smg-7 Is Required for mRNA Surveillance in Caenorhabditis elegans

Corresponding author: Philip Anderson, Department of Genetics, University of Wisconsin, 445 Henry Mall, Madison, WI 53706., andersn{at}facstaff.wisc.edu (E-mail)

Communicating editor: R. K. HERMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Eukaryotic mRNAs that contain premature stop codons are degraded more rapidly than their wild-type counterparts, a phenomenon termed "nonsense-mediated mRNA decay" (NMD) or "mRNA surveillance." Functions of six previously described Caenorhabditis elegans genes, smg-1 through smg-6, are required for NMD. Whereas nonsense mutant mRNAs are unstable in smg(+) genetic backgrounds, such mRNAs have normal stability in smg(-) backgrounds. Previous screens for smg mutations have likely not identified all genes involved in NMD, but efforts to identify additional smg genes are limited by the fact that almost 90% of smg mutations identified in genome-wide screens are alleles of smg-1, smg-2, or smg-5. We describe a modified screen for smg mutations that precludes isolating alleles of smg-1, smg-2, and smg-5. Using this screen, we have identified and cloned smg-7, a previously uncharacterized gene that we show is required for NMD. smg-7 is predicted to encode a novel protein that contains an acidic carboxyl terminus and two probable tetratricopeptide repeats. We provide evidence that smg-7 is cotranscribed with the previously characterized gene lin-45 and show that null alleles of smg-7 confer a temperature-sensitive defect in NMD.

EUKARYOTIC mRNAs that contain premature translation termination codons are usually less stable than their wild-type counterparts, a phenomenon termed "non-sense-mediated mRNA decay" (NMD) or "mRNA surveillance." NMD may protect cells from the deleterious effects of truncated proteins by degrading the aberrant mRNAs that encode them (PULAK and ANDERSON 1993 ; CALI and ANDERSON 1998 ). NMD was first identified in yeast (LOSSON and LACROUTE 1979 ) and has since been described in all eukaryotes tested (for recent reviews, see MAQUAT 1995 ; JACOBSON and PELTZ 1996 ; RUIZ-ECHEVARRIA et al. 1996 ). Because of NMD, the steady-state levels of nonsense mutant mRNAs are substantially reduced relative to normal mRNAs.

Genes whose functions are required for NMD have been described in both Caenorhabditis elegans (smg-1 through smg-6) and Saccharomyces cerevisiae (UPF1, UPF2/NMD2, UPF3; LEEDS et al. 1991 , LEEDS et al. 1992 ; PULAK and ANDERSON 1993 ; CUI et al. 1995 ; HE and JACOBSON 1995 ; LEE and CULBERTSON 1995 ). Loss-of-function alleles of smg or UPF genes eliminate NMD and prevent the normally rapid decay of mRNAs containing nonsense or frameshift mutations. Thus, the smg/UPF genes encode trans-acting factors needed for NMD. NMD is nonessential in both yeast and nematodes, as upf and smg mutants are viable yet defective for NMD. At least one of the proteins required for NMD is both structurally and functionally conserved among yeast and nematodes. SMG-2 of C. elegans is about 50% identical to Upf1p of yeast (M. F. PAGE, B. CARR, K. R. ANDERS and P. ANDERSON, unpublished results). Both SMG-2 and Upf1p are also ~50% identical to a human protein that is likely involved in mammalian NMD (PERLICK et al. 1996 ; APPLEQUIST et al. 1997 ; SUN et al. 1998 ).

Most smg mutations were identified because of their suppression phenotype. smg mutations are allele-specific but gene-nonspecific suppressors of mutations affecting a variety of C. elegans genes (HODGKIN et al. 1989 ). A small fraction of nonsense mutations and certain gene rearrangements that express mRNAs with aberrant 3' untranslated regions are phenotypically suppressed by smg mutations (HODGKIN et al. 1989 ; KUWABARA et al. 1992 ; PULAK and ANDERSON 1993 ; ZARKOWER et al. 1994 ; ROUGVIE and AMBROS 1995 ; BARNES and HODGKIN 1996 ). Genetic screens for smg mutants are likely not saturated. In a genome-wide screen for smg mutations, only one allele of smg-4 and two alleles of smg-6 were isolated (HODGKIN et al. 1989 ). Efforts to identify additional smg genes have been limited by the fact that almost 90% of smg mutations identified in such screens are alleles of smg-1, smg-2, or smg-5. We describe here a modified screen for smg mutants that precludes isolating alleles of smg-1, smg-2, and smg-5. Using this modified screen, we have identified, cloned, and molecularly characterized a previously uncharacterized gene, smg-7, which we show is required for NMD.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and general methods:
C. elegans strains were maintained as previously described (BRENNER 1974 ). All strains were maintained at 20° except where noted. Mutations and chromosomal aberrations used in this study have been previously described (HODGKIN 1988 , HODGKIN 1997 ) or are noted below:

• LGI: dpy-5(e61), fog-1(q180) (BARTON and KIMBLE 1990 ), lev-11(x12), smg-1(r861) (HODGKIN et al. 1989 ), smg-2(r863), smg-5(r860), unc-13(e51), unc-15(e1402), unc-54(r293, r259, r315) (EIDE and ANDERSON 1985A , EIDE and ANDERSON 1985B ; PULAK and ANDERSON 1988 , PULAK and ANDERSON 1993 ), unc-73(e936).

• LG II: dpy-10(e128), mut-5(st701), vab-9(e1744), tra-2(e1209), unc-4(e120).

• LG III: dpy-1(e1), smg-6(r886, r896) (HODGKIN et al. 1989 ).

• LG IV: bli-6(sc16), deb-1(st555), dpy-4(e1166), dpy-9(e12), dpy-13(e184), lag-1(q385) (CHRISTENSEN et al. 1996 ), lin-45(sy96), smg-3(r867, ma117), unc-22(e66), unc-24(e138), unc-33(e204), unc-44(e362).

• LG V: dpy-11(e224), him-5(e1490), smg-4(ma116), unc-46(e177), unc-60(e723), unc-70(e524), unc-76(e911).

• Chromosome deficiencies: nDf41, stDf7, stDf8 (WILLIAMS and WATERSTON 1994 ).

Mutations isolated and mutagens used in this study are as follows:

• LGI: Ethylmethane sulfonate (EMS)-induced, unc-54(r1189, r1190, r1191, r1192); N-ethyl-N-nitrosourea (ENU)-induced, unc-54(r1195); mut-5 induced, unc-54(r1130, r1132, r1174).

• LG III: EMS-induced, smg-6(r1165, r1188); ENU-induced, smg-6(r1178, r1187, r1198, r1203, r1204, r1205, r1206, r1214); diepoxyoctane (DEO)-induced, smg-6(r1171); mut-5-induced, smg-6(r1172, r1217).

• LG IV: EMS-induced, smg-3(r1081, r1163, r1164, r1175, r1176, r1177); ENU-induced, smg-3(r1179, r1180, r1183, r1184, r1186, r1199, r1200, r1201, r1202, r1213, r1215, r1216); smg-7(r1182, r1197); DEO-induced, smg-3(r1168, r1170); mut-5-induced, smg-3(r1149, r1173, r1218); smg-7(r1131).

• LG V: ENU-induced, smg-4(r1181); DEO-induced, smg-4(r1169).

• Chromosomal aberrations isolated in this study, rT1(IV;V) and rDf3.

Identifying new smg mutations:
Strain TR2034 [genotype unc-13(e51) dpy-5(e61) fog-1(q180) unc-54(r293); +/szT1 (I;X) [+ + + unc-54(r293); lon-2(e678)]] is homozygous for unc-54(r293) and heterozygous for szT1. szT1 balances the right half of LG X and the left half of LG I, which includes the wild-type alleles of smg-1, smg-2, and smg-5. TR2034 is paralyzed due to unc-54(r293), a smg-suppressible allele of the unc-54 myosin heavy chain gene (HODGKIN et al. 1989 ). Intragenic and extragenic suppressors of r293 were identified as motile revertants of TR2034. Animals were mutagenized with EMS, DEO, and 1 mM or 5 mM ENU as described previously (BRENNER 1974 ; DE STASIO et al. 1998 ). Spontaneous mutations were isolated in strain TR2038, which is identical to TR2034 but also contains mut-5(st701) vab-9(e1744). Populations were established with 20–40 mutagenized young adults, and subsequent generations were screened for animals that exhibited increased motility and/or egg-laying proficiency. To ensure independence, no more than one mutant was retained from a single population. Each homozygous revertant was outcrossed to wild-type (N2) males, and F₂ progeny were examined for the presence of Unc-54 animals. The absence of such animals indicated that a suppressor was tightly linked to unc-54, while their presence indicated that a suppressor was unlinked. All intragenic suppressors were found to be dominant. All extragenic suppressors were found to be recessive and were tested for complementation of canonical alleles of smg-1 through smg-6 using suppression of unc-54(r293) at 15°, 20°, and 25° as the scored phenotype.

Mapping of smg-7(r1131):
Hermaphrodites of genotype unc-13(e51) dpy-5(e61) fog-1(q180) unc-54(r293); +/szT1 [+ + + unc-54(r293); lon-2(e678)]; mut-5(st701) vab-9(e1744); smg-7(r1131) were mated to N2 males. Wild-type cross-progeny hermaphrodites were picked singly and allowed to self-fertilize. Unc-54 offspring were then picked and allowed to self at 25°, the nonpermissive temperature of r1131. smg-7(r1131) homozygotes were identified among progeny as fully motile animals. Non-Vab non-Dpy offspring were picked, and the outcross to N2 was repeated three additional times, yielding a strain of genotype smg-7(r1131); unc-54(r293). This strain was crossed with N2 males and a smg-7(r1131) unc-54(+) homozygote was identified among the offspring based upon (i) its protruding vulva (pVul) phenotype and (ii) its wild-type motility at 20°. This smg-7 single mutant was used in mapping experiments described in Table 1.

To test nDf41 for complementation of smg-7(r1131), nDf41/+ males were mated to smg-7(r1131) unc-24(e138); unc-54(r293) hermaphrodites at 25°. One-half of the cross-progeny were pVul, indicating failure of nDf41 to complement smg-7(r1131). Approximately one-fourth of the offspring of these heterozygotes were lethal (nDf41 homozygotes), and the remainder were fully motile at 25°, confirming the failure of r1131 to complement nDf41. To test r1131 for complementation of stDf7 and stDf8, smg-7(r1131) unc-24(e138)/++; unc-54(r293)/+ males were crossed with unc-24(e138) unc-22(e66)/stDf7 or stDf8 hermaphrodites. smg-7(r1131) unc-24(e138)/stDf7 or stDf8 cross-progeny were identified as Unc-24 animals whose offspring were approximately one-fourth inviable (Df homozygotes) but never Unc-22. Approximately one-half of such heterozygotes yielded Unc-54 offspring, indicating that both stDf7 and stDf8 complement smg-7(r1131).

Testing suppression of tra-2(e1209) by smg-7(r1131):
dpy-10(e128) + unc-4(e120)/+ tra-2(e1209) + hermaphrodites were mated with N2 males, and single F₁ males were mated with smg-7(r1131) hermaphrodites. Resultant tra-2(e1209)/+; smg-7(r1131)/+ hermaphrodites were selfed at 25°, and smg-7(r1131) homozygotes picked as pVul hermaphrodites. smg-7(r1131); tra-2(e1209) offspring were examined at 20° and 25° for suppression of the fertility and vulval and tail morphology defects characteristic of tra-2(e1209). Animals raised at 20° were self-sterile [essentially identical to tra-2(e1209)], whereas animals raised at 25° were self-fertile and had more normal vulval and tail morphology. smg-1(r861); tra-2(e1209) was included as a control and showed equivalent suppression at 25°. Brood sizes of smg-7(r1131); tra-2(e1209) and smg-1(r861); tra-2(e1209) animals at 25° were less than those reported for smg suppression of e1209 at 20° (i.e., brood size of 1–5 animals at 25° vs. a mean of 11 animals at 20°; HODGKIN et al. 1989 ).

Testing suppression of dpy-5(e61):
dpy-5(e61)/+ males were crossed with fog-1(q180) dpy-5(e61) unc-13(e51) unc-54(r293); smg-7(r1131) females. fog-1(q180) dpy-5(e61) unc-13(e51) unc-54(r293)/+ dpy-5(e61) + +; smg-7(r1131)/+ cross-progeny were picked and selfed at 25°. Unc-54 offspring were picked and selfed at 25°. smg-7(r1131) homozygotes were identified in the next generation by their normal motility. dpy-5(e61) unc-54(r293); smg-7(r1131) homozygotes were identified among these suppressed animals as those that failed to yield Fog or Unc-13 offspring. Such animals were noticeably longer and less Dpy than dpy-5(e61) homozygotes and indistinguishable from dpy-5(e61); smg-3(r867) control animals.

Noncomplementation screen for new smg-7 alleles:
smg-5(r860) unc-15(e1402) unc-54(r293); him-5(e1409) males reared at 15° (e1402 is temperature sensitive) were mutagenized with EMS and mated to unc-54(r293); smg-7(r1131) unc-24(e138) hermaphrodites at 25°. Self-progeny exhibit an Unc-24 phenotype, and, as both smg-5(r860) and smg-7(r1131) are recessive, most cross-progeny [genotype smg-5(r860) unc-15(e1402) unc-54(r293)/++ unc-54(r293); smg-7(r1131) unc-24(e138)/++; him-5(e1490)/+] exhibit an Unc-54 phenotype. Rare cross-progeny carrying a new mutation in smg-7 are fully motile, because the paralysis of unc-54(r293) is suppressed in smg-7(r1131)/smg-7(new) "homozygotes." smg-7(new) can be subsequently distinguished from smg-7(r1131), as smg-7(r1131) is coupled to unc-24(e138) in the screen.

Analysis of rT1:
Tests of linkage: Heterozygous m/+ males (where m denotes one of several recessive visible markers tested) were mated with unc-44(e362) smg-7(r1131) unc-24(e138); +/rT1 (IV;V) [+++;+] hermaphrodites. Cross-progeny of genotype m/+; +/rT1[+;+] animals were isolated and allowed to self, and m/m offspring were picked. Such animals were examined for whether a large number of dead eggs were present among their progeny, as is characteristic for rT1/+ animals. If m is unlinked to rT1, two-thirds of m/m homozygotes are expected to be rT1/+ heterozygotes. If m is linked to rT1, less than two-thirds of m/m homozygotes are expected to be rT1/+ heterozygotes, with the actual proportion determined by the frequency of crossing over between m and the breakpoint of rT1. Markers used in mapping rT1, and the fraction of m/m that were heterozygous for rT1 were as follows:

• LG I: lev-11 (4/4), unc-73 (3/6).

• LG II: none tested.

• LG III: dpy-1 (4/5), dpy-18 (7/9).

• LG IV: unc-24 (0/10).

• LG V: unc-70 (0/9), dpy-11 (0/7).

Tests of recombination suppression:
Heterozygous m1 m2/++ males, where m1 and m2 denote linked markers between which recombination was being tested (see Table 2), were mated to unc-44(e362) lag-1(q385); +/rT1 (IV;V) [++;+] hermaphrodites, and m1 m2; +/rT1 (IV;V) [++;+] heterozygous cross-progeny were isolated. Among the self-progeny of such heterozygotes, the numbers of m1 non-m2 and m2 non-m1 offspring were scored.

RNase protection assays:
RNA isolation and RNase protection assays were performed as previously described (PULAK and ANDERSON 1993 ). Twenty micrograms of total RNA was tested in each assay. RNase-protected products were quantified with a Phosphorimager, and the unc-54 signal was normalized to that of act-1.

Identification, cloning, and mapping of dimorphic Tc1 elements near smg-7:
smg-7(r1131), which was isolated in a strain that contains a high copy number of Tc1, was outcrossed five times to N2 prior to analysis. Dimorphic Tc1 elements linked to smg-7(r1131) were identified on Southern blots and mapped relative to two crossovers between bli-6 and unc-24 (see Figure 2). Two dimorphic Tc1 elements, designated rP14::Tc1 and r1131::Tc1, were inseparable from smg-7 in these tests of linkage. We isolated genomic DNAs flanking rP14::Tc1 and r1131::Tc1 by inverse PCR (OCHMAN et al. 1988 ), cloned them into plasmid vectors, and sequenced the Tc1 insertional junction fragments. When compared to C. elegans genomic and cDNA sequences (WATERSTON et al. 1997 ), rP14::Tc1 proved to derive from cosmid C25A8, while r1131::Tc1 (contained on plasmid clone TR#235) had not yet been sequenced. We refined the genetic positions of rP14::Tc1 and r1131::Tc1 using PCR tests of their inheritance among 20 crossovers in the ~0.4 map unit unc-44 to deb-1 interval, yielding the following distribution of crossovers: unc-44 (15/20) smg-7(r1131::Tc1) (4/20) rP14::Tc1 (1/20) deb-1. Analysis of 23 additional crossovers in the same interval failed to separate r1131::Tc1 from smg-7.

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. A modified screen for smg mutations that precludes isolating alleles of smg-1, smg-2, and smg-5. TR2034 is heterozygous for the reciprocal translocation szT1 and homozygous for unc-54(r293). smg-1, smg-2, smg-5, and fog-1 are located on LG I (left), within the region balanced by szT1. Among the euploid offspring of TR2034, szT1 and fog-1 homozygotes are lethal and sterile, respectively, whereas szT1 heterozygotes are viable and fertile. Aneuploid offspring are inviable. Mobile revertants of TR2034 identified among the F2 offspring are either alleles of unc-54 itself or extragenic suppressors of unc-54(r293). In the absence of crossing over (which is suppressed by szT1), recessive smg mutations must be located on LGs II, III, IV, or V. smg mutations on LGs I (right) and X (left) can in principle be identified, but they must occur concomitantly with an appropriate crossover. See MATERIALS AND METHODS for a complete description of the screen.

View larger version (11K): In this window In a new window Download PPT slide   Figure 2. Linkage map near smg-7. Except for smg-7, genes shown above the chromosome have been previously ordered by a combination of three-factor crosses and deficiency mapping. We mapped smg-7 relative to bli-6, smg-3, lin-45, and deb-1 by three-factor crosses, and relative to stDf7, stDf8, and nDf41 by complementation tests.

Transformation rescue of smg-7:
The insert of full-length cDNA clone TR#288 was amplified by PCR and cloned into the EcoRV site of expression vector pPD49.78 (MELLO and FIRE 1995 ), yielding plasmid clone TR#302. Plasmids TR#302 and pRF4 were microinjected into unc-54(r293); smg-7(r1131) and heritable transformants established. In the absence of heat shock, such transformants have normal motility, indicating that they are non-Smg. A small fraction of animals was slightly slow, suggesting that the transgene partially rescues smg-7(r1131) under non-heat-shock conditions. To test for heat-shock-dependent rescue, a mixed-stage population was heat-shocked at 33° for 50 min, followed by a 30-min incubation at 22°. Roller larvae (those that retain the transforming array) and nonroller larvae (those that have lost it) were isolated singly, raised at 25° for 1–2 days, and scored for motility. All roller animals (10/10) grew to be Unc-54 Egl adults (non-Smg phenotype), while all nonroller animals (24/24) grew to be fully motile, egg-laying proficient adults (Smg phenotype).

Polyclonal anti-SMG-7 antibodies and Western blots:
An 800-bp EcoRV fragment from smg-7 cDNA clone TR#288, corresponding to SMG-7 amino acids 174–448, was cloned in-frame into the SmaI site of pGEX-2T (Pharmacia, Piscataway, NJ) and transformed into the Escherichia coli host strain BL21. The resulting SMG-7/GST fusion protein was induced, found to be insoluble, and purified as inclusion bodies as previously described (WILLIAMS et al. 1995 ). Two New Zealand White rabbits were immunized with 1.0 mg of fusion protein and boosted 2, 6, 10, and 14 wk thereafter. Sera was collected after the fourth boost and affinity purified as previously described (WILLIAMS et al. 1995 ) using both SMG-7/GST- and GST-affinity columns to purify anti-SMG-7 antibodies. Protein samples for Western blots were prepared by boiling 30 adult worms in SDS loading buffer, after which they were run on denaturing SDS/polyacrylamide gels and transferred to immobilon-P membranes (Millipore Corp., Bedford, MA) using a semidry blotter. Membranes were incubated for 1 hr in PBS containing 0.1% Tween-20, 5% powdered milk, and primary and secondary antibodies at appropriate dilutions. Signals were detected by chemiluminescence (ECL system, Amersham Life Sciences, Arlington Heights, IL) according to the manufacturer's instructions.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

unc-54(r293) is one of several previously identified alleles suppressed by mutations of smg-1 through smg-6 (HODGKIN et al. 1989 ). unc-54 encodes myosin heavy chain B, which is required in body-wall muscle for normal locomotion (EPSTEIN et al. 1974 ). unc-54(r293) is a small deletion that deletes the unc-54 polyadenylation signal but does not affect the unc-54 open reading frame (PULAK and ANDERSON 1988 ). unc-54(r293) expresses an unusually large mRNA that is unstable due to the action of NMD. smg mutations phenotypically suppress unc-54(r293) by eliminating NMD. Increased accumulation of unc-54(r293) mRNA in smg mutants causes increased synthesis of a myosin heavy chain B whose amino acid sequence is normal (PULAK and ANDERSON 1993 ).

A modified screen for smg mutations:
We designed a modified screen for smg mutations for two purposes: (i) to preclude isolating additional alleles of smg-1, smg-2, and smg-5, which collectively comprise almost 90% of smg mutations identified in genome-wide screens (HODGKIN et al. 1989 ); and (ii) to facilitate the cloning by transposon tagging of any new smg genes that we might identify. We constructed strain TR2034, which is homozygous for unc-54(r293) and heterozygous for szT1 (see Figure 1). szT1 is a reciprocal translocation that balances the left half of LG I and the right half of LG X (FODOR and DEAK 1985 ). smg-1, smg-2, and smg-5 are all located on the left half of LG I, and szT1 contains the wild-type alleles of those genes. The unrearranged chromosome I of TR2034 contains fog-1(q180), which causes sterility when homozygous (BARTON and KIMBLE 1990 ). TR2034 contains additional mutations coupled to fog-1(q180) and szT1 to mark their presence (see MATERIALS AND METHODS). Isolating recessive alleles of smg genes located on the left half of LG I or the right half of LG X is precluded from this screen, because animals homozygous for szT1 are inviable and those homozygous for fog-1 are sterile. To facilitate isolating smg mutations not previously described, we mutagenized TR2034 with EMS, DEO, or ENU, each of which yields a different spectrum of mutations (ANDERSON 1995 ; DE STASIO et al. 1998 ). In screens for spontaneous smg mutations, the mutator mutation mut-5(st701) was included in the background, because it activates the transposable element Tc1 (MORI et al. 1988 ).

We isolated a total of 49 suppressors of unc-54(r293) using this screen. Eight are dominant and inseparable by recombination from unc-54(r293) (RF < 0.1 map unit; see MATERIALS AND METHODS). The tight linkage and dominance of these mutations suggest that they are intragenic suppressors affecting unc-54 itself. Molecular characterization of these mutations confirms this assignment and will be described elsewhere (S. O'CONNOR, H. SHANG and P. ANDERSON, unpublished results). The remaining 41 suppressors are recessive and unlinked to unc-54. Complementation tests with canonical alleles of smg-1 through smg-6 demonstrated that 23 are alleles of smg-3, 13 are alleles of smg-6, and 2 are alleles of smg-4. The remaining 3 suppressors define a new gene, smg-7, which is described below.

Genetic identification of smg-7:
Three suppressors (r1131, r1182, and r1197) complement canonical alleles of smg-1 through smg-6 yet fail to complement each other. Three lines of genetic evidence demonstrate that these alleles define smg-7, a previously unidentified smg gene. First, smg-7 mutations exhibit the same pattern of allele-specific suppression as other smg mutations. smg-7(r1131) suppresses tra-2(e1209) and dpy-5(e61), but does not suppress unc-54(r308), unc-54(r315), or unc-54(r259). Second, smg-7(r1131, r1182, and r1197) hermaphrodites have protruding vulvae and males have abnormal bursae, morphogenetic defects characteristic of smg mutants. Third, smg-7 maps to a genetic position that is distinct from previously described smg genes (see Figure 2). smg-7 maps between bli-6 and deb-1, ~0.1 map unit to the right of smg-3. Although smg-7 is within the same three-factor interval as smg-3 (see Table 1), the two genes are clearly separable. From a heterozygote of genotype unc-44 + smg-7 unc-24/+ smg-3 + +, one of three Unc-24 non-Unc-44 recombinant chromosomes was a smg-3 smg-7 double mutant. The unc-24 smg-3 smg-7 chromosome fails to complement alleles of both smg-3 and smg-7, whereas the smg-3 and smg-7 alleles involved in the cross fully complement each other. Furthermore, all known alleles of smg-7, including two deletions (see below), fully complement smg-3(r867 and ma117) for both suppression and morphogenetic phenotypes.

smg-7(r1131, r1182, and r1197) are temperature sensitive, in contrast to all previously described smg mutations. At 15° and 20°, unc-54(r293); smg-7 homozygotes are slightly more motile and egg-laying proficient than unc-54(r293); smg(+) controls, but the suppression is very weak. At 25°, however, unc-54(r293); smg-7 animals have normal motility and egg laying. Suppression of unc-54(r293) by smg-7(r1131) is affected by maternal genotype. unc-54(r293); smg-7(r1131) homozygotes recovered from unc-54(r293); r1131/+ heterozygous mothers are less motile (more weakly suppressed) than offspring of unc-54(r293); smg-7(r1131) homozygous mothers. A similar maternal-presence effect has been described for alleles of smg-3, smg-4, and smg-6 (HODGKIN et al. 1989 ).

smg-7 is required for nonsense-mediated mRNA decay:
To test whether smg-7 is required for NMD, we measured the steady-state levels of unc-54 wild-type and nonsense mutant mRNAs in a smg-7(r1131) background. mRNA of unc-54(r315), an amber mutation at unc-54 codon 1263 (out of 1966), is known from previous work to be unstable in smg(+) backgrounds but stable in smg(-) backgrounds (PULAK and ANDERSON 1993 ). The RNase protection experiment shown in Figure 3 demonstrates that smg-7(r1131) eliminates the rapid degradation of unc-54(r315) mRNA. unc-54(r315) mRNA is reduced in this experiment to 11% of wild type in an smg(+) background (Figure 3, lane 4), whereas it is of normal abundance in smg-7(r1131) animals raised at 25° (Figure 3, lane 5). smg-7(r1131) does not significantly affect the quantity of unc-54(+) mRNA (Figure 3, lane 3). These results demonstrate that smg-7(r1131), like mutations affecting smg-1 through smg-6, prevents the rapid decay of nonsense mutant mRNAs.

View larger version (49K): In this window In a new window Download PPT slide   Figure 3. Effect of smg-7(r1131) on nonsense-mediated mRNA decay. RNase protection experiments using total RNA of the indicated strains were performed as previously described (PULAK and ANDERSON 1993 ). Lane 1, wild-type strain N2; lane 2, unc-54(r259), a deletion that removes most of unc-54; lane 3, smg-7(r1131); unc-54(+); lane 4, unc-54(r514), an amber mutation at unc-54 codon 1263; lane 5, unc-54(r315); smg-7(r1131). unc-54 hybridization signals were quantified by phosphorimaging. To adjust for lane-to-lane variation in the amount of assayed RNA, unc-54 hybridization signals were measured relative to those of act-1 and normalized to the quantity of act-1 mRNA found in wild type (lane 1).

Isolating alleles of smg-7 in a noncomplementation screen:
All smg-7 alleles described above are temperature sensitive for suppression. At 25°, r1131, r1182, and r1197 are strong suppressors of unc-54(r293), while at 15° and 20° they are very weak suppressors. Our failure to isolate nonconditional alleles of smg-7 might indicate that, in the absence of SMG-7, NMD is an inherently temperature-sensitive process. Alternatively, the smg-7 alleles isolated above might be weak or altered-function alleles of a gene whose null phenotype is more severe, possibly even lethal. In order to isolate smg-7 alleles in a manner that does not require their viability when homozygous, we isolated additional smg-7 alleles in a "noncomplementation" screen. We knew that we could isolate null alleles of smg-7 heterozygous to smg-7(r1131), because nDf41 deletes smg-7 (see Figure 2) and smg-7(r1131)/nDf41 heterozygotes are viable, fertile, and Smg. We mutagenized smg-7(+) males with EMS and mated them with smg-7(r1131) hermaphrodites at 25°. Additional mutations were present in the parent strains to distinguish self-progeny from cross-progeny, to allow unc-54(r293) homozygous males to mate, to mark the r1131-containing chromosome, and to phenotypically distinguish Smg from non-Smg offspring (see MATERIALS AND METHODS). We isolated two smg-7 alleles in this screen among ~4400 mutagenized genomes. Both of these alleles, r1147 and r1148, are lethal when homozygous.

smg-7(r1147) fails to complement both smg-7(r1131) and the nearby mutation lin-45(sy96). smg-7(r1147) complements smg-3(r867), bli-6(sc16), and deb-1(st555). lin-45 encodes a member of the raf family of serine/threonine kinases, and lin-45 mutants exhibit a zygotic vulvaless and maternal-effect lethal phenotoype due to defects of ras-mediated signal transduction (HAN et al. 1993 ). Three lines of evidence indicate that lin-45 and smg-7 are distinct genes. First, smg-7 viable mutations do not exhibit a Lin-45 phenotype and fully complement lin-45(sy96) for all phenotypes. Second, lin-45(sy96), a strong loss-of-function allele (HAN et al. 1993 ), does not suppress unc-54(r293) or exhibit the morphogenetic phenotypes of smg mutants. Third, we crossed a transgenic array of clone raf-E, which contains lin-45(+) and rescues lin-45(sy96) (HAN et al. 1993 ), into smg-7(r1131); unc-54(r293) and determined that the lin-45(+) array does not rescue smg-7(r1131). Molecular analysis of smg-7 (see below) demonstrates that smg-7(r1147) is a deletion that removes all transcribed sequences of both smg-7 and lin-45.

Two lines of evidence suggest that smg-7(r1148), the second allele isolated in the noncomplementation screen, is a reciprocal translocation involving chromosomes IV and V. In the following discussion, we refer to smg-7(r1148) as rT1. First, rT1 exhibits linkage to markers on two different chromosomes (unc-24 IV, unc-70 V, and dpy-11 V; see MATERIALS AND METHODS). rT1 does not exhibit linkage to lev-11 I, unc-73 I, dpy-1 III, or dpy-18 III. Second, rT1 inhibits recombination on the right arm of LG IV and on the left arm of LG V. We placed rT1 heterozygous to various chromosome IV and V double-mutant chromosomes and measured the apparent frequency of crossing over. As shown in Table 2, rT1 reduces the frequency of crossing over on LG IV (right) and LG V (left), but not LG IV (left) or LG V (right). Such effects are consistent with rT1 being a reciprocal translocation having breakpoints on LG IV near smg-7 and on LG V between dpy-11 and unc-76. rT1 fails to complement both smg-7(r1131) and lin-45(sy96), suggesting that the LG IV breakpoint of rT1 disrupts both of these genes. Molecular analysis of smg-7 demonstrates that the breakpoint of rT1 is located within lin-45 and that all smg-7 sequences are deleted in rT1 (see below).

Transposon tagging of smg-7 and identifying smg-7 genomic and cDNA clones:
smg-7(r1131) was identified as a spontaneous mutation arising in a mut-5(st701) genetic background. As many mut-5-induced mutations are due to insertion of the transposon Tc1 (MORI et al. 1988 ), we investigated whether a Tc1 element is associated with smg-7(r1131). We mapped dimorphic Tc1 elements near r1131 using a collection of 45 crossovers that we isolated in the ~0.4 map unit intervals between bli-6 and unc-24 or between unc-44 and deb-1 (see Figure 2). Two Tc1 elements mapped near smg-7. rP14::Tc1 was separated from smg-7(r1131) by 4 of the 45 crossovers. rP14::Tc1 maps rightward of smg-7, between smg-7 and deb-1. A second Tc1 element was not separated from smg-7(r1131) and, as shown below, is located within smg-7 itself. We designated this element r1131::Tc1 and cloned one of its insertion junctions by inverse PCR using Tc1-specific primers (OCHMAN et al. 1988 ). Using this clone as a hybridization probe, we identified a phage lambda genomic clone of the wild-type region and subsequently subcloned a 1.6-kb EcoRI wild-type fragment into which Tc1 had inserted in smg-7(r1131).

To determine whether r1131::Tc1 was inserted into an expressed sequence, we identified eight cDNA clones that hybridized to the r1131::Tc1 junction fragment from a mixed-stage wild-type cDNA library (BARSTEAD and WATERSTON 1989 ). DNA sequencing revealed that seven of these cDNAs correspond to transcripts of lin-45 (HAN et al. 1993 ), while one (clone TR#274) defined a previously uncharacterized expressed sequence. As described below, further analysis of TR#274 demonstrated that it corresponds to smg-7. Comparing the sequence of TR#274 to genomic sequence of lin-45 (A. GOLDEN and P. STERNBERG, personal communication) indicated that smg-7 is located immediately downstream of and transcribed in the same relative orientation as lin-45 raf. To test whether lin-45 and smg-7 constitute an operon, and to potentially isolate a smg-7 cDNA clone complete to the 5' end, we amplified by RT-PCR the 5' end of smg-7 cDNA using an SL2-specific and a smg-7-specific primer pair, yielding cDNA clone TR#285. Comparing the sequence of TR#285 to lin-45 genomic sequence demonstrated that smg-7 is trans-spliced to SL2 at a 3' splice site 95 nt downstream of the lin-45 poly(A) addition site. We conclude that lin-45 and smg-7 constitute an operon, with smg-7 being a downstream gene and trans-spliced to SL2.

Transformation rescue of smg-7:
Transformation rescue experiments establish that cDNA clones TR#274 and TR#285 correspond to smg-7. We derived a full-length smg-7 cDNA clone by PCR-mediated ligation of TR#274 and TR#285, yielding plasmid clone TR#288. We cloned the insert of TR#288 into the C. elegans expression vector pPD49.78 (MELLO and FIRE 1995 ) downstream of the hsp16-2 promoter, which is inducible by heat shock (STRINGHAM et al. 1992 ). The resulting plasmid (TR#302) was microinjected together with plasmid pRF4, which carried a dominant allele of rol-6 as a marker for successful transformation (MELLO et al. 1991 ), into unc-54(r293) smg-7(r1131) hermaphrodites. Heritable transformants were identified by their roller phenotype and subsequently assessed for their Smg phenotype before and after heat shock. Non-Smg animals are expected to be Unc-54 (indicating absence of r293 suppression), whereas Smg animals are expected to have normal motility (HODGKIN et al. 1989 ). The results are shown in Table 3. Following heat shock and growth at 25°, hsp16-2::smg-7(+) transformants are indistinguishable from unc-54(r293), indicating efficient rescue of smg-7(r1131) (Table 3, line 3). In the absence of heat shock, transformants have normal motility, indicating absence of smg-7(+) expression. Control transformants demonstrate that heat shock of hsp16-2::smg-7(+) does not confer a dominant Unc-54 phenotype (Table 3, line 4). We conclude that cDNA clone TR#288 contains a complete smg-7 coding region.

SMG-7 is a novel acidic protein containing two probable tetratricopeptide repeats:
The DNA sequence and deduced amino acid sequence of cDNA clone TR#288 (GenBank accession no. AF089729) is shown in Figure 4. TR#288 is 1521 nt long, contains SL2 at its 5' end, and an oligo(A) tract at its 3' end. TR#288 hybridizes on Northern blots to a wild-type smg-7 mRNA that we estimate to be 1.6 kb long. Thus, we believe that TR#288 is full length, or very nearly so. TR#288 contains a single long open reading frame predicted to encode a 53,080-D protein. Two features of SMG-7 are noteworthy. First, the carboxyl terminus of SMG-7 is predicted to have a high net negative charge. One-third of the 75 carboxyl terminal amino acids are aspartic acid or glutamic acid residues, and the pI of SMG-7 is predicted to be 4.98. Second, SMG-7 contains two probable tetratricopeptide repeats (TPRs). TPRs form pairs of amphipathic -helices and have been shown to mediate numerous protein-protein interactions in both prokaryotes and eukaryotes (SIKORSKI et al. 1990 ; LAMB et al. 1995 ; DAS et al. 1998 ). While the primary sequence of TPRs is not highly conserved, several features characterize TPR motifs. They are 34 amino acids long and often arranged as tandem clusters. Positions 8, 20, and 27 are preferentially small or small hydrophobic residues, while certain other positions (especially 4, 11, and 24) tend to be bulky hydrophobic residues. Proline residues are often found near the end of helical regions, especially at position 32 near the end of helix B. The two TPR domains of SMG-7 (amino acids 130–163 and 164–197) are arranged in tandem and related (up to 38% identical, 62% similar) to TPR domains of a diverse set of proteins, the most similar of which are shown in Figure 5.

View larger version (80K): In this window In a new window Download PPT slide   Figure 4. Sequence of smg-7 cDNA clone TR#288. Sequences of the trans-spliced leader SL2 are underlined with dashes. The two TPR domains (amino acids 130–163 and 164–197) are boxed. The acidic carboxyl terminus is underlined. Arrowheads indicate the positions of smg-7 introns, and the site of Tc1 insertion in r1131 (nt 368) is indicated by an open triangle. Nucleotide 644 is changed from T to A (circled) in smg-7(r1197).

View larger version (32K): In this window In a new window Download PPT slide   Figure 5. Alignment of SMG-7 TPR domains. TPR1 and TPR2 of SMG-7 are similar to TPR domains of a variety of prokaryotic and eukaryotic proteins, the most similar of which are shown here. A TPR consensus was derived from the 42 TPR domains contained in the Pfam protein domain families database release 2.1 (SONNHAMMER et al. 1998 ) and represents a slight refinement of the previous consensus (LAMB et al. 1995 ). Positions of conservation within TPR domains are numbered above and below the alignments. Positions of identity between SMG-7 TPRs and the aligned or consensus TPRs are shaded. Helices A and B were defined by the crystal structure of TPR domains of protein phosphatase 5 (DAS et al. 1998 ). Aligned sequences are as follows: Borrelia burgdorferi CDC27 (FRASER et al. 1997 ), yeast PEX5 (VAN DER LEIJ et al. 1993 ), yeast SSN6 (SCHULTZ and CARLSON 1987 ), Methanobacterium thermoautotrophicum O-linked GlcNAc (DUBOIS et al. 1997 ), and Drosophila melanogaster PEX5 (GenBank accession no. 2832771).

The smg-7 null phenotype is a temperature-sensitive defect in NMD:
To investigate the nature of viable smg-7 mutations, we amplified genomic fragments of r1131, r1182, and r1197 by PCR and determined their sequences. These data defined the positions of smg-7 introns, which are noted in Figure 4. smg-7(r1131::Tc1) is a Tc1 insertion into a TA dinucleotide at position 368. smg-7(r1197) is a T A transversion at position 644, which changes a TTA (Leucine) codon to a TAA (Stop) codon. We did not detect a sequence change in smg-7(r1182), but we did not sequence the lin-45 to smg-7 intergenic region, mutation of which might affect smg-7 expression.

To confirm that smg-7(r1197) is a null allele and to identify SMG-7 on Western blots, we prepared affinity-purified polyclonal antibodies against an SMG-7 fusion protein expressed in bacteria (see MATERIALS AND METHODS). As shown in Figure 6, this antibody (designated AS74) detects a protein migrating at ~57 kD that is absent in smg-7(r1131, r1182, and r1197) at both permissive (20°) and restrictive (25°) temperatures. AS74 also cross-reacts with an ~59-kD protein that is unaffected by smg-7 mutations and is, therefore, not a product of the smg-7 gene. These results confirm that cDNA clone TR#288 represents smg-7 and establish that smg-7(r1131, r1182, and r1197) eliminate expression of SMG-7. The smg-7 null phenotype, therefore, is a temperature-sensitive defect in NMD.

View larger version (49K): In this window In a new window Download PPT slide   Figure 6. Western blot of SMG-7. Total proteins of wild types (lanes 1 and 5) were probed with anti-SMG-7 antibody AS74. SMG-7 (indicated by arrow) is absent in smg-7(r1131, R1182, and r1197) grown at both 20° and 25° (lanes 2–4 and 6–8). AS74 cross-reacts with a protein slightly larger than SMG-7 that is not affected by smg-7 mutations.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

C. elegans smg genes are required for nonsense-mediated mRNA decay (PULAK and ANDERSON 1993 ). Despite having isolated over 50 smg mutations in previous screens (HODGKIN et al. 1989 ), the distribution of alleles among genes suggested that not all smg genes had been identified. For example, only one allele of smg-4 and two alleles of smg-6 were previously isolated. Discovery of additional smg genes by continuing such screens is inefficient, because almost 90% of recovered mutations are alleles of smg-1, smg-2, or smg-5. We designed a modified screen for smg mutants that precludes isolating alleles of smg-1, smg-2, and smg-5. The screen is based on the fortuitous circumstance that smg-1, smg-2, and smg-5 reside on the left arm of chromosome I and on properties of szT1, a reciprocal translocation involving LG I and LG X. In animals heterozygous for both szT1 and fog-1(q180), recessive mutations on the left arm of LG I and the right arm of LG X are not recovered, because szT1 homozygotes are lethal and fog-1 homozygotes are sterile. smg mutations located outside of the balanced region can be isolated, but not those within the balanced region. As shown in Table 4, the modified screen effectively eliminates smg-1, smg-2, and smg-5 mutations from the recovered sample. While smg-1, smg-2, and smg-5 mutations constituted 87% of previously described alleles, none were isolated among 41 smg mutations described here.

We have used this screen to define smg-7, a previously unidentified smg gene. Like smg-1 through smg-6, smg-7 is required for rapid decay of nonsense mutant mRNAs (Figure 3). smg-7(r1131, r1182, and r1197) are all temperature sensitive, even though r1197 is an ochre nonsense mutation in the middle of the smg-7 open reading frame, and Western blots demonstrate that all three alleles express no detectable SMG-7. We conclude that, in the absence of SMG-7, NMD is an inherently temperature-sensitive process and that SMG-7 is required only for NMD at elevated temperature. The temperature-sensitive character of NMD in smg-7 mutants should prove useful for manipulating C. elegans gene expression. Expression of smg-sensitive mRNAs can be rendered conditional by incorporating smg-7 (or other) temperature-sensitive mutations into appropriate strains. Why were smg-7 mutations not recovered in previous screens? At 20°, the temperature at which this and previous suppressor screens were done, smg-7(r1131, r1182, and r1197) are weak suppressors. Their weak phenotype likely caused smg-7 alleles to be overlooked in previous screens and underrepresented in our modified screen.

SMG-7 is a member of the family of proteins that contain TPRs. TPRs are found in a variety of prokaryotic and eukaryotic proteins, including cell division cycle proteins, hsp90-binding immunophilins, transcription factors, and peroxisomal and mitochondrial import proteins (SIKORSKI et al. 1990 ; GOEBL and YANAGIDA 1991 ; LAMB et al. 1995 ). About 25 different proteins with TPR domains have been identified to date. Individual TPR domains are 34 amino acids long and have a loose consensus sequence. TPRs form paired amphipathic helices (DAS et al. 1998 ), and mutational studies suggest that TPRs mediate homo- and heteromeric protein-protein interactions. TPR domains may coordinate the assembly of multiprotein complexes, examples of which include mitochondrial and peroxisomal import receptor complexes (FRANSEN et al. 1995 ; NEUPERT 1997 ), anaphase promoting complex (LAMB et al. 1994 ; KING et al. 1995 ), and transcription repression complexes (SMITH et al. 1995 ; TZAMARIAS and STRUHL 1995 ). TPR interactions in such complexes can occur between two TPR domains, or between TPR and non-TPR regions.

The presence of TPRs within SMG-7 suggests that they are part of a protein complex. Indeed, recent work demonstrates that SMG-7 and SMG-5 interact with each other. Both SMG-5 and SMG-7 are coimmunoprecipitated with anti-SMG-5 or anti-SMG-7 antibodies, although it is unknown if this interaction is direct or indirect (K. ANDERS and P. ANDERSON, unpublished results). A complex of proteins involving Upf1p, Upf2p, Upf3p, and translation release factors eRF1 and eRF3 has been implicated in yeast NMD (HE and JACOBSON 1995 ; HE et al. 1997 ; CZAPLINSKI et al. 1998 ). Although the primary sequences of SMG-7 and Upf2p are not similar, both proteins have strongly acidic carboxyl termini (CUI et al. 1995 ; HE and JACOBSON 1995 ). The acidic carboxyl terminus of Upf2p has been shown to mediate direct interactions with Upf1p (HE and JACOBSON 1995 ; HE et al. 1997 ). Perhaps SMG-7 will prove to interact with SMG-2, which is homologous to Upf1p, via its acidic carboxyl terminus.

We have shown previously that activity of SMG-7 influences the phosphorylation status of SMG-2 (M. F. PAGE, B. CARR, K. R. ANDERS and P. ANDERSON, unpublished results). In smg-7 mutants, a phosphorylated isoform of SMG-2 accumulates to abnormally high levels. We note several TPR-containing proteins that, in other systems, influence the metabolism of certain phosphoproteins. For example, protein phosphatase 5 (PP5) contains three TPR domains. Protein phosphatase 2A (PP2A, which does not contain TPR domains) has been shown to interact with eRF1, which has in turn been shown to interact with Upf1p of yeast (ANDJELKOVIC et al. 1996 ; CZAPLINSKI et al. 1998 ). Perhaps phosphorylated SMG-2 is a substrate of PP5 or PP2A, and the TPR domain of SMG-7 targets the appropriate phosphatase to its SMG-2 substrate. According to this model, smg-7 mutants would be expected to accumulate phosphorylated SMG-2. P58^IPK is a TPR-containing inhibitor of the double-stranded, RNA-activated, protein kinase PKR (LEE et al. 1994 ). P58^IPK binds to PKR via one of its TPR domains and, in doing so, inhibits PKR activity (GALE et al. 1996 ). If SMG-7 were a negative regulator of the kinase that phosphorylates SMG-2 (analogous to the P58^IPK/PKR relationship), smg-7 mutants might express an overly active SMG-2 kinase, resulting in elevated levels of SMG-2 phosphorylation. Several subunits of anaphase-promoting complex are TPR-containing proteins (LAMB et al. 1994 ). Anaphase-promoting complex functions at specific times during the cell cycle to target certain proteins for ubiquitin-dependent proteolysis (KING et al. 1996 ). If SMG-7 were involved in targeting phosphorylated SMG-2 for proteolysis, a phosphorylated isoform of SMG-2 might accumulate in smg-7 mutants. Refinement of these or other explicit models of smg-7 function must await its further characterization, especially defining the protein(s) with which it interacts.

	FOOTNOTES

¹ Present address: Whitehead Institute for Biomedical Research, 9 Cambridge Center, Cambridge, MA 02142.

	ACKNOWLEDGMENTS

We thank A. Golden and P. Sternberg for providing lin-45 strains and clones, S. O'Connor for help in constructing TR2034, and Michelle Page for editing the manuscript. Strains used for mapping smg-7 were obtained from the Caenorhabditis Genetics Center, a resource supported by the National Institutes of Health (NIH) National Center for Research Resources. This work was supported by an NIH grant (GM-50933) to P.A., by a Howard Hughes Medical Institute Predoctoral Fellowship to B.M.C., and by the University of Wisconsin Training Grant in Genetics.

Manuscript received August 3, 1998; Accepted for publication October 26, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ANDERSON, P., 1995  Mutagenesis. Methods Cell Biol. 48:31-58[Medline].

ANDJELKOVIC, N., S. ZOLNIEROWICZ, C. VAN HOOF, J. GORIS, and B. A. HEMMINGS, 1996  The catalytic subunit of protein phosphatase 2A associates with the translation termination factor eRF1. EMBO J. 15:7156-7167[Medline].

APPLEQUIST, S. E., M. SELG, C. RAMAN, and H. M. JACK, 1997  Cloning and characterization of HUPF1, a human homolog of the Saccharomyces cerevisiae nonsense mRNA-reducing UPF1 protein. Nucleic Acids Res. 25:814-821[Abstract/Free Full Text].

BARNES, T. M. and J. HODGKIN, 1996  The tra-3 sex determination gene of Caenorhabditis elegans encodes a member of the calpain regulatory protease family. EMBO J. 15:4477-4484[Medline].

BARSTEAD, R. J. and R. WATERSTON, 1989  The basal component of the nematode dense body is vinculin. J. Biol. Chem. 264:10177-10185[Abstract/Free Full Text].

BARTON, M. K. and J. KIMBLE, 1990  fog-1, a regulatory gene required for specification of spermatogenesis in the germ line of Caenorhabditis elegans.. Genetics 125:29-39[Abstract].

BRENNER, S., 1974  The genetics of Caenorhabditis elegans.. Genetics 77:71-94[Abstract/Free Full Text].

CALI, B. M. and P. ANDERSON, 1998  mRNA surveillance mitigates dominance in C. elegans. Mol. Gen. Genet. 260:176-184[Medline].

CHRISTENSEN, S., V. KODOYIANNI, M. BOSENBERG, L. FRIEDMAN, and J. KIMBLE, 1996  lag-1, a gene required for lin-12 and glp-1 signaling in Caenorhabditis elegans, is homologous to human CBF1 and Drosophila Su(H).. Development 122:1373-1383[Abstract].

CUI, Y., K. W. HAGAN, S. ZHANG, and S. W. PELTZ, 1995  Identification and characterization of genes that are required for the accelerated degradation of mRNAs containing a premature translational termination codon. Genes Dev. 9:423-436[Abstract/Free Full Text].

CZAPLINSKI, K., M. J. RUIZ-ECHEVARRIA, S. V. PAUSHKIN, X. HAN, and Y. WENG ET AL, 1998  The surveillance complex interacts with the translation release factors to enhance termination and degrade aberrant mRNAs. Genes Dev. 12:1665-1677[Abstract/Free Full Text].

DAS, A. K., P. T. W. COHEN, and D. BARFORD, 1998  The structure of the tetratricopeptide repeats of protein phosphatase 5: implications for TPR-mediated protein-protein interactions. EMBO J. 17:1192-1199[Medline].

DE STASIO, E., C. LEPHOTO, L. AZUMA, C. HOLST, and D. STANISLAUS et al., 1998  Characterization of revertants of unc-93(e1500) in C. elegans induced by N-ethyl-N-nitrosourea. Genetics 147:597-608[Abstract].

DUBOIS, J., T. ALDREDGE, R. BASHIRZADEH, D. BLAKELY, and R. COOK et al., 1997  Complete genome sequence of Methanobacterium thermoautotrophicum delta H: functional analysis and comparative genomics. J. Bacteriol. 179:7135-7155[Abstract/Free Full Text].

EIDE, D. and P. ANDERSON, 1985a  The gene structures of spontaneous mutations affecting a Caenorhabditis elegans myosin heavy chain gene. Genetics 109:67-79[Abstract/Free Full Text].

EIDE, D. and P. ANDERSON, 1985b  Transposition of Tc1 in the nematode C. elegans.. Proc. Natl. Acad. Sci. USA 82:1756-1760[Abstract/Free Full Text].

EPSTEIN, H., R. WATERSTON, and S. BRENNER, 1974  A mutant affecting the heavy chain of myosin in Caenorhabditis elegans.. J. Mol. Biol. 90:291-300[Medline].

FODOR, A. and P. DEAK, 1985  The isolation and genetic analysis of a C. elegans translocation (szT1) strain bearing an X-chromosome balancer. J. Genet. 64:143-157.

FRANSEN, M., C. BREES, E. BAUMGART, J. C. VANHOOREN, and M. BAES et al., 1995  Identification and characterization of the putative human peroxisomal C-terminal targeting signal import receptor. J. Biol. Chem. 270:7731-7736[Abstract/Free Full Text].

FRASER, C. M., S. H. W. M. CASJENS, G. G. SUTTON, R. A. CLAYTON, and R. LATHIGRA et al., 1997  Genomic sequence of a Lyme disease spirochete, Borrelia burgdorferi.. Nature 390:580-586[Medline].

GALE, M., S.-L. TAN, M. WAMBACH, and M. G. KATZE, 1996  Interaction of the interferon-induced PKR protein kinase with inhibitory proteins P58^IPK and vaccinia virus K3L is mediated by unique domains: implications for kinase regulation. Mol. Cell. Biol. 16:4172-4181[Abstract].

GOEBL, M. and M. YANAGIDA, 1991  The TPR snap helix: a novel protein repeat motif from mitosis to transcription. Trends Biochem. Sci. 16:173-177[Medline].

HAN, M., A. GOLDEN, Y. HAN, and P. W. STERNBERG, 1993  C. elegans lin-45 raf gene participates in let-60 ras-stimulated vulval differentiation. Nature 363:133-140[Medline].

HE, F. and A. JACOBSON, 1995  Identification of a novel component of the nonsense-mediated mRNA decay pathway by use of an interacting protein screen. Genes Dev. 9:437-454[Abstract/Free Full Text].

HE, F., A. H. BROWN, and A. JACOBSON, 1997  Upf1p, Nmd2p, and Upf3p are interacting components of the yeast nonsense-mediated mRNA decay pathway. Mol. Cell. Biol. 17:1580-1594[Abstract].

HODGKIN, J., 1988 Genetics, pp. 491–584 in The Nematode Caenorhabditis elegans, edited by W. B. WOOD. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HODGKIN, J., 1997 Genetics, pp. 881–1047 in C. elegans II, edited by D. L. RIDDLE, T. BLUMENTHAL, B. J. MEYER and J. R. PRIESS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HODGKIN, J., A. PAPP, R. PULAK, V. AMBROS, and P. ANDERSON, 1989  A new kind of informational suppression in the nematode Caenorhabditis elegans.. Genetics 123:301-313[Abstract/Free Full Text].

JACOBSON, A. and S. W. PELTZ, 1996  Interrelationships of the pathways of mRNA decay and translation in eukaryotic cells. Annu. Rev. Biochem. 65:693-739[Medline].

KING, R. W., J. M. PETERS, S. TUGENDREICH, M. ROLFE, and P. HIETER et al., 1995  A 20S complex containing cdc27 and cdc16 catalyses the mitosis-specific conjugation of ubiquitin to cyclin B. Cell 81:279-288[Medline].

KING, R. W, R. J. DESHAIES, J. M. PETERS, and M. W. KIRSCHNER, 1996  How proteolysis drives the cell cycle. Science 274:1652-1659[Abstract/Free Full Text].

KUWABARA, P. E., P. G. OKKEMA, and J. KIMBLE, 1992  tra-2 encodes a membrane-protein and may mediate cell communication in the Caenorhabditis elegans sex determination pathway. Mol. Biol. Cell 3:461-473[Abstract].

LAMB, J. R., W. A. MICHAUD, R. S. SIKORSKI, and P. S. HIETER, 1994  Cdc16p, Cdc23p and Cdc27p form a complex essential for mitosis. EMBO J. 13:4321-4328[Medline].

LAMB, J. R., S. TUGENDREICH, and P. HIETER, 1995  Tetratricopeptide repeat interactions: to TPR or not to TPR? Trends Biochem. Sci. 20:257-259[Medline].

LEE, B. S. and M. R. CULBERTSON, 1995  Identification of an additional gene required for eukaryotic nonsense mRNA turnover. Proc. Natl. Acad. Sci. USA 92:10354-10358[Abstract/Free Full Text].

LEE, T. G., N. TANG, S. THOMPSON, J. MILLER, and M. G. KATZE, 1994  The 58,000-dalton cellular inhibitor of the interferon-induced double-stranded RNA-activated protein kinase (PKR) is a member of the TPR repeat family of proteins. Mol. Cell. Biol. 14:2331-2342[Abstract/Free Full Text].

LEEDS, P., S. W. PELTZ, A. JACOBSON, and M. R. CULBERTSON, 1991  The product of the yeast UPF1 gene is required for rapid turnover of mRNAs containing a premature translational termination codon. Genes Dev 5:2303-2314[Abstract/Free Full Text].

LEEDS, P., J. M. WOOD, B. S. LEE, and M. R. CULBERTSON, 1992  Gene products that promote mRNA turnover in Saccharomyces cerevisiae.. Mol. Cell. Biol. 12:2165-2177[Abstract/Free Full Text].

LOSSON, R. and F. LACROUTE, 1979  Interference of nonsense mutations with eukaryotic messenger RNA stability. Proc. Natl. Acad. Sci. USA 76:5134-5137[Abstract/Free Full Text].

MAQUAT, L. E., 1995  When cells stop making sense: effects of nonsense codons on RNA metabolism in vertebrate cells. RNA 1:453-465[Abstract].

MELLO, C., and A. FIRE, 1995 DNA transformation, pp. 451–482 in Caenor