Degradation of the Separase-cleaved Rec8, a Meiotic Cohesin Subunit, by the N-end Rule Pathway*

The Ate1 arginyltransferase (R-transferase) is a component of the N-end rule pathway, which recognizes proteins containing N-terminal degradation signals called N-degrons, polyubiquitylates these proteins, and thereby causes their degradation by the proteasome. Ate1 arginylates N-terminal Asp, Glu, or (oxidized) Cys. The resulting N-terminal Arg is recognized by ubiquitin ligases of the N-end rule pathway. In the yeast Saccharomyces cerevisiae, the separase-mediated cleavage of the Scc1/Rad21/Mcd1 cohesin subunit generates a C-terminal fragment that bears N-terminal Arg and is destroyed by the N-end rule pathway without a requirement for arginylation. In contrast, the separase-mediated cleavage of Rec8, the mammalian meiotic cohesin subunit, yields a fragment bearing N-terminal Glu, a substrate of the Ate1 R-transferase. Here we constructed and used a germ cell-confined Ate1−/− mouse strain to analyze the separase-generated C-terminal fragment of Rec8. We show that this fragment is a short-lived N-end rule substrate, that its degradation requires N-terminal arginylation, and that male Ate1−/− mice are nearly infertile, due to massive apoptotic death of Ate1−/− spermatocytes during the metaphase of meiosis I. These effects of Ate1 ablation are inferred to be caused, at least in part, by the failure to destroy the C-terminal fragment of Rec8 in the absence of N-terminal arginylation.

The N-end rule pathway is a set of proteolytic systems whose unifying feature is the ability to recognize and polyubiquitylate proteins containing degradation signals called N-degrons, 6 thereby causing degradation of these proteins by the proteasome (Fig. 1, A and B) (1)(2)(3)(4)(5)(6)(7)(8). The main determinant of an N-degron is either an unmodified or chemically modified destabilizing N-terminal residue of a protein. The identity of the next residue, at position 2, is often important as well. A second determinant of an N-degron is a protein's internal Lys residue. It functions as the site of a protein's polyubiquitylation and tends to be located in a conformationally disordered region (4,9,10). Recognition components of the N-end rule pathway are called N-recognins. In eukaryotes, N-recognins are E3 ubiquitin (Ub) ligases that can target N-degrons (Fig. 1, A and B). Bacteria also contain the N-end rule pathway, but Ub-independent versions of it (11)(12)(13)(14)(15)(16).
Regulated degradation of proteins and their natural fragments by the N-end rule pathway has been shown to mediate a strikingly broad range of biological functions, including the sensing of heme, nitric oxide (NO), oxygen, and short peptides; the control, through subunit-selective degradation, of the input stoichiometries of subunits in oligomeric protein complexes; the elimination of misfolded or otherwise abnormal proteins; the degradation of specific proteins after their retrotranslocation to the cytosol from mitochondria or other membrane-enclosed compartments; the regulation of apoptosis and repression of neurodegeneration; the regulation of DNA repair, transcription, replication, and chromosome cohesion/segregation; the regulation of G proteins, autophagy, peptide import, meiosis, immunity, fat metabolism, cell migration, actin filaments, cardiovascular development, spermatogenesis, and neurogenesis; the functioning of adult organs, including the brain, muscle, testis, and pancreas; and the regulation of leaf and shoot development, leaf senescence, and many other processes in plants (Fig. 1, A and B) (see Refs. 4 -7 and references therein).
Nt-arginylation is mediated by the Ate1-encoded arginyltransferase (Arg-tRNA-protein transferase; R-transferase), a component of the Arg/N-end rule pathway and one subject of the present study ( Fig. 1A) (28 -36). Alternative splicing of mouse Ate1 pre-mRNAs yields at least six R-transferase isoforms, which differ in their Nt-arginylation activity (28,31). R-transferases are sequelogous (similar in sequence (37)) throughout most of their ϳ60-kDa spans from fungi to mammals (4). R-transferase can arginylate not only N-terminal Asp and Glu but also N-terminal Cys, if it has been oxidized to Cys-sulfinate or Cys-sulfonate, through reactions mediated by NO, oxygen, and N-terminal Cys-oxidases (5,6,30). Consequently, the Arg/N-end rule pathway functions as a sensor of NO and oxygen, through the conditional oxidation of N-terminal Cys in proteins such as the Rgs4, Rgs5, and Rgs16 regulators of G proteins in mammals (30,38) and specific transcriptional regulators in plants (reviewed in Refs. 5 and 6). By now, more than 200 distinct proteins (including natural protein fragments) have been either shown or predicted to be Nt-arginylated (21, 22, 39 -41). Many, possibly most, Nt-arginylated proteins are conditionally or constitutively short-lived substrates of the Arg/N-end rule pathway (Fig. 1A).
In mammals and other multicellular eukaryotes, the bulk of cohesin-mediated confinement of sister chromatids is removed during prophase through openings of cohesin rings that do not involve proteolytic cuts. Sister chromatids continue to be held together until the end of metaphase, to a large extent through still intact cohesin rings at the centromeres. Once the bipolar spindle attachment of chromosomes is achieved at the end of metaphase, the activated separase cleaves the mammalian Rad21 subunit (a sequelog of yeast Scc1) in the closed cohesin complexes, resulting in their opening and allowing the separation of sister chromatids (45,48,58,65). At least in mammals, the Rad21 cohesin subunit can be cleaved in vivo not only by separase but also by calpain-1 (a Ca 2ϩ -activated protease) (58) and by caspases as well (56,57).
During meiosis, in which cohesins also play essential roles, the mitosis-specific cohesin subunit Rad21 is replaced by the sequelogous (similar in sequence (37)), meiosis-specific Rec8 subunit (43, 59, 66 -70). (A second meiosis-specific kleisin-type subunit, called Rad21L, is expressed during early meiosis and disappears afterward (71).) Meiotic DNA replication is followed by two rounds of cell division to produce four haploid daughter cells. During the first meiotic cell division cycle (meiosis I), replicated homologous chromosomes pair and recombine with each other. The pairs of modified (recombined) homologous chromosomes are separated at the end of meiosis I, yielding two diploid daughter cells (72,73). During meiosis II, the replicated sister chromatids of each chromosome are pulled apart to produce haploid daughter cells. In the testis of male mice, meiosis I and II take place in meiotic spermatocytes, leading to the formation of haploid spermatids and later mature sperm cells (43,59,66,74).
In S. cerevisiae, the separase-generated C-terminal fragment of the Scc1 cohesin subunit bears N-terminal Arg. This fragment of Scc1 forms late in mitosis upon the activation of separase and is rapidly destroyed by the Arg/N-end end rule pathway (62). A failure to eliminate this (normally short-lived) Scc1 fragment (e.g. in a ubr1⌬ mutant that lacks the Arg/Nend rule pathway) results in chromosome instability (62). The C-terminal fragment of Scc1 retains a physical affinity for the rest of the cohesin complex (62). Therefore, the plausible (but not proven) cause of chromosome instability in ubr1⌬ cells is an interference with cohesin mechanics by the metabolically stabilized, cohesin-bound C-terminal fragment of Scc1. Because the yeast Scc1 fragment bears N-terminal Arg, the fragment's degradation by the Arg/N-end rule pathway does not require Nt-arginylation (62). In mammals, however, the separase-generated C-terminal fragments of the Rad21 subunit of mitotic cohesin and the Rec8 subunit of meiotic cohesin bear N-terminal Glu, a substrate of the Ate1 R-transferase (Fig. 1, A and C) (47,75,76).
In the present work, we constructed an Ate1 Ϫ/Ϫ mouse strain in which the ablation of Ate1 was confined to germ cells. We show that the separase-generated C-terminal fragment of Rec8, a subunit of meiotic cohesin, is a short-lived physiological substrate of the Arg/N-end rule pathway and that the degradation of this Rec8 fragment requires its Nt-arginylation. These and related results suggest that a failure to destroy this fragment in arginylation-lacking spermatocytes of Ate1 Ϫ/Ϫ mice contributes to a greatly reduced male fertility of these mice, due to the observed arrest and apoptotic death of Ate1 Ϫ/Ϫ spermatocytes at the end of meiosis I. These results expand the already large set of functions of the Arg/N-end rule pathway. Together with earlier data about Ate1 Ϫ/Ϫ and Ubr2 Ϫ/Ϫ mice (33,77), the present findings also indicate that perturbations of the Arg/N-end rule pathway may be among the causes of infertility in humans.

Experimental Procedures
Mutant Mouse Strains-Conditional (cre-lox-based) Ate1 flox/flox mice were described (33). Tnap-Cre mice, in which the Cre recombinase is selectively expressed in primordial germ cells (78), were purchased from Jackson Laboratory (Bar Harbor, ME). Germ-cell specific Ate1 Ϫ/Ϫ mouse strains were constructed in the present study through matings of Ate1 flox/flox mice (33) and Tnap-Cre mice (78). All animal experiments were approved by the Animal Research Panel of the Committee on Research Practice of the University of the Chinese Academy of Science.
Antibodies-Mouse anti-Sycp3 monoclonal antibody (SC-74569) and mouse anti-PLZF monoclonal antibody (SC-28319) were from Santa Cruz Biotechnology, Inc. (Dallas, TX). Rabbit anti-Sycp3 polyclonal antibody (ab150292) and rabbit anti-FIGURE 1. The mammalian N-end rule pathway and the separase cleavage site in Rec8, a meiosis-specific cohesin subunit. See the Introduction for descriptions of the pathway's mechanistic aspects and biological functions. Amino acid residues are denoted by single-letter abbreviations. A, the Arg/N-end rule pathway. It targets proteins for degradation through their specific unacetylated N-terminal residues. A yellow oval denotes the rest of a protein substrate. R-transferase, Ate1 arginyltransferase; primary, secondary, and tertiary, mechanistically distinct classes of destabilizing N-terminal residues; type 1 and type 2, two sets of primary destabilizing N-terminal residues, basic (Arg, Lys, and His) and bulky hydrophobic (Leu, Phe, Trp, Tyr, Ile, and Met followed by a bulky hydrophobic residue (⌽)), respectively. These sets of N-terminal residues are recognized by two distinct substrate-binding sites of N-recognins, the pathway's E3 ubiquitin ligases. B, the Ac/N-end rule pathway. It targets proteins through their Nt-acetylated residues. The red arrow on the left indicates the cotranslational removal of the N-terminal Met residue by Met-aminopeptidases (MetAPs). N-terminal Met is retained if a residue at position 2 is larger than Val. C, alignments of amino acid sequences near the main separase cleavage site in mammalian Rec8, between Arg 454 and Glu 455 of mouse Rec8. The consensus sequence of this cleavage site (its P4 -P1 residues) in both mitotic (Rad21) and meiotic (Rec8) kleisin-type cohesin subunits is also shown. Conserved residues of mammalian Rec8 near the cleavage site are shown in yellow. The conserved P4, P3, P1, and P1Ј residues of mammalian Rec8 at the cleavage site are in red, orange, green, and blue, respectively.
Male Fertility Assays-Breeding assays with wild-type and Ate1 Ϫ/Ϫ male mice were carried out as described previously (79). Briefly, each examined male mouse (8 -9 weeks old) was caged with two wild-type (CDI strain) female mice (6 -8 weeks old), and their vaginal plugs were checked every morning. The number of pups produced by each pregnant female was counted within a week after birth. Each male was tested through at least six cycles of this breeding assay.
Epididymal Sperm Count-The cauda epididymis was dissected from adult mice. Sperm was released by cutting the cauda epididymis into pieces, placing them in 1 ml of phosphate-buffered saline (PBS), and incubating for 10 min at 37°C. Thereafter, 10-l samples were transferred to a hemocytometer for counting sperm cells.
Rabbit reticulocyte-based degradation assays were carried out using the TNT T7 coupled transcription/translation system (Promega, Madison, WI), largely as described previously (21)(22)(23). Nascent proteins translated in the extract were pulse-labeled with L-[ 35 S]methionine (0.55 mCi/ml, 1,000 Ci/mmol, MP Biomedicals, Santa Ana, CA) for 10 min at 30°C in a total volume of 30 l. Labeling was quenched by the addition of 0.1 mg/ml cycloheximide and 5 mM unlabeled methionine, bringing the final reaction volume to 40 l. Samples of 10 l were removed at the indicated time points, and the reaction was terminated by the addition of 80 l of TSD buffer (1% SDS, 5 mM dithiothreitol (DTT), 50 mM Tris-HCl, pH 7.4), snap-frozen in liquid nitrogen, and stored at Ϫ80°C until use. Following the collection of all time points, samples were heated at 95°C for 10 min and then diluted with 1 ml of TNN buffer (0.5% Nonidet P-40, 0.25 M NaCl, 5 mM Na-EDTA, 50 mM Tris-HCl, pH 7.4) containing 1ϫ Complete protease inhibitor mixture (Roche Diagnostics) and immunoprecipitated using 5 l of anti-FLAG M2 magnetic beads (Sigma-Aldrich). Samples were incubated with rotation at 4°C for 4 h, followed by three washes in TNN buffer, a wash in 10 mM Tris-HCl, pH 8.5, and resuspension in 20 l of SDS-sample buffer. Samples were heated at 95°C for 5 min and fractionated by SDS-10% PAGE, followed by autoradiography and quantification, using a Typhoon-9500 Imager and ImageQuant (GE Healthcare).
Reticulocyte extract-based assays in which Rec8 f , its derivatives, and the f DHFR-HA-Ub K48R reference protein were detected by immunoblotting were carried out similarly, except that no [ 35 S]Met/Cys labeling was involved. The synthesis-deubiquitylation-degradation of URT-based test fusions was allowed to proceed for 1 h at 30°C either in the absence of added dipeptides or in the presence of either Arg-Ala (1 mM) or Ala-Arg (1 mM), followed by SDS-PAGE and immunoblotting with anti-FLAG antibody (Fig. 5E).
Histology and Immunohistochemistry-For histological analyses, testes and epididymis were fixed, after dissection, in 4% paraformaldehyde (formaldehyde, HCHO) (Solarbio (Beijing, China), P1110) overnight at 4°C, followed by standard procedures (79). Briefly, tissues were dehydrated through a series of ethanol washes and thereafter embedded in paraffin wax. Sections (5 m thick) were produced using a microtome. Deparaffinized and rehydrated sections were stained with hematoxylin and eosin for histological observations. For immunohistochemical assays, deparaffinized and rehydrated sections were rinsed three times at room temperature in PBS (pH 7.4), followed by antigen retrieval by boiling for 15 min in 10 mM sodium citrate, pH 6.0. Sections were then incubated for 10 min at room temperature with 3% H 2 O 2 , followed by blocking of each section by incubating it for 30 min at room temperature in 5% bovine serum albumin (BSA; Sigma). Sections were incubated with a primary antibody at 4°C overnight, followed by incubation with an HRP-conjugated secondary antibody at 37°C for 1 h. Sections were then stained with 3,3Ј-diaminobenzidine according to the manufacturer's instructions (Zhong Shan Jin Qiao, ZL1-9018), and nuclei were stained with hematoxylin. Negative controls were processed identically but with-out the primary antibody. Sections were examined using a Nikon 80i inverted microscope with a charge-coupled camera.
Spread of Spermatocyte Nuclei-Spermatocyte nuclei were prepared as described previously (81) with modifications. Briefly, testes were washed in PBS after dissection from mice. The tunicae were removed, and adherent extratubular tissues were removed by rinsing the seminiferous tubules with PBS at room temperature. The tubules were placed in a hypotonic extraction buffer (50 mM sucrose, 17 mM sodium citrate, 0.5 mM DTT, 0.5 mM PMSF, 3 mM Tris-HCl, pH 8.2) for 30 -60 min. Thereafter, ϳ1 inch of tubule was shredded to pieces by fine tipped forceps in 20 l of 0.1 M sucrose (pH adjusted to pH 8.2 by NaOH) on a clean glass slide. Another 20 l of sucrose solution was then added, and a slightly cloudy suspension was prepared, using a pipettor. Residual tubular tissues were removed. The suspension was transferred onto a new precoated (3-aminopropyl-triethoxysilane (Zhong Shan Jin Qiao, ZL1-9002)) glass slides, each of them containing on the surface 0.1 ml of freshly made (and filtered through a 0.22-m Rephile filter (Rephile (Shanghai, China), RJP3222SH)) solution of 1% paraformaldehyde (PFA), 0.15% Triton X-100 (with pH adjusted to 9.2 using 10 mM sodium borate, pH 9.2). Each slide was gently rocked to mix the initial suspension with PFA solution, followed by drying for at least 2 h in a closed box with high humidity. To stain the resulting nuclei spreads, slides were washed with 0.4% Photoflo (Eastman Kodak Co.) three times and with PBS three times and then blocked in 5% BSA for 1 h, incubated with a primary antibody in 1.5% BSA and 0.3% Triton X-100 overnight at 4°C, and then incubated with secondary antibody in PBS for 1 h at 37°C. The resulting slides were washed in PBS, and nuclei were stained with 4Ј,6-diamidino-2phenylindole (DAPI). Slides were examined using an LSM 780/ 710 microscope (Zeiss).
TUNEL Assays-TUNEL assays were carried out using the In Situ Cell Death Detection Kit (Roche Diagnostics, 11684795910) as described by us previously (82). Briefly, sections of testis were heated at 60°C for 2 h, followed by washing in xylene and rehydration through a graded series of washes with ethanol and double-distilled water. Thereafter, the sections were treated with proteinase K for 15 min at room temperature and rinsed twice with PBS. After adding the TUNEL reaction mixture, slides were incubated in a humidified atmosphere for 60 min at 37°C in the dark, followed by staining with DAPI (82). All experiments were repeated at least three times, with S.D. values shown.

Results
Germ Cell-specific Ablation of the Ate1 R-transferase Strongly Decreases Fertility of Male Mice-Probing sections of mouse testis with affinity-purified antibody to mouse Ate1 indicated the presence of the Ate1 R-transferase in the testis, particularly in spermatocytes and spermatogonia (precursors of spermatocytes) (Fig. 2A). These immunohistochemical results were in agreement with in situ hybridization data about Ate1 expression in spermatocytes (77). To produce mouse strains in which Ate1 was selectively ablated in primordial germ cells, we mated the previously constructed Ate1 flox/flox mice (33) with Tnap-Cre mice expressing Cre recombinase from the primordial germ cell-specific Tnap promoter (78) (see "Experimental Procedures"). Immunoblotting analyses of testis extracts from the resulting Tnap-Ate1 Ϫ/Ϫ mice versus Ate1 flox/flox (wild-type) mice with anti-Ate1 antibody indicated a dramatic decrease of Ate1 in Tnap-Ate1 Ϫ/Ϫ testes (Fig. 2B). Inasmuch as spermatogonia and spermatocytes (in which Tnap-Cre was selectively expressed) comprise a large fraction but not the entirety of testicular cells, these results (Fig. 2B) indicated that the Ate1 R-transferase was either completely or nearly completely absent from spermatogonia and spermatocytes of Tnap-Ate1 Ϫ/Ϫ mice.
Fertility of Tnap-Ate1 Ϫ/Ϫ and "wild-type" Ate1 flox/flox males was assessed by mating three males of each strain with wildtype females. For each male mouse, at least six plugged females were collected, and the pregnancy rates were recorded. Only ϳ9% of plugged females became pregnant after mating with Tnap-Ate1 Ϫ/Ϫ male mice, in comparison with a pregnancy rate of ϳ78% after mating with Ate1 flox/flox males (Fig. 2D). In addition, the average number of pups born to females that were mated with Tnap-Ate1 Ϫ/Ϫ males was only ϳ1.3, in contrast to ϳ7.3 pups that were born, on average, to Ate1 flox/flox females mated to Ate1 flox/flox males (Fig. 2E).
These results (Fig. 2) indicated that the Ate1 R-transferase was required for normal fertility levels in male mice. Given very low but still non-zero fertility of Ate1 Ϫ/Ϫ males (Fig. 2, D and E) as well as complete or nearly complete absence of the Ate1 R-transferase from germ cells in Tnap-Ate1 Ϫ/Ϫ testes (Fig. 2B), it is formally possible that the total absence of the Ate1-mediated arginylation in spermatogonia and spermatocytes is still compatible with a low but non-zero probability of sperm maturation. The alternative and a priori more likely interpretation is that the low but detectable sperm maturation that underlies the residual fertility of Tnap-Ate1 Ϫ/Ϫ mice (Fig. 2, D and E) is made possible by rare spermatocytes of Tnap-Ate1 Ϫ/Ϫ testes that retained at least one copy of the intact Ate1 flox gene.
The Absence of Arginylation Is Compatible with Early Stages of Germ Cell Development-Using an antibody to the promyelocytic leukemia zinc finger (Plzf), a spermatogonia-specific marker, we observed similar levels of Plzf in presumptive spermatogonia in either the absence or presence of Ate1 (ϳ56 Plzf- positive cells, on average, per section of seminiferous tubule in both Tnap-Ate1 Ϫ/Ϫ and Ate1 flox/flox mice) (Fig. 3, A and C). Formation of the synaptonemal complex and synapsis of chromosomes during the prophase of meiosis I in spermatocytes is accompanied by expression of the synaptonemal complex proteins 1 and 3 (Sycp1 and Sycp3) (83)(84)(85). Antibodies to Sycp1 and Sycp3 stained meiotic chromosomes indistinguishably in chromosome spreads of either Tnap-Ate1 Ϫ/Ϫ or Ate1 flox/flox spermatocytes (Fig. 3B). Similar immunofluorescence assays with Mlh1, a marker for chromosome crossovers (86 -89), also showed no significant differences between Tnap-Ate1 Ϫ/Ϫ and Ate1 flox/flox spermatocytes (Fig. 3, B, D, and E).
In addition, we could readily identify, cytologically, every major stage of the prophase of meiosis I in both Tnap-Ate1 Ϫ/Ϫ and Ate1 flox/flox spermatocytes, including leptotene, zygotene, pachytene, and diplotene, and there were no statistical significant differences between the two genotypes vis-á-vis the percentages of each stage of prophase I (Figs. 3 (E and F) and 4 (A  and B)). The average numbers of spermatocytes per seminiferous tubule section in the prophase of meiosis I were also similar between Tnap-Ate1 Ϫ/Ϫ and Ate1 flox/flox mice (ϳ61 and ϳ60 spermatocytes, respectively) (Fig. 3E). These results indicated that (at resolution levels of our assays) premetaphase stages of germ cell development did not require the Ate1 R-transferase.
Lack of Arginylation Causes Metaphase Arrest in Meiosis I-In contrast to the absence of detectable defects in the progression of Ate1-lacking Tnap-Ate1 Ϫ/Ϫ spermatocytes through the prophase of meiosis I, we found these cells to be arrested at the metaphase of meiosis I, followed by their death through apoptosis (Figs. 3 (G and H) and 4 (C, D, and F)). In stage XII seminiferous tubules, metaphase I spermatocytes could be identified by their highly condensed (hematoxylin-stained) chromatin (74) (Fig. 3, G and H). Both metaphase and later stage (anaphase) spermatocytes (indicated by arrowheads and arrows, respectively) could be observed in Ate1-containing tubules of Ate1 flox/flox testes (Fig. 3G). However, no anaphase spermatocytes could be detected in tubules of Ate1-lacking Tnap-Ate1 Ϫ/Ϫ testes, whose relative content of metaphase spermatocytes was ϳ9.9%, in contrast to ϳ2.6% of such cells in Ate1 flox/flox testes (Fig. 3, G and H). In agreement with these results, it was easy to detect spindle bodies marked by ␣-tubulin (consistent with the arrest of metaphase I spermatocytes) in Tnap-Ate1 Ϫ/Ϫ testes (Fig. 4E). We also examined, by immunoblotting, the expression of cyclin B1, a marker of metaphase (90). The level of cyclin B1 was significantly increased in Tnap-Ate1 Ϫ/Ϫ testes (Fig. 2J), yet another indication of metaphase arrest of Ate1-lacking spermatocytes in meiosis I.
Apoptotic Death of Arginylation-lacking Spermatocytes in Metaphase of Meiosis I-The terminal TUNEL assay was used to measure the extent of apoptosis of Tnap-Ate1 Ϫ/Ϫ versus Ate1 flox/flox spermatocytes (Fig. 4, C, D, and F). In Tnap-Ate1 Ϫ/Ϫ testes, on average ϳ33% of spermatocytes were overtly apoptotic (TUNEL-positive), in comparison with ϳ9% of such cells in Ate1-containing Ate1 flox/flox testes (Fig. 4C). In addition, only ϳ1.8 apoptotic cells/seminiferous tubule section were found, on average, in Ate1 flox/flox testes, versus ϳ8.3 apoptotic cells in Tnap-Ate1 Ϫ/Ϫ testes (Fig. 4F). Together, these results are likely to account for the observed massive decrease in the content of mature sperm cells in the testes of Ate1-lacking Tnap-Ate1 Ϫ/Ϫ males and the resulting very low fertility of these mice (Fig. 2, C-E). As described below, the metabolic stabilization of a natural fragment of the meiosis-specific Rec8 cohesin subunit is likely to be at least a significant, and possibly the major, reason for this functional consequence of Ate1 ablation in germ cells.
Arginylation-mediated Degradation of the Separase-produced Rec8 Fragment-Rec8 is the main meiosis-specific cohesin subunit of the kleisin family. Mouse Rec8 is sequelogous to both yeast and mammalian kleisin-type cohesin subunits (see Fig. 1C and Introduction). The cleavage of mouse Rec8 by the separase would be expected to generate a 15-kDa C-terminal fragment bearing N-terminal Glu, a secondary destabilizing residue and a substrate of the Ate1 R-transferase (Fig. 1, A and  C). Although separase can cleave mouse Rec8 in vitro at more than one location, the cleavage between Arg-454 and Glu-455 is by far the predominant one (Fig. 1C) (76).
To determine whether the separase-generated Rec8 fragment was a substrate of the Arg/N-end rule pathway, we used both steady-state and pulse-chase assays. Immunoblotting of extracts from wild-type and Tnap-Ate1 Ϫ/Ϫ mouse testes with antibody to a C-terminal region of Rec8 showed the presence of both the full-length endogenous Rec8 protein and its fragment. The latter species migrated, upon SDS-PAGE, at a position close to the one expected for the 15-kDa Glu 455 -Rec8 fragment (Figs. 1C and 5A, lane 2). Strikingly, however, whereas this endogenous Rec8 fragment was abundant in extracts from Tnap-Ate1 Ϫ/Ϫ testes, it was virtually absent in wild-type extracts (Fig. 5A, lane 1 versus lane 2).
A parsimonious interpretation of these results is that the separase-generated 15-kDa Glu 455 -Rec8 fragment (Fig. 1C) (76) was arginylated by the Ate1 R-transferase in wild-type cells and thereafter rapidly destroyed by the "downstream" part of the Arg/N-end rule pathway, whereas in Tnap-Ate1 Ϫ/Ϫ cells, the Glu 455 -Rec8 fragment was long-lived, because it could not be arginylated (Figs. 1A and 5A); hence, the virtual absence of the endogenous Glu 455 -Rec8 fragment in wild-type testes at steady state (Fig. 5A, lane 1) and its accumulation in Tnap-Ate1 Ϫ/Ϫ testes (Fig. 5A, lane 2). Interestingly, the level of the full-length Rec8 protein was also considerably higher in Tnap-Ate1 Ϫ/Ϫ testes than in wild-type ones, at equal total protein loads (see Fig. 5A and "Discussion").
Degradation of the Glu 455 -Rec8 fragment was also assayed directly, using 35 S-pulse-chases and the URT, derived from the Ub fusion technique (Fig. 5C) (4,21,22,91,92). Cotranslational cleavage of a URT-based Ub fusion by deubiquitylases that are present in all eukaryotic cells produces, at the initially equimolar ratio, both a test protein with a desired N-terminal residue and the reference protein f DHFR-Ub R48 , a FLAG-tagged derivative of the mouse dihydrofolate reductase (Fig. 5C). In URTbased pulse-chase assays, the labeled test protein is quantified by measuring its levels relative to the levels of a stable reference at the same time point during a chase. In addition to being more accurate than pulse-chases without a built-in reference, URT also makes it possible to detect and measure the degradation of the test protein before the chase (i.e. during the pulse) (17,91,92).
URT-based 35 S-pulse-chases with C-terminally FLAGtagged Glu 455 -Rec8 f and its derivatives were performed in a transcription-translation-enabled rabbit reticulocyte extract, which contains the Arg/N-end rule pathway and has been extensively used to analyze this pathway (4,21,22). The indicated f DHFR-Ub R48 -X 455 -Rec8 f URT fusions (where X represents Glu, Val, or Arg-Glu) were labeled with [ 35 S]Met/Cys in reticulocyte extract for 10 min at 30°C, followed by a chase, immunoprecipitation with a monoclonal anti-FLAG antibody, SDS-PAGE, autoradiography, and quantification (Fig. 5, B-D). The logic of these assays involves a comparison between the degradation rates of a protein bearing a destabilizing N-terminal residue and an otherwise identical protein with an N-terminal residue, such as Val, which is not recognized by the Arg/Nend rule pathway (Fig. 1A).
The Glu 455 -Rec8 f fragment was short-lived in reticulocyte extract (t1 ⁄ 2 of 10 -15 min) (Fig. 5, B and D). Moreover, ϳ30% of pulse-labeled Glu 455 -Rec8 f was degraded during the pulse (i.e. before the chase) in comparison with the otherwise identical Val 455 -Rec8 f , which was also completely stable during the chase (Fig. 5, B and D). We also constructed and examined Arg-Glu 455 -Rec8 f . This protein was a DNA-encoded equivalent of the posttranslationally Nt-arginylated Glu 455 -Rec8 f fragment of Rec8. As would be expected, given the immediate (cotranslational) availability of N-terminal Arg in the DNA-encoded Arg-Glu 455 -Rec8 f , this protein was destroyed by the Arg/N-end Procedures"). Also indicated in D and E, are the molecular masses of key protein species, the C-terminally FLAG-tagged Rec8 fragment (16 kDa) and the N-terminally FLAG-tagged reference protein DHFR-Ub (33 kDa). An asterisk denotes a protein band that cross-reacted with anti-FLAG antibody. rule pathway even more rapidly than the already short-lived Glu 455 -Rec8 f . Specifically, nearly 80% of Arg-Glu 455 -Rec8 f was eliminated during the 10-min pulse (before the chase), in comparison with ϳ30% of Glu 455 -Rec8, with the long-lived Val 455 -Rec8 f control being a part of the reference set (Fig. 5, B and D).
The results of pulse-chase analyses (Fig. 5, B-D) were in agreement with other measurements, in which the synthesisdeubiquitylation-degradation of f DHFR-Ub R48 -X-Rec8 f fusions in reticulocyte extract was allowed to proceed for 1 h, followed by detection of Glu 455 -Rec8 f , of other test proteins, and of the f DHFR-Ub R48 reference protein by SDS-PAGE and immunoblotting with anti-FLAG antibody (Fig. 5E). In these assays, the samples were incubated in reticulocyte extract either without added dipeptides or with 1 mM Arg-Ala (RA), bearing N-terminal Arg, a type 1 primary destabilizing residue, or with 1 mM Ala-Arg (AR), bearing N-terminal Ala, a residue that is not recognized by N-recognins of the Arg/N-end rule pathway (Figs. 1A and 5E).
Ubr1 and Ubr2, the two sequelogous (47% identical) and functionally overlapping 200-kDa N-recognins (E3 Ub ligases) of the mammalian Arg/N-end rule pathway, have several substrate-binding sites. These sites recognize (bind to) specific classes of N-degrons and specific internal (non-N-terminal) degrons. The two substrate-binding sites that recognize N-degrons are the type 1 site, which specifically binds to the N-terminal basic residues Arg, Lys, or His, and the adjacent but distinct type 2 site, which specifically binds to the N-terminal bulky hydrophobic residues Leu, Phe, Tyr, Trp, and Ile (4,6,93,94). Dipeptides bearing, for example, type 1 N-terminal residues can competitively and selectively inhibit the binding of Ubr1/Ubr2 to a type 1 N-degron but not to a type 2 N-degron in a test protein (4,6,(95)(96)(97).
In agreement with the rapid degradation of the Glu 455 -Rec8 f fragment in 35 S-pulse-chase assays (Fig. 5, B-D), the synthesisdeubiquitylation-degradation of the f DHFR-Ub R48 -Glu 455 -Rec8 f fusion for 1 h in reticulocyte extract either in the absence of added dipeptide or in the presence of Ala-Arg (which does not bind to Ubr1/Ubr2) did not result in a detectable accumulation of the Glu 455 -Rec8 f fragment, at its expected (roughly 15 kDa) position in SDS-PAGE-based immunoblots (Fig. 5E, lanes  8 and 10). In striking contrast, the same assay but in the presence of the Arg-Ala dipeptide (which competitively inhibits the recognition of type 1 N-degrons) resulted in a prominent protein band at the expected position of Glu 455 -Rec8 f (Fig. 5E, lane  9 versus lanes 8 and 10). Given the URT-based design of f DHFR-Ub R48 -Glu 455 -Rec8 f (Fig. 5C), that protein band was inferred to be the Glu 455 -Rec8 f fragment that had been metabolically stabilized by the Arg-Ala dipeptide. In contrast, the larger Met 206 -Rec8 f fragment, corresponding to a putative (and at most a minor) separase cleavage site in the full-length Rec8 protein, was metabolically stable irrespective of the presence or absence of the Arg-Ala dipeptide (Fig. 5E, lanes 5-7 versus lanes 8 -10). The same was true of full-length Rec8 f . It should also be mentioned that full-length Rec8 f was not converted into a smaller fragment in reticulocyte extract, indicating (as would be expected) the absence of active separase in that extract (Fig. 5E,  lanes 2-4).

Discussion
The arginyltransferase (R-transferase) Ate1 is a component of the Arg/N-end rule pathway of protein degradation. Ate1 utilizes Arg-tRNA as a cosubstrate to arginylate N-terminal Asp, Glu, or (oxidized) Cys of a targeted protein substrate. The resulting N-terminal Arg is recognized by E3 ubiquitin ligases (N-recognins) of the Arg/N-end rule pathway (see Fig. 1A and Introduction). In the present study, we constructed an Ate1 Ϫ/Ϫ mouse strain in which the ablation of Ate1 was confined to germ cells. We used this and other experimental tools to characterize Glu 455 -Rec8, a specific C-terminal fragment of the kleisin-type, meiosis-specific Rec8 subunit of mouse cohesin. Rec8 is cleaved, late in meiosis I (see Introduction), by a nonprocessive protease called separase. The main C-terminal fragment of that cleavage, Glu 455 -Rec8, bears N-terminal Glu, a substrate of Ate1 (Fig. 1, A and C). We have shown that the mouse Glu 455 -Rec8 fragment is a short-lived substrate of the Arg/N-end rule pathway and that the degradation of this fragment requires its Nt-arginylation by the Ate1 R-transferase (Fig. 5).
In S. cerevisiae, Scc1/Rad21/Mcd1 is the mitotic counterpart of the mammalian meiotic Rec8 subunit of cohesin. Similarly to the mouse Glu 455 -Rec8 fragment, the separase-generated C-terminal fragment of yeast Scc1 is also a short-lived substrate of the Arg/N-end rule pathway, and the failure to destroy this fragment in ubr1⌬ cells (which lack the Arg/N-end rule pathway) results in chromosome instability (62). The separase-generated C-terminal fragment of yeast Scc1 retains, in part, the physical affinity of Scc1 for the rest of the cohesin complex (62). The mouse Glu 455 -Rec8 fragment would also be likely to interact in vivo with the rest of meiotic cohesin. If so, the failure to arginylate the Glu 455 -Rec8 fragment in arginylation-lacking spermatocytes of Ate1 Ϫ/Ϫ mice and hence the failure to destroy this fragment (Figs. 1A and 5) would be expected to interfere with cohesin mechanics. This (at present hypothetical) interference would account, at least in part, for the observed arrest and apoptotic death of Ate1 Ϫ/Ϫ spermatocytes at the end of meiosis I and the resulting strong decrease in the fertility of Ate1 Ϫ/Ϫ males (Fig. 2, C-L).
Immunoblotting analyses of the arginylation-dependent degradation of the endogenous Glu 455 -Rec8 fragment showed that this fragment was virtually absent, at steady state, in wild-type mouse testes but accumulated in Ate1 Ϫ/Ϫ testes (Fig. 5A). Interestingly, the steady-state level of the full-length Rec8 protein was also significantly higher in Ate1 Ϫ/Ϫ testes than in wild-type ones at equal total protein loads (Fig. 5A). A plausible but unproven interpretation of this result is that the cleavage of the full-length mouse Rec8 by separase may be subject to a product-mediated inhibition of separase in Ate1 Ϫ/Ϫ spermatocytes if the product, Glu 455 -Rec8, is no longer eliminated by the Ate1-dependent arginylation branch of the Arg/N-end rule pathway (Figs. 1A and 5).
Although selective ablation of Ate1 in mouse germ cells nearly abrogates the fertility of male mice (Fig. 2, C-E), it is unlikely that specific cases of human infertility can be caused by unconditionally null mutations of human Ate1, inasmuch as global mouse Ate1 Ϫ/Ϫ mutants are late embryonic lethals (29). In addition, a post-natal ablation of mouse Ate1, in adult mice (using cre-lox technology and a ubiquitously expressed Cre recom-binase), although compatible with mouse viability, causes a variety of abnormal phenotypes, including the loss of fat and hyperkinetic behavior (33). Nevertheless, partially active (hypomorphic) mutants of the human Ate1 R-transferase might underlie some, currently obscure, cases of human infertility.
The same disposition would obtain if the human Arg/N-end rule pathway were to be partially inactivated "downstream" of the Ate1 R-transferase, at the level of the pathway's Ub ligases (Fig. 1A). For example, unconditional Ubr1 Ϫ/Ϫ mice, lacking one of two major N-recognins, Ubr1 and Ubr2, are viable and fertile while exhibiting some abnormal phenotypes (98). The analogous Ubr2 Ϫ/Ϫ mice, lacking the second major N-recognin (it is structurally and functionally similar to Ubr1), are also viable (in some strain backgrounds) but exhibit male infertility (77). This infertility is similar to the low fertility phenotype of Ate1 Ϫ/Ϫ mice in the present study, because in both cases the infertility is caused by apoptotic death, in meiosis I, of either Ubr2 Ϫ/Ϫ or Ate1 Ϫ/Ϫ spermatocytes (Figs. 2 (C and D) and 4 (C, D, and F)). A parsimonious interpretation of these results is that a failure to rapidly destroy the separase-generated Glu 455 -Rec8 fragment is a common mechanistic denominator of both Ate1 Ϫ/Ϫ and Ubr2 Ϫ/Ϫ infertility phenotypes.
No human Ubr2 Ϫ/Ϫ mutants have been identified so far. In contrast, human patients with the previously characterized Johansen-Blizzard syndrome have been shown to be null Ubr1 Ϫ/Ϫ mutants (99). It is unknown whether or not human Johansen-Blizzard syndrome patients are fertile, in part because the overall phenotype of human Johansen-Blizzard syndrome is more severe than the analogous phenotype of Ubr1 Ϫ/Ϫ mice. Abnormal phenotypes of human Johansen-Blizzard syndrome (Ubr1 Ϫ/Ϫ ) patients include anatomical malformations, an insufficiency and inflammation of the acinar pancreas, mental retardation, and deafness (4,99).
Although separase can cleave mouse Rec8 in vitro at more than one site, the cleavage between Arg-454 and Glu-455, resulting in the Glu 455 -Rec8 fragment, is by far the predominant one (Fig. 1C) (76). Mammalian Rad21, the mitosis-specific counterpart of the meiotic Rec8 cohesin subunit, is also cleaved by separase, late in mitosis. Similarly to the cleavage of the meiosis-specific Rec8, the separase-mediated cleavage of the mammalian mitotic Rad21 subunit also yields the C-terminal fragment of Rad21 bearing N-terminal Glu (75). However, in contrast to the present results with the Glu 455 -Rec8 fragment of meiotic Rec8 (Fig. 5, B and D), our recent analyses of the N-terminal Glu-bearing mitotic Rad21 fragment using 35 S-pulse-chases and the URT method (Fig. 5C) indicated that this fragment was long-lived. 7 Further analyses of these unexpected (and therefore particularly interesting) results regarding the apparently stable mitotic Glu-Rad21 fragment vis-à-vis the short-lived meiotic Glu 455 -Rec8 fragment (Fig. 5)