The Ubiquitin Ligase Ubr2, a Recognition E3 Component of the N-End Rule Pathway, Stabilizes Tex19.1 during Spermatogenesis

Ubiquitin E3 ligases target their substrates for ubiquitination, leading to proteasome-mediated degradation or altered biochemical properties. The ubiquitin ligase Ubr2, a recognition E3 component of the N-end rule proteolytic pathway, recognizes proteins with N-terminal destabilizing residues and plays an important role in spermatogenesis. Tex19.1 (also known as Tex19) has been previously identified as a germ cell-specific protein in mouse testis. Here we report that Tex19.1 forms a stable protein complex with Ubr2 in mouse testes. The binding of Tex19.1 to Ubr2 is independent of the second position cysteine of Tex19.1, a putative target for arginylation by the N-end rule pathway R-transferase. The Tex19.1-null mouse mutant phenocopies the Ubr2-deficient mutant in three aspects: heterogeneity of spermatogenic defects, meiotic chromosomal asynapsis, and embryonic lethality preferentially affecting females. In Ubr2-deficient germ cells, Tex19.1 is transcribed, but Tex19.1 protein is absent. Our results suggest that the binding of Ubr2 to Tex19.1 metabolically stabilizes Tex19.1 during spermatogenesis, revealing a new function for Ubr2 outside the conventional N-end rule pathway.


Introduction
The N-end rule pathway is a ubiquitin-dependent proteolytic system [1,2]. In this pathway, the stability of proteins is defined by their N-terminal amino acids that are distinguished into stabilizing and destabilizing residues. The latter constitute so-called Ndegrons, which are signatures for degradation of short-lived proteins. Destabilizing residues include basic (Arg, Lys, and His) and bulky hydrophobic (Phe, Tyr, Trp, Leu, and Ile) residues. An N-degron can also be created by either endoproteolytic cleavage or modifications of a pre-N-degron (Cys, Asn, Asp, Gln, or Glu) through a series of N-terminal modifications [2]. Cysteine at position 2 (after methionine) is a unique type of destabilizing residue in mammalian cells. If N-terminally exposed, Cys can be oxidized to Cys-O 2 (H) or Cys-O 3 (H) before being arginylated by the arginine (R)-transferase ATE1 [3][4][5]. The N-degron is recognized by a family of UBR box-containing E3 ligases [6]. The mammalian genome encodes at least four UBR members (Ubr1, Ubr2, Ubr4 and Ubr5) characterized by the UBR box, a ,70-residue zinc finger-like domain [2,6]. N-end rule substrates known to contain N-degrons include a set of cardiovascular GPCR regulators (RGS4, RGS5, and RGS16) in mammals [3,5], D. melanogaster DIAP1 [7], and S. cerevisiae cohesin component Scc1 [8]. Substrates targeted through internal degradation signals include histone H2A in mouse spermatocytes [9], S. cerevisiae Cup9 (a transcriptional repressor of the peptide transporter Ptr2) [10], and mammalian c-Fos [11].
Genetic studies have revealed that the N-end rule pathway plays an important role in many biological processes, including cardiac development, angiogenesis, and meiosis. Ate1-deficient mice die at fetal stages due to cardiovascular defects [4]. Mutations in the human UBR1 gene cause Johanson-Blizzard syndrome, which is characterized by exocrine pancreatic insufficiency, multiple malformations and mental retardation [12]. Disruption of Ubr2 in mice causes spermatogenic defects and female lethality [13]. Ubr2 localizes to meiotic chromatin regions and functions together with the ubiquitin conjugating (E2) enzyme HR6B in histone H2A ubiquitylation during male meiosis [9].
We previously identified Tex19.1 (also known as Tex19) as a gene with germ cell-specific expression in the testis [14]. Disruption of Tex19.1 causes defects in spermatogenesis [15]. Here we demonstrate that Tex19.1 forms a stable complex with Ubr2 during spermatogenesis. In Ubr2-deficient germ cells, Tex19.1 protein is absent despite abundant Tex19.1 mRNA, suggesting that Ubr2 is required for stabilization rather than degradation of Tex19.1 during spermatogenesis.

Tex19.1 forms a stable complex with Ubr2
Mouse Tex19.1 is a small protein (351 aa) with a coiled-coil domain, which is known to mediate protein-protein interactions [14]. To identify potential binding partners of Tex19.1, we performed immunoprecipitation (IP) experiments with testicular protein extracts using a Tex19.1-specific antibody that we generated. One prominent protein band (,200 kD) was found in the co-immunoprecipitated proteins from wild type testes but not Tex19.1 2/2 testes (Fig. 1A). Mass spectrometry analysis identified this band as Ubr2, one of the recognition E3 components of the N-end rule pathway [13].
To verify the interaction between Tex19.1 and Ubr2, we performed co-immunoprecipitation experiments followed by Western blot analysis with specific antibodies. The abundance of Ubr2 in testis was too low to be detected in total testicular extract (Fig. 1B). However, Ubr2 was readily detectable in the protein fraction immunoprecipitated with the anti-Tex19.1 antibody from wild type but not from Tex19.1-deficient testes (Fig. 1B). Likewise, Tex19.1 was co-immunoprecipitated in the reciprocal IP with the anti-Ubr2 antibody (Fig. 1C), demonstrating that Tex19.1 and Ubr2 are associated with each other in the testis. Co-transfection of NIH3T3 cells followed by coimmunoprecipitation and Western blot analysis further support the interaction between Tex19.1 and Ubr2 ( Fig. 2A, Lane 1).

Interaction of Ubr2 with Tex19.1 is arginylationindependent
In the mammalian N-end rule pathway, Cys at position 2 (after Met) is a unique type of destabilizing residue. Cleavage of the first Met by Met aminopeptidase exposes Cys at position 2. Tex19.1 from all species examined bears an N-terminal Met 1 -Cys 2 sequence, a putative pre-degron that can destabilize a substrate through oxidation and arginylation [3][4][5]. To test whether the binding of Tex19.1 by Ubr2 requires the N-terminal arginylation of Tex19.1 Cys 2 , we generated two different mutant Tex19.1 proteins in which cysteine (C2) was replaced with either glycine (G2) or valine (V2). Co-transfection and co-IP experiments using NIH3T3 cells showed that Ubr2 binds to both Tex19.1 mutant proteins, demonstrating that the interaction between Ubr2 and Tex19.1 does not require Cys 2 ( Fig. 2A, Lanes 2 and 3).
Ubr2 binds to the evolutionarily conserved N-terminal region of Tex19.1 The human genome contains only one TEX19 gene. Human TEX19 protein consists of only 164 aa and is homologous to the N-terminal part of mouse Tex19.1 protein (aa 1-162). This partial conservation suggests that the N-terminal half of Tex19 interacts with Ubr2. To test this hypothesis, we generated two partial murine Tex19.1 proteins consisting of either the N or the Cterminal half: Tex19.1N (aa 1-163) and Tex19.1C (aa 164-351). Co-IP analyses of these proteins with Ubr2 demonstrated that Ubr2 binds to the evolutionarily conserved N-terminal half but not the unconserved C-terminal half of Tex19 (Fig. 2B).

Ubr2 interacts with Tex19.2
Murine Tex19.1 forms a two-gene family with its sequence homologue Tex19.2, which is separated by only 27 kb from Tex19.1, but transcribed independently. While Tex19.1 is ex- Figure 1. Tex19.1 interacts with Ubr2 in the testis. Testicular protein extracts prepared from 20-day-old wild type and Tex19.1 2/2 mice were used for co-immunoprecipitation (IP) experiments. (A) Identification of Tex19.1-associated proteins from testis by mass spectrometry. Tex19.1-associated proteins were co-immunoprecipitated from testicular extracts with affinity-purified antibody, and analyzed by SDS-PAGE and SYPRO Ruby staining. To confirm specificity, IP of proteins from Tex19.1 2/2 testes was performed in parallel. A second differentially expressed band (lower mass, indicated by arrow) was identified as actin by mass spectrometry. (B) Co-IP of Ubr2 with Tex19.1 from testis. IP was performed with the anti-Tex19.1 antibody and probed with the anti-Ubr2 antibody. Note that Ubr2 was too low in abundance in the total testicular extract to be detected by western blot analysis. Myh11 (myosin heavy chain 11) served as a loading control. (C) Reciprocal Co-IP experiment. Tex19.1 was co-immunoprecipitated with anti-Ubr2 but not control antibody. Bands indicated by asterisks in the IP are likely to be antibody light/heavy chains or non-specific species. Molecular mass standards are shown in kilodaltons. doi:10.1371/journal.pone.0014017.g001 pressed specifically in germ cells, Tex19.2 expression is restricted to testicular somatic cells [16]. Using co-transfection and co-IP assays, we found that Ubr2 also binds to Tex19.2 (Fig. 3).
Tex19.1-null mouse mutant phenocopies Ubr2-null mutant Ubr2 2/2 mice exhibit spermatogenic defects and embryonic lethality preferentially affecting females [13]. Because Tex19.1 and Ubr2 form a stable complex in the testis, we next investigated the consequence of loss of this interaction in the Tex19.1-null mice. We disrupted the Tex19.1 gene by homologous recombination in embryonic stem (ES) cells (Fig. 4A). Tex19.1 consists of three exons with the entire coding region (ORF) residing in the last exon. In the targeted allele, the ORF of Tex19.1 was replaced by the selection marker (Neo), rtTA and LacZ. Western blot analysis confirmed the absence of the Tex19.1 protein in Tex19.1 2/2 testes (Fig. 4B). The testes of sterile XX Y* mice completely lack germ cells. The absence of Tex19.1 in XX Y* testes (Fig. 4B) demonstrates that Tex19.1 is germ cell-specific in the testis, consistent with previous studies [14][15][16][17]. In contrast with Tex19.1, Tex19.2 is expressed in somatic cells of the testis [16]. Therefore, our Western blot data ( Fig. 4B) also shows that our antibody is specific to Tex19.1 and does not recognize Tex19.2, due to limited amino acid sequence homology (58% identity) between Tex19.1 and Tex19.2.
Disruption of Tex19.1 resulted in sharply reduced testis size. The weight of Tex19.1 2/2 testes (104648 mg/pair, n = 6 mice, p,0.001) from 2-3 month old mice was half (50%) that of Tex19.1 +/2 testes (210634 mg/pair). Notably, the testes from Tex19.1 2/2 males varied dramatically in size (Fig. 4C). Likewise, the sperm output in the cauda epididymus from Tex19.1 2/2 mice varied greatly, ranging from 3.0610 5 to 1.2610 7 , but correlated well with the testis size. In most Tex19.1-deficient testes, spermatogenesis appeared to be blocked at the meiotic stage with abundant zygotene or early pachytene-like spermatocytes but few postmeiotic germ cells (Fig. 4E). However, the severity of meiotic defects varied greatly from mouse to mouse, and even from tubule to tubule in the same testis. The most severe meiotic defect was meiotic arrest at early pachytene stage. In testes with less severe defects, round and elongating spermatids were present but at a greatly reduced number (data not shown). It is intriguing but unclear why the Tex19.1-deficient mice exhibit such a large degree of phenotypic variation.
In summary, Tex19.1 2/2 mouse mutant that we generated confirms the spermatogenic phenotypes of a previously reported Tex19.1 mutant [15]. In addition, our study has uncovered a novel finding that the embryonic lethality in Tex19.1-null mice preferentially affects females. Collectively, the Tex19.1 2/2 mouse mutant exhibits abnormal spermatogenesis, defective chromosomal synapsis, and incomplete penetrance of embryonic lethality preferentially affecting females, all of which are present in the Ubr2 2/2 mouse mutant [13]. The heterogeneity of spermatogenic defects was also observed in Ubr2-deficient mice. The testis weight of Ubr2-deficient testis ranged from 30 to 70% of the wild type controls and some Ubr2-deficient males produced sperm [13]. Thus, consistent with the interaction between Tex19.1 and Ubr2, these two mouse null mutants phenocopy each other.

Depletion of the Tex19.1 protein in Ubr2-deficient testis
Ubr2 is an ubiquitin E3 ligase of the N-end rule pathway [13]. If Tex19.1 were an in vivo substrate of Ubr2 and thus targeted for degradation, one would expect increased abundance of Tex19.1 protein in Ubr2-deficient testes. To test this possibility, we first analyzed Tex19.1 transcript abundance in Ubr2 2/2 and wild type testes and found comparable levels (Fig. 5A). However, Tex19.1 protein was not detected in Ubr2 2/2 testes by Western blot analysis (Fig. 5B). The absence of Tex19.1 in Ubr2 2/2 testes was not due to loss of germ cells, since Sycp2, another meiosis-specific protein, was present in Ubr2 2/2 testes. Furthermore, we detected abundant Tex19.1 in Sycp2 2/2 testes exhibiting meiotic arrest (Fig. 5B) [20]. We then performed immunofluorescence analysis of testis sections to investigate the localization of Tex19.1 protein throughout meiosis in wild type compared to Ubr2-deficient germ cells. In early wild type germ cells, including meiotic spermatocytes, we observed that Tex19.1 localizes to the cytoplasm but not the nucleus. In contrast, Tex19.1 was not detected in post-meiotic spermatids, which express Acrv1, a component of the acrosomes (Fig. 5C) [21]. Consistent with our Western blot data, Tex19.1 protein was clearly absent from Ubr2-deficient germ cells at all stages (Fig. 5C). Therefore, the absence of Tex19.1 may be a major causative mechanism underlying spermatogenic defects in Ubr2-deficient mice. These results suggest that Ubr2 causes stabilization rather than degradation of the Tex19.1 protein in testes.

Discussion
Ubr2 is the recognition E3 component of the N-end rule pathway, in which an ubiquitin ligase recognizes a destabilizing Nterminal residue as an essential component of N-degron. It has been shown that Ubr2 plays a role in transcriptional silencing of meiotic chromosomes through ubiquitination of histone H2A [9,13]. Here, we provide several lines of evidence that Ubr2 plays a novel function outside the N-end rule pathway: protein stabilization. Firstly, we demonstrate that Ubr2 forms a stable complex with Tex19.1 in testis. If Tex19.1 were an enzymatic substrate of Ubr2 and thus destined for ubiquitin-dependent degradation, its association with Ubr2 might be too transient to be detected by co-IP. In fact, although Ubr2 protein was not detectable in the total testicular extracts by Western blot, it was readily co-immunoprecipitated with Tex19.1, suggesting the formation of a stable protein complex. Secondly, Tex19.1 bears an N-terminal cysteine. Therefore, Tex19.1 might be a substrate of ATE1-dependent arginylation and Ubr2-dependent ubiquitylation. However, our results show that the Ubr2-Tex19.1 interaction does not require the evolutionarily conserved Cys 2 residue, a putative arginylation substrate. Thirdly, the Tex19.1 mouse mutant phenocopies the Ubr2 null mutant, underscoring the physiological relevance of the Tex19.1-Ubr2 interaction. Lastly, the Tex19.1 protein is absent rather than more abundant in Ubr2deficient testes. As Ubr2 binds to Tex19.1, the most parsimonious explanation is that Ubr2 stabilizes Tex19.1 metabolically. However, we cannot exclude the possibility that Ubr2 might also play a role in the translation of the Tex19.1 mRNA.
The notion that Ubr2 plays a non-canonical function in protein turnover is further supported by the study of RECQL4 [22]. RECQL4, a putative DNA helicase, is mutated in the Rothmund-Thomson and RAPADILINO syndromes [23,24]. Although RECQL4 forms a stable complex with both Ubr1 and Ubr2, RECQL4 is not ubiquitylated and is a long-lived protein in Hela cells, suggesting that it might also be stabilized by Ubr1/Ubr2 under certain conditions [22]. In our study, Ubr2 is required for the stability of Tex19.1. However, it is unclear how Ubr2 prevents Tex19.1 from degradation. One possibility is that the stable binding of Ubr2 to Tex19.1 blocks the accessibility of Tex19.1 by other E3 ligases.

Generation of anti-Tex19.1 and anti-Ubr2 polyclonal antibody
The entire mouse Tex19.1 coding region was cloned into the pQE30 vector (QIAGEN). The 6xHis-Tex19.1 fusion protein was expressed in M15 bacteria, affinity-purified with Ni-NTA beads, and eluted in 8 M urea according to the manufacture's standard purification protocol (QIAGEN). The N-terminal 100 aa of mouse Ubr2 was expressed as a GST fusion protein in E. coli using the pGEX4T-1 vector and affinity purified with glutathione Sepharose. Each of the recombinant proteins was used to immunize two rabbits according to the company's standard protocol (Cocalico Biologicals, Inc). The anti-Tex19.1 antiserum (serum 2109) was used for western blot (1:500). The anti-Ubr2 antiserum (serum 2186) was used for western blot (1:500).
Targeted inactivation of the Tex19.1 gene In the Tex19.1 targeting construct, two homologous arms (2.1 kb each) were amplified from a Tex19.1-positive BAC clone (RP23-400P17) by high-fidelity PCR and were subcloned to flank the rtTA2S-M2-IRES-LacZ-PGK-Neo knockin/selection cassette (Fig. 4). The rtTA2S-M2 fragment was PCR amplified from the pUHDrtTA2S-M2 plasmid [25]. Hybrid V6.5 XY ES cells (C57BL/66129/sv) were electroporated with the linearized Tex19.1 targeting construct (pUP77-29A/NotI) and were cultured in the presence of G418 (350 mg/ml). Electroporation was performed in a 0.4-cm Bio-Rad cuvette with the Bio-Rad Gene Pulser Xcell unit (Voltage, 400 v; Capacitance, 25 mF; Resistance, infinite; Expected time constant, 0.4 msec). Seven days after electroporation, 384 G418-resistant ES cell clones were picked. Screening of 96 clones by PCR identified twelve ES cell clones in which homologous recombination had occurred. PCR was performed with a forward primer upstream of the left arm and a reverse primer in the knockin/Neo cassette or a forward primer in the knockin/Neo cassette and a reverse primer downstream of the right arm. One targeted ES cell clone (3H6) was injected into B6C3F1 (Taconic) blastocysts that were subsequently transferred to the uteri of pseudopregnant ICR females. Male chimeras were bred with C57BL/6J females and germ-line transmission of the Tex19.1 knockout/knockin allele was obtained. Mice from mixed strain backgrounds (C57BL/66129/sv) were used in this study. In the Tex19.1 mutant mice, the expression of rtTA and LacZ is expected to be under the control of the endogenous Tex19.1 promoter. Offspring were genotyped by PCR of tail genomic DNA with the following primers: wild type (311 bp), ATG-GATCCTGTCCCCCAGTCAGCGTT and GCGTCGACTT-AGCACATAAAGGGACCCCAAT; mutant (450 bp), AAG-TCGACATGTCTAGACTGGACAAGAG and CCTCCAA-TACGCAGCCCAGTGTAA. ,3 mm-long tail biopsy was digested in 200 ml proteinase K buffer (10 mM Tris, 10 mM EDTA, 10 mM NaCl, 0.2% SDS, pH 8.0) at 55uC overnight. Genomic DNA was precipitated with 200 ml iso-propanol, washed once with 1 ml of 70% ethanol, air died and resuspended in 200 ml 1xTE buffer. 1 ml of genomic DNA was used in a 20 ml PCR genotyping reaction (94uC, 30 seconds; 55uC, 30 seconds; 72uC, 45 seconds; 35 cycles). Full details of the study were approved by the Institutional Animal Care and Use Committee (IACUC protocol # 802780) of the University of Pennsylvania.