 6, if the male was heterozygousfora crossover suppressor (hT2(Z;ZIZ) then all the wild-type progeny carried the duplication. If the male didnotcarrya crossover suppressor, the wild types had to be progeny testedto
+
+;+/
K. S. McKim and A. M. Rose
126 TABLE 7 Duplication stability: 111. Male sperm W t herDpy herDuplication Wt male maphrodite Dpy malemaphrodite
hDp2 hDp3 hDp4 hDp6 hDpl2 hDpZ0 hDpZ2
46 168 17 58 166 175 165
36 105 9 37 51 73 122
598 278 219 192 84 50 693
All crosses were dpy-5; hDpx male dite.
X
583 294 249 182 199 163 674
Frequency of D p sperm
0.07 0.32 0.05 0.20 0.43 0.54 0.17
dpy-5; unc-36 hermaphro-
recovered with strains carrying the unstable duplications but not with strains carrying the stable duplications (such as sDp2 and h D p l 5 ) (datanot shown). Mosaic worms (having a Dpy or Unc phenotype) have been recovered from strains carrying an unstable duplication and unc-ll dpy-5, unc-57 dpy-5 or dpy-5 dpy1 4 chromosomes. Isolation of mosaics withunc-74 dpy5 or unc-63 dpy-5 chromosomes is described later. If a duplication was lost early in germ line development, an unc dpy; op[++] hermaphrodite would lack the duplication in its germ line and produce no wild-type progeny. hDp2, hDp4, hDp59, hDp62 and hDp72 strains frequently produced germ line mosaics that segregatedno Dp-containing progeny. This could indicate a higher level of somatic loss compared to most other duplications, even those with similar stabilities when measured by duplication recovery in the gametes. Duplicationstability i n themale: We have tested duplication stability in the male by crossing dpy-5 +/ dpy-5dpy-14;hDpx males to dpy-5;unc-36 hermaphrodites and scoring the cross progeny. The Dpy (non-Unc)progenydidnot receive theduplication while the wild-type progeny resulted fromfertilization with a duplication carryingsperm. The results are shown in Table 7. The duplication recovery in the male sperm was similar to thatobserved in hermaphrodite oocytes and considerably lower thanforhermaphroditesperm. For example, hDp3 was found in 32% of the male sperm compared to 41 % of the hermaphrodite gametes. hDp4 was found in 5% of the male sperm, 2% in the hermaphrodite oocytes and 40% of the hermaphrodite sperm. In general, more loss was detected in male sperm andhermaphrodite oocytes than in hermaphrodite sperm. As others have shown (HERMAN, MADL and KARI 1979; DELONG,CASSON and MEYER 1987), some duplications tend to segregate from the single X chromosome in males. Segregation of the D p from the X chromosome would produce an excess of wild-type males ( D p ; X O ) and Dpy hermaphrodites (nullo-Dp; X X ) over wild-type hermaphrodites ( D p ; X X ) and Dpy
males (nullo-Dp; X O ) . We observed this in our experiments with h D p l 2 and hDp20 (Table 7). In h D p 2 , 3 , 4 , 6 and hDp22 males an excess of wild-type males over hermaphrodites was observed but not the excess of Dpy hermaphrodites over males. In these cases it was possible that the segregation distortion could not be observed among the Dpy progeny because there were too many non-Dp worms. Spontaneous shortening of duplications: Several duplications have been observed to spontaneously shorten. These duplication derivatives were isolated onthe basis of an exceptional phenotype, either by chance or in selective screens (see below). For example, our unc dpy; Dpstrains rarely segregated Unc worms that produced either no Uncs in their progeny (and thus were probably a genetic mosaic) or produced both Unc and Dpy Unc progeny. This latter class may have been the result of a deletion in the duplication chromosome to uncover the unc mutation, or a new mutation either on the duplication or elsewhere in the genome. Fourteen of these Unc strains were analyzed by complementationtesting and allof them were found to be the result of a loss of material from one end of the duplication and not through the addition of markers via mutation or recombination. The shortening events were not limited to the left end of sDp2. hDp3O and hDp77 were dpy-l4(-); the result of loss of sequences from the right end of sDp2 (Figure 3). The spontaneous duplications and their progenitors are shown in Figure 7. T o determine the frequency at which duplications shorten, we used a selective system to screen a large number of worms for exceptional individuals. Using the procedure described in MATERIALS AND METHODS, the progeny of unc-74 dpy-5; hDpx or unc-63 dpy-5; hDpx hermaphrodites were screened for levamisole resistant Uncs. The recovery of these duplications is summarized in Table 8. Because the worms were screened in the F2 generation, it could not be ensured that each event was independent. Some Po hermaphroditessegregated morethanone Unc F2. It is possible some of the shortening events occurred premeiotically. For this reason the numbers in Table 8 may not reflect the per chromosome shortening frequency. We have observed different chromosomes shortening at different rates. There is a correlation between spontaneous shorteningrate andmitotic stability. T h e more stable duplications shortened at a lower rate. For example, sDp2, which is mitotically very stable, spontaneously shortenedata low frequency. In a levamisole selection experimept, unc-74 dpy-5; sDp2 hermaphrodites segregated two Dp(unc-74(-)) derivatives in 127,000 chromosomes screened. hDp2, which is mitotically unstable as judged by the produc-
127
Duplications in C . elegans
gamma I
h Dp5
hDp3 gamma
hDp 7 gamma
hDp75 gamma
hhDDpp724
hDp2O
gemma
gamma
h 0025
h D p 76
1
~
~
h D ph2D6 ph2D7 ph3D0 p 5 9
~
c
FIGURE7.-Pedigree of spontaneous duplications. All the spontaneous shortened duplications and their progenitors areshown. I f gamma radiation was used, this is indicated, otherwise the duplications were spontaneous.
DISCUSSION
TABLE 8
Recovery of spontaneously shortened duplication strains
Duplication
Total progeny
72 000 unc-74 dpy-5; hDp2 unc-74 dpy-5; hDp5 116 000 unc-63 dpy-5; hDp20 134 000 unc-63 dpy-5; hDp23 41 000 unc-74 dpy-5; sDp2 127 000
Shortened duplications
30 3 1 18 2
Frequency
4.2 X 2.6 X 7.4 X 2.2 X 1.6 X
Mosaics
54 3
1 18 2
tion of mosaics and reduced recovery in the gametes, shortened at a much higher frequency. In at least one case, a duplication which had previously shortened continued to shorten at a high frequency. An hDp3 strain segregated hDp23 which subsequently segregated even shorter duplication chromosomes (Figure 7 and Table 8). In addition to recovering shorter duplications, these experiments recovered mosaics. These worms were either Unc non-Dpy or Unc semi-Dpy and segregated only Dpy Unc progeny or wild types and Dpy Unc progeny (Table 8). sDp2 and ILXLszTl haveno effect on recombination frequency: These two duplications were examined to see if they have any effects on recombination. Strains carrying the duplication in addition to normal chromosomes heterozygousfor cis-linked markers were scored and the recombination frequencies were calculated as described in MATERIALS AND METHODS (Table 9). sDp2 had no effect on recombination in both the dpy-5 unc-13 interval and the let-362 dpy-5 interval. In the dpy-5 unc-13 interval, ILXLszTl had no detectable effect on recombination frequency.
The original goals of this research were to isolate rearrangement breakpointson chromosome I and use these to map mutations and study chromosome behavior. Sixty-two duplications were recovered following gamma radiation and 14 were isolated as a result of spontaneous events. These duplications and five deficiencies have divided the left 20 m.u. of chromosome I into 24 regions. An array of duplications facilitates the analysis of a large chromosomal region: In the process of identifying essential genes in the sDp2 region (HOWELL et a l . 1987), over 500 EMS induced mutations have been isolated and arebeing mappedand complementation tested (HOWELL1989; J. MCDOWALL and A. ROSE, unpublished results). This would be an enormous task without rearrangement breakpoints to position the mutations into smaller intervals before undertaking complementation testing.By using the mapping techniques and crosses described in this paper, any recessive mutation can be efficiently mapped to an interval. While the spontaneous shortening of duplications couldproduce false negative results, we have avoided this problem through frequent testing of strains with appropriate markers. Using the methods described here, a set of nested duplications could be isolated and used to intensively characterize any genetic region in C. elegans. A large number of breakpoints can be easily generated in any region having an appropriatepre-existing duplication. As a tool for the geneticdissection of a region, duplication mapping may be an attractive alternativeto the conventional deficiency mapping (SIGURDSON, SPANIER and HERMAN 1984; ROSENBLUTH et al. 1988). The isolation of duplication breakpoints is fast and easy.
K. S. McKim and A. M. Rose
128
TABLE 9 Effects of duplications on recombination Genotype
dpy-5 unc-13 unc-29/+++; sDp2' dpy-5 unc-13 unc-29/+++'
Wt
4
1279 2678
let-362 dpy-5 unc-13/+++; Dpy-5Unc-13 sDp2 20 652 149 1406 let-362 dpy-5 ~ n c - 1 3 / + + + ~ Dpy-5Unc-13 6 1769 dpy-5 unc-13/++; I"XLszTI dpy-5 ~ n c - 1 3 / + + ~ 52 4775
Recombinants
Dpy-5 10 Unc-29 34 Dpy-5 21Unc-29
Dpy-5 Dpy-5
I11.U.
(C.1.P
1.3 (0.3-3.1) 1.4 (0.7-2.4) 1.9 (1.3-2.6) 1.2 (0.8-2.6) 15.0 (9.3-24.3) 15.4 (13.0-17.9) 1.5 (0.7-3.3) 1.6 (1.3-1.9)
5 1 Unc-I3 (3.1. = 95% confidence interval.
' This data was collected by A . M . HOWELL. ' Data fronl MCKIM, HOWELLand ROSE (1988). 'I Data from HOWELL et al. ( 1 987).
Furthermore, unlike deficiencies whose viability decreases with increasing size, larger duplications are more stable (see below). Having such large aberrations is a useful tool to quickly localizeunmapped mutations to a large region; that being either the region covered or not covered by the duplication. Subsequent experiments can map themutationmore precisely using additional breakpoints. Since deficiencies are generally smaller, more experiments are required tomap a mutation to a region of comparable size. In addition, most of the duplications used in this study have one end in common with sDp2. Because only one breakpoint varies, mapping is a binary procedure. In theory the same could be said of nested deficiencies, but such a collection is not easily isolated and can only be effective for a small region due to the smaller size of deficiencies. The nested duplications are a by-product of the method used to generate them and there is no restriction on the size of region to be analyzed. We have mapped the duplication breakpoints with respect to some known visible and lethal mutations in the regim. In addition, the duplication breakpoints can be used as genetic landmarks that can be placed on the physical map being constructed (COULSON et al. 1986,1988). Within the sDp2 regionapproximately 2000 kB of DNA has been placed on the physical map. At present, this DNA is aligned to the genetic map at six points (STARRet al. 1989;J. BABITY, unpublished results). The duplications described here are being used by us and others to position cosmids on the genetic map of chromosome I . Shortening duplications decreases mitotic stability: We observed considerable variation in the mitotic stability of the chromosomeI duplications; some were lost more frequently than others. We measured instability as adecrease in thenumber of Dp carrying gametes and an increase in the frequency of genetically mosaic worms. The isolation of genetically mosaic worms results from mitotic loss of the duplica-
tions. It is not known if there is significant loss during meiosis. HERMAN,ALBERTSON and BRENNER(1976) observed considerable amounts of premeiotic duplication lossin cytogenetic observations of duplication strains. This behavior has been interpreted as a consequence of the holokinetic structure of C . elegans chromosomes (ALBERTSON and THOMSON 1982). During mitosis and meiosis, individual spindles are observed to attach at many sites along the chromosome. ALBERTSON and THOMSON (1 982) proposed the efficiency of spindle attachment was proportional to kinetochorelength. Thus smaller chromosomefragments, with their smaller kinetochores, would sometimes have insufficient spindle attachment for proper segregation. Consistent with this hypothesis, size is an important determinant of chromosome I duplication stability. Our results, however, show there are significant exceptions to a strict relationship between duplication size and stability. hDp14, hDp20 and hDp73 are more stable than expected but this iseasily explained by their attachment to other chromosomes or fragments. Lesseasily explained are hDp2,hDp4,hDp62 and hDp72. The stabilities found in this group aresurprisingly low for their size. hDp2, for example, is 30% less stable than three other duplications of the same size. One possibility is that different structures of the duplications result in different stabilities. For example, linear and circular chromosomesare expected to have different mitotic properties. (MCCLINTOCK1938; LEIGH 1976). Another exception to a strict size-stability correlation was found by determining the effects of deleting certain regions from theduplications. Deletion of the dpy-5 unc-I3 region had small effects on duplication stability whereas a deletion of the unc-57 unc-74 region significantly reduced duplication stability. For example, there is little difference between sDp2,
Duplications in C. elegans hDp 12 and hDp 15whereas there aresignificant differences between the three groups hDpl8 and hDp3, 5 , IO and hDp4, 6 , 11. The greater loss of stability that accompanied deletion in the unc-11 dpy-5 region when compared to similar sized deletions in the dpy-5 unc1 3 region was not theresult of deleting larger amounts of DNA. The dpy-5 unc-13 region contains at least as many genes and as much or more DNA than the unc11 dpy-5 region (HOWELLet al. 1987; KIM and ROSE 1987; HOWELL 1989; STARR et al. 1989). Since size is not the only determinant of stability, what other factors could be involved? The first possibility, differences in structure, has already been discussed. Second, a threshold level of chromosome size may be necessary for stability. If a size threshold is important, then deletionsof chromosome material do not have significant effects above the size threshold. Duplications like hDp3 would be smaller thanthe threshold and be in the range where the probability of spindle attachment is sensitive to size changes. A third possibility is the DNA sequences responsible for stability may not be evenly distributed along the chromosome. These sequences need not be kinetochore attachment sites. If DNA sequences are an important factor,then hDp6 isless stable than hDp3 because sequences conferring mitotic stability have been deleted. Our results do not unambiguously resolve these possibilities. In fact, any combination of these may be influencing the stability of a given duplication. Some free X chromosome duplications have higher levels of loss in the oocyte line than in sperm (HERMAN, ALBERTSON and BRENNER 1976). We observed a lower frequency of duplication recovery in the hermaphrodite oocyte and the male sperm than in the hermaphroditesperm.Hermaphroditespermare made before oocytes and undergo fewer cell divisions (KIMBLEand HIRSH 1979).Thiscould explain the different rates of loss. The more frequent and rapid cell divisions of oogenesis could provide more opportunityforduplication loss. Similarly, male sperm undergomore cell divisions thanhermaphrodite sperm. hDplO was recovered at equal frequencies in bothhermaphroditegerm lines. Some duplications may be less sensitive to loss than others during the cell divisions of oogenesis and male spermatogenesis. The more unstable duplications will be useful for mosaic analysis. HERMAN(1984)has used unstable duplications of the X chromosome to create genetically mosaic worms. This allows one to ask questions regarding the tissue specificity and cell autonomous expression of a gene. sDp2 is too stable for mosaic analysis butstrains with shorter duplications frequently produce mosaic worms. With hDp2 or hDp23, the frequency of mosaics was approximately one in 2500 progeny when dpy-5 and a levamisole resistant unc were used. This is similar to the frequency of
129
genetic mosaics recovered by HERMAN(1984).We plan to use mosaic analysis as part of the analysis of genes in the left half of chromosome I . Duplication chromosomes spontaneously break: A striking aspect of our results was the spontaneous shortening of the duplications. We observed the spontaneous shortening of five duplications in addition to the two events from sDp2. Previously, HERMAN (1984) observed spontaneous shortening of one X chromosome duplication in C. eleguns. Possibly allduplications will spontaneously shorten.Whetherspontaneous shortening at the frequencies observed here occurs in normal chromosomesin a euploid genetic background is not known. Depending on their structure,however, different duplications shorten at different rates. Mitotically stable duplications (sDp2,hDp5,hDp20) shorten less frequentlythan unstable duplications (hDp2, hDp23). The correlation between mitotic stability and spontaneous shortening leads us to suggest that mitotic/meiotic segregation problems cause both duplication loss and breakage. It is possible the unstable duplications have difficulty aligning with the spindle during nuclear divisions (meiotic or mitotic). If the chromosome is not properly aligned, the chromatids may not be equally divided between the sister cells, lost altogether or undergo a breakage event(see below). ALBERTSON and THOMSON (1982) provided cytological evidence for this type of behavior during mitosiswith thefree X chromosomeduplication mnDp2. They observed the duplication often lying on the outside of the metaphase plate or lagging at anaphase. The improperalignment of duplicationchromosomes during nuclear division may provide an opportunity for breakage. Aholokinetic chromosome could be pulled apart if the differentspindles from thesame pole do not attach to the same chromatid. Misalignmentmight allow spindles from opposite poles to attach to thesame chromatid. If two adjacent spindles were pulling in opposite directions the duplication chromosome might break apart and the result would be a shorter duplication. Improper spindle attachment to normal chromosomes probably occurs but the rate at which this occurs or at which exceptional events are recovered is not known. The reason for the elevated frequency of spontaneous shortening for some of the duplications could be because they misalign on the metaphase plate more often than sDp2. Simply being aunivalent poses problems for a chromosome. Our duplications arenot engaged in exchange events and should be univalents during meiosis. Other authors have observed that univalents in corn (MILLER 1963), wheat (SEARS1952) and Drosophilamelanogaster (SANDLERand BRAVER1954; CARPENTER1973; BAKER1975) are unstable. The univalent may divide equationally at meiosis I,or
130
K. S . McKim and A. M. Rose
either chromatid may be lost at meiosis I or 11, or the univalent may undergo misdivision.Misdivision occurs when a metacentric univalent attempts to separate transversely at the centromere resulting in telocentric chromosomes. Duplication breakagefrequency was either low (sDp2, hDp5 and hDp20) or high (hDp2 and hDp23). There were no intermediate values. It is possible that duplication shortening only occurs at one of the two frequencies. We have no reason to doubt that normal chromosomes may break at the frequency observedin the low frequency (sDp2) class. Furthermore, it is possible that shortening is not entirely the result of mitotic instability, but could also be a consequence of aparticularchromosomestructure.Chromosome breakage is observed with dicentric chromosomes produced by recombination events involving inversion or circular chromosomes (MCCLINTOCK 1938, 1941; HABER, THORBURN and ROGERS1984).It shouldbe emphasized that in addition to missing part of the normal chromosome, these duplications do not pair and recombine with chromosome I . Therefore, we have not addressed the influence of pairing and recombination on the spontaneousshorteningfrequency of a duplication. Although it is assumed genetic mosaics are produced through duplication loss (HERMAN 1984), it is also possible a genetic mosaic could be generated by a breakage event in the lineage under study [as observed in maize (MCCLINTOCK 1938, 1941)l. Mosaics produced through a breakage eventmay be a significant fraction of the total number based on our observations that the recovery of mosaics and breakage events occurs at similar frequencies. A mosaic produced through a breakage event can give different results than amosaic produced through completeloss. In a chromosome loss event, all alleles on the duplication are lost from that lineage. In a breakage event, some of the alleles would remain. Two examples of spontaneous chromosome shortening have been observed in D. melanogaster. BIESSMANN and MASON(1988) isolated terminal deficiencies of the X chromosome which lost sequences from their distal ends at a rate of 75 bp per generation. LEVIS (1989) isolated terminal deletionsof chromosome 3R. The deletion chromosomesalso lost approximately 75 bp per generation. In both of these cases, the deficiencies were recovered in amutagenicgenetic background. These experiments did not report attempts to recover large deletion events. It is possible that our duplications are also shortening by a small amount. We have not doneexperiments to detect theseevents. Broken chromosomes lacking a proper telomereare mitotically unstable in maize (MCCLINTOCK 1941) and yeast (HABER, THORBURN and ROGERS1984) anduntil recently (see above) have not been isolated in Dro-
sophila (MULLERand HERSKOWITZ 1954; ROBERTS 1975). The structures of natural C. elegans telomeres or theends of duplication chromosomes are unknown. Also unknown are the behavioral characteristics of broken, unrepaired C. elegans chromosomes. By analogy with Drosophila, it is reasonable to postulate that in the process of shortening sDp2 or ILXLszTl with gamma radiation, shortened chromosomes were capped with telomerecarryingfragments of other chromosomes originating from the same gamma radiation exposure (WILLIAMSON and PARKER 1976). In the caseof spontaneous breakage events, however, capping in this way cannot be proposed. The free ends created by a breakage event may be repaired by the acquisition of telomerecontaining sequences from another genomic location; involving eitherintrachromosomal or interchromosomalrearrangements (MCCLINTOCK 1941; HABERand THORBURN 1984; RUDINand HABER 1988). Supportingthis proposal is the observation that some shortening events are associated with attachmenttoanotherchromosome (HERMANandKARI 1989; K. MCKIM,unpublished results). Alternatively, a stable end could be created through thede novo addition of a telomereas proposed in Tetrahymena (KING and YAO 1982). Still another solution to the telomere problem is a ring chromosome. Based on deletion analysis of the type described here, evidence in favor of the existence of ring chromosomes has been collected by us (K. MCKIM, unpublished results) and others (C. P. HUNTER andW. B. WOOD,personal communication). Not all duplications, however, are rings. sDp2 appears to be linear because deletions of one end usually do not carry over to the other end. Themost radical possibility is that there is no healing event at all. The solution to the structure of the duplication telomere awaits the cloning of a duplication end. Meioticbehavior of duplications: Two observations support theassertion that duplications of the left half of chromosome I have no pairing activity during meiosis. The first is the failure of sDp2 to recombine with the normal homolog(ROSE,BAILLIE and CURRAN 1984). The second is our observation that these duplications have no influence ontherecombination frequency between the normal homologs in the duplicated regions. These observations are in contrast to another duplication of chromosome I , sDp1, which carries sequences on the other (right) end of chromosome I and does compete for pairing and recombination with the normal homologs (ROSE, BAILLIE and CURRAN1984). While we conclude the left end of the chromosome does not containsequences which can bringthechromosomestogetherforpairing (“homologrecognition”), this region may influence the distribution of recombination events along chromosome. Based on a studyof chromosome V deficien-
Duplications in C. elegans cies, ROSENBLUTH, JOHNSEN and BAILLIE (1990) have proposed a pairingmodel requiring synapsis initiation at both endsof C. eleguns chromosomes. In that study, it was found that recombinationwas reduced adjacent to internal deficiencies. The gamma ray induced lethal mutations h61 and h655 described here may be similar to these. The insertion of chromosome I material into the X chromosome (hDpl4)alters X chromosome recombination frequencies. T o the left of the insertion site recombination is either unaffected or slightly increased. T o the right, however, recombination is decreased tenfold. This suppression of recombination to the right of the insertion agrees with the predictions of our model of chromosome pairing in C. eleguns (McKIM, HOWELL and ROSE1988). We proposed the homolog recognition information for each chromosome is located at or near the one end of each chromosome and that synapsis initiates at this site. The region for homolog recognition on the X chromosome was proposed to be at theleft (unc-I ) region (HERMAN and KARI 1989; MCKIM, HOWELL and ROSE1988). In the case of the hDpl4 chromosome, the progression of pairing which would have begun at the left end might have been disrupted by chromosome I sequences. Recombination is normal on the side of the insertion nearestthe homolog recognitionsite because pairing can proceed efficiently up to the insertion site. The insertionapparentlycannot efficiently be “jumped” by the pairing process resulting in the disruption ofsynapsis. The disruption of pairing and recombination on one side of a heterology is analogous to that observed for translocations [described in MCKIM, HOWELLand ROSE(1988)l. Pairing and recombination occur on one half of a reciprocal translocation but cannot initiate on the other half because there is no homolog recognition information there. Similarly, HERMAN,ALBERTSONand BRENNER (1 976)observed a polareffect on recombination with the duplication mnDpl ( V ; X ) . This duplication is attached to chromosome V and lowers recombination on the left half. If the right arm of chromosome V contains the homologrecognition sequences which initiate the pairing process as we proposed previously (McKIM, HOWELLand ROSE 1988), then the mnDpl situation would be the same as hDpl4.Unfortunately there is no accurate positioning of mnDpl on chromosome V . In summary, we can now state several features contributing to the function and organization of chromosome I . Despite carrying large amounts of chromosome I sequence, sDp2 and ILXLszTl have no obvious effect on chromosome I meiotic recombination. Thisargues against these duplications having any pairing behavior which would compete with pairing of the normal homologs. Homolog recognition and
131
pairingappears to initiate at the right end of the chromosome I with other sequences, such as some near the left end, setting the frequency and distribution of recombination events. The inability of these duplications to recombine with chromosome I allows themtobe used effectively as balancers for lethal mutations and as a source of breakpoints for genetic mapping. The duplications described hereare less stable during nuclear division thananormalchromosome. We have investigated the natureof the chromosome structures governing segregation fidelity. In addition to size, the stability of a duplication is dependent on other, as yet undescribed, structural features. We have described some of the consequences of an unstable duplication. For example, an unstable duplication is more prone to breakage events than a mitotically stable one. Further analysis, perhaps at a cytogenetic or molecular level, may help to elucidate the structure of these chromosomes and the relationship of these structures to chromosomebehavior. We wish to thank DAVIDL. BAILLIE,ANN MARIE HOWELL and RAJAE. ROSENBLUTH for valuable advice during this work and to our reviewers for comments on the manuscript. We also are grateful to T. STARR forisolating the duplications marked “b” in Table 3. Some of thestrains used wereprovided by the Caenorhabditis Genetics Center, which was supported by contract NO1-9-2113 between the National Institutes of Health and the Curators of the University of Missouri. K.McK. was supported by a Medical Research Council (MRC) (Canada) Studentship. This work was supported by grants from the MRC and the Natural Sciences and Engineering Research Council of Canada to A.M.R.
LITERATURE CITED ALBERTSON, D. G . , and J. N. THOMSON,1982 T h e kinetochores of Caenorhabditis elegans. Chromosoma 86: 409-428. BAKER,B. S., 1975 Paternal loss (pal): ameiotic mutant in D. melanogaster causing loss of paternal chromosomes. Genetics 8 0 267-296. H., and J. M. MASON,1988 Progressive loss of DNA BIESSMANN, sequences from terminal chromosome deficiencies in Drosophila melanogaster. EMBO J. 7: 1081-1086. BRENNER, S., 1974 T h e genetics of Caenorhabditiselegans. Genetics 77: 7 1-94. CARPENTER, A. T . C., 1973 A mutantdefective in distributive disjunction in Drosophila melanogaster. Genetics 73: 393-428. CLARK,D. V., T . M. ROGALSKI, L. M. DONATIand D. L. BAILLIE, 1988 T h e unc-Z2(IV)region of Caenorhabditiselegans: genetic analysis of lethal mutations. Genetics 1 2 9 345-353. COULSON, A., J. SULSTON, S. BRENNER and J. KARN,1986 Towards a physical map of the genome of the nematode Caelzorhabditis elegans. Proc. Natl. Acad. Sci. USA 83: 7821-7825. COULSON, A., R. WATERSTON, J. KIFF,J. SULSTON and Y.KOHARA, 1988 Genome linking with yeast artificial chromosomes. Nature 335: 184-186. CROW,E. L., and R. S. GARDNER, 1959 Confidence intervals for the expectation of a poisson variable. Biometrika 4 6 441-453. DELONG,L., L. P. CASSONand B. J. MEYER,1987 Assessment of X chromosome dosage compensation in Caenorhabditis elegans by phenotypic analysis of lin-14. Genetics 117: 657-670. EDGLEY,M. L., and D. L. RIDDLE,1987 Caenorhabditiselegans, pp. 351-365 in Genetic Maps 1987, Vol. 4, edited by S. J. O’BRIEN.Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
132
K. S. McKim and A. M. Rose
FERGUSON, E. L., andH. R. HORVITZ,1985 Identification and characterization of 22 genes that affect the vulval cell lineages of the nematode Caenorhabditis elegans. Genetics 1 1 0 17-72. FODOR,A., and P. DEAK,1985 The isolation and genetic analysis of a C. elegans translocation ( s z T I ) strain bearing an X-chromosonle balancer. J. Genet. 64: 143-157. HABER, J. E., and P. C. THORBURN, 1984 Healing of broken linear dicentric chromosomes in yeast. Genetics 106: 207-226. HABER, J. E., P. C. THORBURN and D. ROGERS,1984 Meiotic and mitoticbehavior of dicentric chromosomes in Saccharomyces cerevisiae. Genetics 106 185-205. HERMAN, R. K., 1984 Analysis of genetic mosaics of the nematode Caenorhabditis elegans. Genetics 108: 165-180. HERMAN,R. K., D. G. ALBERTSONand S. BRENNER,1976 Chromosomerearrangements in Caenorhabditiselegans. Genetics 83: 91-105. HERMAN, R. K., and C. K. KARI,1989 Recombinationbetween small X chromosome duplications and the X chromosome in Caenorhabditis elegans. Genetics 121: 723-737. HERMAN, R. K., J. E. MADLand C. K. KARI,1979 Duplications in Caenorhabditis elegans. Genetics 92: 419-435. HORVITZ,H. R., S. BRENNER,J. HODCKINandR.HERMAN, 1979 Auniform geneticnomenclatureforthenematode Caenorhabditis elegans. Mol. Gen. Genet. 175: 129-133. HOWELL,A. M., 1989 Essential genes in a region of chromosome I in Caenorhabditis elegans. Ph.D. thesis, University of British Columbia, Vancouver, B.C., Canada. and A. M. ROSE, HOWELL,A. M., S. G. GILMOUR,R. A. MANCEBO 1987 Genetic analysis of a large autosomal region in Caenorhubditis elegansby the use of a free duplication. Genet. Res. 49: 207-21 3. KIM, J.S., and A. M. ROSE,1987 The effect of gamma radiation on recombination in Caenorhabditis elegans. Genome 2 9 457462. KIMBLE, J., and D. HIRSH, 1979 Postembryonic cell lineages of the hermaphrodite and male gonads in Caenorhabditis elegans. Dev. Biol. 7 0 396-417. KING, B. O., and M.-C. YAO, 1982 Tandemly repeated hexanucleotide in Tetrahymena rDNA free end is generated from a single copy during development. Cell 31: 177-182. LEIGH,B., 1976 Ring chromosomes and radiation induced chromosome loss, pp. 505-528 in The Genetics and Biology of Droand E. NOVITSKI. sophila, Vol. lb, edited byM. ASHBURNER Academic Press, London. LEVIS,R. W., 1989 Viable deletions of a telomere from a Drosophila chromosome. Cell 5 8 791-801. LEWIS,J. A.,C. H. WU, H. BERGand J. H. LEVINE,1980 T h e genetics of levamisole resistance in the nematode C.elegans. Genetics 95: 905-928. MCCLINTOCK, B., 1938 T h e production of homozygous deficient tissue by means of aberrant mitotic behavior of ring-shaped chromosomes. Genetics 23: 3 15-376. MCCLINTOCK, B., 1941 T h e stability of broken ends of chromosomes in Zea mays. Genetics 26: 234-282. MCKIM,K. S., A . M. HOWELLand A. M. ROSE, 1988 T h e effects of translocations on recombination frequencyin Caenorhabditis elegans. Genetics 120: 987-1001. MENEELY,P. M., and R. K. HERMAN,1981 Suppression and function of X-linked lethal and sterile mutationsin Caenorhabditis elegans. Genetics 97: 65-84.
MILLER,0..1963 Cytological studies of asynaptic in maize. Genetics 48: 1445-1466. MULLER,H. J., and I. H. HERSKOWITZ,1954 Concerningthe healing of chromosome ends produced by breakage in Drosophila melanogaster. Am. Nat. 88: 177-208. ROBERTS,P. A., 1975 Insupport of thetelomereconcept. Genetics 80: 135-142. ROSE,A. M., 1980 Genetic studies on the gene coding for paramyosin in Caenorhabditis elegans: unc-15 and the adjacent region. Ph.D. thesis, Simon Fraser University, Burnaby, B.C., Canada. ROSE,A. M., and D. L. BAILLIE,1979 Effect of temperature and parental age on recombination and nondisjunction in Caenorhabditis elegans. Genetics 92: 409-41 8. ROSE,A. M., and D. L. BAILLIE,1980 Genetic organization of the region around unc-15(1),a gene affecting paramyosin in Caenorhabditis elegans. Genetics 96: 639-648. ROSE,A. M., D. L. BAILLIEand J. CURRAN,1984 Meiotic pairing behavior of two free duplications of linkage group I in Caenorhabditis elegans. Mol. Gen. Genet. 1 9 5 52-56. ROSENBLUTH, R. E., and D. L. BAILLIE,1981 Analysis of a reciprocal translocation, eTI ( I I k V ) , in Caenorhabditis elegans. Genetics 99: 415-428. ROSENBLUTH, R. E., C. CUDDEFORD and D. L. BAILLIE,1985 Mutagenesis in Caenorhabditis elegans. 11. A spectrum of mutational events induced for 1500 R of gamma-radiation. Genetics 109: 493-51 l . ROSENBLUTH, R. E., R. C. JOHNSEN and D. L. BAILLIE, 1990 Pairing for recombination in Caenorhabditis elegans: a model based on the effect of deficiency-heterozygosity of recombination. Genetics (in press). R. C. JOHNSEN,L. M. ADDIROSENBLUTH, R. E., T . M . ROGALSKI, SON and D. L. BAILLIE,1988 Genomicorganization in Caenorhabditiselegans: deficiency mapping on linkage group V (left). Genet. Res. 52: 105-118. RUDIN,N., and J. E. HABER,1988 Efficient repair of HO-induced chromosomal breaks in Saccharomyces cerevisiae by recombination between flanking homologous sequences. Mol. Cell. Biol. 8: 3918-3928. SANDLER, L., and G. BRAVER,1954 T h e meiotic loss of unpaired chromosomes in Drosophilamelanogaster. Genetics 3 9 365377. SEARS,E. R., 1952 Misdivision ofunivalents in commonwheat. Chromosoma 4 S.535-550. SIGURDSON,D. C., G. J. SPANIER and R. K. HERMAN, 1984 Caenorhabditis elegans deficiency mapping. Genetics 108: 331-345. STARR,T., A. M. HOWELL,J. MCDOWALL,K. PETERSand A. M. ROSE,1989 Isolation and mapping of DNA probes within the linkage group I gene cluster of Caenorhabditis elegans. Genome 32: 365-372. WARD,S., and J. S. CARREL,1979 Fertilization and sperm competition in the nematode Caenorhabditis elegans. Dev. Biol. 73: 304-321. WILLIAMSON, J. H., and D. R.PARKER, 1976 Recombination between the X and Y chromosomes, pp. 701-720 in The Genetics and Biology of Drosophila, Vol. lb, edited by M. ASHBURNER and E. NOVITSKI.Academic Press, London. Communicating editor: W. M. GELBART
Chromosome I Duplications in Caenorhabditis elegans Kim S. McKim and Ann M. Rose Department of Medical Genetics, University of British Columbia, Vancouver, BritishColumbia, Canada Manuscript received April 26, 1989 Accepted for publication September 29, 1989 ABSTRACT We have isolated and characterized 76 duplications of chromosome I in the genome of Caenorhabditis elegans. The region studied is the 20 map unit left half of the chromosome. Sixty-twoduplications were induced with gamma radiation and 14 arose spontaneously. The latter class was apparently the result of spontaneous breaks within the parental duplication. The majority of duplications behave as if they are free. Threeduplications are attached to identifiable sequences from other chromosomes. The duplication breakpoints have been mapped by complementation analysis relative to genes on chromosome I. Nineteen duplication breakpoints and seven deficiency breakpoints divide the left half of the chromosome into 24 regions. We have studied the relationship between duplication size and segregational stability. While size is an important determinant of mitotic stability, it is not the only one. We observed clear exceptions to a size-stability correlation. In additionto size, duplication stability may be influenced by specific sequences or chromosome structure. The majority of the duplications were stable enough to be powerful tools for gene mapping. Therefore the duplications described here will be useful in the genetic characterization of chromosome I and the techniques we have developed can be adapted to otherregions of the genome.
HARACTERIZATION of theCaenorhabditis eleguns genome is progressing at a rapid rate. The progress has been recently stimulated by the physical mapping of contigs which are tied to the genetic map (COULSON et al. 1986, 1988). Parallel to the physical mapping are the genetic mapping effortsof C. elegans researchers who are isolating mutations and finding their positions on both the genetic and physical maps. A few labs have undertaken the task of genetically characterizing large regions of the genome with regard to the organizationof essential genes (MENEELY and HERMAN 1981; SIGURDSON, SPANIERand HERMAN 1984;CLARK et al. 1988; ROSENBLUTH et al. 1988). The present study was undertaken to provide two contributions toward our understanding of chromosome I organization. The first was to make rearrangements to be used in the mapping of genes in a large region of chromosome I . Previously HOWELL et al. (1987) described a system for identifying complementation groups on the left half of chromosome I using the duplication sDp2. The large number of mutations that have been isolated requires the generation of rearrangements in order to increase the mapping resolution and reduce the number of complementation tests. The ultimate goal of these studies is to correlate the physical and genetic maps; that is to identify the mutant phenotype for coding regions in the cloned DNA.
C
T h e publication costs of this article were partly defrayedby the payment of page charges. This article nust therefore be hereby marked "advertisement" in accordance with I8 U.S.C. 91734 solely to indicate this fact. Genetics 124: 1 1.5- 132 (January, 1990)
The second contribution of this study was to use these rearrangements to study the physicalbasisof chromosomebehavior.Earlier studies with translocations (ROSENBLUTH and BAILLIE1981; MCKIM, HOWELL and ROSE1988) and duplications (HERMAN, MADLand KARI 1979; HERMAN and KARI 1989; ROSE, BAILLIEand CURRAN 1984)have demonstrated the utility of this approach. A large numberof rearrangement breakpointshave been isolated as gamma radiation-induced or spontaneous shorter derivatives of preexisting duplications. We used two duplications to isolate our derivatives. The first duplication, sDp2(Z;f), was initially characterized by ROSE, BAILLIEandCURRAN(1984)and et al. (1 987) to recoverlethal mutaused by HOWELL tions. The second duplication is a chromosome segregated from a reciprocal translocation strain. This chromosome, I L X L s z T l (McKIM, HOWELLand ROSE 1988), is one half of the s z T I ( I ; X ) translocation (FoDOR and DEAK1985) and is used here in addition to the euploid genetic background.I L X L s z T l is somewhat a largerthan sDp2 and is attachedatoneendto fragment of the X chromosome.Neither of these duplications recombine with the normal chromosome I. This paperdescribes the characterization of 76 new duplications with regardtothe genetic position of their breakpoints and their meiotic/mitotic properties. Most of these duplications are free; that is they are not translocated toanother chromosome. The duplications contribute at least 19 breakpoints, which
116
K. S. McKim and A. M. Rose
in combination with seven deficiency breakpoints, divide the left half of chromosome Z into at least 24 regions. Using these duplications we have studied the relationship between chromosome size and segregationalstability. In addition, we have observed that free duplicationsspontaneouslybreakwith a frequency that correlates with their segregational instability. MATERIALS AND METHODS
General: Nematodes were grown and maintained on nematode growth medium streaked with Escherichia coli strain OP50 (BRENNER 1974). Unless otherwise noted, all experiments were carried out at 20" (ROSEand BAILLIE1979). The standard wild-type strain, N2, and some mutant strains Simon derived from N2 were obtained from D. L. BAILLIE, Fraser University, Burnaby, Canada or from the Caenorhabditis Genetics Center at the University of Missouri, Columbia, Missouri. fog-1 (q253ts) was received from K. BARTON, University of Wisconsin. sup-I I (n403n682)and nDf25 were received from R. H. HORVITZat MIT. spe-ll(hc90) was received from D. SHAKES, Carnegie Institutionof Washington. The nomenclature used in describing these strains is the uniform system outlined in HORVITZet al. (1979). The wild-type phenotype is oftenabbreviated Wt. The following mutant genes and alleles were used: I dpy-5(e61); dpy-5(h660); bli-3(e767);b M ( e 9 3 7 ) ; dpy14(e188);egl-30(n686); egl-30(n715);fog-l (q253ts); speI I (hc90); lin-6(h92); lin-I 7(n671);sup-I I (n403n682); unc-1I(e47); unc-I3(e450); unc-I4(e57); unc-29(e403); unc-35(e259); unc-37(e262); unc-38(e264); unc-40(e271); unc-55(e402); unc-57(e406); unc-63(e384); unc73(e936); unc-74(~19); unc-89(e1460). 111 unc-36(e251). X dpy-3(e27);dpy-7(e88); dpy-7(sc27); dpy-8(e130); lonZ(e678); unc-3(e151); und"(e112). Most of the EMS induced lethal mutations on chromosome I were isolated and described by ROSE and BAILLIE (1980) and HOWELL et al. (1 987).Also included in this study are three mutations, let-373(h234), let-374(h25I) and let375(h259),mapped to the hDf6 region by HOWELL(1989). One additional mutation, let-400(h269),has been positioned between dpy-14 and unc-I3 by T. STARR(personal communication). The reciprocal translocation s z T l ( 1 ; X ) was originally described by FODORand DEAK(1985) and further characterized byMCKIM, HOWELLand ROSE(1988). The reciprocal translocation hT2(I;Ill) was isolated by K. PETERS. AS a heterozygote it suppresses recombination on chromosome I from the left end to the right of unc-29. The dpy-5(h660) mutation is linked to hT2(I;Zll) and was induced by J. BABITY. The deficiencies used in this study were hDf6 (McKIM, HOWELL and ROSE1988), nDf23, nDf24 and nDf25 (FERGUSON and HORVITZ 1985)andsDf4 (ROSE1980; HOWELLet al. 1987). Some characteristics of sDpZ(1;f) (ROSE, BAILLIEand CURRAN 1984) and ILXLszTI (McKIM, HOWELL and ROSE1988) aredescribed in RESULTS. Mutagenesis with gamma radiation:The screening protocols are described in RESULTS. Gravid worms were treated with either 1500 or 3000 rad of gamma radiation from a cobalt-60 source (as recommended by ROSENBLUTH, CUDDEFORD and BAILLIE1985). The dose rate varied between 9.0 and 7.5 rad/sec. Levamisole selection: Some shortened derivative duplications were selected on the basisof resistance to 1 mM levamisole usingplates prepared as described in LEWISet al.
(1980). In general, two or three unc-74 dpy-5; Dpx or unc63 dpy-5; Dpx hermaphrodites were placed on each small Petri plate. When the Fz progeny were larvae, they were washed off in M 9 buffer (BRENNER 1974) and seeded onto new large (100 X 15 mm) plates containing 1mM levamisole. Two or three days later the plates were screened for levamisole resistant non-Dpy worms. Survivors were placed onto normal plates to observe their progeny. Recombinationmapping: Recombination frequencies between pairs of markers were determined by scoring the self progeny of cis-heterozygous hermaphrodites under the conditions described by ROSEand BAILLIE (1 979).For standard crosses, the total number of progeny was calculated as 4/3 (the number of Wt plus one recombinant class). The recombination frequency, p , was calculated using the formula p = 1 - (1 - ZR)", where R is the fraction of marked (non-Wt) recombinant individuals over total progeny (BRENNER 1974). Map distances are reported as map units (m.u. = loop). Confidence limits of 95% were calculated using the statistics of CROWand GARDNER (1959). Special consideration was given to the calculation of recombination in crossesinvolving duplications. When scoring recombination from the progeny of dpy-x unc-y/+ Dp(1;f ) hermaphrodites, the calculation was modified for the fraction of gametes which carried the duplication. In most cases this was 41-42% (see RESULTS). The recombinant frequency was calculated as
+;
p
= [(0.17W - 0.170)
- ((0.170 - 0.17W)'
- 4(0.09W + 0.03D)(0.74D))"]/2(0.09W + 0.030)
+
where W was the number of wild types (or Wt Unc-13 with sDp2). Depending on the cross, D was the number of Dpy or Dpy Unc progeny. The frequency at which a duplication was found in the germ line was calculated by scoring the progeny of m/m/ Dpx hermaphrodites; where m was dpy-5, unc-I1 or unc-13. The fraction of duplication bearing gametes was x/(2 x); where x = Wt/mutant. In these calculations the duplication homozygotes were not considered as they were not scored. These worms were usually small and slow growing. Complementation tests: Visible mutation with duplication [Dp(Z;f)]:Mostof the mapping of the duplications with respect to visible markers was done using markers linked to either dpy-5, dpy-14 or unc-13. The exception to this was the mapping of Dp[unc-1I (+) dpy-5(-)] duplications. This mapping was done using visible mutations linked to unc-11. T o facilitate mapping, most duplications isolated with chromosomes other than dpy-5dpy-14 were resegregated into dpy-5 dpy-14;Dp strains. In order to map a mutation (unc-x),the progeny of uncx/+; Dp(Z;f)hermaphrodites were examined to see if their unc-x; Dp progeny were Unc (i.e., Dp is unc-x(-)) or wild type ( D p is unc-x(+)). Most parental hermaphrodites were constructed by crossing unc-x bearing males to a duplication stock. For example, unc-x dpy-5 hermaphrodites were crossed to unc-11 dpy-14; O / s z T I ( I ; X ) [ + ;lon-21 males. The only resulting male progeny (+ unc-xdpy-5 +/uric-I1 ++ dpy-14) were crossed to duplication strains. In crosses to dpy-5 dpy-14;h D p z ( 1 ; f )strains, most wild-typehermaphrodite progeny were unc-x dpy-5 +/+ dpy-5 dpy-14; hDpz. If the duplication was [dpy-14(+)], then + dpy-5 dpy-14 / unc-1I + dpy-14; hDpz hermaphrodites were also recovered. These possibilities were distinguished by examining the resulting self fertilization progeny. Upon examining the progeny of the desired (unc-x dpy-5 +/+ dpy-5 dpy-14; hDpz) hermaphrodites, Dpy-5 and Unc-x progeny were observed if the duplication did not carry unc-x(+)(in the ratio 4 Wt: 2 Dpy-
+
Duplications in C. elegans
117
5: 2 Dpy-14: 2 Unc-x: 1 Dpy-5 Dpy-14: 1 Dpy-5 Unc-x). No used to indicate themarkers which are deleted. Unc-X progenywereobserved if the duplication carried Dp[dpy-5(+) dpy-14(-)]refers to a duplication which unc-x(+). The only exceptions to the above procedure recarries the wild-type dpy-5 allele but is deleted in the sulted from recombinant chromosomes. dpy-14 region. This is not meant to indicate the duIn crosses of unc-x dpy-5 +/uric-1 1 dpy-14 males to unc-11 dpy-5; h D p z ( 1 ; f ) strains (where hDpz was u n c - l l ( - ) plication carries a dpy-14 mutation. dpy-5(+)),most wild-type hermaphrodite progenywere uncInduceddeletion of duplicationchromosomes: 1 1 dpy-5 dpy-5 dpy-14; hDpz(1; f ). For crosses tounc-11 Duplications of different sizes with one end in comdpy-5; hDpz(1;f ) strains (where hDpz was u n c - l l ( + ) dpymon would provide a linear array of breakpoints for 5 ( - ) ) , the mutation to be tested was linked to unc-11 and rapidmapping of mutations. T o generate such an brought in through the male of genotype unc-11 unc-x +/ ++ dpy-5. Usingsimilarmethodology, dpy-14 unc-13; array, we used gamma radiationto shortenpreexisting hDpz[dpy-5(-) unc-I3(+)] duplications were crossed touncduplications. Shortened duplications were recovered x dpy-14 +/++ unc-13 males when mapping a visible mutaby screening for the exposureof mutant phenotypes tion. previously rescued by the duplication. The protocol In some of the complementation experimentsunc-x dpyfor one such experiment is diagrammed in Figure 1. 5 / + dpy-5; hDpz hermaphroditeswere generated by crossing duplicationbearingmalestohomozygous mutanthersDp2 derivatives: In the first screen (Figure l), demaphrodites. For example,h D p l 8 males were generated by rivatives of sDp2 were recovered which had breaks in ; males todpy-5 dpy-14; crossing d p y - 5 ; O / s z T l ( I ; X ) [ +lon-21 the unc-ll-dpy-5 region. T h e Po had a wild-type pheh D p l 8 hermaphrodites.Wild-type malesfromthiscross notype and segregated Wts and Dpy Uncs. T h e F I (dpy-5 + /dpy-5 dpy-14; hDpl8)were crossed to unc-x dpy-5 progeny of gammaradiation treated unc-11dpy-5/ hermaphrodites to generate unc-x dpy-5/ + dpy-5; hDpl8 hermaphrodites. worms were screened for rare Uncunc-11 dpy-5; sDp2 sDp.2-balancedlethal with duplication [Dp(Z;f ) ] : Lethal 11 and Dpy-5 individuals. Because no recombination bearing strains (i.e., dpy-5 let-x unc-13; sDpZ(1; f )) were has been observed between the normal chromosomes crossed tounc-11 dpy-14;O/ szTl (I;X)[ lon-21 males. The and CURRAN 1984), the rare and sDp2 (ROSE,BAILLIE resulting male progeny (+ dpy-5 let-x unc-l3/unc-ll dpy-14 +) were then crossed to dpy-5 dpy-14; hDpz hermaUnc-1 1s and Dpy-5s could have resulted from: (1) phrodites. Maleprogenycarrying sDp.2 have a distinctive shorter duplications derived from a break between phenotype and do not mate. Wild-type hermaphrodite progu n c - l l ( + ) and dpy-5(+) on sDp2, (2) a point mutation eny from this cross were dpy-5 let-x unc-13/dpy-5 dpywithin unc-11 or dpy-5 on sDp2, (3) somatic loss of dpy-5 dpy-14 /unc-11 + dpy-141 hDpz if 14 +/ hDpz or sDp2 resulting in a mosaic worm, or (4) intragenic hDpz was dpy-14(+). These twoclasseswere distinguished by examining their progeny. If the duplication carried letrevertants.Thirteen exceptional events were rex(+), Dpy-5 and Unc-13 progeny were observed in the ratio covered. T h e complementation testing described be6 Wt: 2 Dpy-5: 1 Dpy-5Dpy-14: 2 Unc-13: 2 Dpy-14. low revealed 12 class (1) mutations and one class (2) hDp[dpy-5(-) unc-l3(+)] strains were tested by generating mutation (sDp2 [ dpy-5( h585)l). Four duplications with dpy-5 let-x unc-131 dpy-14 unc-13; hDpz hermaphrodites and looking for viable Dpy-5 progeny. These hermathe right end and eightduplications with the left end unc-13; phrodites were recovered from crossing dpy-14 of sDp2 deleted were recovered(Table1).Intwo let-x unc-13 hDpz hermaphrodites to dpy-14 ++/dpy-5 cases, two duplications were recovered fromthe same males. Po. hDp3 and hDp7 were isolated from the same Po Lethal with. lethal or deficiency: These complementation tests requiredthat the two lethal mutations be tightly linked and hDp5 and hDp21 were isolated from the same Po. to a common visible marker. Mostof the lethal mutations hDp3 and hDp7 have identical breakpoints while the or deficiencies isolatedin this lab were linked to dpy-5. Any latter two are different. Both of these situations could that were not had a dpy-5 marker put on the chromosome have beencaused by a premeiotic mutationwhich was, by recombination.Complementationtestsweredone by in the case of hDp21, accompanied by a spontaneous crossing one lethal heterozygote to another lethal or deficiency heterozygote and looking for Dpy-5 males and fertile postmutation shortening (see below). Dpy-5 hermaphrodites in the F1 (or Dpy-5 Unc-13 if both A similar approach was used to obtainbreaks in the lethals were also linkedto unc-13). dpy-5 dpy-14 region. In this screen, dpy-5 dpy-l4/dpy5 dpy-14; sDp2 worms were treated with gamma raRESULTS diation and their F1 progeny screened for rareDpy-5 and Dpy-14 worms. T h e same possibilities as outlined T h e primary goal of this research was to identify a above apply to this screen. Eight Dpy-14 worms were set of mutations which could be used in the physical has recovered(Table 2). Complementationtesting characterization of chromosome I in C. elegans. We describe here theisolation of duplications with breaks shown all of these to be shortened derivatives of sDp2. in the left half of chromosome I . These were mapped Inaddition,one unlinkeddominant Dpy mutation with respect to known visible and lethal mutations in ( h 6 3 0 ) was recovered. Unlike theprevious screen, no this region. Finally we have characterized these dupliduplications with the left end (dpy-5) deleted were cations with respect to segregational stability and recovered. structural integrity. When describing duplication ILXLszTl derivatives: A third screen using a different chromosomes, the following terminology has been duplication, ZLXLszTl, was used. This duplication is
+
++
+/+
+; +
+
+
+
+
++
+
+
+
+
+
K. S. McKim and A. M. Rose
118 sDpZ(Z;f)
unc;ll dpy-5 unc- 11 dp7-5 Treat with gamma r a d i a t i o n
I
FIGURE 1 .-Protocol for isolating shortened duplications. This figureillustrates the experiment summarized in Table 1.
1-
= =
" "
unc-11 dpy-5
11
unc-
dpy-5
"
"
unc-11 dpy-5
unc-11 dpy-5 "
"
Uric- 11
DPY - 5 TABLE 1
TABLE 2
Isolation of duplications breaking between unc-11 and dpy-5
Isolation of duplications breaking between dpy-5 and dpy-14
Dose
1500 R 3000 R
Dp[ull
+ d5-]
1 3
Dp[ull-d5+] Chromosomes
4 4
16,000 7,900
Rate
3.1 8.9
X X
one of the translocation chromosomes fromszTl (I;X) that has been recovered in a strain where it exists in addition to the euploid complement of chromosomes (McKIM, HOWELL andROSE 1988). The nomenclature describes thestructure of this chromosome. I"X"szT1 carries the portion of chromosome I left of unc-29 ( I L )joined to a fragment of the X-chromosome (X"). This X-chromosome portion carriesunc-I (+) and dpy-3(+) butnot unc-20(+) (McKIM, HOWELLand ROSE1988). dpy-5 unc-13; I"X"szT1 worms are wildtype and Him (High incidence of males) due to X nondisjunction. The dpy-5unc-13; I"X"szT1 worms were treated with gamma radiation and their F1 progeny were screened forrare Dpy-5 and Unc-13 individuals. Twenty-four Unc-13 and 18Dpy-5 hermaphrodites were recovered (Table 3). Complementation mapping (below) showed 23 of the Uncs and all of the Dpys were shorter derivatives of ILXLszTI.One Unc13 strain carried no dpy-5 marker. Two dpy-5(+) dpy1 4 ( - ) duplications, hDp40 and hDp63, caused the hermaphrodites to be sick and slow growing. dpy-5 dpy-14; Dp strains could not be constructed for these two duplications and therefore no extensive mapping of their breakpoints was done. Unlike the sDp2 screen of this region, we recovered duplications with the left end (dpy-5) deleted. Included in Table 3 as entries in columns two and three is a worm which segregated Unc-13, Dpy-5, Wt and Dpy-5 Unc-13progeny. The Wts were picked and these segregated more Wt, Unc-13, Dpy-5, and Dpy-5 Unc-13 worms. We were able to isolate two
Dose
1500 R 3000 R
D p [ d 5 + d l 4 - ] D p [ d Y d l 4 + ] Chromosomes
5 3
0 0
4500 2500
Rate
1.1 X lo-' 1.2 X 10"
duplications from the Wt strain, the Unc-13 worms carried an unc-13(-) duplication ( h D p 6 4 ) while the Dpy-5 worms carried a dpy-5(-) duplication ( h D p 7 3 ) . Since hDp64 and hDp73 have similar breakpoints, they could be products of the same mutational event. The recovery frequency for theILXLszTlderivative duplications (Table 3) is approximately the same as that for the sDp2 derivative duplications in the dpy-5 dpy-14 screen (Table 2). The duplications with breaks in the unc-I I dpy-5 region were recovered four times less frequently (Table 1) thanduplications with breaks in the dpy-5 unc-13 region. Spontaneous derivatives: A fourth source of duplications was the spontaneous shortening of various duplications. Their isolation is described laterunder "SegregationAnalysis." We have characterized 14 of these events and their analysis is reported with the analysis of the gamma ray induced duplications. Mapping duplication breakpoints: sDp2-balanced mutations: HOWELLet al. (1987) used sDp2 to isolate and characterize 58 EMS inducedmutations in 40 essential genes. These authors also reported the isolation of 29 gamma ray induced mutations of essential genes. None of these mutations were found to delete more than one locus. In the present study the latter mutations are indicated by allele numbers h 1 7 I and below. Genes identified by an EMS induced mutation are indicated by their name. All sDp2-rescued mutations were recovered and maintained by providing the wild-type gene copy in the form of the free duplication. Chromosome Z genes identified by HOWELLet
Duplications in C . eleguns TABLE 3 Isolation of duplications breaking betweendfiy-5 and unc-I3 Dose
1500 R a b 3000 R a b
Dp[dS+u13-] Dp[dS-u13+]
Chromosomes
Rate
46006 6
4 3
4600
2.2 X lo-’ 1.9 X lo-’
9 3
8 3
5000 4700
3.4 X lo-’ 1.3 X lo-’
al. (1 987),ROSEand BAILLIE(1 980) andEDGLEY and RIDDLE(1 987)were tested for complementationwith our duplications. Translocation(szTI(Z;X))balancedmutations: We previously described the properties of the translocation s z T l ( Z ; X ) (McKIM, HOWELLand ROSE 1988) a reciprocal interchange between chromosomes Z and X . Thatreport describedaset of chromosome I gamma ray induced lethal mutations balanced by s z T l ( Z ; X ) . Only one deficiency, hDf6, was found in that set. The rest were point mutations at the resolution of the mapping. Mutations designated h546 and above are from this set. Mapping strategy: Complementationtesting using autosomal duplications requires observation of the F2 generation to determine if the duplication carries the wild-type copy ofthe gene. The object of our protocol was to selectively recover duplication-bearing strains from a cross between a duplication strainand a strain carrying the mutation to be tested. The F2 progeny from these crosses were observed to see if the duplication complemented the tested mutation. The following protocol allowed for rapid mapping of large numbers of mutations. If the male was of the correct genotype, then in a cross to a duplication hermaphrodite,all the Wt progeny carried the duplication and at least half of these also carried the mutation to be tested. We found the pseudolinkage betweenchromosomes Z and X in s z T I ( Z ; X ) heterozygotes to be useful for producing therequired males. s z T l ( Z ; X ) heterozygous males have a Lon-2 phenotype (FODORand DEAK1985) and carry only therearranged X chromosome. These males cannot sire hermaphrodite progeny with a paternal normal chromosome Z. For the same reason these males sire only male progeny which carry the paternal normal chromosome I. The Lon males were used togenerate males truns-heterozygous forthe mutation a to be tested and anothermutation b in the duplicationstrain (MATERIALS AND METHODS). The mutation to betested was linked to a common marker c. The trans-heterozygous males ( a c b ) were then crossed to duplication strains ( c b; Dp).With the exception of rare recombinants, at least half the wildtype hermaphrodite progeny carried both the duplication and themutation to be tested ( a c +/+ cb; Dp).
+/++
119
If the duplication complementedthe mutation a , then a non-mutant a c; Dp strain could be established. Mappingresults: Duplications hDp8, hDp9, hDpl0 and h D p l 1 were mapped using mutations linked to unc-11 (e47).These four have a single breakpoint in the unc-73 dpy-5 region. Their left endpoint is probably the same as sDp2 because all four complement bZi-3. The rest of the duplications were mapped with mutations linked to dpy-5(e61), dpy-l4(e188)or unc13(e450). Considerable mapping against t h e essential genes of HOWELLet al. (1987) in this region was carried out. We also mapped 16 gamma ray induced mutations of essential genes [ 11 isolated with szTI (McKIM, HOWELLand ROSE 1988) and five isolated with sDp2 (HOWELL et aZ. 1987)Jwith the duplications; these are indicated by their allele designation ( h ) . The results of this mapping combinedwith the deficiencies in the region are shown in Figures 2, 3, 4 and 5 . Mapping of genes with respect to thedeficiencies sDf4 and hDf6 is published elsewhere (HOWELL et al. 1987; MCKIM,HOWELLand ROSE1988; HOWELL 1989). T h e duplications in combination with sDf4 and hDf6 have subdividedthe 4 m.u. unc-I1 unc-13region into 18 intervals. None of the duplications tested had a break in the unc-37-dpy-14 interval (Figures 4 and 5), even though this interval is recombinationally 1/5 of thedpy-5 unc-I3 region (EDGLEY and RIDDLE1987). Twelve of 41 ILXLszTl duplications broke in the dpyI 4 u n c - 1 3 region; an interval genetically 1/5 the size of the dpy-5 unc-13 region (ROSEand BAILLIE1979). We have mapped these duplication breakpoints with five mutations in the dpy-14 unc-13 region (Figures 4 and 5). Ofthe12 duplicationbreakpoints in this interval, seven had a breakpoint similar to sDp2 (between unc-14 and Zet-75) and the other five had different breakpoints. The breakpoint of hDp62 is now the closestphysical marker on the right of dpy-14. Two of the duplications breaking between let-75 and unc-13,hDp64 and hDp73, were isolated in the F, worm and, as mentionedearlier,might have been reciprocal products of the same mutational event. Spontaneous shortening of duplications in the region to the left of dpy-5 (see below) provided some new breakpoints. All duplications which break between unc-38lunc-63 and dpy-5 were of spontaneous origin. Most duplications appeared to result from a single break in the selected regionproducingaterminal deletion (Figures 2, 3, 4 and 5). At least five of the gamma ray induced duplications, however, contained morethanonebreakpoint(Figure 2). hDp69 and hDp71, two ZLXLszTl[dpy-5(-)unc-13(+)] duplications, carry an internal deletion around dpy-5. These two duplications are dpy-5(-) but complemented Zet362, unc-I1 and unc-74. hDpl8, in addition to being deleted for the dpy-I4 region, was also found to be
K. S. McKim and A. M. Rose
120
A
1 -
h61,h167,h17J,h655 blt-3 egi-30 unc-35
let-362 I
ltn-6
1%"-17
I
sup-ll
I
I
unc-11
f0q-J
$7,-5
I
=+ =+ hDpl0
-4
I
unc-40bit-4 l
l
h654 h601 dpy-14 unc-13 unc-29 I
1
I
1
-
hDp69 hDp71
B ILXLSZTI
==iFt unc-l d q - 3 unc-fO
ion-2
ldfy-8
dpr-7
- - Unf-3
X
H
2 I.U.
FIGURE2.-(A) Duplications and deficiencies on the left half of chromosome I . Chosen examples of duplications are shown here. All the markers on the map except bli-?, egl-30 and unc-35 were tested for complementation with each duplication. When it is not known if a duplication complements a marker, the duplication is drawn with a gapped double line. Gamma induced lethal mutations h61, h167, h171 to the left of hDp18.(B) Genetic map of the X chromosome. T h e X chromosome portion of ILXLszTland the and h655 were only. mapped .. positions of X-linked markers used in this study are shown.
deficient for the left end of the chromosome. hDpl8 complemented unc-11 and fog-I but not sup-11, linI7,lin-6 or let-362. Thus, atleast one unselected break occurred in theregion between sup-I I and fog-I. hDp62 and hDp72, two of the I L X L s z T l [dpy-5(+)unc13(-)] derivatives, failed to complement egl-30, let362 and lin-6. In addition, hDp72 complemented supI I but hDp62 did not. N o other distal markers have been tested but, like hDpl8, these duplications probably have a terminal deletioncaused by a second break to the left of unc-1I. Some of the gamma ray induced lethal mutations mapped in this study were not separable from dpy-5 or unc-13 in recombination experiments(R. MANCEBO and K. MCKIM, unpublished results). Duplication mapping of these mutations has positioned some of
them into unexpectedplaces on the map. Four gamma ray induced lethal mutations (h61, h167, h171, h655) recombinationally inseparable from dpy-5 have been mapped to the left end of the chromosome because they failed to complement hDpl8 while complementing hDp12. These four mutations fully complement each other. Two other mutations (h601, h654)recombinationally inseparable from dpy-5 mapped to the right of unc-13 based on complementation with I L X L s z T I . These two mutationswerecomplementationtested against the deficiencies in the unc-29 region. h654 maps inside nDf24 and nDf25 but outside nDf23. h601 is inside all three deficiencies. h654 and h601 fully complement each other and unc-29. Allsix of these gamma ray inducedmutations
h549 Eet-361
121
Duplications in C. elegans h550
Zet-359
unc-11 I
Zet-353 Zct-356 let-354 Eet-366 Zet-374 let-373
h565
let-351
unc-73 unc-89 unc-74 unc-57 let-375 I
1
I
I
/
let-364 unc-38 unc-63
let-363' let-371 spe-11 dpy-5
I
I
I
unc-40 I
b12-4 I
unc-37 dpy-14 UnC-13 !
I
I
sDf4 I
I
hD022*. ?5*
l"--1 m.u.
FIGURE3.-Position of breakpoints of sDp2 derivative duplications which break in the unc-11 dpy-5 region. Genes andgamma ray-induced mutations of essential genes (designated by allele number)are indicated above the line. Mapping of the essential genes with respect to sDf4, hDf6,hDp3, hDp2O and h D p 2 2 is the work of HOWELL (1989). When it is not known if a duplication complements a marker, the duplication is drawn with a gapped double line. All duplications were tested for complementation with all the other markers on the map. Duplications marked with an asterisk arose spontaneously in a duplication strain. The rest were gamma-ray induced.
should have been recombinationally separated from dpy-5.h61 and h655 suppressed recombination in trans to a bli-3 unc-11 chromosome while h60I and h654 suppressed recombination in trans to a unc-1 I dpy-14 chromosome. These results suggest these four mutations disrupt sequences required for normal recombination. Gamma ray induced lethal mutations were complementationtested against known EMS mutations of essential genes. h549 was shown by HOWELL (1989) to be an allele of let-354. HOWELL(1989) also tested h550 and h565 for allelism with known genes in the hDf6 to dpy-5 interval. No allelism was found indicating h550 and h565 define new loci. We have found that h590 is an allele of let-400 and h563 is an allele of let-378. h170 and h546 fail to complement each other but have not been tested against other lethal mutations in the region. Among the mutations mapping near the left end of the chromosome (Figure 2), h17I was found to be allelic to let-362. h D p l 4 ( I ; X ) :The duplication hDpl4 (Figure 4)was inserted into the X chromosome. T h e insertion event did not disrupt an essential gene on theX since hDpl4 homozygotes were viable. Hermaphrodites of the gen-
otype dpy-5/dpy-5;hDpl4/hDpl4 were veryslow growing; they had a generation timeof 6 instead of 4 days for hDpl4 heterozygotes. These homozygous duplicationstrains also had a Him phenotype, and were never observedto segregate Dpy-5 progeny. dpy5 / d p y - 5 ; h D p l 4 / 0males were fertileand contributed the duplication to all their hermaphrodite progeny. The insertion event was associated withloss of the sequences at the right end of sDp2 (bli-4-dpy-14) but no detectable loss occurred at the left end of sDp2. let-362,egl-30 and unc-35 were complemented by hDpl4 (Figure 2). T h e insertion site was positioned by three factor mapping the Him phenotype. T h e Him mutation must have been tightly linked to the duplication since the duplication and theHim phenotype always segregated together. The Him mutation was mapped to the right of dpy-3 and unc-20 but to the left of dpy-8 (Table 4). T h e single non-Him Dpy-3 recombinant could have been the result of adouble crossover event,one between dpy-3 and unc-20, and the second between unc-20 and the insertion. In the unc-20 dpy-8 experiment, more Unc-20 recombinants had the duplication than Dpy-8 recombinants. It may be in error to con-
122
K. S. McKim and A. M. Rose
dpy-5
h3 70 h54 6 h563
unc-40
h55G h564
I
I
I
I
let-86 bli-4
-.Y
dpy-14 unc-37
I
I
8-'\r: 7 2 2
I
I
I
let-75 unc-13 I
I
I
I L I L ~2 TI
I
hDp36:43,64
1
sDp2, hDp38,@,60 hDp58
I
hDp62
i
hDpl9,33,35,39,54,56 hDp14,16,18,31,32,34
2 hDp12,17,55,61, 72 hDp37
I hDp13,,15,41,57
FIGURE4,"Position of breakpoints of sDp2 and ILXLszT1derivative duplications which break in the dpy-5 unc-13 region. Most of the duplications extend to the left end of the chromosome (Figure 2). This was tested by complementation to let-362. With the exception of hDpl8, all of the gamma ray induced sDpZ[dpy-5(+) dpy-l4(-)] derivatives (hDps 12-19) carried this locus. Nine ILXLsrT1[dpy-5(+)unc13(-)] derivatives (hDp31, 32, 33, 34, 35, 36,37, 38,58) tested complemented let-362. hDf62 and hDp72 have left end deletions (Figure 2). See the legend of Figure 3 for more details.
unc-14 h590 let-400 dpy - 5
bli-4
unc-40
ZLXLszTl hDp48,49,52,53,7
I I==
I
hDp47,67,68,69,70
-
hDp44,51
t
I
hDp45,46,66 hDp50 65
=I+
'
hDp73
H
FIGURE5.-Position of breakpoints of dpy-5(-) derivatives of ILXLszT1. Most of these duplications do not carry any left end sequences; they are terminal deletions. Six ILXLszT1[dpy-5(-) unc-13(+)]duplications (hDp44, 45, 48, 49,53, failed 65) to complement let-362. As shown in Figure 2, hDp6Y and hDp71 are internal deletions of ILXLszTI.See the legend of Figure 3 for more details.
Duplications in C. eleguns
123
TABLE 4
TABLE 5
Mapping of hDpl4(I;X)
Duplication Stability:I. Hermaphrodite Gametes
Genotype"
Wtb
Recombi- Fraction nants' m.u.Him
(C.1.y
dpy-3 unc-2U/hDpl4
389 19 Dpy-3 18/19 6.3 (4.6-8.9) 12 Unc-2Od 0112 7 Dpy-5 017 10.6 (4.4-22.4)
unc-20 dpy-blhDp14
273 27 Dpy-8 3/27 16 Unc-20 13/16
dpy-7 unc-3/hDp14
248Unc-3 3 313 2 D ~ y - 7 ~ 012
dpy-7 unc-3/++;hDpl6
296 36 Unc-3
ND'
(0.9-3.9) 1.8
All strains were heterozygous for dpy-5(e61) on chromosome I except in the h D p l 6 experiment in which dpy-5 was homozygous. Male progeny not included. ' C.I. = 95% confidence interval. One-fourth of this classwas not observed because of dpy-5 segregating. ' m.u. were not calculated because these markers flank the insertion site.
clude from this that the insertion site maps closer to dpy-8 because, as described next,the insertion disrupts the normal distribution of recombination on the X chromosome. The insertion of chromosome I material disrupted the normal distribution of recombination events on the X chromosome (Table 4). T o the left of the insertion, recombinationwas normal or slightly higher than normal. The dpy-3 unc-20 interval was 6.3 m.u. in duplication heterozygotes compared to 5.0 m.u. on the normal map (McKIM, HOWELL and ROSE 1988). In the dpy-3 unc-20 experiment, the Dpy-5 recombinants were the result of recombination between unc20 and the insertion. In these experimentsdpy-5(e61) was heterozygous on chromosome I , therefore, in the calculation of this distance the number of Dpy-5 recombinants was multiplied by four. The 10.6m.u. observed in the u n c - 2 0 h D p l 4 interval is only slightly higher than the 7.2 m.u. normally observed in the unc-20dpy-8 region (McKIM, HOWELLand ROSE 1988). In contrast, to the right of the insertion recombination was greatly reduced. In h D p l 4 heterozygotes, recombination in the dpy-7 unc-3 interval was reduced tenfold from 21 m.u. (EDGLEYand RIDDLE1987) to 1.8 m.u. As control, a recombination in dpy-5; h D p 1 6 ( I ; f ) ; dpy-7unc-3/ hermaphrodites was scored (Table 4). h D p l 6 is of comparable size to h D p l 4 but is free. This result indicates the hyperploidy in h D p l 4 strains did not cause the reduction in X chromosome recombination frequency. Segregation analysis of the duplications: Duplication stability in the hermaphrodite: All of the duplications from the screens described in Tables 1 and 2, some of those from Table 3, andtwo of spontaneous origin, were tested for segregation stability. Strains of the genotype dpy-5/dpy-5/hDpx[dpy-5(+)unc-13(-)] or u n c - l l l u n c - 1 1 ; h D p x [ u n c - l l ( + ) d p y - 5 ( - ) ] or unc-
++
Duplication
Wt"
dpy-5; sDP2 563 1067 dpy-5; hDp2 374 dpy-5; hDp3 dpy-5; hDp4 339 351 dpy-5; hDp6 183 dpy-5; hDp20 dpy-5; hDp22 302 dpy-5; hDp23 45 1 613 dpy-5; hDpl.2 dpy-5; hDpl4 231 (27) dpy-5; hDpl5 834 464 dpy-5; h D p l 8 dpy-5; ILXLszTl 401 (30) dpy-5; hDp34 268 38 1 dpy-5; hDP36 dpy-5; hDp37 730 dpy-5; hDp54 751 dpy-5; hDp57 209 151 dpy-5; hDp62 828 unc-1 I ; sDp2 574 unc-11; hDplU unc-11; hDpll 525 (16) I unc-13; I ~ ~ ~ S Z T 438 unc-13; hDp49 597 unc-13; hDp50 500 348 unc-13; hDp66 277 (15) unc-13; hDp69 553 unc-13; hDp73 a
Dpy or Unc"
Frequency of Dp gametes
405 1782 328 529 693 122 520 415 434 82 (32) 59 1 352 273 (20) 206 283 668 488 128 814 570 628 934 305 (6) 45 1 480 590 276 (3) 248
0.4 1 0.23 0.36 0.24 0.20 0.43 0.22 0.35 0.41 0.58 0.41 0.40 0.42 0.39 0.40 0.35 0.43 0.45 0.08 0.42 0.31 0.22 0.42 0.40 0.34 0.23 0.33 0.53
Male progeny scored are in parentheses.
13/unc-13; hDpx[dpy-5(-) unc-13(+)] were constructed. The ratio of wild type to mutant progeny was used to calculate the gametic frequency of each duplication (see MATERIALS AND METHODS). These results are summarized in Table 5 and Figure 6. The results from dpy-5;sDp2,dpy-5;ILXLszT1,unc-11; sDp2 and unc-13; ILXLszTl were not significantly different and thus experiments with different markers and duplications are comparable. Every duplication tested is notreported because those with similar breakpoints usually gave similar results. T h e calculations for gamete frequency assume the contributions fromthespermand oocyte were equal. As shown below, this was not always the case. The data in Table 5 does not include duplication homozygotes, which were slow growing, small and clear. sDp2 homozygotes were smaller than other D p homozygotes and only rarelyproducedprogeny. Homozygotes of sDp2 derivatives deleted at the left end also had this phenotype. Dp[dpy-5(+) dpy-14(-)] strains, however, produced homozygotes which were fertile but still small and clear. While wild types made upalargerfraction of the progenyfrom D p ( I ; f ) homozygotes than from hemizygotes, on the orderof three to five fold, this corresponded to a D p gamete frequency of only 60-70%. Thus it can be concluded
K. S. McKim and A. M. Rose
124 blt-3 let-362 I
s u Ipm - 1-16 I
Itn-17 I
fog-!
I
tmc-!I I
dpy-5
unc-74 unc-57 I
1
I
unc-40 I
rDp2
biz-4 1
unc-13 dpy-14 I
1
[417.; 417.1
hop12 [417.; 427.; 437.1 hDpl8 [407.; 327.;]
I
hop2
[232; varmble; i7.1
hDp3
[362; 212;
hDp4
[24%; 1.27.; 521
327.1
-
hDp20 [437.; -; 547.1 hDp22 [ Z Z ; - ;
li7.1
hDp62 [SZ]
+"" &!O
[317.; 377.1
4
" " " "
h D p l l [227,; 2.421
"i
" "
&69
r337.1
k"---L 1 m.u.
FIGURE6.-The relationship between the extent of duplicated material and the stability of the chromosome. Within the brackets, the percentages indicate the fraction of gametes carrying a duplication. If only a single percentage is shown, this refers to the average recovery of hermaphrodite sperm and oocyte as measured from self fertilization experiments. When more than one percentage is shown, each number is derived from a different assay and presented in the form: [hermaphrodite self-fertilization;oocyte; male sperm]. See the text for details.
that theduplications do not segregate fromeach other in hermaphrodite homozygotes. Most duplications derivedfromeither sDp2 or ILXLszTI in which part of the dpy-5 unc-13 region was deletedhad similar recovery frequencies;approximately 40% of the gametes carried the duplication. This included hDpl8 which is deleted at both endsof sDp2. Significant differences in duplication stability sDp2 or ILXLszTl were found with comparedto hDpl4,hDp62 and hDp73. hDpl4 was expected to be very stable since it is inserted into the X chromosome. hDp73 was alsovery stable. Strainscontaining this duplication segregated worms which grew to adults but were sterile. Like hDpl4, hDp73 may be an insertion into another chromosome, but unlike hDpl4,the insertion disrupted an essential gene. hDp62 is very unstable. It is also missing the left end of the chromosome but this alone cannot explain its instability because other duplications missing the left end, such as hDpl8,are much more stable. While most duplications with breaks in the dpy-5 unc-13region had similar stabilities, duplications with breaks in the unc-11 dpy-5 region varied with respect to their recovery frequency in gametes. Most duplications with breaks in this region had reduced stabilities. In addition,duplications carrying the unc-75unc57 region were often more stable than those deleting it. Thus unlike the dpy-5 unc-I3 region where deletion of chromosomal material had only small effects on duplication stability, deletions in the unc-1I dpy-5 region had significant effects on stability. hDp20 was much more stable than expected for its
size; it breaks between unc-57 and unc-63 but was found in 43% of the hermaphrodite gametes. Other duplications of similar size (e.g., hDp6) were recovered in fewer (20%) gametes. The hDp20 chromosome also contains part of chromosome V (K. MCKIM,unpublished results). These added sequences may confer the added stability to hDp20. hDp2 includes unc-74 but was found to be unstable like unc74(-) duplications (Table 5 and below). This may result fromadifference in thestructure of hDp2 compared to hDp3,5,and hDp7. The four unc-1I (+) dpy-5(-) duplications were of two classes. hDp9 and hDpl0 were found in approximately 31% of the gametes while hDp8 and hDpl I were found in approximately 22% of the gametes. We do not know how the stabilities of these four duplications relates to their structure because we have not precisely defined their breakpoints. The stabilities of ILXLszTl derivatives which had of chromosome lost the left end [dpy-5(-) unc-13(+)] I were less than sDp2 but greater than or equal to those sDp2 derivatives with breaks in the unc-II dpy5 region. This level of stability could be explained if the X chromosome sequences are attached to the right end of ILXLszTl. We have no otherevidence indicating to which end of ILXLszTI the X chromosome material is attached. Assaying for the Him phenotype was not informative because, while ILXLszTl strains were Him, most of the derivatives were not. Only the two duplications which carried an internal deletion around dpy5 (hDp69 and hDp71) had a Him phenotype. Assuming the X nondisjunction phenotype of ILXLszTI and
Duplications in C.elegans TABLE 6 Duplication Stability:11. Oocytes Dpy-5 hermaWt hermaphrodite Wt male
112
phrodite
125
see if the duplication or a recombinant chromosome was present. When hT2(Z;ZZZ)[+ dpy-5 +;+]/uric-1 I dpy-14; males were crossed to dpy-5 dpy-14; sDp2 hermaphrodites, 4 % 1 of the cross progeny carried the duplication. Thus the contribution from sperm and oocyte in the sDp2 hermaphrodite was equal. The smaller hDpl5 was recovered in slightly fewer oocytes. Similar experiments with hDpl2 showed it was also recovered in similar frequencies in the two hermaphrodite germ lines. In a cross with hDpl8 hermaphrodites, 32% of the cross progeny carried the duplication. Thisreduced recovery frequency of hDp18 compared to hDpl2 (42%) could be attributed to the left end deletion of hDpl8. For hDp[unc-ll(-) dpy-5(+)] strains, unc-1I dpy14/dpy-5 unc-13 males were crossed to unc-11 dpy-5; hDpx hermaphrodites. hDp4 was found in only 1.2% of the oocytes (Table 6). When the same cross was done with hDp3,21% of the oocytes carried the duplication. In self fertilization experiments, hDp3 was found in 36% of the gametes (Table 5). Knowing the hermaphrodite oocyte and male spermduplication frequencies, we calculated that hDp3 was found in 41 % of the hermaphrodite sperm.hDp4 was found at a similar levelin hermaphrodite sperm. Therefore, the low frequency of oocytes carrying hDp3 or hDp4 accounted for the reduction from 4 1% in sDp2 self fertilization experiments to 36% in hDp3 and 24% in the hDp4 self-fertilization experiments. hDp2 was lost frequently during oogenesis, but the results were inconsistent. When +dpy-5+unc-l3/unc11 +dpy-14 males were crossed to seven dpy-5 dpy14; hDp2 hermaphrodites,thefrequency of hDp2bearing oocytes varied from1to25%(datanot shown). This is probably because hDp2 is significantly more unstable in the somatic cells than hDp3. The frequency of hDpl0 and hDpl1 in the oocytes hT2(Z;IZZ)[++ was tested by crossing bli-3 unc-11 dpy-5;+] males to unc-11dpy-5;hDpx hermaphrodites. hDpl0 was recovered in 32% of the oocytes (Table 6); a rate similar to the self-fertilization frequency (Table 5). This indicates that duplication loss, while greater than that for sDp2,occurred equally in both germ lines. In contrast, hDpll was recovered in only 2.4% of the oocytes; a rate tenfold lower thanthe selffertilization frequency. In the matingexperiments, we oftennoted selffertilization progeny amongst cross progeny. Our experimental procedure allowed us to observe with certainty that self-fertilization progeny were produced amongst cross progeny. Either the duplication hermaphrodites multiply mated or the male sperm was not always used preferentially tohermaphrodite sperm asis normally found (WARD and CARREL 1979). The instability of these duplications results in part from somatic loss.Mosaic worms were frequently
+
+
Frequency of Dpy-5 male Dp oocytes"
dpy-5 hT2/unc-l I dpy-14 X dpy-5 dpy-14; ~ D p 2 ~ 118 66 0.41 dpy-5 hT2/unc-ll dpy-14 X dpy-5 dpy-14; hDpl2 23 32 0.42 38 39 dpy-5 hTZ/unc-II dpy-14 X dpy-5 dpy-14; hDp15 58 92 51 97 0.37 dpy-5 hT2/unc-ll dpy-14 X dpy-5 dpy-14; hDpl8 62 56 132 117 0.32 dpy-5 unc-13/unc-ll dpy-I4 X unc-11 dpy-5; hDp3 53' N DC 261 248 0.21 dpy-5 unc-13/unc-l1 dpy-14 X unc-11 dpy-5; hDp4 3' N DL 243 0.012 258 bli-3 unc-llldpy-5 hT2 X unc-11 dpy-5; h D p l 0 52 16 17 50 0.32 bli-3 unc-1 lldpy-5 hT2 X unc-11 dpy-5; hDpl1 3 0 0.024 71 53 a Fraction of duplication carrying oocytes calculated on the basis of Wt and Dpy progeny.
Dp oocytes= Dp worms/Total progeny. Dp worms = 2 X Wt except in the sDp2 experiment
+
where Dp worms = Wt. Total progeny = Dp worms (2 X Dpy). The crosses are indicated above the data. The male genotype is on the left and the hermaphrodite genotype in on the right. ' These experiments required progeny testing o i the Wt progeny because no balancer was used in the male. Males were not progeny tested and thus not included in the data.
the stability of the Dp[dpy-5(-) unc-13(+)] chromoX chromosome sequences somes was duetothe (McKIM, HOWELL and ROSE1988), the failure of most Dp[dpy-5(-) unc-13(+)] strains to be Him may have resulted from their reduced size. With regard to comparing the effects of deleting DNA in the unc-11 dpy-5 region to the dpy-5 unc-13 regions, hDp69 is of interest because it is a deletion of ZLXLszTlin the unc-11 dpy-5 region but still carries the flanking regions intact. Its stability was similar to hDp3 and hDpl0 despite the fact it contains more DNA (Figure 2). This may be because of the deletion around dpy-5. Most of the reduction in stability of these duplications can be accounted for by loss during oogenesis. T h e duplications were usually recovered at a higher frequency in hermaphrodite sperm than oocyte. The fraction of oocytes carryingaduplicationchromosome was assayed by crossing heterozygous males to duplicationhermaphroditesandscoringthe cross progeny.In the crosses described below, thefrequency of Wt progeny (which carried the duplication) was equivalent tothe frequency of oocytes which carried the duplication (Table 6). In Table 6, if the male was heterozygousfora crossover suppressor (hT2(Z;ZIZ) then all the wild-type progeny carried the duplication. If the male didnotcarrya crossover suppressor, the wild types had to be progeny testedto
+
+;+/
K. S. McKim and A. M. Rose
126 TABLE 7 Duplication stability: 111. Male sperm W t herDpy herDuplication Wt male maphrodite Dpy malemaphrodite
hDp2 hDp3 hDp4 hDp6 hDpl2 hDpZ0 hDpZ2
46 168 17 58 166 175 165
36 105 9 37 51 73 122
598 278 219 192 84 50 693
All crosses were dpy-5; hDpx male dite.
X
583 294 249 182 199 163 674
Frequency of D p sperm
0.07 0.32 0.05 0.20 0.43 0.54 0.17
dpy-5; unc-36 hermaphro-
recovered with strains carrying the unstable duplications but not with strains carrying the stable duplications (such as sDp2 and h D p l 5 ) (datanot shown). Mosaic worms (having a Dpy or Unc phenotype) have been recovered from strains carrying an unstable duplication and unc-ll dpy-5, unc-57 dpy-5 or dpy-5 dpy1 4 chromosomes. Isolation of mosaics withunc-74 dpy5 or unc-63 dpy-5 chromosomes is described later. If a duplication was lost early in germ line development, an unc dpy; op[++] hermaphrodite would lack the duplication in its germ line and produce no wild-type progeny. hDp2, hDp4, hDp59, hDp62 and hDp72 strains frequently produced germ line mosaics that segregatedno Dp-containing progeny. This could indicate a higher level of somatic loss compared to most other duplications, even those with similar stabilities when measured by duplication recovery in the gametes. Duplicationstability i n themale: We have tested duplication stability in the male by crossing dpy-5 +/ dpy-5dpy-14;hDpx males to dpy-5;unc-36 hermaphrodites and scoring the cross progeny. The Dpy (non-Unc)progenydidnot receive theduplication while the wild-type progeny resulted fromfertilization with a duplication carryingsperm. The results are shown in Table 7. The duplication recovery in the male sperm was similar to thatobserved in hermaphrodite oocytes and considerably lower thanforhermaphroditesperm. For example, hDp3 was found in 32% of the male sperm compared to 41 % of the hermaphrodite gametes. hDp4 was found in 5% of the male sperm, 2% in the hermaphrodite oocytes and 40% of the hermaphrodite sperm. In general, more loss was detected in male sperm andhermaphrodite oocytes than in hermaphrodite sperm. As others have shown (HERMAN, MADL and KARI 1979; DELONG,CASSON and MEYER 1987), some duplications tend to segregate from the single X chromosome in males. Segregation of the D p from the X chromosome would produce an excess of wild-type males ( D p ; X O ) and Dpy hermaphrodites (nullo-Dp; X X ) over wild-type hermaphrodites ( D p ; X X ) and Dpy
males (nullo-Dp; X O ) . We observed this in our experiments with h D p l 2 and hDp20 (Table 7). In h D p 2 , 3 , 4 , 6 and hDp22 males an excess of wild-type males over hermaphrodites was observed but not the excess of Dpy hermaphrodites over males. In these cases it was possible that the segregation distortion could not be observed among the Dpy progeny because there were too many non-Dp worms. Spontaneous shortening of duplications: Several duplications have been observed to spontaneously shorten. These duplication derivatives were isolated onthe basis of an exceptional phenotype, either by chance or in selective screens (see below). For example, our unc dpy; Dpstrains rarely segregated Unc worms that produced either no Uncs in their progeny (and thus were probably a genetic mosaic) or produced both Unc and Dpy Unc progeny. This latter class may have been the result of a deletion in the duplication chromosome to uncover the unc mutation, or a new mutation either on the duplication or elsewhere in the genome. Fourteen of these Unc strains were analyzed by complementationtesting and allof them were found to be the result of a loss of material from one end of the duplication and not through the addition of markers via mutation or recombination. The shortening events were not limited to the left end of sDp2. hDp3O and hDp77 were dpy-l4(-); the result of loss of sequences from the right end of sDp2 (Figure 3). The spontaneous duplications and their progenitors are shown in Figure 7. T o determine the frequency at which duplications shorten, we used a selective system to screen a large number of worms for exceptional individuals. Using the procedure described in MATERIALS AND METHODS, the progeny of unc-74 dpy-5; hDpx or unc-63 dpy-5; hDpx hermaphrodites were screened for levamisole resistant Uncs. The recovery of these duplications is summarized in Table 8. Because the worms were screened in the F2 generation, it could not be ensured that each event was independent. Some Po hermaphroditessegregated morethanone Unc F2. It is possible some of the shortening events occurred premeiotically. For this reason the numbers in Table 8 may not reflect the per chromosome shortening frequency. We have observed different chromosomes shortening at different rates. There is a correlation between spontaneous shorteningrate andmitotic stability. T h e more stable duplications shortened at a lower rate. For example, sDp2, which is mitotically very stable, spontaneously shortenedata low frequency. In a levamisole selection experimept, unc-74 dpy-5; sDp2 hermaphrodites segregated two Dp(unc-74(-)) derivatives in 127,000 chromosomes screened. hDp2, which is mitotically unstable as judged by the produc-
127
Duplications in C . elegans
gamma I
h Dp5
hDp3 gamma
hDp 7 gamma
hDp75 gamma
hhDDpp724
hDp2O
gemma
gamma
h 0025
h D p 76
1
~
~
h D ph2D6 ph2D7 ph3D0 p 5 9
~
c
FIGURE7.-Pedigree of spontaneous duplications. All the spontaneous shortened duplications and their progenitors areshown. I f gamma radiation was used, this is indicated, otherwise the duplications were spontaneous.
DISCUSSION
TABLE 8
Recovery of spontaneously shortened duplication strains
Duplication
Total progeny
72 000 unc-74 dpy-5; hDp2 unc-74 dpy-5; hDp5 116 000 unc-63 dpy-5; hDp20 134 000 unc-63 dpy-5; hDp23 41 000 unc-74 dpy-5; sDp2 127 000
Shortened duplications
30 3 1 18 2
Frequency
4.2 X 2.6 X 7.4 X 2.2 X 1.6 X
Mosaics
54 3
1 18 2
tion of mosaics and reduced recovery in the gametes, shortened at a much higher frequency. In at least one case, a duplication which had previously shortened continued to shorten at a high frequency. An hDp3 strain segregated hDp23 which subsequently segregated even shorter duplication chromosomes (Figure 7 and Table 8). In addition to recovering shorter duplications, these experiments recovered mosaics. These worms were either Unc non-Dpy or Unc semi-Dpy and segregated only Dpy Unc progeny or wild types and Dpy Unc progeny (Table 8). sDp2 and ILXLszTl haveno effect on recombination frequency: These two duplications were examined to see if they have any effects on recombination. Strains carrying the duplication in addition to normal chromosomes heterozygousfor cis-linked markers were scored and the recombination frequencies were calculated as described in MATERIALS AND METHODS (Table 9). sDp2 had no effect on recombination in both the dpy-5 unc-13 interval and the let-362 dpy-5 interval. In the dpy-5 unc-13 interval, ILXLszTl had no detectable effect on recombination frequency.
The original goals of this research were to isolate rearrangement breakpointson chromosome I and use these to map mutations and study chromosome behavior. Sixty-two duplications were recovered following gamma radiation and 14 were isolated as a result of spontaneous events. These duplications and five deficiencies have divided the left 20 m.u. of chromosome I into 24 regions. An array of duplications facilitates the analysis of a large chromosomal region: In the process of identifying essential genes in the sDp2 region (HOWELL et a l . 1987), over 500 EMS induced mutations have been isolated and arebeing mappedand complementation tested (HOWELL1989; J. MCDOWALL and A. ROSE, unpublished results). This would be an enormous task without rearrangement breakpoints to position the mutations into smaller intervals before undertaking complementation testing.By using the mapping techniques and crosses described in this paper, any recessive mutation can be efficiently mapped to an interval. While the spontaneous shortening of duplications couldproduce false negative results, we have avoided this problem through frequent testing of strains with appropriate markers. Using the methods described here, a set of nested duplications could be isolated and used to intensively characterize any genetic region in C. elegans. A large number of breakpoints can be easily generated in any region having an appropriatepre-existing duplication. As a tool for the geneticdissection of a region, duplication mapping may be an attractive alternativeto the conventional deficiency mapping (SIGURDSON, SPANIER and HERMAN 1984; ROSENBLUTH et al. 1988). The isolation of duplication breakpoints is fast and easy.
K. S. McKim and A. M. Rose
128
TABLE 9 Effects of duplications on recombination Genotype
dpy-5 unc-13 unc-29/+++; sDp2' dpy-5 unc-13 unc-29/+++'
Wt
4
1279 2678
let-362 dpy-5 unc-13/+++; Dpy-5Unc-13 sDp2 20 652 149 1406 let-362 dpy-5 ~ n c - 1 3 / + + + ~ Dpy-5Unc-13 6 1769 dpy-5 unc-13/++; I"XLszTI dpy-5 ~ n c - 1 3 / + + ~ 52 4775
Recombinants
Dpy-5 10 Unc-29 34 Dpy-5 21Unc-29
Dpy-5 Dpy-5
I11.U.
(C.1.P
1.3 (0.3-3.1) 1.4 (0.7-2.4) 1.9 (1.3-2.6) 1.2 (0.8-2.6) 15.0 (9.3-24.3) 15.4 (13.0-17.9) 1.5 (0.7-3.3) 1.6 (1.3-1.9)
5 1 Unc-I3 (3.1. = 95% confidence interval.
' This data was collected by A . M . HOWELL. ' Data fronl MCKIM, HOWELLand ROSE (1988). 'I Data from HOWELL et al. ( 1 987).
Furthermore, unlike deficiencies whose viability decreases with increasing size, larger duplications are more stable (see below). Having such large aberrations is a useful tool to quickly localizeunmapped mutations to a large region; that being either the region covered or not covered by the duplication. Subsequent experiments can map themutationmore precisely using additional breakpoints. Since deficiencies are generally smaller, more experiments are required tomap a mutation to a region of comparable size. In addition, most of the duplications used in this study have one end in common with sDp2. Because only one breakpoint varies, mapping is a binary procedure. In theory the same could be said of nested deficiencies, but such a collection is not easily isolated and can only be effective for a small region due to the smaller size of deficiencies. The nested duplications are a by-product of the method used to generate them and there is no restriction on the size of region to be analyzed. We have mapped the duplication breakpoints with respect to some known visible and lethal mutations in the regim. In addition, the duplication breakpoints can be used as genetic landmarks that can be placed on the physical map being constructed (COULSON et al. 1986,1988). Within the sDp2 regionapproximately 2000 kB of DNA has been placed on the physical map. At present, this DNA is aligned to the genetic map at six points (STARRet al. 1989;J. BABITY, unpublished results). The duplications described here are being used by us and others to position cosmids on the genetic map of chromosome I . Shortening duplications decreases mitotic stability: We observed considerable variation in the mitotic stability of the chromosomeI duplications; some were lost more frequently than others. We measured instability as adecrease in thenumber of Dp carrying gametes and an increase in the frequency of genetically mosaic worms. The isolation of genetically mosaic worms results from mitotic loss of the duplica-
tions. It is not known if there is significant loss during meiosis. HERMAN,ALBERTSON and BRENNER(1976) observed considerable amounts of premeiotic duplication lossin cytogenetic observations of duplication strains. This behavior has been interpreted as a consequence of the holokinetic structure of C . elegans chromosomes (ALBERTSON and THOMSON 1982). During mitosis and meiosis, individual spindles are observed to attach at many sites along the chromosome. ALBERTSON and THOMSON (1 982) proposed the efficiency of spindle attachment was proportional to kinetochorelength. Thus smaller chromosomefragments, with their smaller kinetochores, would sometimes have insufficient spindle attachment for proper segregation. Consistent with this hypothesis, size is an important determinant of chromosome I duplication stability. Our results, however, show there are significant exceptions to a strict relationship between duplication size and stability. hDp14, hDp20 and hDp73 are more stable than expected but this iseasily explained by their attachment to other chromosomes or fragments. Lesseasily explained are hDp2,hDp4,hDp62 and hDp72. The stabilities found in this group aresurprisingly low for their size. hDp2, for example, is 30% less stable than three other duplications of the same size. One possibility is that different structures of the duplications result in different stabilities. For example, linear and circular chromosomesare expected to have different mitotic properties. (MCCLINTOCK1938; LEIGH 1976). Another exception to a strict size-stability correlation was found by determining the effects of deleting certain regions from theduplications. Deletion of the dpy-5 unc-I3 region had small effects on duplication stability whereas a deletion of the unc-57 unc-74 region significantly reduced duplication stability. For example, there is little difference between sDp2,
Duplications in C. elegans hDp 12 and hDp 15whereas there aresignificant differences between the three groups hDpl8 and hDp3, 5 , IO and hDp4, 6 , 11. The greater loss of stability that accompanied deletion in the unc-11 dpy-5 region when compared to similar sized deletions in the dpy-5 unc1 3 region was not theresult of deleting larger amounts of DNA. The dpy-5 unc-13 region contains at least as many genes and as much or more DNA than the unc11 dpy-5 region (HOWELLet al. 1987; KIM and ROSE 1987; HOWELL 1989; STARR et al. 1989). Since size is not the only determinant of stability, what other factors could be involved? The first possibility, differences in structure, has already been discussed. Second, a threshold level of chromosome size may be necessary for stability. If a size threshold is important, then deletionsof chromosome material do not have significant effects above the size threshold. Duplications like hDp3 would be smaller thanthe threshold and be in the range where the probability of spindle attachment is sensitive to size changes. A third possibility is the DNA sequences responsible for stability may not be evenly distributed along the chromosome. These sequences need not be kinetochore attachment sites. If DNA sequences are an important factor,then hDp6 isless stable than hDp3 because sequences conferring mitotic stability have been deleted. Our results do not unambiguously resolve these possibilities. In fact, any combination of these may be influencing the stability of a given duplication. Some free X chromosome duplications have higher levels of loss in the oocyte line than in sperm (HERMAN, ALBERTSON and BRENNER 1976). We observed a lower frequency of duplication recovery in the hermaphrodite oocyte and the male sperm than in the hermaphroditesperm.Hermaphroditespermare made before oocytes and undergo fewer cell divisions (KIMBLEand HIRSH 1979).Thiscould explain the different rates of loss. The more frequent and rapid cell divisions of oogenesis could provide more opportunityforduplication loss. Similarly, male sperm undergomore cell divisions thanhermaphrodite sperm. hDplO was recovered at equal frequencies in bothhermaphroditegerm lines. Some duplications may be less sensitive to loss than others during the cell divisions of oogenesis and male spermatogenesis. The more unstable duplications will be useful for mosaic analysis. HERMAN(1984)has used unstable duplications of the X chromosome to create genetically mosaic worms. This allows one to ask questions regarding the tissue specificity and cell autonomous expression of a gene. sDp2 is too stable for mosaic analysis butstrains with shorter duplications frequently produce mosaic worms. With hDp2 or hDp23, the frequency of mosaics was approximately one in 2500 progeny when dpy-5 and a levamisole resistant unc were used. This is similar to the frequency of
129
genetic mosaics recovered by HERMAN(1984).We plan to use mosaic analysis as part of the analysis of genes in the left half of chromosome I . Duplication chromosomes spontaneously break: A striking aspect of our results was the spontaneous shortening of the duplications. We observed the spontaneous shortening of five duplications in addition to the two events from sDp2. Previously, HERMAN (1984) observed spontaneous shortening of one X chromosome duplication in C. eleguns. Possibly allduplications will spontaneously shorten.Whetherspontaneous shortening at the frequencies observed here occurs in normal chromosomesin a euploid genetic background is not known. Depending on their structure,however, different duplications shorten at different rates. Mitotically stable duplications (sDp2,hDp5,hDp20) shorten less frequentlythan unstable duplications (hDp2, hDp23). The correlation between mitotic stability and spontaneous shortening leads us to suggest that mitotic/meiotic segregation problems cause both duplication loss and breakage. It is possible the unstable duplications have difficulty aligning with the spindle during nuclear divisions (meiotic or mitotic). If the chromosome is not properly aligned, the chromatids may not be equally divided between the sister cells, lost altogether or undergo a breakage event(see below). ALBERTSON and THOMSON (1982) provided cytological evidence for this type of behavior during mitosiswith thefree X chromosomeduplication mnDp2. They observed the duplication often lying on the outside of the metaphase plate or lagging at anaphase. The improperalignment of duplicationchromosomes during nuclear division may provide an opportunity for breakage. Aholokinetic chromosome could be pulled apart if the differentspindles from thesame pole do not attach to the same chromatid. Misalignmentmight allow spindles from opposite poles to attach to thesame chromatid. If two adjacent spindles were pulling in opposite directions the duplication chromosome might break apart and the result would be a shorter duplication. Improper spindle attachment to normal chromosomes probably occurs but the rate at which this occurs or at which exceptional events are recovered is not known. The reason for the elevated frequency of spontaneous shortening for some of the duplications could be because they misalign on the metaphase plate more often than sDp2. Simply being aunivalent poses problems for a chromosome. Our duplications arenot engaged in exchange events and should be univalents during meiosis. Other authors have observed that univalents in corn (MILLER 1963), wheat (SEARS1952) and Drosophilamelanogaster (SANDLERand BRAVER1954; CARPENTER1973; BAKER1975) are unstable. The univalent may divide equationally at meiosis I,or
130
K. S . McKim and A. M. Rose
either chromatid may be lost at meiosis I or 11, or the univalent may undergo misdivision.Misdivision occurs when a metacentric univalent attempts to separate transversely at the centromere resulting in telocentric chromosomes. Duplication breakagefrequency was either low (sDp2, hDp5 and hDp20) or high (hDp2 and hDp23). There were no intermediate values. It is possible that duplication shortening only occurs at one of the two frequencies. We have no reason to doubt that normal chromosomes may break at the frequency observedin the low frequency (sDp2) class. Furthermore, it is possible that shortening is not entirely the result of mitotic instability, but could also be a consequence of aparticularchromosomestructure.Chromosome breakage is observed with dicentric chromosomes produced by recombination events involving inversion or circular chromosomes (MCCLINTOCK 1938, 1941; HABER, THORBURN and ROGERS1984).It shouldbe emphasized that in addition to missing part of the normal chromosome, these duplications do not pair and recombine with chromosome I . Therefore, we have not addressed the influence of pairing and recombination on the spontaneousshorteningfrequency of a duplication. Although it is assumed genetic mosaics are produced through duplication loss (HERMAN 1984), it is also possible a genetic mosaic could be generated by a breakage event in the lineage under study [as observed in maize (MCCLINTOCK 1938, 1941)l. Mosaics produced through a breakage eventmay be a significant fraction of the total number based on our observations that the recovery of mosaics and breakage events occurs at similar frequencies. A mosaic produced through a breakage event can give different results than amosaic produced through completeloss. In a chromosome loss event, all alleles on the duplication are lost from that lineage. In a breakage event, some of the alleles would remain. Two examples of spontaneous chromosome shortening have been observed in D. melanogaster. BIESSMANN and MASON(1988) isolated terminal deficiencies of the X chromosome which lost sequences from their distal ends at a rate of 75 bp per generation. LEVIS (1989) isolated terminal deletionsof chromosome 3R. The deletion chromosomesalso lost approximately 75 bp per generation. In both of these cases, the deficiencies were recovered in amutagenicgenetic background. These experiments did not report attempts to recover large deletion events. It is possible that our duplications are also shortening by a small amount. We have not doneexperiments to detect theseevents. Broken chromosomes lacking a proper telomereare mitotically unstable in maize (MCCLINTOCK 1941) and yeast (HABER, THORBURN and ROGERS1984) anduntil recently (see above) have not been isolated in Dro-
sophila (MULLERand HERSKOWITZ 1954; ROBERTS 1975). The structures of natural C. elegans telomeres or theends of duplication chromosomes are unknown. Also unknown are the behavioral characteristics of broken, unrepaired C. elegans chromosomes. By analogy with Drosophila, it is reasonable to postulate that in the process of shortening sDp2 or ILXLszTl with gamma radiation, shortened chromosomes were capped with telomerecarryingfragments of other chromosomes originating from the same gamma radiation exposure (WILLIAMSON and PARKER 1976). In the caseof spontaneous breakage events, however, capping in this way cannot be proposed. The free ends created by a breakage event may be repaired by the acquisition of telomerecontaining sequences from another genomic location; involving eitherintrachromosomal or interchromosomalrearrangements (MCCLINTOCK 1941; HABERand THORBURN 1984; RUDINand HABER 1988). Supportingthis proposal is the observation that some shortening events are associated with attachmenttoanotherchromosome (HERMANandKARI 1989; K. MCKIM,unpublished results). Alternatively, a stable end could be created through thede novo addition of a telomereas proposed in Tetrahymena (KING and YAO 1982). Still another solution to the telomere problem is a ring chromosome. Based on deletion analysis of the type described here, evidence in favor of the existence of ring chromosomes has been collected by us (K. MCKIM, unpublished results) and others (C. P. HUNTER andW. B. WOOD,personal communication). Not all duplications, however, are rings. sDp2 appears to be linear because deletions of one end usually do not carry over to the other end. Themost radical possibility is that there is no healing event at all. The solution to the structure of the duplication telomere awaits the cloning of a duplication end. Meioticbehavior of duplications: Two observations support theassertion that duplications of the left half of chromosome I have no pairing activity during meiosis. The first is the failure of sDp2 to recombine with the normal homolog(ROSE,BAILLIE and CURRAN 1984). The second is our observation that these duplications have no influence ontherecombination frequency between the normal homologs in the duplicated regions. These observations are in contrast to another duplication of chromosome I , sDp1, which carries sequences on the other (right) end of chromosome I and does compete for pairing and recombination with the normal homologs (ROSE, BAILLIE and CURRAN1984). While we conclude the left end of the chromosome does not containsequences which can bringthechromosomestogetherforpairing (“homologrecognition”), this region may influence the distribution of recombination events along chromosome. Based on a studyof chromosome V deficien-
Duplications in C. elegans cies, ROSENBLUTH, JOHNSEN and BAILLIE (1990) have proposed a pairingmodel requiring synapsis initiation at both endsof C. eleguns chromosomes. In that study, it was found that recombinationwas reduced adjacent to internal deficiencies. The gamma ray induced lethal mutations h61 and h655 described here may be similar to these. The insertion of chromosome I material into the X chromosome (hDpl4)alters X chromosome recombination frequencies. T o the left of the insertion site recombination is either unaffected or slightly increased. T o the right, however, recombination is decreased tenfold. This suppression of recombination to the right of the insertion agrees with the predictions of our model of chromosome pairing in C. eleguns (McKIM, HOWELL and ROSE1988). We proposed the homolog recognition information for each chromosome is located at or near the one end of each chromosome and that synapsis initiates at this site. The region for homolog recognition on the X chromosome was proposed to be at theleft (unc-I ) region (HERMAN and KARI 1989; MCKIM, HOWELL and ROSE1988). In the case of the hDpl4 chromosome, the progression of pairing which would have begun at the left end might have been disrupted by chromosome I sequences. Recombination is normal on the side of the insertion nearestthe homolog recognitionsite because pairing can proceed efficiently up to the insertion site. The insertionapparentlycannot efficiently be “jumped” by the pairing process resulting in the disruption ofsynapsis. The disruption of pairing and recombination on one side of a heterology is analogous to that observed for translocations [described in MCKIM, HOWELLand ROSE(1988)l. Pairing and recombination occur on one half of a reciprocal translocation but cannot initiate on the other half because there is no homolog recognition information there. Similarly, HERMAN,ALBERTSONand BRENNER (1 976)observed a polareffect on recombination with the duplication mnDpl ( V ; X ) . This duplication is attached to chromosome V and lowers recombination on the left half. If the right arm of chromosome V contains the homologrecognition sequences which initiate the pairing process as we proposed previously (McKIM, HOWELLand ROSE 1988), then the mnDpl situation would be the same as hDpl4.Unfortunately there is no accurate positioning of mnDpl on chromosome V . In summary, we can now state several features contributing to the function and organization of chromosome I . Despite carrying large amounts of chromosome I sequence, sDp2 and ILXLszTl have no obvious effect on chromosome I meiotic recombination. Thisargues against these duplications having any pairing behavior which would compete with pairing of the normal homologs. Homolog recognition and
131
pairingappears to initiate at the right end of the chromosome I with other sequences, such as some near the left end, setting the frequency and distribution of recombination events. The inability of these duplications to recombine with chromosome I allows themtobe used effectively as balancers for lethal mutations and as a source of breakpoints for genetic mapping. The duplications described hereare less stable during nuclear division thananormalchromosome. We have investigated the natureof the chromosome structures governing segregation fidelity. In addition to size, the stability of a duplication is dependent on other, as yet undescribed, structural features. We have described some of the consequences of an unstable duplication. For example, an unstable duplication is more prone to breakage events than a mitotically stable one. Further analysis, perhaps at a cytogenetic or molecular level, may help to elucidate the structure of these chromosomes and the relationship of these structures to chromosomebehavior. We wish to thank DAVIDL. BAILLIE,ANN MARIE HOWELL and RAJAE. ROSENBLUTH for valuable advice during this work and to our reviewers for comments on the manuscript. We also are grateful to T. STARR forisolating the duplications marked “b” in Table 3. Some of thestrains used wereprovided by the Caenorhabditis Genetics Center, which was supported by contract NO1-9-2113 between the National Institutes of Health and the Curators of the University of Missouri. K.McK. was supported by a Medical Research Council (MRC) (Canada) Studentship. This work was supported by grants from the MRC and the Natural Sciences and Engineering Research Council of Canada to A.M.R.
LITERATURE CITED ALBERTSON, D. G . , and J. N. THOMSON,1982 T h e kinetochores of Caenorhabditis elegans. Chromosoma 86: 409-428. BAKER,B. S., 1975 Paternal loss (pal): ameiotic mutant in D. melanogaster causing loss of paternal chromosomes. Genetics 8 0 267-296. H., and J. M. MASON,1988 Progressive loss of DNA BIESSMANN, sequences from terminal chromosome deficiencies in Drosophila melanogaster. EMBO J. 7: 1081-1086. BRENNER, S., 1974 T h e genetics of Caenorhabditiselegans. Genetics 77: 7 1-94. CARPENTER, A. T . C., 1973 A mutantdefective in distributive disjunction in Drosophila melanogaster. Genetics 73: 393-428. CLARK,D. V., T . M. ROGALSKI, L. M. DONATIand D. L. BAILLIE, 1988 T h e unc-Z2(IV)region of Caenorhabditiselegans: genetic analysis of lethal mutations. Genetics 1 2 9 345-353. COULSON, A., J. SULSTON, S. BRENNER and J. KARN,1986 Towards a physical map of the genome of the nematode Caelzorhabditis elegans. Proc. Natl. Acad. Sci. USA 83: 7821-7825. COULSON, A., R. WATERSTON, J. KIFF,J. SULSTON and Y.KOHARA, 1988 Genome linking with yeast artificial chromosomes. Nature 335: 184-186. CROW,E. L., and R. S. GARDNER, 1959 Confidence intervals for the expectation of a poisson variable. Biometrika 4 6 441-453. DELONG,L., L. P. CASSONand B. J. MEYER,1987 Assessment of X chromosome dosage compensation in Caenorhabditis elegans by phenotypic analysis of lin-14. Genetics 117: 657-670. EDGLEY,M. L., and D. L. RIDDLE,1987 Caenorhabditiselegans, pp. 351-365 in Genetic Maps 1987, Vol. 4, edited by S. J. O’BRIEN.Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
132
K. S. McKim and A. M. Rose
FERGUSON, E. L., andH. R. HORVITZ,1985 Identification and characterization of 22 genes that affect the vulval cell lineages of the nematode Caenorhabditis elegans. Genetics 1 1 0 17-72. FODOR,A., and P. DEAK,1985 The isolation and genetic analysis of a C. elegans translocation ( s z T I ) strain bearing an X-chromosonle balancer. J. Genet. 64: 143-157. HABER, J. E., and P. C. THORBURN, 1984 Healing of broken linear dicentric chromosomes in yeast. Genetics 106: 207-226. HABER, J. E., P. C. THORBURN and D. ROGERS,1984 Meiotic and mitoticbehavior of dicentric chromosomes in Saccharomyces cerevisiae. Genetics 106 185-205. HERMAN, R. K., 1984 Analysis of genetic mosaics of the nematode Caenorhabditis elegans. Genetics 108: 165-180. HERMAN,R. K., D. G. ALBERTSONand S. BRENNER,1976 Chromosomerearrangements in Caenorhabditiselegans. Genetics 83: 91-105. HERMAN, R. K., and C. K. KARI,1989 Recombinationbetween small X chromosome duplications and the X chromosome in Caenorhabditis elegans. Genetics 121: 723-737. HERMAN, R. K., J. E. MADLand C. K. KARI,1979 Duplications in Caenorhabditis elegans. Genetics 92: 419-435. HORVITZ,H. R., S. BRENNER,J. HODCKINandR.HERMAN, 1979 Auniform geneticnomenclatureforthenematode Caenorhabditis elegans. Mol. Gen. Genet. 175: 129-133. HOWELL,A. M., 1989 Essential genes in a region of chromosome I in Caenorhabditis elegans. Ph.D. thesis, University of British Columbia, Vancouver, B.C., Canada. and A. M. ROSE, HOWELL,A. M., S. G. GILMOUR,R. A. MANCEBO 1987 Genetic analysis of a large autosomal region in Caenorhubditis elegansby the use of a free duplication. Genet. Res. 49: 207-21 3. KIM, J.S., and A. M. ROSE,1987 The effect of gamma radiation on recombination in Caenorhabditis elegans. Genome 2 9 457462. KIMBLE, J., and D. HIRSH, 1979 Postembryonic cell lineages of the hermaphrodite and male gonads in Caenorhabditis elegans. Dev. Biol. 7 0 396-417. KING, B. O., and M.-C. YAO, 1982 Tandemly repeated hexanucleotide in Tetrahymena rDNA free end is generated from a single copy during development. Cell 31: 177-182. LEIGH,B., 1976 Ring chromosomes and radiation induced chromosome loss, pp. 505-528 in The Genetics and Biology of Droand E. NOVITSKI. sophila, Vol. lb, edited byM. ASHBURNER Academic Press, London. LEVIS,R. W., 1989 Viable deletions of a telomere from a Drosophila chromosome. Cell 5 8 791-801. LEWIS,J. A.,C. H. WU, H. BERGand J. H. LEVINE,1980 T h e genetics of levamisole resistance in the nematode C.elegans. Genetics 95: 905-928. MCCLINTOCK, B., 1938 T h e production of homozygous deficient tissue by means of aberrant mitotic behavior of ring-shaped chromosomes. Genetics 23: 3 15-376. MCCLINTOCK, B., 1941 T h e stability of broken ends of chromosomes in Zea mays. Genetics 26: 234-282. MCKIM,K. S., A . M. HOWELLand A. M. ROSE, 1988 T h e effects of translocations on recombination frequencyin Caenorhabditis elegans. Genetics 120: 987-1001. MENEELY,P. M., and R. K. HERMAN,1981 Suppression and function of X-linked lethal and sterile mutationsin Caenorhabditis elegans. Genetics 97: 65-84.
MILLER,0..1963 Cytological studies of asynaptic in maize. Genetics 48: 1445-1466. MULLER,H. J., and I. H. HERSKOWITZ,1954 Concerningthe healing of chromosome ends produced by breakage in Drosophila melanogaster. Am. Nat. 88: 177-208. ROBERTS,P. A., 1975 Insupport of thetelomereconcept. Genetics 80: 135-142. ROSE,A. M., 1980 Genetic studies on the gene coding for paramyosin in Caenorhabditis elegans: unc-15 and the adjacent region. Ph.D. thesis, Simon Fraser University, Burnaby, B.C., Canada. ROSE,A. M., and D. L. BAILLIE,1979 Effect of temperature and parental age on recombination and nondisjunction in Caenorhabditis elegans. Genetics 92: 409-41 8. ROSE,A. M., and D. L. BAILLIE,1980 Genetic organization of the region around unc-15(1),a gene affecting paramyosin in Caenorhabditis elegans. Genetics 96: 639-648. ROSE,A. M., D. L. BAILLIEand J. CURRAN,1984 Meiotic pairing behavior of two free duplications of linkage group I in Caenorhabditis elegans. Mol. Gen. Genet. 1 9 5 52-56. ROSENBLUTH, R. E., and D. L. BAILLIE,1981 Analysis of a reciprocal translocation, eTI ( I I k V ) , in Caenorhabditis elegans. Genetics 99: 415-428. ROSENBLUTH, R. E., C. CUDDEFORD and D. L. BAILLIE,1985 Mutagenesis in Caenorhabditis elegans. 11. A spectrum of mutational events induced for 1500 R of gamma-radiation. Genetics 109: 493-51 l . ROSENBLUTH, R. E., R. C. JOHNSEN and D. L. BAILLIE, 1990 Pairing for recombination in Caenorhabditis elegans: a model based on the effect of deficiency-heterozygosity of recombination. Genetics (in press). R. C. JOHNSEN,L. M. ADDIROSENBLUTH, R. E., T . M . ROGALSKI, SON and D. L. BAILLIE,1988 Genomicorganization in Caenorhabditiselegans: deficiency mapping on linkage group V (left). Genet. Res. 52: 105-118. RUDIN,N., and J. E. HABER,1988 Efficient repair of HO-induced chromosomal breaks in Saccharomyces cerevisiae by recombination between flanking homologous sequences. Mol. Cell. Biol. 8: 3918-3928. SANDLER, L., and G. BRAVER,1954 T h e meiotic loss of unpaired chromosomes in Drosophilamelanogaster. Genetics 3 9 365377. SEARS,E. R., 1952 Misdivision ofunivalents in commonwheat. Chromosoma 4 S.535-550. SIGURDSON,D. C., G. J. SPANIER and R. K. HERMAN, 1984 Caenorhabditis elegans deficiency mapping. Genetics 108: 331-345. STARR,T., A. M. HOWELL,J. MCDOWALL,K. PETERSand A. M. ROSE,1989 Isolation and mapping of DNA probes within the linkage group I gene cluster of Caenorhabditis elegans. Genome 32: 365-372. WARD,S., and J. S. CARREL,1979 Fertilization and sperm competition in the nematode Caenorhabditis elegans. Dev. Biol. 73: 304-321. WILLIAMSON, J. H., and D. R.PARKER, 1976 Recombination between the X and Y chromosomes, pp. 701-720 in The Genetics and Biology of Drosophila, Vol. lb, edited by M. ASHBURNER and E. NOVITSKI.Academic Press, London. Communicating editor: W. M. GELBART
