Rpn10-Mediated Degradation of Ubiquitinated Proteins Is Essential for Mouse Development

Gene targeting of Rpn10.A targeting vector for Rpn10-null mice was constructed by replacing exons 2 to 8 with a neomycin resistance gene (neo) cassette. For Rpn10a knock-in mice (with, at the same time, conditional deletion of UIM domains), exons 7 to 10 were replaced with the corresponding cDNA of Rpn10a, with a polyadenylation signal attached at its 3′ end. A neo cassette was inserted at the 3′ end of the cDNA. LoxP sequences were inserted at the 5′ end of the cDNA and the 3′ end of the neo cassette. TT2 embryonic stem cells were screened as described previously (27). For Southern blot analysis, genomic DNA was digested with EcoRI for Rpn10 knockout or with EcoRV for Rpn10a knock-in and was hybridized with the probes shown in Fig. 1A and 2A, respectively. EIIa-Cre and Alb-Cre were purchased from The Jackson Laboratory. PCR primers used for mouse genotyping are listed in Table 1. Mice were housed in pathogen-free facilities, and the experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of the Tokyo Metropolitan Institute of Medical Science.

• Open in new tab
• Download powerpoint

FIG. 2.

Generation and analysis of Rpn10a knock-in mice. (A) Schematic representation of the targeting vector and the targeted allele of the Rpn10 gene. Light shaded box, the Rpn10a cDNA fragment corresponding to amino acids 219 to 376 with a polyadenylation signal. Dark shaded box, Rpn10a exon 7 splicing acceptor sequences followed by a stop codon and polyadenylation signal. Triangles, loxP sequences. Open box, probe used for Southern blot analysis. Arrows indicate positions of PCR primers. (B) Southern blot analysis of genomic DNAs extracted from mouse tails. Wild-type and mutant alleles are detected as 12.4- and 7-kb bands, respectively. (C) PCR analysis of genomic DNA extracted from the tails of wild-type, Rpn10^a/+, and Rpn10^a/a mice. The amplified fragments derived from wild-type and Rpn10a alleles are indicated. (D) RT-PCR analysis of Rpn10 splice variant transcripts. (E) Liver lysates from 13-week-old mice were immunoblotted with antibodies against the indicated proteins. (F) Lysates from Rpn10^+/+ and Rpn10^a/a livers were fractionated by glycerol gradient centrifugation (10 to 40% glycerol from fraction 1 to fraction 30). (Left) An aliquot of each fraction was used for an assay of chymotryptic activity of proteasomes using succinyl-Leu-Leu-Val-Tyr-7-amino-4-methyl-coumarin (Suc-LLVY-AMC) as a substrate in the absence (top) or presence (bottom) of 0.025% sodium dodecyl sulfate (SDS). (Right) Immunoblot analysis of each fraction was performed using antibodies against the indicated proteins.

RNA isolation, reverse transcription, and real-time PCR.Expression levels of Rpn10 variants were determined by reverse transcription-PCR (RT-PCR) as described previously (18). Specific primers for each variant are listed in Table 2. For real-time PCR analysis, total RNAs were isolated from the livers of 5-week-old mice by using an RNAspin minikit (GE Healthcare), reverse transcribed to cDNA using a Transcriptor first-strand cDNA synthesis kit (Roche), and subjected to real-time PCR using the LightCycler 480 system (Roche). PCR primers and universal probes (Roche), which are listed in Table 3, were designed according to the Universal Probe Assay Design Center (http://www.roche-applied-science.com/sis/rtpcr/upl/adc.jsp ). Glucuronidase beta (GUSβ) was used for normalization. Real-time PCR data were analyzed by the ΔΔC_T method.

Immunological analysis.Mouse livers were homogenized and subjected to immunoblotting and immunoprecipitation as described previously (13). The antibodies against Rpn1, Rpn3, Rpt6, the VWA domain of Rpn10 [Rpn10(N)], USP14, polyubiquitin, and actin have been described previously (13). Polyclonal antibodies against a6, a7, Rpn6, Rpn7, Rpt3, Rpt5, the UIM domain of Rpn10 [Rpn10(C)], and mHR23B (all sequences were derived from mice) were raised in rabbits by using recombinant proteins expressed in and purified from strain BL21RIL (Novagen) as His₆ fusion proteins of α6 (residues 152 to 263), α7 (residues 157 to 255), Rpn6 (residues 1 to 162), Rpn7 (residues 1 to 137), Rpt3 (residues 1 to 100), and Rpt5 (residues 1 to 131) and as glutathione S-transferase fusion proteins of Rpn10 (residues 255 to 376) and mHR23B (full length), respectively. For immunoprecipitation, a liver homogenate from an Rpn10^a/a or an Rpn10^a/a:Alb mouse (the two mice had equal Suc-LLVY hydrolyzing activities) was immunoprecipitated with an anti-Rpt6 antibody.

Glycerol gradient analysis.Mouse liver homogenates were clarified by centrifugation at 20,000 × g and subjected to 10 to 40% (vol/vol) linear glycerol gradient centrifugation (22 h, 83,000 × g) as described previously (16).

Assay of proteasome activity.The assays of proteasome chymotryptic peptidase activity, degradation of recombinant ³⁵S-labeled ornithine decarboxylase (ODC), and degradation of polyubiquitinated ³⁵S-labeled cIAP1 protein have been described previously (13, 16).

Histological examination.Embryos in utero were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned. Sections were stained with Mayer's hematoxylin, followed by eosin staining.

Culture of blastocysts.Blastocysts were flushed out from pregnant female uteri at embryonic day 3.5 (E3.5) and were cultured in M16 medium (Sigma-Aldrich) at 37°C under 5% CO₂ on a gelatin-coated chambered coverglass (Nalgene).

Loss of Rpn10 causes early embryonic lethality in mice.To determine the significance of Rpn10 in mammals, we generated Rpn10-null mice by replacing exons 2 to 8 with a neomycin resistance gene (Fig. 1B to D). Rpn10-heterozygous (Rpn10^+/−) mice were born without any gross abnormality and were fertile. These mice were intercrossed to produce Rpn10^−/− mice. The progeny did not contain surviving Rpn10^−/− pups, suggesting that the absence of Rpn10 during embryogenesis is lethal. To investigate this issue further, we examined, at various stages of development (mainly E6.5 to E7.5), embryos that had been produced in timed intercrosses. In normal E6.5 embryos, a cylinder-like two-layered cellular structure was observed. However, Rpn10-deficient embryos failed to form this structure at E6.5 (Fig. 1E, left) and were resorbed by E7.5 (Fig. 1E, right).

To analyze further the defects associated with Rpn10 deficiency, we isolated blastocysts from Rpn10^+/− intercrosses at E3.5. Rpn10^−/− blastocysts were identified by PCR at the expected Mendelian frequency (Fig. 1F, left; also data not shown), representing relatively mild phenotypes compared to those of Rpt3 and Rpt5 knockout mice, which did not develop beyond the 8-cell stage (33). When these blastocysts were cultured in vitro, most of the Rpn10^−/− blastocysts hatched from the zona pellucida, spread trophoblastic cells with proliferating inner cell masses (ICMs), which form the future embryonic ectoderm, and grew on the gelatin-coated glass, like wild-type blastocysts (Fig. 1F, center). However, ICMs of Rpn10^−/− blastocysts could not expand beyond 48 h of culture and detached from the trophoblastic cells before 96 h of culture, in contrast to wild-type ICMs (Fig. 1F, right). These results indicate that Rpn10 is essential for embryonic development beyond blastocyst formation; presumably it is involved in the expansion of the embryonic ectoderm after implantation.

Rpn10a is sufficient for the development of mice.In vertebrates, Rpn10 has five splice variants named Rpn10a to Rpn10e. Rpn10a is the conventional isoform expressed throughout development and throughout the body, while Rpn10b to Rpn10e are expressed at specific developmental stages or in specific organs (Fig. 1A). These facts raise the possibility that the diversity of Rpn10 plays a role in development in vertebrates. To test this hypothesis and to clarify the roles of these vertebrate-specific isoforms, we generated Rpn10a knock-in mice. The Rpn10 isoforms are generated by different splice acceptor and donor usages of a genomic locus that corresponds to exons 7 to 10 of the Rpn10a isoform, which encode the major part of the two UIM domains of Rpn10 (18). Therefore, a targeting vector was designed to replace a genomic locus with the corresponding cDNA sequences of Rpn10a and to disrupt the expression of other isoforms (Fig. 2A to C). The inserted Rpn10a cDNA was flanked with loxP sequences to enable the generation of mice expressing Rpn10 lacking UIM domains (Rpn10ΔUIM). Mice heterozygous for the Rpn10a knock-in allele (Rpn10^a/+ mice) were born healthy and fertile without noticeable pathological phenotypes. Rpn10^a/a mice, obtained by intercrossing Rpn10^a/+ mice, were born healthy at Mendelian frequency, were fertile, and grew apparently normally without any gross abnormality (data not shown). RT-PCR analysis of RNAs from newborn mice demonstrated loss of Rpn10b to Rpn10e isoforms in Rpn10^a/a mice, while all the isoforms were expressed in wild-type mice (Fig. 2D). The protein levels of Rpn10 as well as other proteasome subunits in Rpn10^a/a mice were similar to those in the wild type (Fig. 2E), and the expressed Rpn10a was incorporated normally into 26S proteasomes, like that expressed in wild-type mice (Fig. 2F, right). The proteasome activity of the Rpn10^a/a liver, assessed with fluorogenic peptides, was nearly equal to that of the wild-type liver (Fig. 2F, right). The proteasome activities of adult brains were also comparable in Rpn10^a/a and wild-type mice (data not shown). These results indicate that vertebrate-specific isoforms of Rpn10 do not play an important role in development and that the conventional isoform Rpn10a is sufficient for life, at least under normal circumstances. However, it is possible that isoforms Rpn10b to Rpn10e are involved in the degradation of specific target proteins or play a role in a process other than ubiquitin-mediated proteolysis, defects in which might become apparent only under certain conditions.

Mice deficient in UIM domains exhibit embryonic lethality but survive longer than Rpn10-null mice.In genetic analyses using yeast and moss, lack of the UIM domains of Rpn10 displayed modest phenotypes compared to null mutations, thus questioning the physiological significance of the UIM domains of Rpn10 (6, 9). To examine the role of these domains in mice, we generated Rpn10ΔUIM-expressing mice by Cre recombinase-mediated excision of the UIM domain-coding region (Fig. 2A, bottom, and Fig. 3A). By crossing Rpn10^a/a mice with EIIa-Cre transgenic mice, in which the expression of Cre recombinase appears from the zygote stage (22), we obtained mice harboring an Rpn10 gene encoding Rpn10ΔUIM protein throughout the body, including germ cells (Rpn10^ΔUIM/+ mice). Rpn10^ΔUIM/+ mice were born at the expected Mendelian frequency but exhibited slightly retarded growth and maturation compared to wild-type mice or even to Rpn10^+/− mice (data not shown), implying that incorporation of Rpn10 protein lacking UIM domains into 26S proteasomes exerted a somewhat dominant-negative effect. However, these mice were fertile and showed no obvious phenotypes other than slow growth. Rpn10^ΔUIM/+ mice were intercrossed to produce Rpn10^ΔUIM/ΔUIM mice. The progeny did not contain any surviving Rpn10^ΔUIM/ΔUIM pups, suggesting that the absence of the UIM domains of Rpn10 was incompatible with embryogenesis. Examination of embryos at various developmental stages revealed that the development of Rpn10^ΔUIM/ΔUIM embryos was normal before E6.5 (data not shown) but appeared to be delayed at E8.5 (Fig. 3B and C). At E9.5, development arrested at a stage corresponding to E8.5 of the wild type; the turning process that results in a fetal position, normally seen at the transition from the 6-somite to the 8-somite stage, was not initiated (Fig. 3B and C). However, we could not find specific morphological defects in the embryos, such as disturbed formation of heart tubes, which are often associated with the failure of turning seen in other knockout mice such as GATA4 knockout mice (47). Intriguingly, Rpn10^ΔUIM/ΔUIM embryos developed to an advanced stage compared to Rpn10^−/− embryos, indicating that the VWA domain of Rpn10 rescued development from E6.5 to E9.5. These results suggest that the VWA domain alone plays some roles in proteasome function but that the UIM domain-dependent function of proteasomes is still required for mouse development, especially at the turning stage.

• Open in new tab
• Download powerpoint

FIG. 3.

Developmental arrest of mice deficient in the UIM domains. (A) PCR analysis of genomic DNA extracted from wild-type, Rpn10^ΔUIM/ΔUIM, and Rpn10^ΔUIM/+ embryos at E8.5. The amplified fragments derived from wild-type, ΔUIM/+, and ΔUIM/ΔUIM alleles are indicated. (B) Morphology of Rpn10 embryos. Genotypes of embryos were determined by PCR. The longitudinally arranged panels represent littermates. Note the lack of turning of mutant embryos at E9.5. (C) Rpn10^+/+ (top) and Rpn10^ΔUIM/ΔUIM (bottom) embryos at E8.5 and E9.5 in utero were sagittally (E8.5) and transversely (E9.5) sectioned and stained with hematoxylin and eosin. Genotypes of embryos were deduced morphologically. Scale bars, 200 μm. ac, amniotic cavity; al, allantois; am, amnion; ch, chorion; en, endocardium; epc, ectoplacental cone; exo, exocoelom; fg, foregut; hf, head fold; hg, hindgut; ht, heart tube; mc, myocardium; mes, mesoderm; ne, neuroectoderm; nf, neural fold; no, notochord; nt, neural tube; vys, visceral yolk sac.

UIM domain deficiency in the liver is associated with impaired degradation of ubiquitinated proteins.To determine the biochemical basis of the significance of the UIM domain of Rpn10, we generated mice that expressed Rpn10ΔUIM exclusively in postnatal hepatocytes by crossing Rpn10^a/a mice with transgenic mice that expressed Cre recombinase under the control of the albumin (Alb) promoter (28). Rpn10^a/a:Alb mice, which expressed Rpn10ΔUIM proteins instead of Rpn10a proteins in the liver postnatally, were born without any abnormal appearance or developmental defect. We first confirmed the deletion of the UIM domains of Rpn10 in the liver. In the Rpn10^a/a:Alb liver, no full-length Rpn10a proteins were detected. Instead, as expected, a truncated form of Rpn10 appeared, which could be detected by an anti-Rpn10 antibody raised against the VWA domain of Rpn10 [Rpn10(N)] but not by an anti-Rpn10 antibody raised against the UIM domain of Rpn10 [Rpn10(C)] (Fig. 4A). This Rpn10ΔUIM species was incorporated correctly into 26S proteasomes, consistent with the findings of studies with yeast (Fig. 4B). Interestingly, immunoblot analysis of liver lysates revealed that protein levels of subunits of the 20S proteasome (α6, α7), the base (Rpn1, Rpt3, Rpt5, Rpt6), and the lid (Rpn3, Rpn6, Rpn7), as well as levels of some proteasome-interacting proteins (mHR23B, USP14), were all increased in the Rpn10^a/a:Alb liver (Fig. 4A). The up-regulation of proteasome subunits led to approximately twofold increases in the levels of 20S and 26S proteasomes and in proteasome-specific peptidase activities in Rpn10^a/a:Alb liver lysates relative to those for Rpn10^a/a liver lysates (Fig. 4C). To examine the reason for the increased proteasome levels, we quantified relative mRNA levels of proteasome subunits by real-time PCR analysis. We noted 1.8- to 2.5-fold increases in levels of mRNAs of proteasome subunits and proteasome-interacting proteins in the Rpn10^a/a:Alb liver relative to those in the Rpn10^a/a liver, indicating that transcription of overall proteasome-related genes was up-regulated in the Rpn10^a/a:Alb liver (Fig. 4D). Despite the elevated amounts of proteasomes and the consequently increased peptidase activities, accumulation of polyubiquitin-conjugated proteins was noted in the Rpn10^a/a:Alb liver (Fig. 4E). To test whether degradation of native proteins was impaired in the Rpn10^a/a:Alb liver, we measured the degradation rates of two types of proteasome substrates in vitro. One is ODC, which is degraded by 26S proteasomes in a ubiquitin-independent but antizyme-dependent manner (26). The other is cIAP1 protein, a RING finger type ubiquitin ligase that ubiquitinates itself for degradation by 26S proteasomes in a ubiquitin-dependent manner (37). The degradation rate of ODC was increased in lysates of the Rpn10^a/a:Alb liver, and this increase correlated with the increase in the level of the 26S proteasome (Fig. 4F, left). In contrast, the degradation rate of ubiquitinated cIAP proteins was markedly reduced in the Rpn10^a/a:Alb liver, although the amounts of 26S proteasomes and mHR23B were larger than those in the Rpn10^a/a liver (Fig. 4F, right). These results indicate that the UIM domain of Rpn10 plays an important role in the recognition and degradation of ubiquitinated proteins in the mouse liver. It is likely that increased transcription of proteasome-related genes is a feedback regulation mechanism to compensate for the impaired degradation and accumulation of ubiquitinated proteins, as was also observed for the Rpn10-deficient fly (40, 48).

• Open in new tab
• Download powerpoint

FIG. 4.

UIM domain deficiency in the liver causes impaired degradation of ubiquitinated proteins. (A) Homogenates from 6-week-old Rpn10^a/a and Rpn10^a/a:Alb mouse livers were immunoblotted with the indicated antibodies. (B) The homogenates for which results are shown in panel A were fractionated by 10 to 40% glycerol gradient centrifugation and immunoblotted with the indicated antibodies. The faint Rpn10a bands detected in blots with anti-Rpn10(C) and anti-Rpn10(N) (asterisk) in the panels for Rpn10^a/a:Alb liver are presumably derived from non-albumin-expressing cells in the liver. (C) The peptide-hydrolyzing activity of each fraction from panel B was measured as for Fig. 2F. (D) Real-time RT-PCR was used to measure the expression of transcripts encoding proteasome-related genes in the livers of 6-week-old Rpn10^a/a and Rpn10^a/a:Alb mice. Data represent levels of transcripts in Rpn10^a/a:Alb liver relative to those in Rpn10^a/a liver and are means ± standard deviations from experiments with three pairs of littermates. (E) The homogenates for which results are shown in panel A were immunoblotted with an anti-ubiquitin (Ub) antibody. (F) Ubiquitin-independent and -dependent protein-degrading activities of proteasomes. Homogenates from 6-week-old mouse livers were subjected to an in vitro protein degradation assay. Antizyme-dependent degradation of ³⁵S-labeled ODC (left) and ubiquitin-dependent degradation of ³⁵S-labeled cIAP1 (right) were measured. Data are means ± standard deviations from triplicate experiments. (G) Homogenates from Rpn10^a/a and Rpn10^a/a:Alb mouse livers were immunoprecipitated with an anti-Rpt6 antibody and subjected to immunoblotting with the indicated antibodies.

Previous reports showed that human homologues of yeast Rad23 bind to proteasomes via the second UIM domain of human Rpn10 (7, 17, 25, 46), whereas Rad23 binds directly to Rpn1 in yeast (5, 32). To assess the significance of the UIM domains of Rpn10 in recruiting Rad23 species to proteasomes in mammals, we immunoprecipitated 26S proteasomes from liver lysates with an anti-Rpt6 antibody and compared the amount of proteasome-associated mHR23B (a mouse homologue of yeast Rad23) in Rpn10ΔUIM liver to that in Rpn10^a/a liver by immunoblotting (Fig. 4G). Since Rpn10ΔUIM liver contained increased levels of proteasomes and proteasome activity (Fig. 4A and C), the amounts of proteasomes loaded were adjusted for the peptide-hydrolyzing activities of the lysates. Although the band intensities of proteasome subunits (Rpt6, Rpn1, Rpn2, Rpn6) were comparable between the genotypes, the amount of mHR23B in Rpn10ΔUIM proteasomes was approximately 60% lower than that in Rpn10^a/a proteasomes (Fig. 4G), consistent with the previous observations that the UIM domain of Rpn10 recruits Rad23 species in mammals (7, 17, 25, 46). However, this result also indicates that the UIM domain is not the sole receptor for Rad23 species in mammals, because a portion of mHR23B remained associated with Rpn10ΔUIM proteasomes (Fig. 4G). It should be noted that an increased amount of USP14 was detected in Rpn10ΔUIM proteasomes, consistent with the observation with yeast that ubiquitin stress enhances the loading of proteasomes with Ubp6, which is known to bind to 26S proteasomes via Rpn1 (14, 24). It is possible that USP14 competes with mHR23B for binding to proteasomes, specifically to Rpn1, although the binding of USP14 and mHR23B to mammalian Rpn1 has yet to be established. Therefore, we cannot evaluate the exact contribution of the UIM domain in accepting Rad23 in mammalian proteasomes. Whether ubiquitinated proteins are recognized directly by the UIM domains of Rpn10 or delivered to the UIM domain by Rad23 species, these results demonstrate that deletion of the UIM domains of Rpn10 causes insufficient delivery of ubiquitinated proteins to proteasomes, resulting in impairment of their degradation in mammals.