C. elegans dss-1 is functionally conserved and required for oogenesis and larval growth

Background Dss1 (or Rpn15) is a recently identified subunit of the 26S proteasome regulatory particle. In addition to its function in the protein degradation machinery, it has been linked to BRCA2 (breast cancer susceptibility gene 2 product) and homologous DNA recombination, mRNA export, and exocytosis. While the fungal orthologues of Dss1 are not essential for viability, the significance of Dss1 in metazoans has remained unknown due to a lack of knockout animal models. Results In the current study deletion of dss-1 was studied in Caenorhabditis elegans with a dss-1 loss-of-function mutant and dss-1 directed RNAi. The analysis revealed an essential role for dss-1 in oogenesis. In addition, dss-1 RNAi caused embryonic lethality and larval arrest, presumably due to loss of the dss-1 mRNA maternal contribution. DSS-1::GFP fusion protein localised primarily in the nucleus. No apparent effect on proteasome function was found in dss-1 RNAi treated worms. However, expression of the C. elegans dss-1 in yeast cells deleted for its orthologue SEM1 rescued their temperature-sensitive growth phenotype, and partially rescued the accumulation of polyubiquitinated proteins in these cells. Conclusion The first knockout animal model for the gene encoding the proteasome subunit DSS-1/Rpn15/Sem1 is characterised in this study. In contrast to unicellular eukaryotes, the C. elegans dss-1 encodes an essential protein, which is required for embryogenesis, larval growth, and oogenesis, and which is functionally conserved with its yeast and human homologues.


Background
Dss1 (or Rpn15) is a small (70-amino-acids in humans), conserved protein that has been implicated in several cellular functions. Although it was originally isolated as a candidate gene for the split-hand-split-foot syndrome (DSS1, deleted in the split hand/split foot SHFM1) [1], the connection with the syndrome has not been confirmed [2,3]. Instead, it has emerged that Dss1 is a subunit of the 26S proteasome [4][5][6][7]. The proteasome is a multi-protein complex responsible for the degradation of polyubiquitinated target proteins. It consists of a 20S core, responsible for the degradation of proteins inside the core channel, and a 19S regulatory particle, containing about 20 proteins arranged into a base and lid structure. The 19S regulatory particle binds ubiquitin-tagged substrates, unfolds the substrates and transfer them into the proteolytic core, but the precise role of many of the subunits is not yet known [8,9]. Within the cell, protein degradation must be regulated with precision. Emerging data indicates that the number of proteasomes and their subunit composition can vary depending on cellular conditions [9]. Various proteasome-interacting proteins have been discovered, and post-translational modifications of proteasome subunits may have an effect on proteolytic activity. Alternative regulatory particles to 19S exist that possibly function in ubiquitin-independent proteolysis, and the proteasome has also been implicated in tasks separate from protein degradation, e.g. transcription [10,11].
Dss1 binds to the 19S regulatory particle both in mammalian cells and in fission and budding yeast [4][5][6][7]. This binding is possibly mediated by lid components Rpn3 and Rpn7, with separate binding sites in Dss1 for both proteins [12,13]. Genetic and biochemical data from yeast suggest that Dss1 is involved in the proteolytic function of the proteasome. Accumulation of polyubiquitinated proteins occurs in sem1 mutants, the Dss1 homologue in Saccharomyces cerevisiae. Synergistic interactions are seen between double mutants of sem1 and a lid subunit, rpn10 [4][5][6]. In addition to a role in protein degradation, Dss1 is involved in DNA repair. It binds to BRCA2, a breast cancer susceptibility gene product and a component of the homologous recombination machinery [14,15]. Loss of either BRCA2 or Dss1, both in mammalian and fungal cells, results in defects in homologous recombination [16,17]. Conflicting evidence exists whether the effect of Dss1 is mediated by regulation of BRCA2 stability [17,18]. It is possible though that the Dss1 effect on homologous recombination is dependent on its role in the proteasome complex as proteasomes have been shown to bind to double-stranded breaks, sites of homologous recombination, and specifically BRCA2 binds the proteasome lid components Rpn3 and Rpn7, similar to Dss1 [5,13]. Moreover, proteasome activity is required for DNA repair. However, Dss1 does not regulate BRCA2 affinity for the proteasome, as the BRCA2-Rpn7 interaction is not dependent on Dss1 [13]. Dss1 has also been shown to be required for mRNA export in fission yeast [19]. Finally, both SEM1 overexpression and deletion in S. cerevisiae can suppress exocyst mutations, suggesting that Dss1 is directly or indirectly involved in protein secretion [20]. Sem1 in yeast is dispensable at 24°C, the normal growth temperature of yeast, despite the accumulation of polyubiquitinated proteins. However, SEM1 deleted cells are temperature-sensitive ceasing to grow at elevated temperatures [14]. In addition, loss of Sem1 triggers or enhances a cell differentiation process, pseudohyphal growth, in diploid S. cerevisiae cells [20]. In higher eukaryotes, there is currently no in vivo data about the significance of Dss1. Attempts to identify mutations in the DSS1 coding region of a selected set of breast cancer patients did not support a direct role for Dss1 in tumorigenesis [21]. Here we present the first characterization of dss-1 in a multicellular organism, the nematode Caenorhabditis elegans. We show that dss-1 is functionally conserved from fungi to metazoans, and required for oogenesis and intestinal function. In addition, dss-1 is important for embryogenesis and larval growth. Taken together, the results show, for the first time, that dss-1 is essential for the development of metazoan animals.

C. elegans dss-1 is expressed throughout development and encodes a nuclear protein
The C. elegans dss-1 (Y119D3B.15) encodes a 82 aa protein that is 55% similar and 34% identical to the product of the S. cerevisiae SEM1 gene, and 71% and 46%, respectively, to the human DSS1 gene product [20]. We examined the expression of the dss-1 transcript and DSS-1::GFP fusion protein in worm tissues. It has previously been shown that human and mouse dss1 are expressed ubiquitously in several adult and fetal tissues [1]. Similarly, the C. elegans dss-1 transcript was present throughout larval development and in adult worms ( Figure 1A). To observe the cellular localisation of DSS-1, transgenic worms were generated expressing DSS-1::GFP fusion protein from 4 kb of dss-1 promoter sequence. This fusion protein localised to the intestinal epithelium and to unidentified neuronal cells in the head region (data not shown). No expression was seen in the gonad, presumably because of germ line silencing of transgene expression. Subcellularly the strongest expression was seen in the nucleus with diffuse staining in the cytoplasm ( Figure 1B). This is similar to the reported Dss1 localisation in cultured human and yeast cells [14,22].

dss-1 is required for fertility
To study the significance of dss-1 we examined the phenotype of dss-1 mutant worms. A deletion allele for dss-1 was obtained from the National Bioresource Project, Japan. dss-1(tm370) is a loss-of-function allele with a 1466 bp deletion completely removing exon 2 ( Figure 2A). The dss-1(tm370) homozygous worms grew up into adults but produced no offspring ( Figure 2B). We tested whether the lack of progeny were due to a defect in sperm formation by mating dss-1 mutant worms with N2 males. No offspring were obtained from mutant worms indicating that wild type sperm cannot rescue the sterility of dss-1(tm370) (n = 7). The sterility is not due to compromised gonad development as a two-arm gonad with developing germ cells was present in 98% of mutant worms examined (n = 169). Gonad development was further studied by crossing the mutants to the lag2::GFP marker strain. This marker expresses GFP in the distal tip cells (DTC) of the gonad [23]. Comparable numbers of dss-1(tm370) homo-and heterozygous worms showed a GFP signal in both distal tip cells (data not shown). However, the formation of oocytes was defective in the dss-1(tm370) homozygous worms ( Figure 2C-E). Only 23% of 3-day-old mutant worms had any oocytes (n = 30) in contrast to 100% of the heterozygous worms of similar age containing oocytes (n = 18). The difference was partially due to a delay in oogenesis, as the percentage of mutant worms with oocytes increased to 50% when the worms were examined at 3.5 days (n = 72), and to 73% at 4-6 days after egg-laying (n = 67). However, the oocyte formation was abnormal as also in the older mutant worms the number of oocytes was always less than in control worms, with an average of two oocytes per gonad arm ( Figure 2H). The oocytes were generally smaller and aberrant-looking. Occasionally embryo-like multicellular structures were visible in the uterus (25% in 3.5-day-old and 57% in 4-6day-old worms) ( Figure 2F and 2G). Their number per gonad arm was low, and no eggs were laid by the dss-1(tm370) mutants. Sperm were present in the spermatheca of the mutant worms in numbers comparable to the heterozygous control worms (83% of dss-1(tm370)/dss-1(tm370) mutants had sperm versus 94% dss-1/+ in 3-day-old worms). This suggested that spermatogenesis was superficially normal. However, some delay in sperm formation could be witnessed as spermatocytes were still found in 35% of 3-day-old dss-1 mutants, but in none of the control worms ( Figure 2I and 2I'). In 3.5-day-old mutants spermatocytes were no longer seen.
The growth rate of the dss-1 mutants was examined by counting the progeny of individual dss-1/sC1 heterozygotes grown for 4 days (Table 1) (the sC1 balancer carrying the dpy-1(s2170) mutation was used to balance the dss-1(tm370) mutation. Homozygous sC1 worms were identified by the dumpy phenotype). No embryonic or larval lethality was evident. The expected ratios of fertile (dss-1/sC1, 48% vs. expected 50%) and dumpy adults (sC1/sC1, 26% vs. expected 25%) were observed. The remaining dss-1(tm370)/dss-1(tm370) worms were classified as either sterile or young adults indicating that the mutant worms seemed to grow at comparable rates to their heterozygous siblings. When grown at 25°C or 26°C, the results were similar except that the adult mutant worms were smaller than mutants grown at lower temperatures (data not shown). dss-1(tm370) homozygotes grown at room temperature were thinner than dss-1/+ heterozygotes, with often a slightly protruding vulva ( Figure  2B). The lumen of their intestine was generally (82%, n = 138) expanded ( Figure 2J and 2K) and filled with undigested bacteria giving the worms a somewhat shiny appearance. This was often accompanied by a twisted pharynx ( Figure 2L). In dss-1/+ heterozygotes the intestinal lumen was also expanded in 36% of the worms examined (n = 78), but to a much milder degree than in the homozygotes. The lifespan of homozygous dss-1 mutants was shorter than that of heterozygous or wild type N2 controls (Table 2). At 25°C the earlier death of the mutants was slightly more pronounced.

dss-1 is required for embryogenesis and larval growth
RNAi against the dss-1 transcript was performed with bacterial feeding by placing N2 L4 larva onto plates. Phenotypes observed in the dss-1 RNAi F1 progeny mimicked the dss-1(tm370) phenotypes in a temperature-sensitive manner. When RNAi was performed at 15°C and 20°C no phenotype was evident. At temperatures between 24.5°C and 26°C sterile worms were found so that at lower temperatures only a fraction of the dss-1 RNAi treated worms were sterile (data not shown), whereas at 26°C no offspring were produced. At the same temperatures, but with vector-treated control RNAi, the worms were fertile (Figure 3A). Similar to the dss-1(tm370) mutants, oogenesis was defective with only 18% (n = 40) of the worms having oocytes ( Figure 3B-D). Very few embryo-like structures were found in the uterus ( Figure 3E). Instead, 32% of the gonads contained abnormal vacuolar structures ( Figure  3F). DAPI staining revealed emo-like cells in the gonad    Figure 3G and 3G') and uterus. The emo phenotype is caused by endomitotic DNA replication resulting in polyploidy, and is identified by strongly DAPI positive nuclei. In order to follow gonadal development, RNAi was also performed in a lag2::GFP strain. Two DTC cells were found in equivalent numbers in both control (88%, n = 25) and dss-1 (88%, n = 65) RNAi worms ( Figure 3H and 3I).
In addition to mimicking the dss-1(tm370) mutant phenotype, dss-1 RNAi treated worms had additional phenotypes, most likely due to depletion of the maternal transcript. At 26°C a major portion of the RNAi worms showed embryonic lethality and larval arrest ( Figure 3A). Some of the worms arrested at L1/L2 stages and some later at L3/L4. 5-10% of the worms had deformed epithelia, either in the tail or the head ( Figure 3J and 3K). These worms were usually arrested at L1 or L2. Those dss-1 RNAi worms that survived until L3 rarely had a deformed head or tail, and no dss-1 RNAi adult survivors had epithelial defects. Similar to the dss-1(tm370) mutants 40% of the surviving RNAi worms had a bloated intestine with undigested bacteria ( Figure 3L).
We examined Wormbase [25] for mutants with similar defects in oogenesis as in dss-1 mutants, and looked for proteins that might play a role in similar processes as DSS- Young adult N2 and dss-1 homo-and heterozygous worms of equivalent age were placed at either RT or 25°C, and transferred onto fresh places when required. The number of survivors was followed for four days. The results are a representative of two (RT) and three (25°C) independent experiments. The progeny of three individual dss-1(tm370)/sC1 worms were followed for four days. sC1 balancer carries the recessive dumpy-1 mutation. The results are a representative of three independent experiments.   with dss-1 mutants for two reasons. First, the kgb-1 mutant phenotype has many similarities with the dss-1 homozygotes. The mutants are sterile, with no mature oocytes or embryos, and, in older worms, exhibit a protruding vulva and bloated appearance [26]. This sterility is temperaturesensitive at 26°C similar to the dss-1 RNAi phenotype. Secondly, KGB-1, a MAP kinase, is believed to act in proteolytic degradation of an oogenesis-specific protein. It binds GLH-1, an RNA helicase required for fertility [26], and in kgb-1 mutants GLH-1 protein levels are elevated [27]. Biochemical evidence indicates that KGB-1 regulates GLH-1 degradation by phosphorylation [27]. We postulated that DSS-1 and KGB-1 might both be regulatory factors required for the proteolysis of GLH-1 and/or other proteins required for oogenesis. kgb-1(um3) uteri were generally void of embryos (compare Figure 4E to Figure  2D). The gonad arms were disorganized with no wellformed rachis (data not shown). Occasionally the kgb-1 proximal gonad contained a mixture of more differentiated oocytes surrounded by less developed germ cells (Figure 4F). Although the dss-1 mutant gonads were in general more organized, upon closer inspection we found some similar mixtures of more mature oocytes surrounded by less mature germ cells ( Figure 4G). A prominent feature of the kgb-1 mutant gonads is the emo phenotype ( Figure  4H). In younger dss-1 mutants no DAPI-positive emo cells were visible either in the gonad or the uterus (n = 33). However, in older dss-1 mutants 20% of the gonads had emo-like cells (n = 24) ( Figure 4I). In addition, in 71% of the dss-1 mutants DAPI-positive emo-like structures were found in the uterus, possibly resulting from unfertilised embryos ( Figure 4J). In kgb-1 mutants 95% had emo cells in the gonad (n = 20), but only 20% in the uterus.

dss-1 RNAi results in sterility and growth defects
Examination of GLH-1 protein levels by Western analysis showed no significant differences in the amount of GLH-1 between dss-1 homozygotes, dss-1/sC1 heterozygotes, and N2 control worms (data not shown). This indicates that DSS-1 is not directly involved in the regulation of GLH-1 turnover.

Proteasomal activity is unaffected by dss-1 RNAi
Most of the proteasome subunits are necessary for development, and their depletion by RNAi in C. elegans results in early lethality [28]. One exception is RPN-10, a component of the proteasome RP lid structure, which is not required for viability in yeast or C. elegans [29,30]. Instead, rpn-10 null worms are sterile at 25°C implying similarity to dss-1(tm370) mutants [30]. However, unlike dss-1 mutants, the main cause of sterility for rpn-10 mutants is the absence of sperm formation, while oocytes seem to form normally ( Figure 4K-M and [30]).
In yeast RPN10 shows synthetic interactions with the sem1/dss1 deletion indicating involvement of both pro-teins in the same cellular process, protein degradation [4][5][6]. We tested whether similar genetic interactions between dss-1 and rpn-10 occur in C. elegans. With N2 or rrf-3 worms no effect was seen when RNAi against dss-1 was performed at 20°C. Similarly no phenotype was obtained either when dss-1 RNAi was performed at 20°C in an rpn-10(tm1349) background (data not shown), indicating a lack of a synthetic interaction. The experiment could not be done at higher temperatures due to the temperaturesensitive sterility of rpn-10 worms.
To examine whether DSS1 affects the proteolytic activity of the proteasome, the chymotrypsin-like activity of the proteasome was measured in whole worm extracts of dss-1 RNAi worms. Similar levels of activity were detected in extracts from dss-1 RNAi or control vector-treated worms. When the chymotrypsin-like activity was assigned as 100% (standard deviation 3.77 and 2.68 in two separate experiments) in extracts from control vector-treated worms, the activity in the dss-1 RNAi treated samples was 111.0% +/-3.76 and 108.4% +/-13.65, respectively. Accordingly, Western blot analysis revealed no difference in the levels of proteasome subunits in the dss-1 RNAi and control vector-treated worms (data not shown).

C. elegans dss-1 can rescue SEM1 deletion defects in yeast
In order to further study the function of DSS-1, we made use of S. cerevisiae cells deleted for the dss-1 orthologue SEM1. These cells are temperature-sensitive for growth [14], and accumulate polyubiquitinated proteins [4,6]. Yeast cells deleted for SEM1 were transformed with an empty control plasmid or plasmids containing either dss-1 or the yeast SEM1. Transformants were tested for growth at elevated temperatures and for their ability to process polyubiquitinated proteins. The temperature-sensitive phenotype caused by the deletion of SEM1 was rescued both by the C. elegans dss-1 and by SEM1 ( Figure 5A). At the same time a partial, but consistent rescue by the C. elegans dss-1 of accumulation of polyubiquitinated proteins was observed in SEM1 deleted cells ( Figure 5B). Quantification showed that on average reintroduction of SEM1 to sem1∆ cells rescued 75-80% of the accumulated proteins whereas rescue by dss-1 was 50% after 3 h of incubation and 20% after 5 h (data not shown). Taken together, the growth phenotype and rescue from polyubiquitinated protein accumulation suggest that DSS-1 protein is functionally conserved.

Discussion
Our analysis of the loss-of-function phenotype of C. elegans dss-1 shows that dss-1 is essential for the development of a multicellular animal. In dss-1(tm370) mutant animals oogenesis is severely affected, and food absorption of the intestinal membrane appears malfunctional. Depletion of the dss-1 transcript by RNAi results in increased embryonic lethality and larval growth arrest, indicating the importance of dss-1 also at early stages of nematode development.

dss-1 and fertility
The germ line development in dss-1(tm370) animals was slightly delayed. At three days after egg-laying, control dss-1 heterozygotes had both mature oocytes and sperm, whereas dss-1 homozygotes still had spermatocytes and relatively few animals had oocytes at all. Approximately 12 h later sperm had matured in the majority of the mutants, and the proportion of adult worms with oocytes continued to increase. This delay could be intrinsic to the germ line itself. Alternatively, the sterility could be a sec-ondary consequence caused by the general sickness of the worms, possibly mediated by the intestine. For example, lipoproteins secreted by the intestine are known to affect germ line development [31,32]. The 4 kb dss-1 promoter drives expression mainly in the intestine ( Figure 1B). We attempted to rescue the dss-1 mutant with the DSS-1::GFP fusion protein expressed under the 4 kb promoter but were not successful (data not shown). Since transgenic expression is often silenced in the germ line [33] this suggests though that DSS-1 expression solely in the intestine is not sufficient for rescuing dss-1 oogenesis defects, and that its expression in the germ line may be required for oogenesis.
Although oocyte formation seemed to increase with aging of the dss-1 mutant worms it remained defective. Normal numbers of oocytes were not achieved, and the formed oocytes were not capable of fertilization, even with wild type sperm. C. elegans BRC-2 is required for proper progression of meiosis, and brc-2 mutants have abnormal diakinesis (chromosome condensation) [24]. Similarly, abnormal chromosome condensation was found occasionally in dss-1 mutants. However, abnormal diakinesis does not explain completely the oogenesis defects seen in dss-1 mutant worms as in brc-2 mutants oocytes progress through fertilization producing embryos that die, something not occurring to the same degree in dss-1 mutants. Thus, DSS-1 is likely to have a role in oogenesis independent from its function with BRC-2.
kgb-1 and dss-1 mutants have strongly overlapping phenotypes [26] and Figure 4), and both proteins have been linked to the protein degradation machinery. In kgb-1 mutants the increased levels of the GLH-1 RNA helicase appear to result in a defective germ line [27]. In dss-1 mutants GLH-1 levels were normal suggesting that kgb-1 and dss-1 are not involved in the same degradation process. It is possible that the oogenesis defects in dss-1 mutants might be caused by the aberrant regulation of degradation of another essential protein required for oogenesis, in combination with errors in homologous recombination.

dss-1 is essential for growth and survival
RNAi against dss-1 causes growth defects and embryonic lethality not seen in dss-1 mutants. This could be caused by targeting of secondary mRNAs, homologous to dss-1.
We find this explanation unlikely as no homologous targets have been identified either by Wormbase [25] or by our own attempts. An alternative possibility is that the progeny obtained from dss-1 RNAi treated worms are escapers from the oogenesis defect, and that the embryonic lethality and larval arrest observed are caused indirectly by partially defective oogenesis. However, as the parental generation of RNAi treated worms showed no C. elegans dss-1 can partially rescue yeast sem1 mutation Figure 5 C. elegans dss-1 can partially rescue yeast sem1 mutation. A. S. cerevisiae sem1 mutants have a growth defect at restrictive temperature (vector). Expression of either C. elegans dss-1 or the yeast SEM1 in sem1 mutants rescues the growth defect. B. Accumulation of polyubiquitinated proteins is seen in sem1 mutants transformed with an empty vector both at permissive and restrictive temperature at two different time points. Expression of yeast SEM1 in sem1 mutant cells reduces efficiently the accumulation of polyubiquitinated proteins, whereas expression of C. elegans dss-1 has only a modest effect. Lower panels have been blotted with anti-Sso1/2 antibody as a control for equal loading. signs of sterility and laid eggs comparably to the control worms ( Figure 3A), we do think this is a probable option. Therefore, we believe that the depletion of the maternal transcript provided by the heterozygous mothers of dss-1 mutants results in the early defects seen in the RNAi treated worms. This suggests that the dss-1 transcript is an essential component of embryogenesis and larval development. It is feasible that these effects are mediated by the involvement of DSS-1 in proteasome function. The proteasome is required for development, as RNAi against the majority of the C. elegans proteasome subunits has a strong, early lethal effect [28].
Accumulation of bacteria in the intestine is seen in the dss-1 mutants and dss-1 RNAi treated worms. This could be caused by defective absorption by the intestinal membrane or alternatively, by problems in secretion of digestive enzymes. The fact that the yeast homologue of DSS-1, Sem1, has also been linked functionally to exocytosis suggests that DSS-1 participates directly or indirectly in secretion [20].
The dss-1 mutant worms have a slightly reduced lifespan. In addition, stress such as heat appears to enhance the phenotypic effects seen in dss-1 mutants or RNAi worms. This is in line with the role of DSS-1 in protein degradation since the proteasome function has previously been linked to aging and stress [9,34].

The cellular function of DSS-1
Dss1 is known to participate in the proteasome complex, and either directly or indirectly via proteolytic activity in BRCA2 function and DNA repair [4][5][6][7]14,15]. In addition, it has been implicated in mRNA export and regulation of exocytosis [19,20]. What does the analysis of the loss-offunction phenotype of dss-1 in C. elegans tell us of its function in a multicellular organism?
The proteasome is an essential component of the housekeeping machinery inside a cell, and perturbing its function generally results in lethality [28,35]. As such, the significant portion of embryonic lethality and larval growth defects seen in dss-1 RNAi treated worms are indicative of an essential function. Whether this function relates to the proteasome is uncertain. We did not observe an apparent effect on proteasome activity or on the amount of proteasome subunits in whole worm extracts when dss-1 was knocked down by RNAi. In addition, dss-1 homozygous mutants survive into adulthood with only specific defects. The homozygous dss-1 mutant embryos are likely to be rescued by expression of the maternal dss-1 transcript. The lack of a lethal phenotype in the adult homozygotes can indicate either that DSS-1 is not required for proteasome activity in adults, or that the requirement of DSS-1 for proteasome activity is restricted to specific tissues or specific subsets of proteins. Indeed, recent work in cultured mammalian cells showed that depletion of dss-1 mRNA by RNAi has a minimal effect on proteolytic activity [36]. Taken into account that the dss-1 depletion in their experiment was not complete, it is possible that in metazoans Dss1 is not necessary for general proteasome activity. In order to resolve these issues, methods to measure local or substrate-specific proteasome activity in the dss-1 mutant worms are required.
The fertility defects seen in the dss-1 mutants differed from the defects seen when rpn-10, another proteasome subunit, was mutated [30]. Either, DSS-1 and RPN-10 have independent tasks in the proteasome complex, or, alternatively, RPN-10 acts redundantly with another subunit, thus explaining the mild oogenesis defects seen in rpn-10 mutants. Indeed, RNAi against rpn-10 has no effect, whereas combined RNAi against rpn-10 and rpn-12 is lethal [28]. We did not observe any synthetic effect between rpn-10 and dss-1 suggesting that RPN-10 and DSS-1 are not required for the same molecular process.
The function of dss-1 has been at least partially conserved during evolution, as we have shown that both the human and nematode homologues can complement the growth and differentiation phenotypes of the yeast sem1 mutation ( Figure 5, [20]). However, the C. elegans dss-1 can not fully rescue the accumulation of polyubiquitinated proteins in sem1 mutant yeast cells. sem1 mutant cells accumulate polyubiquitinated proteins already at the permissive temperature but their growth compared to SEM1 expressing cells is indistinguishable. It is possible that the homology between DSS-1 and the yeast Sem1 protein is only sufficient for partial functionality in proteasome activity, but that even this incomplete activity is enough to sufficiently reduce the amount of polyubiquitinated proteins to allow growth at the restrictive temperature. Alternatively, the temperature-sensitive growth defect could result from a proteasome-independent function of Sem1 that is not conserved between yeast and nematode. We do not think that this function would be related to homologous recombination. First, in yeast no BRCA2 homologue has been identified. Second, loss of SEM1 in S. cerevisiae does not cause increased sensitivity to UV-irradiation or to chemicals that induce doublestranded breaks, or cause significant differences in the capability of these cells to repair gapped plasmids ( [5], J. Jäntti, unpublished). It is thus likely that the growth complementation in yeast is based on functions of DSS-1 in processes other than homologous recombination.

Conclusion
C. elegans dss-1 is functionally conserved with its yeast and human homologues. Although no direct effect on the proteasome function was observed in the dss-1 knockdown worms, the efficient growth rescue and the partial rescue of accumulated polyubiquitinated proteins in yeast suggest that C. elegans DSS-1 is likely to play a role in proteasome function. It is possible that this role is spatially or temporally restricted. In summary, in contrast to unicellular organisms, dss-1 in a multicellular animal is an essential gene, which is required for oogenesis, embryogenesis, and larval growth.

Genetics
Worms were grown at room temperature on NGM plates unless otherwise indicated [37].
The following worm strains used: dss-1(tm370) allele was outcrossed once when balanced against sC1. Three types of worms were obtained from the balanced strain: sterile dss-1 homozygotes, fertile dss-1/ sC1 heterozygotes, and fertile dumpy sC1 homozygotes. For crosses of dss-1(tm370) mutants with wild type males, L3 or L4 non-dumpy larvae from the progeny of tm370/sCi heterozygote worms were mated on individual plates with three N2 males each. The 52 larvae tested were either hetero-or homozygous for dss-1. After scoring for the presence of progeny, the parent hermaphrodite was genotyped by PCR.

RNAi
For dss-1 RNAi construct, dss-1 cDNA was cloned from yk115f10 (Yuri Kohara) into Pst I/Xho I sites of L4440 (Andrew Fire). RNAi was performed by bacterial feeding [38] either in Bristol N2, rrf-3(pk1426), or rpn-10(tm1349) strains. Both L4 and L1 larvae were used as starting stages for the RNAi experiments. RNAi starting with L1 larvae were tested at different temperatures, but no phenotypes were observed. Phenotypes were obtained using N2 L4 larvae at 24.5°C -26°C. At 26°C the RNAi phenotype was most severe (without the control worms being affected) so this temperature was chosen for phenotypic analysis. L4 larvae were placed on bacterial RNAi plates and allowed to lay eggs for 24 h after which they were transferred to a fresh set of RNAi plates. After an additional 24 h the worms were removed from RNAi plates. The progeny from both plate sets was followed for three to five days.

DAPI staining
As the dss-1 mutant gonads were very fragile, either disintegrating or remaining inside the worm carcass upon dissection, DAPI staining was performed on whole animals. Worms were fixed in 3% PFA and postfixed in methanol. DAPI (Molecular Probes, USA) was used at 0.5 µg/ml.

Proteasome activity assay
A fluorogenic peptide substrate assay was performed to determine the chymotrypsin-like peptidase activity of the proteasome in whole worm extracts. Whole worm extracts were prepared from frozen worm pellets resuspended in lysis buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 5 mM EDTA, and 2 mM DTT) and subjected to sonication and centrifugation. 25 µg of whole animal extracts was incubated with 140 µM Suc-LLVY-AMC (Calbiochem) in proteasome activity assay buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 5 mM EDTA, and 5 mM ATP) at room temperature.
The fluorescence intensity was measured in triplicate samples at 380 nm excitatory and 455 nm emission wavelengths using a FLUOstar Optima Mircoplate Reader (BMG Labtech). The assay was performed in the absence and presence of the proteasome inhibitor MG132 (10 µM, Peptides Interrnational) to calculate the proteasome-specific activity, after which the L4440 sample was assigned an activity value of 100%. The results are from two independent RNAi experiments.

Western blotting
For GLH-1 immunoblotting 50 adult worms of each genotype were lysed in 25 µl of 1× Laemmli sample buffer. The samples were run on a 12% SDS-PAGE gel and transferred onto a nitrocellulose membrane. Rabbit anti-GLH-1 antibody (kind gift from Susan Strome) was used at a 1:1000 dilution. The membrane was reblotted with mouse anti-α-tubulin (Sigma T5168, dilution 1:1000) to check for equal loading. For detection of proteasome subunits in RNAi-treated worms, extracts were prepared from frozen worm pellets by sonication in lysis buffer supplemented with 0.5 mM NEM, 10 µM MG132 and complete protease inhibitor cocktail (Roche Diagnostics). Equal amounts of proteins were subjected to Western blotting with antibodies against proteasome α-subunits (BIOMOL International, PW8195) and α-tubulin.

Yeast sem1 mutant rescue
Yeast strain H2187 (MATa ura3 sem1::kanMX) was transformed with the empty yeast expression plasmid pVT102U or pVT102U containing dss-1 or sem1 cDNA. Transformants were patched on SCD-ura plates, grown overnight at 24°C and replicated using velvet to different temperatures. The growth of the patches was followed for three days. For analysis of accumulation of polyubiquitinated proteins, the above transformants were grown overnight in SCD-ura at 24°C. The cultures were diluted to OD 600 0.2 and regrown to OD 600 0.6-1.0. The cultures were divided in two and the growth was continued either at 24°C or at 37°C. Samples were collected at 1.5, 3 and 5 hours after the temperature shift. Cells were lysed in 2% SDS, the total protein concentration was measured (BCA Protein Assay Kit, Pierce) and equal amounts of the proteins were subjected to SDS-PAGE and Western blotting with an anti-polyubiquitin antibody (BIOMOL International, PW8805) and anti-Sso1/2 antibody [40]. Quantification of the band intensities was done by scanning the films with GS-710 densitometer (Bio-Rad) using the Quantity One program, version 4.0 (Bio-Rad). The optical density of the whole lane for each sample was quantified. The polyubiquitin band intensities were normalized to the Sso1/2 band intensities.