Caenorhabditis elegans

Proc. Natl. Acad. Sci. USA Vol. 87, pp. 2901-2905, April 1990 Genetics

Genes that can be mutated to unmask hidden antigenic determinants in the cuticle of the nematode Caenorhabditis elegans (surface antigen/immunogenetics/genetic mapping/radioiodination/extracellular matrix)

SAMUEL M. POLITZ*t, MARIO PHILIPP*, MIGUEL ESTEVEZ§¶, PETER J. O'BRIEN§11, AND KARL J. CHIN§* *Department of Biology and Biotechnology, Worcester Polytechnic Institute, Worcester, MA 01609; tMolecular Parasitology Group, New England Biolabs, 32 Tozer Road, Beverly, MA 01905; and tSchool of Applied Biology, Georgia Institute of Technology, Atlanta, GA 30332

Communicated by William B. Wood, December 26, 1989 (received for review September 19, 1989)

ABSTRACT Rabbit antisera directed against a mixture of proteins solubilized from the wild-type adult Caenorhabdiis ekgans cuticle were used to isolate mutants, induced by ethyl methanesulfonate treatment, that exhibit alterations in surface antigenicity by immunofluorescence. Genetic mapping and complementation data for four such mutations define two genes, srf-2(I) and srf-3(IV). The mutant phenotypes observed by immunofluorescence appear to result from unmasking of antigenic determinants that are normally hidden in the wildtype cuticle. In support of this hypothesis, surface radioiodination experiments indicate that components labeled on the wild-type surface are missing or less readily labeled on the surface of srf-2 and srf-3 mutants.

Nematodes form a diverse phylum composed of free-living and parasitic species. The nematode cuticle is the organism's major contact with, and barrier to, the environment. All nematodes experience a life cycle involving four postembryonic molts; at each molt, a new cuticle is synthesized and the old one is shed. Several nematode parasites cause diseases that are among the most prevalent and debilitating of man (1). Parasitic nematodes expose the host to a complex mixture of antigens and elicit a correspondingly complex immune response (2), yet the infections generally remain chronic (3). How nematode parasites survive in the adverse host immune environment is poorly understood. The nematode surface is antigenic in the infected host. The appearance of specific surface antigens is dynamic, with changes occurring at and between molts (4, 5). However, little is understood of the genetic mechanisms underlying surface antigen expression. The cuticle of the free-living nematode Caenorhabditis elegans has been studied by classical genetics. Mutations affecting cuticle structure include those that alter gross morphology (6-9), some of which are in collagen genes (1013), and heterochronic mutations that cause expression of adult-specific cuticle features in larvae or vice versa (14-16). In addition, an adult-specific surface antigenic class is missing in certain C. elegans varietal strains; this marker maps to a locus designated srfl(II) (17). Little is known genetically about a characteristic feature of the nematode cuticle-i.e., its layered ultrastructural organization (e.g., ref. 18). We have isolated and analyzed several C. elegans surface antigen mutants that may have implications for understanding the layering pattern and the immune response to nematode parasites. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

2901

MATERIALS AND METHODS Preparation of Immunogen and Antisera. Rabbit antisera directed against wild-type adult C. elegans cuticle proteins solubilized by SDS and 2-mercaptoethanol (2-ME) were prepared as described (19). Antiserum Adsorption. Rabbit no. 74 serum was used, except that in srf-1 srf-2 epistasis experiments, a 1:1 (vol/vol) mixture of no. 74 antiserum and the antiserum that defined srf-1 (17) was used. Adsorption of antiserum with wild-type fourth stage larvae (L4s) or adults was performed as described (17). The number for adsorption was 1250 N2 L4s per ,ul of serum or 1650 N2 adults per Al serum. Immunofluorescent Staining. Nematodes were stained and examined as described (17) and surface immunofluorescence was photographed in Pyrex spot plate wells under 0.1 M NaN3 anesthesia. Mutant Selection by Immunofluorescent Screening. F1 progeny of ethyl methanesulfonate (EMS)-treated animals (6) were forced to produce F2 dauer larvae on egg white plates (18). After purification by SDS treatment (18) and storage in M9 buffer (6) at 16°C, F2 dauer larvae were allowed to develop synchronously into L4s on Escherichia coli lawns for 17.5 hr at 20°C. Alternatively, synchronous F2 descendents of EMStreated animals were obtained by allowing eggs laid by F1 animals in a 60- to 120-min period (8) to develop into L4s for 36-40 hr at 20°C. Up to 10,000 synchronous F2 larvae were stained for immunofluorescence with L4-adsorbed antiserum and examined; only mutant larvae with alterations in surface antigenicity were expected to stain (17). Immunofluorescent Li to L4 larvae were individually picked. False positive clones were eliminated by rescreening. Only independent mutants were saved. Strains. Wild-type (strain N2) and mutant strains (6) of C. elegans var. Bristol were obtained from the Caenorhabditis elegans Genetics Center, Columbia, MO. sqt-J(scJOJ)II, sqt-J(e1350), sqt-J(sc97), sqt-1(sclOO), sqt-2(sc3)II, sqt3(sc8)V, sqt-3(sc24), sqt-3(sc63), rol-8(scJ5)II, and rol8(sc98) were obtained from R. S. Edgar (University of California, Santa Cruz) and were described in ref. 9. linJJ(n389)I was described in ref. 20. A strain containing sDf2(IV) (21) was genotypically sDf2/nTJ(IV); +/nTl(V) (22, 23). Abbreviations: 2-ME, 2-mercaptoethanol; LI, L3, and L4, first, third, and fourth larval stage, respectively; EMS, ethyl methanesulfonate; PI, protease inhibitors. tTo whom reprint requests should be addressed. VPresent address: Division of Biological Sciences, University of Missouri, Columbia, MO 65211. IPresent address: Lilly Research Labs, Indianapolis, IN 46285.

2902

Genetics: Politz et al.

Genetic Methods. Stocks were grown and genetic crosses were performed at 20'C using standard methods (6). The surface antigen marker was identified by antibody screening (immunofluorescence staining with adult-adsorbed antiserum). Visible mutant phenotypes were used as markers to select segregant clones or individuals for evaluation of the surface antigen marker. All mutants were back-crossed twice to wild type. Mutants were tested for linkage against a set of autosomal uncoordinated markers. Complementation between surface antigen mutations was scored by antibody screening of animals heterozygous for two surface antigen markers. To obtain heterozygotes of sDf2 and srf-3(yjJO) for complementation testing, srf3(yjJO)/unc-22 srf-3(yjJO) males were mated with sDf2/ nTl(IV); +/nTJ(V) hermaphrodites. Twitcher F1 outcross progeny contained sDf2 opposite unc-22 srf-3(yjJO). For epistasis testing, double mutants carrying srf-J and srf-2 mutations were isolated as Dpy Unc segregants of unc4 srf-J(yjl)/dpy-5 srf-2 hermaphrodites. srf-J(yjl) was described previously (17). Surface Iodination. Adult wild-type and srf-2(yj262) worms were grown from dauer larvae as described (17). Synchronous srf-3(yjlO) adult worms were grown from eggs freed from worms by 0.05 M NaOH/0.2% NaCIO4 treatment as described (17). Five thousand to 10,000 live adult worms were radiolabeled with 1.0 mCi of Na125I (IMS-30; Amersham; 1 Ci = 37 GBq), using Iodo-Gen (Pierce) (24). After a 5-min reaction at room temperature, 20 ,ul of saturated tyrosine in 0.01 M sodium phosphate/0.15 M NaCl (PBS) was added. Then worms were washed repeatedly with PBS by

centrifugation. Extraction of Radiolabeled Worms. Worms either were disrupted on ice in a glass/glass homogenizer containing PBS with the protease inhibitors (PI) described previously (24) plus 5 mM N-ethylmaleimide and 5 ,uM pepstatin (PBS-PI) or were solubilized by boiling for 30 min in PBS-PI with 1% SDS/1.Oo 2-ME/8 M urea. Extracts were then centrifuged for 30 min at 14,000 x g and the supernatants were used in all experiments. Immunoprecipitation of Radiolabeled Antigens. Radiolabeled worm extracts in PBS-PI (10-24 IlI, 70,000-110,000 cpm) were incubated overnight at 4°C with 2.5 ,l of immune or preimmune serum from rabbit no. 74 in 50 1.d of PBS-PT. Ten microliters of sheep anti-rabbit IgG serum (Polyclonal Sera Lab., Cambridge, MA) was added; precipitates were washed with PBS and solubilized by boiling for 30 min in PBS with 1% SDS/1.0%o 2-ME/8 M urea. SDS/PAGE. Radiolabeled worm extracts and immunoprecipitates were analyzed by standard methods (25) on 10-20%o acrylamide linear gradient gels (Enprotech, ISS, Hyde Park, MA). Dried gels were autoradiographed on Kodak X-Omat films using intensifying screens.

RESULTS Isolation and Phenotypic Characterization of Mutants. Antiadult cuticle antiserum previously adsorbed with [As binds to wild-type adults but not to wild-type larvae (17). Here mutants were isolated as immunofluorescent larvae that were recognized by L4-adsorbed antiserum. Apparent frequency of such mutants was 4 x 10-4 per F1 (six mutants) when isolated via the L3 developmental pathway and 1.6 x 10-3 (seven mutants) when isolated via the post-dauer developmental pathway (Materials and Methods). We describe here detailed genetic analyses of four of these mutants. Three strains carrying mutations yj262, yj422, and yjl33 were isolated as post-dauer L4s and appeared similar to wild type in gross morphology. The strain carrying yjlO was isolated after growth through the L3 developmental pathway. yjlO animals had normal morphology but were somewhat

Proc. Natl. Acad. Sci. USA 87 (1990)

fragile to normal handling, and yjJO dauer larvae were killed by the SDS treatment usually used for dauer purification (Materials and Methods). Thus mutations like yjJO could not readily have been isolated after passage through the dauer developmental pathway. All stages of the four mutant stocks tested antigen-positive with L4-adsorbed antiserum (data not shown). Because this reagent binds wild-type adults specifically (17), the mutant phenotypes might result from heterochronic expression of an adult cuticle at earlier stages (14). Therefore, L4s from the four mutant strains were examined under Nomarski optics to determine if adult-specific lateral alae were present at an earlier developmental stage. However, alae were observed only on adults, as expected for wild type. The binding of L4-adsorbed antibodies by mutant larvae could alternatively result from exposure of antigens that are present, but are normally hidden, in the wild-type cuticle. To test for this, antiserum was adsorbed with wild-type adults to remove all antibodies capable of reaction with the wild-type surface. Mutant and wild-type animals were rescreened by indirect immunofluorescence with this adult-adsorbed serum. Wild-type animals showed little staining with adultadsorbed antibodies (Fig. la), as expected. In contrast, all four mutant strains showed a significant level of immunofluorescent staining with adult-adsorbed antibodies (e.g., Fig. 1 b and c). All stages in the mutant strains were antigen-positive. This result indicates the presence of a class of antigenic determinants on the mutant surface that is not available for antibody binding on the wild-type surface. Strains carrying mutations in genes sqt-1, sqt-2, sqt-3, and rol-8 that affect gross morphology were also tested for

FIG. 1. Indirect immunofluorescence staining of wild-type and surface antigen mutant strains of C. elegans. Live nematodes were stained with adult-adsorbed rabbit anticuticle antibodies and photographed. (a) Wild type (strain N2). (b) srf-2(yj262). (c) srf-3(yjlO). (Bar in a = 500 ,um.)



staining with adult-adsorbed antibodies. These strains included the mutant phenotypes Lon, Dpy, Left Rol, Right Rol, and pseudo-wild type (9). In contrast to the surface antigen mutants, none of these morphological mutants (listed in Materials and Methods) showed levels of immunofluorescence significantly higher than the wild-type control (data not shown). Genetic Analysis of Mutations. Recognition of the mutant surface by adult-adsorbed antibodies was uniform; penetrance was >98% in srf-2(yj262) adults and 99.2% in srf-3 adults. All of the mutant phenotypes assessed by antibody screening were recessive to the wild-type phenotype in heterozygotes. Complementation between mutants was measured in heterozygotes in all pairwise combinations, using surface immunofluorescence as marker. Most complementation results were scored in males, but results with hermaphrodites were similar. One complementation group, hereafter designated srf-2, contained yj262, yjl33, and yj422. The remaining mutation, yjJO, complemented the other three and defines srf-3. As expected for recessive mutations, in crosses between males heterozygous for one mutation and hermaphrodites homozygous for a second, noncomplementing mutation, approximately one-half of the cross-progeny were antigenpositive.

The three srf-2 mutations showed linkage to unc-13(eSJ)I. The fourth mutation, srf-3(yjlO), showed linkage to unc24(e138)IV. The linkage data support the complementation data indicating that two complementation groups are represented. As expected for recessive mutations, in cases where unlinked unc markers were assorting independently from the surface antigen marker, approximately one-fourth ofthe Unc segregants were antigen-positive. A map position for srf-2(yj262) just to the right of lin-J I (I) (Fig. 2) was determined from two-factor (Table 1) and threefactor (Table 2) crosses. Two-factor crosses indicated that srf-3(yjJO) is located to the right of unc-22 (Table 1). In deficiency mapping experiments, srf-3(yjlO) complements sDf2. Of 83 animals of genotype unc-22 srf-3(yjJO)/sDf2 screened with adult-adsorbed antibodies, all were antigennegative (non-mutant). Because sDf2 fails to complement 4 map units

dpy-5 unc-13

lin-ll

unc-75

ID Ir

srf- 2 unc-24

IVR

_

unc-22

I

unc-31

Iet-97

sDf2 srf-3 3 map units

FIG. 2. Partial genetic maps of the right arms of C. elegans linkage groups I and IV showing positions of srf-2 I and srf-3 IV, relevant genetic loci, and sDf2 IV. The scale for linkage group I is shown above the maps and the scale for linkage group IV is shown below the maps. Positions of all loci on the lines except for let-97 are from ref. 26. Positions of sDf2 IV and let-97 IV are from ref. 23. Mapping of srf-2 and srf-3 is described in the text. The bar length for sf-2 position represents the 95% confidence interval inferred from a three-factor cross with flanking markers lin-il and unc-75. The bar length for srf-3 position represents the 95% confidence interval for the distance from unc-22 measured in a two-factor cross and takes into account that srf-3 complements sDf2.

Table 1. Two-factor crosses No. of positive Parental non-Unc genotype recombinants +

+

unc-75 srf-2 +

+

unc-24 srf-3 +

2903

Total no. scored

Map distance, cM

36

1751

3.15 ± 1.04

21

1286

2.48

±

1.08

+

0.45 ± 0.44 4 1346 unc-22 srf-3 Non-Unc segregants of the parental genotypes shown were stained with adult-adsorbed antibodies, and antigen-positive recombinants and the total number non-Uncs were scored. Map distances (in centimorgans) were calculated from the expression p = 1 - [(1 - 3 k)]1/2 (6), where k is the ratio of antigen-positive nonmarker progeny to total nonmarker progeny. Ninety-five percent confidence intervals were calculated from the formula 1.96 [(pq/N)1/2], where p = antigen-positive non-Unc individuals/total non-Unc individuals, q = 1 - p, and N = total non-Unc individuals.

let-97, it follows that srf-3(yjlO) is located to the right of let-97 (Fig. 2). Testing of Epistasis Between srf-1 and srf-2. srf-2 or srf-3 mutant phenotypes evaluated by immunofluorescence represent a gain of antigenicity compared to wild type. In contrast, the srf-J mutant phenotype previously identified by failure to bind L4-adsorbed antibodies (17) represents the loss of an adult-specific antigen class. To test whether a srf-J mutation causing loss of antigens from the adult surface prevents expression of the srf-2 mutant phenotype, double mutant strains containing srf-J(yjl) and srf-2 were constructed and tested for antibody binding with L4-adsorbed and adult-adsorbed antibodies by immunofluorescence. Results are shown for srf-2(yj262) in Table 3; results were similar for srf-2(yj133) (data not shown). Complementation testing confirmed that the putative double contained both srf-J and srf-2 mutations (Table 3, lines 2 and 3). The double mutant was antigen-positive with L4-adsorbed and adult-adsorbed antibodies (Table 3, line 1), similar to srf-2(yj262) alone (Table 3, line 4). In control tests, binding of adult-adsorbed antibodies to either wild type or srf-1 alone was negative (e.g., Table 3, lines 5 and 6), as expected. Thus a srf-J mutation does not prevent expression of a srf-2-like mutant phenotype. Surface Composition of Wild-Type and Mutant Strains. To analyze in molecular terms the phenotypes defined by immunofluorescence, wild-type (N2), srf-2(yj262), and srf3(yjJO) adult worms were surface radiolabeled by lodoGen-mediated iodination. Between 93% and 98% of the radioactivity bound to the worms could be solubilized in SDS/2-ME/urea buffers. This material was analyzed by SDS/PAGE (Fig. 3A). Some differences appeared between the three strains in the labeled bands above Mr 30,000, most notably an Mr 51,000 component not readily detectable in the extract of srf-3 worms (Fig. 3A, track 3). More striking differences were found below Mr 30,000. Wild-type worm Table 2. Three-factor crosses No. of recombinants Antigen- Antigenpositive negative

Recombinant Parental genotype phenotype + unc-13 1 11 dpy-5 Dpy non-Unc Unc non-Dpy + + srf-2(yj262) 8 0 + + Vul Unc 2 10 lin-ll + srf-2(yj262) unc-75 + + 1 10 Homozygous recombinant clones were tested with adult-adsorbed antibodies.

2904



Table 3. Epistasis testing of srf-J and srf-2 No. positive/total* L4Adultadsorbed adsorbed

Test

1 2 3 4 5

Partial genotype srf-2(yj262) I; srf-1 Jff srf-2(yj262)/+; srf-iP srf-2(yj262); srf-1/+l* srf-2(yj262) Wild type

serum

serum

37/37 0/66 36/56 73/73 20/20 0/40 L4s 5/49

67/67 0/42 24/46 50/50 0/54

0/32 srf-J entries indicate adult animals scored except where L4s are indicated. tIn these tests, genotypes also included dpy-5(e61) linked to srf2(yj262) and unc-4(e120) linked to srf-1 (not shown). *It is assumed that -50%o of the non-Unc non-Dpy progeny in this test had the genotype shown (see text for details of construction). 6

*All

extracts showed bands at Mr 19,400 (a doublet of Mr 19,40020,000 in some preparations) and Mr 15,000 and intense bands running at Mr 9700 and 5600 (Fig. 3A, track 1). In contrast, extracts of radiolabeled srf-2(yj262) or srf-3(yjlO) mutants only showed intense bands at Mr 6200 and 6700, respectively (Fig. 3A, tracks 2 and 3). A Mr 29,000 component was detectable in extracts of srf-2(yj262) (Fig. 3A, track 2) but much less so in the other two strains. Tracks 1-3 of Fig. 3A were loaded with samples containing the same amount of trichloroacetic acid-precipitable radioactivity. Homogenization of wild-type adults in PBS-PI solubilized 50-60%o of the bound radioactivity. Predictably, this extract contained fewer bands than the corresponding SDS/ 2-ME/urea extract (Fig. 3B, track 1). Most prominent were the characteristic radiolabeled molecules of Mr 9700 and 5600. The Mr 9700 and 5600 wild-type molecules could be immunoprecipitated with whole, unadsorbed rabbit anticuticle antiserum (Fig. 3B, track 2) but not with normal rabbit serum (Fig. 3B, track 3). These results demonstrate that surface-labeled wild-type molecules are recognized by anti-

B

A 1

2

3

9.w

12 3 -200

200-

-97.4

97.468--

43-

-68

--43

v

29-

-29

18.4-

-18.4

14,3-

r,.0

-14.3

"

is

_

6.0

3.0-

FIG. 3. Autoradiographs of 10-20% gradient SDS/polyacrylamide gels showing surface radioiodinated molecules of wild-type and mutant C. elegans adults. (A) 125I-labeled worms were extracted with SDS/2-ME/urea-containing buffers plus PI. Tracks: 1, wild type (N2) adults; 2, srf-2(yj262) adults; 3, srf-3(yjlO) adults. Tracks 1-3 were loaded with the same amount of trichloroacetic acidprecipitable radioactivity. (B) Wild-type labeled adult worms were also extracted by homogenization in PBS-PI and the extract was immunoprecipitated by incubation with serum from rabbit no. 74 followed by sheep anti-rabbit IgG antiserum. Tracks: 1, total PBS-PI extract of labeled wild-type adults; 2, material immunoprecipitated by incubation with rabbit no. 74 serum; 3, material immunoprecipitated with normal rabbit serum. Tracks 2 and 3 show results from reactions with the same input amount of radioactive extract. Molecular weights are given as Mr x 10-3.

bodies in the anticuticle antiserum. However, no molecules were differentially precipitated by immune and normal sera from PBS-PI extracts of srf-2(yj262) (unpublished data). Immunoprecipitation of srf-3(yjJO) extracts was not attempted. It has been impossible to perform analogous immunoprecipitation experiments with adult-adsorbed antibodies, because at a dilution similar to that resulting from the adsorption procedure, even whole antiserum fails to precipitate detectable amounts of radioactivity.

DISCUSSION

srf-2 and srf-3 mutants are unlike others altered in cuticle structure. They show no morphological abnormalities by immunofluorescence or Nomarski optics, and the srf-2 and srf-3 mutant phenotypes are not shared with the morpholog-

ical mutants we have tested. Two lines of evidence support the idea that srf-2 and srf-3 mutants are not heterochronic mutants. (i) They carry surface antigens that are not available for antibody binding on any wild-type stage and do not appear to express adult-specific alae at an earlier developmental stage. (it) A srf-J mutation does not prevent expression of a srf-2-like phenotype in srf-J srf-2 double mutants, as would have been expected if this phenotype were caused by heterochronic expression of an adult-specific antigen class. This apparent noninteraction also suggests that srf-2 is very different in activity from srf-J. In classical immunogenetics, expression of a cell-surface antigen is generally associated with the presence of a dominant allele (27). srf-J mutations fit this model, with the recessive srf-J phenotype corresponding to a stage-specific loss of surface antigenicity (17). In contrast, srf-2 and srf-3 mutant alleles cause an apparent gain of surface antigenicity. Yet all of the mutant phenotypes are recessive. How can the apparent gain of antigenicity in srf-2 and srf-3 mutants be reconciled with their recessive character? The parent antiserum was raised against a mixture of wild-type cuticle proteins. However, the adsorption experiments indicate that the antigens present on the mutant surface are not present on the wild-type surface. An explanation could be that the mutations cause loss of components from the cuticle surface, thereby exposing antigenic determinants that are normally hidden in the wild-type cuticle. The mutations would be recessive because the activity supplied by one wild-type allele is sufficient for normal cuticle formation. Two independent experiments are consistent with this hypothesis. First, in radioiodination experiments, several surface components of wild-type adult worms either were not detectable or were labeled much less efficiently in extracts of

srf-2(yj262) or srf-3(yjJO) mutants. Conversely, a Mr 29,000

component labeled on srf-2(yj262) worms was not readily detected on wild type or srf-3(yjlO), and Mr 6200 and 6700 components were readily detected only on srf-2(yj262) and srf-3(yjJO) worms, respectively. Labeled components solubilized in PBS-PI extracts of wild-type worms were immunoprecipitated by the same rabbit anticuticle serum used to derive adsorbed antibody reagents for the genetics experiments described here. Second, in preliminary experiments, a mouse monoclonal antibody that reacts specifically with the wild-type Li (first larval stage) surface does not bind detectably to the surface of srf-2 or srf-3 LUs in immunofluorescence tests (unpublished data). Both of these experiments indicate that antigenic determinants exposed in wild type are less readily detected on the mutant surface. The present data do not distinguish this simple model from a more complex one in which a rearrangement of cuticle layers occurs such that hidden antigens are exposed and normally exposed antigens are hidden. However, the surface labeling experiments do tend to eliminate models in which


Genetics: Politz et al. accessibility of internal antigenic determinants increases without a concomitant loss of a component from the cuticle surface. At present, it is unknown whether any of the surface radiolabeled molecules in srf-2 and srf-3 mutants correspond to newly exposed antigens observed by immunofluorescence. Indeed, reduced, not enhanced, radiolabeling of some surface components was observed in the mutants. Because 93-98% of the radioactivity associated with the worms was solubilized by boiling in SDS/2-ME/urea, it is unlikely that major exposed molecules are insoluble to extraction and therefore escape detection. However, some newly exposed molecules may be extractable but not label efficiently. Thus, although the surface iodination experiments have identified wild-type surface antigens, it may be that the antigens newly exposed in the mutants have not yet been identified. Surface antigens are logical candidates for development of vaccines against parasitic nematode infections. In fact, antibodies and effector cells capable of recognizing specific nematode surface antigens have been described and characterized (reviewed in ref. 4). In some cases, these effectors of the immune response are capable of killing parasites in vitro (2830). In one case, partial protective immunity has been conferred by immunization with a nematode surface antigen (31). Surface antigen expression during postembryonic development in parasitic nematodes is dynamic, with changes occurring at the molts, where an entire cuticle is lost and a new one is synthesized, and between molts (reviewed in refs. 4, 32). In some cases, these changes appear to be significant with respect to recognition of specific stages by the host immune system. The surface alterations in the C. elegans mutants described here are of interest with respect to these considerations. It appears that these changes in the antigenicity and composition of the cuticle surface can occur without concomitant changes in body shape or the expression of stage-specific features. We suggest that the layered organization of the cuticle allows for a potentially powerful adaptive mechanism-i.e., the composition of the cuticle surface can: be varied without affecting gross morphology. Of four mutations examined in detail here, three are in the same gene, srf-2 I. Two additional mutations show linkage to chromosome I (unpublished data), suggesting that they may also be srf-2 alleles. Thus, the total number of genes that can be mutated to give this phenotype may be small. Interestingly, although the phenotypes of srf-2(yj262) and srf-3(yjJO) appear identical by immunofluorescence, the patterns of molecules radiolabeled on the surfaces of srf-2 and srf-3 are somewhat different. Thus, mutations that expose hidden structural determinants may have different effects on cuticle structure.

EMS-induced mutations are likely to be single-base substitutions, and thus small genetic changes can apparently cause distinct changes in surface antigenicity, with loss and gain of surface antigenic determinants occurring simultaneously. Mutations causing such changes need not, therefore, accumulate over long periods of evolutionary history in order to be significant. In fact, sympatric nematode species differing in a surface antigen have been observed. The filarial nematodes Brugia malayi and Brugia pahangi are sympatric in mosquitoes but infect human and cat vertebrate hosts, respectively. B. malayi expresses a stage-specific surface epitope that is undetectable in B. pahangi (33). If different surface antigen phenotypes are recognized differentially by a host's immune system, the acquisition or loss of surface antigens could lead to rapid speciation by a sympatric or parapatric mode (ref. 34; cited in ref. 35). All of the important parasitic nematode species reproduce by obligate male-female mating. Thus, recessive mutations like srf-2 and srf-3 would persist in parasitic populations, being

2905

revealed in the phenotype only when random mating of two carriers produced homozygous mutant progeny. An effective vaccine or immune response must target the antigens displayed by all phenotypes represented in an infecting population. Therefore investigating the genetics of surface antigenicity and resistance to the host immune response may prove fruitful. The present results based on C. elegans genetic analysis have allowed inferences about the relationship between genotype and surface antigen phenotype to be made without assumptions about the molecular nature of the surface antigens themselves. Combining these methods with molecular analyses of surface antigens may be a key to understanding the successful survival strategies of parasitic nematodes. We thank R. S. Edgar for strains, Mark Edgley for suggesting the use of, and providing, a lin-li him-8 double mutant, and Ed Hedgecock and Joan C. Politz for discussion. This research was supported by National Science Foundation Grant DCB8510567 and U.S. Public Health Service Grant RO1 GM40339 to S.M.P. 1. Poinar, G. O., Jr. (1983) The Natural History of Nematodes (Prentice-Hall, Englewood Cliffs, NJ). 2. Ogilvie, B. M. & De Savigny, D. (1982) in Immunology of Parasitic Infections, eds. Cohen, S. & Warren, K. S. (Blackwell, London),

pp. 715-757. 3. Wakelin, D. (1988) in Genetics of Resistance to Bacterial and Parasitic Infection, eds. Wakelin, D. M. & Blackwell, J. M. (Taylor & Francis, London), pp. 153-224. 4. Philipp, M. T. & Rumjaneck, F. D. (1984) Mol. Biochem. Parasitol. 10, 245-268. 5. Almond, N. M. & Parkhouse, R. M. E. (1986) in Current Topics in Microbiology and Immunology, ed. Parkhouse, R. M. E. (Springer, Berlin), Vol. 120, pp. 173-203. 6. Brenner, S. (1974) Genetics 77, 71-94. 7. Higgins, B. J. & Hirsh, D. (1977) Mol. Gen. Genet. 150, 63-72. 8. Cox, G. N., Laufer, J. S., Kusch, M. & Edgar, R. S. (1980) Genetics 95, 317-399. 9. Kusch, M. & Edgar, R. S. (1986) Genetics 113, 621-639. 10. Kramer, J. M., Johnson, J. J., Edgar, R. S., Basch, C. & Roberts, S. (1988) Cell 55, 555-565. 11. von Mende, N., Bird, D. M., Albert, P. S. & Riddle, D. L. (1988) Cell 55, 567-576. 12. Kramer, J. M., Cox, G. N. & Hirsh, D. (1982) Cell 30, 599-606. 13. Cox, G. N., Kramer, J. M. & Hirsh, D. (1984) Mol. Cell. Biol. 4, 2389-2395. 14. Ambros, V. R. & Horvitz, H. R. (1984) Science 226, 409-416. 15. Ambros, V. R. & Horvitz, H. R. (1987) Genes Dev. 1, 398-414. 16. Ambros, V. (1989) Cell 57, 49-57. 17. Politz, S. M., Chin, K. J. & Herman, D. L. (1987) Genetics 117, 467-476. 18. Cox, G. N., Kusch, M. & Edgar, R. S. (1981) J. Cell Biol. 90, 7-17. 19. Politz, S. M., Politz, J. C. & Edgar, R. S. (1986) J. Nematol. 18, 303-310. 20. Trent, C., Tsung, N. & Horvitz, H. R. (1983) Genetics 104, 619647. 21. Moerman, D. G. & Baillie, D. L. (1981) Mutat. Res. 80, 273-279. 22. Ferguson, E. L. & Horvitz, H. R. (1985) Genetics 110, 17-72. 23. Clark, D. V., Rogalski, T. M., Donati, L. M. & Baillie, D. L. (1988) Genetics 119, 345-353. 24. Philipp, M. & Davis, T. B. (1986) J. Immunol. 136, 2621-2627. 25. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 26. Hodgkin, J., Edgley, M., Riddle, D. L. & Albertson, D. G. (1988) in The Nematode Caenorhabditis elegans, ed. Wood, W. B. (Cold Spring Harbor Lab., Cold Spring Harbor, NY), pp. 560-580. 27. Hildemann, W. H. (1970) Immunogenetics (Holden-Day, Oakland,

CA).

28. Subrahmanyam, D., Rao, Y. V. G. B., Mehta, K. & Nelson, D. S. (1976) Nature (London) 260, 529-530. 29. Kazura, J. W. & Grove, D. I. (1978) Nature (London) 247,588-589. 30. Mackenzie, C. D., Jungery, M., Taylor, P. M. & Ogilvie, B. M. (1980) Eur. J. Immunol. 10, 594-601. 31. Grencis, R. K., Crawford, C., Pritchard, D. I., Behnke, J. M. & Wakelin, D. (1986) Parasite Immunol. 8, 587-596. 32. Philipp, M., Davis, T. B., Storey, N. & Carlow, C. K. S. (1988) Annu. Rev. Microbiol. 42, 685-716. 33. Carlow, C. K. S., Franke, E. D., Lowrie, R. C., Jr., Partono, F. & Philipp, M. (1987) Proc. Natl. Acad. Sci. USA 84, 6914-6918. 34. Bush, G. L. (1975) Annu. Rev. Ecol. Syst. 6, 339-364. 35. Raff, R. A. & Kaufman, T. C. (1983) Embryos, Genes, and Evolution (Macmillan, New York), p. 91.