The EMAPII Cytokine Is Released from the Mammalian Multisynthetase Complex after Cleavage of Its p43/proEMAPII Component*

Endothelial-monocyte-activating polypeptide II (EMAPII) is an inflammatory cytokine released under apoptotic conditions. Its proEMAPII precursor proved to be identical to the auxiliary p43 component of the aminoacyl-tRNA synthetase complex. We show here that the EMAPII domain of p43 is released readily from the complex after in vitro digestion with caspase 7 and is able to induce migration of human mononuclear phagocytes. The N terminus ofin vitro-processed EMAPII coincides exactly with that of the mature cytokine isolated from conditioned medium of fibrosarcoma cells. We also show that p43/proEMAPII has a strong tRNA binding capacity (K D = 0.2 μm) as compared with its isolated N or C domains (7.5 μm and 40 μm, respectively). The potent general RNA binding capacity ascribed to p43/proEMAPII is lost upon the release of the EMAPII domain. This suggests that after onset of apoptosis, the first consequence of the cleavage of p43 is to limit the availability of tRNA for aminoacyl-tRNA synthetases associated within the complex. Translation arrest is accompanied by the release of the EMAPII cytokine that plays a role in the engulfment of apoptotic cells by attracting phagocytes. As a consequence, p43 compares well with a molecular fuse that triggers the irreversible cell growth/cell death transition induced under apoptotic conditions.

In higher eukaryotic organisms, from Drosophila to mammals, the nine aminoacyl-tRNA synthetases specific for amino acids Glu, Pro, Ile, Leu, Met, Gln, Lys, Arg, and Asp are associated within a multienzyme complex containing three auxiliary proteins, as well (1). The p38 auxiliary component contributes a scaffold protein for the assembly of the complex (2,3). The p18 subunit of the complex might be an anchor for transient association of elongation factor EF-1H (4). The p43 subunit is an RNA-binding protein (5) based on a classical oligonucleotide-oligosaccharide binding fold (6, 7) that might play a role of a cofactor for aminoacylation (8). Whereas the p18 and p38 proteins are always recovered as components of the multisynthetase complex, p43 or p43-like domains are widespread in evolution and distributed in the three kingdoms of the tree of life. They have been described as polypeptide appendices of MetRS 1 (8), PheRS, or TyrRS (9) or as discrete proteins interacting with aminoacyl-tRNA synthetases and/or tRNAs, Trbp in bacteria (10,11), Arc1p in yeast (12), or p43 in ciliated protozoan (13) or in metazoan species (5,7).
Unexpectedly, a human protein homologous to the C-terminal moiety of hamster p43 was reported to have cytokine activities (14,15). The endothelial-monocyte-activating polypeptide II (EMAPII) has been isolated from methylcholanthrene A-induced fibrosarcoma cells. EMAPII is a proinflammatory cytokine that stimulates chemotactic migration of polymorphonuclear granulocytes and mononuclear phagocytes and induces tissue factor activity on endothelial cells. The C-terminal domain of human or bovine TyrRS, which is related to EMAPII, displays identical cytokine activities (16,17). EMAPII is expressed constitutively in all cell types as a ϳ35-kDa precursor and is further processed to an ϳ18-kDa mature form upon onset of apoptosis (18,19). Consistent with its maturation under apoptotic conditions, the precursor polypeptide obtained by in vitro transcription was shown to be a substrate for apoptotic proteases of the caspase family (20).
Because the p43 component of the multisynthetase complex is identical to proEMAPII, we wondered whether EMAPII could be processed from its complex-associated precursor. To address the consequences of p43 cleavage, we equally considered its involvement as a general RNA binding domain or as a cytokine. The results provide strong evidence for a dual role of p43 that can be identified with a molecular fuse. The function of p43 as a cofactor of aminoacyl-tRNA synthetases is lost upon cleavage and release of EMAPII, its C-domain with cytokine activity. Because proEMAPII is p43, a protein involved in translation and therefore ubiquitous to all cell types and tissues, previous reports dealing with the relative abundance of EMAPII mRNA or protein should be considered cautiously. They address, as well, the level of p43 in those tissues where active protein synthesis is required for tissue remodeling.

MATERIALS AND METHODS
Purification of the Multisynthetase Complex from Mouse Liver-Livers (160 g) from 120 mice were homogenized in a Waring Blendor (2 ϫ 15 s) after addition (1 ml per g) of extraction buffer (50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 0.1 mM EDTA, 1 mM DTT, and 10% glycerol) containing protease inhibitors (1 mM diisopropyl fluorophosphate, 2 mM phenylmethylsulfonyl fluoride). Extract was cleared by centrifugation at 30,000 ϫ g for 30 min and subjected further to high speed centrifugation at 260,000 ϫ g for 2 h. The supernatant was applied on a 1,700-ml (5 ϫ 85 cm) Sephacryl S-400 HR column (Amersham Pharmacia Biotech) equilibrated in Buffer A (75 mM potassium phosphate buffer, pH 7.5, 10 mM 2-mercaptoethanol, and 10% glycerol) and developed at a flow rate of 4 ml/min. Fractions with LysRS activity were combined and applied on a 30-ml (1.6 ϫ 15 cm) tRNA-Sepharose column (21) developed at a flow rate of 1 ml/min. The complex was eluted with a linear gradient of potassium phosphate buffer (75 to 350 mM). After a 4-fold dilution with a solution containing 2 mM DTT and 10% glycerol, fractions were applied on a 1-ml Resource Q column (Amersham Pharmacia Biotech) equilibrated in Buffer B (25 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM DTT, 10% glycerol) and developed at 1 ml/min with a linear gradient of KCl from 50 to 500 mM. Fractions were dialyzed against 25 mM potassium phosphate, pH 7.5, 2 mM DTT, and 55% glycerol and stored at Ϫ20°C.
Expression and Purification of p43 and of Derivatives Thereof-Human p43 (h-p43), as well as its N-terminal (h-p43N) or C-terminal domains (h-p43C; EMAPII), were expressed in Escherichia coli with the pET-28b expression system (Novagen). The h-p43 cDNA was produced by polymerase chain reaction between the two oligonucleotides p43-3 (5Ј-cccccatggcaaataatgatgctgttctgaagagac) and p43-2 (5Ј-cccctcgagtttgattccactgttgctca), which introduced an NcoI site with the ATG initiation codon and an XhoI site in frame with the vector sequences encoding the His tail, respectively. The cDNA encoding the equivalent region of murine p43 (m-p43) was also introduced into the pET28b vector after amplification between oligonucleotides p43-m1 (5Ј-cccccatggcaacgaatgatgctgt) and p43-m2 (5Ј-cccctcgagtttaattccactattggccat). The h-p43N cDNA was amplified between oligonucleotides p43-3 and p43-160 (5Јcccctcgaggtcggcacttccagctattga). The h-p43C construct has been described previously (5). A derivative without a His tail at the C terminus was constructed after amplification between p43-1 (5Ј-cccccatggccaagccaatagatgtttccc) and p43-20 (5Ј-cccctcgagttatttgattccactgttgctca), which introduced a stop codon. The nucleotide sequence of the constructs was checked by DNA sequencing.
Expression and purification of p43 variants bearing a His tag was conducted essentially as described previously (5), following two chromatographic steps on a nickel-nitrilotriacetic acid Superflow matrix (Qiagen) and on a SOURCE 15S column (Amersham Pharmacia Biotech.). Purified proteins were stored at Ϫ20°C after dialysis against 20 mM Tris-HCl, pH 7.0, 50 mM NaCl, 2 mM DTT, 55% glycerol.
For purification of h-p43C without His tag, cell extract obtained as described above in 20 ml of 30 mM Tris-HCl, pH 7.0, 30 mM KCl, 0.1 mM EDTA, 2 mM DTT, 10% glycerol was applied to a 1.6 ϫ 13-cm column of S-Sepharose Fast Flow (Amersham Pharmacia Biotech) equilibrated in 20 mM Tris-HCl, pH 7.0, 30 mM NaCl, 1 mM DTT. The protein was eluted (at 300 mM) by a linear gradient of NaCl (30 to 500 mM) in the same buffer. To remove contaminating nucleic acids, fractions were dialyzed against 20 mM Tris-HCl, pH 7.0, 15 mM NaCl, 1 mM DTT and applied to a 2.0 ϫ 9.5-cm SOURCE 15Q column equilibrated in the same buffer. The material recovered in the flow-through fraction was concentrated on a 1-ml Resource 15S column, eluted stepwise with 500 mM NaCl, and stored at Ϫ20°C in 25 mM potassium phosphate buffer, pH 7.5, 2 mM DTT, and 55% glycerol. Protein concentrations were determined by using calculated absorption coefficients of 0.257, 0.260, 0.068, and 0.432 A 280 units/mg Ϫ1 /cm 2 , respectively, for h-p43, m-p43, h-p43N, and h-p43C.
Purification of Mouse EMAPII after Caspase 7 Digestion of the Complex-Mouse complex (500 g) was digested with caspase 7 (10,250 units; 247,600 units/mg (22)) in 6 ml of CFS buffer containing 0.01% Tween 20. After 90 min of incubation at 37°C, the mixture was placed on ice, diluted with 6 ml of 2-fold concentrated Buffer A (Buffer A is 20 mM Tris-HCl, pH 7.5, 15 mM NaCl, 1 mM DTT, 0.01% Tween 20, 10% glycerol), and applied on a 1-ml Resource Q column equilibrated in Buffer A. The flow-through fraction was applied to a 0.8-ml Mini-S column equilibrated in Buffer B (20 mM Tris-HCl, pH 7.5, 2 mM DTT, 0.01% Tween 20) and developed with a linear gradient of NaCl from 0 to 350 mM. EMAPII, eluted at an NaCl concentration of 120 mM, was concentrated by ultrafiltration on MICROSEP (Pall Filtron; 3-kDa molecular mass cutoff). The N-terminal sequence of EMAPII was determined by automated Edman degradation using an Applied Biosystems 473 sequencer.
Preparation of Antibodies and Western Blotting-Polyclonal antibodies were raised in rabbit against purified h-p43N and h-p43C following repeated injections at 2-week intervals of 0.5 mg of homogeneous protein emulsified with complete (first injection) or incomplete Freund's adjuvant. Other antibodies directed to murine EMAPII (18) or to components of the multisynthetase complex (23) were as described. After SDS-PAGE (24), Western blotting was conducted essentially as described (25), using polyvinylidene difluoride transfer membranes (Hybond-P; Amersham Pharmacia Biotech), goat anti-rabbit IgG conjugated with peroxidase, and the ECL detection reagents.
Monocyte Chemotactic Assay-Monocytes were isolated from buffy coats of healthy donors in a two-step procedure by density-gradient centrifugation and elutriation. Peripheral blood mononuclear cells were obtained from citrated venous blood (buffy coats) from healthy donors according to the method of Boyum (26). Blood was diluted 1:3 in Hanks' buffered saline, and 25 ml of this cell suspension was layered over 15 ml of Ficoll-Paque (Amersham Pharmacia Biotech). After centrifugation at 600 ϫ g for 20 min cells at the interphase containing peripheral blood mononuclear cells were collected, and cell suspension was washed twice with Hanks' buffered saline to remove platelets. For further monocyte isolation cells were resuspended in 15 ml of Hanks' buffered saline and placed into the sample tube of a JE-6B elutriator rotor (Beckman). At a constant rotor speed of 2300 rpm flow rate was increased successively from 7 ml/min to 18 ml/min. Monocytes were reproducibly eluted in fractions 5 to 9, containing more than 90% monocytes (CD14 ϩ cells) as determined by fluorescence-activated cell sorter analysis with an anti-CD14 monoclonal antibody. After isolation cells were centrifuged (1000 rpm; 5 min), and monocytes were cultivated not longer than 24 h in serum-free macrophage medium, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin, at a concentration of 2 ϫ 10 6 cells/ml. To avoid cell adhesion to the plastic surface hydrophobic culture plates (In Vitro Systems) were used.
Chemotaxis of monocytes was investigated using the method of Quinn et al. (27) as modified recently (14). Briefly, isolated monocytes were placed in the upper chamber, and the test substances were placed in the lower chamber. Chemotactic assays were performed for 1 h of incubation, unmigrated cells on the filter surface were removed, and migrated cells were stained with Giemsa stain. Cells in at least six power fields were counted for each condition assessed.
Sedimentation Equilibrium-Ultracentrifugation experiments were conducted and analyzed as described previously (8,28). Sedimentation equilibrium data were fitted to theoretical models taking into account molecular weights of 35,418, 17,218, and 19,398, respectively, for the monomers of p43, p43N, and p43C and v values of 0.735, 0.734, and 0.733 at 4°C.
Aminoacylation Assay-Initial rates of tRNA aminoacylation were measured at 25°C in 0.1 ml of 20 mM imidazole-HCl buffer, pH 7.5, 150 mM KCl, 0.5 mM DTT, 5 mM MgCl 2 , 3 mM ATP, 60 M 14 C-labeled amino acid (50 Ci/mol; PerkinElmer Life Sciences), and saturating amounts of tRNA, as previously described (29). Total brewers' yeast tRNA (Roche Molecular Biochemicals) or partially purified beef tRNA were used as tRNA substrates. The incubation mixture contained catalytic amounts (1 to 10 nM) of enzymes appropriately diluted in 10 mM Tris-HCl, pH 7.5, 10 mM 2-mercaptoethanol containing bovine serum albumin at 4 mg/ml. One unit of activity is the amount of enzyme producing 1 nmol of aminoacyl-tRNA/min at 25°C.

RESULTS
The p43 Component of the Multisynthetase Complex Is a Substrate of Caspase 7-The p43 component of the multisynthetase complex is the homolog of proEMAPII, the precursor of the EMAPII cytokine (5). Previous studies showed that [ 35 S]labeled mouse proEMAPII produced in an in vitro transcription/translation system is a substrate of caspase 7 and, to a lesser extent, of caspase 3 (20). Because the only known cellular species of p43 is that associated within the multisynthetase complex (5), we surmised that if p43 is proEMAPII, the precursor of the cytokine, the caspase-7 cleavage site on p43 should be readily accessible in the complex. On the contrary, if the p43 component of the complex would be protected against caspase digestion, it could not be the direct precursor of EMAPII.
Homogeneous mouse multisynthetase complex was incubated in the presence of mouse caspase 7, and time course of p43 digestion was monitored by Western blotting with anti-EMAPII antibodies directed to the C-terminal moiety of p43 (Fig. 1). After 90 min of incubation, the p43 polypeptide (80 ng of protein taking into account that 2 molecules of p43 (2 ϫ 40 kDa) are associated per molecule of complex (1.5 MDa)) was almost entirely converted into EMAPII, an 18-kDa polypeptide. In contrast, after a 90-min digestion with caspase 7 of recombinant mouse p43 (60 ng of the dimeric protein) expressed in E. coli (see below), only ϳ50% of the p43 polypeptide was converted into EMAPII. We concluded that the p43 component of the complex is a substrate of caspase 7 and that its association within the complex facilitates its cleavage. Caspase 3 or caspase 8 did not efficiently cleave the complex-associated form of p43 or the isolated subunit (not shown).
To test the behavior of the other components of the complex toward caspase 7 treatment, the multisynthetase complex subjected to controlled proteolysis was analyzed by SDS-PAGE followed by Coomassie staining (Fig. 2) and Western blotting (see Fig. 2 and below) using antibodies directed to individual components. Complete analysis revealed that IleRS, LeuRS, MetRS, ArgRS, p38, and probably GlnRS were not affected by the addition of caspase 7. In contrast, GluProRS (163 kDa), LysRS (68 kDa), and AspRS (57 kDa), in addition to p43, were cleaved by caspase 7 and converted to polypeptides of 105 ϩ 60 kDa for GluProRS, 66 and 61 kDa for LysRS, and 55 kDa for AspRS. As shown in Fig. 2, when Z-DEVD-CMK, a potent and specific inhibitor of caspase 7, was added in the incubation mixture, cleavage of p43, GluProRS, LysRS, and AspRS was completely abolished.
Caspase 7 Treatment Releases EMAPII from the Complex-If EMAPII is a cleavage product of the complex-associated form of p43, then it should be readily released from the complex after caspase 7 treatment. The mouse complex was incubated with caspase 7 as described above and subjected to size fractionation on a Superose 12 column (Fig. 3). A major peak was eluted with an apparent mass corresponding to the native complex (fraction B). Two additional minor peaks were observed corresponding to the elution volumes of proteins of ϳ100 kDa (fraction C) and of small proteins eluting near the inclusion volume of the column (fraction E). Fraction F corresponded to the elution of components of the incubation buffer in the total bed volume of  the column. Column fractions were analyzed by Western blotting to investigate the association state of the native and truncated components of the complex after caspase treatment. By using antibodies directed to the N-terminal or C-terminal domain of p43, we identified a ϳ16or 18-kDa polypeptide, respectively, in fractions B or E. Therefore, caspase 7 digestion of p43 led to two discrete polypeptides. The N-terminal polypeptide behaves as a complex-associated entity, and EMAPII is released as a soluble monomeric protein.
Concerning the other components of the complex, only ProRS is released from the complex after caspase treatment. It is recovered in fraction C (Fig. 3B) as a dimeric protein with an apparent mass of ϳ100 kDa (Fig. 3A). The N-terminal moiety of bifunctional GluProRS, including GluRS and the linker region made of three repeated units (as assessed by Western blotting with antibodies directed to the GluRS domain or to the repeated units), remained associated with the complex. Similarly, the removal of a small polypeptide from LysRS and AspRS did not impair their association with the other components of the complex.
The consequences of caspase 7 treatment of the complex on the activity of its enzymatic components were appraised by measuring their tRNA aminoacylation capacity in the presence of saturating amounts of crude yeast or beef tRNA. Using these assay conditions, only ProRS activity was shown to be affected, with a 40% reduced initial velocity. This partial loss of activity could be related to the instability of the dimeric enzyme, which proved to readily dissociate into inactive monomers when released from the bifunctional GluProRS (30).
The EMAPII Domain Released by Caspase 7 Has Cytokine Activities-To establish that the EMAPII polypeptide initially isolated by Stern and co-workers (14) from conditioned medium of murine methylcholanthrene A fibrosarcoma cells does indeed correspond to the C-terminal domain of p43, substantial amounts of this polypeptide were isolated to perform its structural and functional characterization. Preparative digestion of the murine multisynthetase complex (500 g of complex) was conducted with 10,000 units of caspase 7. Because two monomers of EMAPII (mass of the monomer ϭ 18 kDa) could be isolated by molecules of complex (1500 kDa), a maximum of 12 g of EMAPII were expected. To prevent absorption of this polypeptide on glassware, 0.01% Tween 20 was included in all buffers. EMAPII was isolated to homogeneity following two chromatographic steps on Resource Q and Mini-S columns and concentrated by ultrafiltration. Analysis by SDS-PAGE revealed a single polypeptide of 18 kDa (Fig. 4B), and a single N-terminal amino acid sequence was obtained by Edman degradation, XKPIDA . . . (the identity of the first amino acid residue was not determined). It precisely matches the sequence of the mouse p43 protein starting from residue 145, following the Asp 144 residue from the ASTD sequence, corresponding to the caspase 7 cleavage site. It also coincides with the N-terminal sequence of the EMAPII cytokine initially isolated from medium of fibrosarcoma cells (14).
EMAPII has been shown to induce the migration of mononuclear phagocytes (MPs) and polymorphonuclear leukocytes FIG. 3. Probing the behavior of individual components of the complex after cleavage by caspase 7. A, purified multienzyme complex from mouse (75 g) was subjected to caspase 7 treatment (2200 units) followed by size fractionation on a Superose 12 HR 10/30 column equilibrated in 25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 1 mM DTT. Elution was monitored at 220 nm. According to the elution profile, fractions were combined to give samples A and B, corresponding to complex-associated proteins, and C-F, corresponding to free species. B, samples A-F were concentrated by ultrafiltration on Centrisart-C4 (10,000 molecular weight cutoff; Sartorius) and analyzed by Western blotting using antibodies directed to the p43C or p43N region of p43, to the repeated units (GluRS) or ProRS domain (ProRS) of GluProRS, and to LysRS, AspRS, MetRS, ArgRS, or p38. Lanes T and T ϩ corresponded to control samples of the multisynthetase complex before (T) or after (T ϩ ) caspase treatment, recovered before size exclusion chromatography. FIG. 4. Expression of p43 and p43-related polypeptides. A, the murine multisynthetase complex (m-Cx), corresponding to the natural form of p43/proEMAPII, the recombinant murine (m-p43) or human (h-p43) p43 polypeptides produced in E. coli with a C-terminal His tag, the (B) p43N and p43C and p43C-H segments of human p43 produced in E. coli with (p43N and p43C) or without (p43C-H) a C-terminal His tag, and the EMAPII polypeptide isolated after in vitro caspase treatment of the mouse multisynthetase complex (EMAPII) were analyzed by SDS-PAGE on a 12 (A) or 15% (B) polyacrylamide gel and visualized by Coomassie staining. and to stimulate the production of tissue factor by MPs and the release of myeloperoxidase from polymorphonuclear leukocytes. The cytokine activity of the EMAPII product derived from the multisynthetase complex by in vitro processing was assessed by its capacity to induce migration of human MPs (Fig. 5). EMAPII derived from the multisynthetase complex induced migration of MPs in a biphasic dose-response manner typical for chemokines. The chemotactic response was already significant at a concentration as low as 10 pg/ml, reached a maximum at about 1 ng/ml, and decreased to less significant values at an EMAPII concentration of 100 ng/ml. This dose response is very similar to that originally observed with EMAPII purified from tumor cell supernatants (14). When used at a protein concentration of 1 ng/ml, the recombinant cytokine (p43Ct) displayed a similar chemotactic activity (Fig. 5). The recombinant native human proEMAPII produced in E. coli (h-p43) also induced migration of MPs (results not shown).
The N and C Domains of p43 Contribute a Bipartite tRNA Binding Site-We previously reported that the C-terminal domain of p43 is a monomer and displays a weak nonspecific tRNA binding capacity (K D ϭ ϳ40 M; see Ref. 5). To determine whether cleavage of p43 into EMAPII could modify the tRNA binding property of p43, we expressed and purified to homogeneity different forms of p43 from different origins. Murine and human p43 (m-p43 and h-p43) (Fig. 4A), as well as the human p43N and p43C moieties corresponding to the two polypeptides

FIG. 5. EMAPII derived from the multisynthetase complex induces chemotactic migration of monocytes.
Peripheral blood monocytes were added to the upper compartment of a modified Boyden chamber. Samples of EMAPII derived from the complex after cleavage with caspase 7 (EMAPII), of recombinant EMAPII (p43C), and of medium alone were added to the lower compartments at the indicated concentrations in ng/ml. The chambers were incubated for 1 h at 37°C and then migrated cells were stained and counted. Results show mean Ϯ S.E. of migrated cells from at least three independent experiments. In each experiment cells were counted in at least six representative high power fields. The statistical significance of the observed values was evaluated on the basis of the number of cells in the individual experiments as calculated from an unpaired t test using the InStat 2.01 program. generated by caspase 7 treatment (Fig. 4B), were expressed in E. coli with a C-terminal His tag. A C-terminal fragment deprived of a His tag was also isolated (p43C-H).
A double-hybrid screen of interactions between components of the complex showed that p43 associated with itself (2). In addition, multisynthetase complexes purified to homogeneity contain two copies of the p43 polypeptide per molecule of complex. These results suggested that p43 could be a dimer, even though p43C is a monomer when analyzed by gel filtration (5). The oligomeric structure of p43 was determined by sedimentation equilibrium (Fig. 6). When p43 was subjected to centrifugation equilibrium at an initial protein concentration of 38 M, experimental data could be fitted to a single species with a molecular mass of 70,912 Da. Taking into account a calculated molecular mass of 35,418 Da for the monomer, we concluded that p43 is a dimer in solution. The C-terminal moiety of p43, p43C, behaved exclusively as a monomer of 18,843 Da (theoretical mass of a monomer, 19, 398 Da). In contrast, the Nterminal moiety of p43, p43N, gave a more complex sedimentation pattern. When analyzed at an initial protein concentration of 150 M, experimental data were fitted to a dimer-tetramer equilibrium, with K D ϭ 20 Ϯ 10 M, taking into account the theoretical mass of 17,218 Da of the monomer. Therefore, the native species of p43 is likely to form a dimer through its N-terminal segment.
We analyzed the ability of p43 to form stable complexes with various tRNAs. Radiolabeled in vitro-transcribed tRNAs (yeast tRNA Asp or tRNA i Met ; human tRNA Lys or tRNA Arg ) were incubated with increasing amounts of p43 (0.06 -4 M; monomer concentration), and free and bound tRNA species were analyzed by a gel retardation assay (Fig. 7). p43 formed a stable complex with tRNA with an apparent K D of ϳ0.2 M (expressed as monomer concentration). The finding that two band shifts were observed suggests that p43 can bind one or two tRNAs per dimer. The interaction with p43 does not require the L-shaped structure of tRNA. Indeed, we observed that a minihelix mimicking the acceptor-T⌿C stem-loop region of tRNA produced a similar interaction pattern (result not shown). In contrast with the robust interaction displayed by full-length p43, its isolated p43N or p43C domains formed weak complexes with tRNAs, with apparent K D of 7.5 and 40 M, respectively. As a control, we observed that p43C and p43C-H (a derivative without His tag) have indistinguishable tRNA binding capacities. Therefore, the two domains of p43 are likely to synergistically contribute a potent tRNA binding site on p43. DISCUSSION Here we showed that the p43 component of the mammalian multisynthetase complex is a substrate for caspase 7, an apoptotic protease. The free recombinant, as well as the natural complex-associated p43 species, are substrates of caspase 7. However, we found that the p43 entity associated within the complex was more efficiently processed into EMAPII than its soluble form. This result suggests that the site of cleavage is made more accessible to the caspase when p43 is forced into a conformation suited for its association with the other components of the complex. Although the EEVD sequence at position 175 to 178 of murine p43 would be an ideal site for caspase 7 according to its known preferred peptide substrates (31) the 141 ASTD2S 145 sequence is used in vitro by caspase 7 (this study), and the N terminus of the in vivo-generated EMAPII product also starts at residue Ser 145 (14). We recently reported the crystal structure of human EMAPII (7). Its two-domain architecture builds a pseudodimer. The 100-amino acid residue N-terminal domain forms an open ␤-barrel related to the 60amino acid residue C-terminal domain by a degenerated 2-fold symmetry. The EEVD peptide belongs to strand ␤1 that forms the first strand of the oligonucleotide-oligosaccharide binding fold ␤-barrel domain. This tetrapeptide appears to be buried in the protein core and is therefore not accessible for recognition by caspase. In contrast, the released Ser 145 at the N terminus is protruding from the compact structure.
Following cleavage at the 141 ASTD2S 145 site, the EMAPII cytokine lost its propensity to associate with the complex and is released as a monomer. The N terminus of p43 remains bound to the complex. These results are in full agreement with our  32 P-Labeled in vitro-transcribed tRNA Asp was incubated with human p43 or with its isolated p43N or p43C polypeptides at the concentrations indicated. After electrophoresis at 4°C on a 6% native polyacrylamide gel, the mobility shift of tRNA was visualized by autoradiography. previous analysis of protein-protein interactions involved in the assembly of the complex. We showed by a two-hybrid analysis that p43 is able to interact with p38, the scaffold component of the complex, but also with GlnRS and ArgRS (2). Furthermore, in vitro studies revealed that the isolated Nterminal moiety of p43 associates with ArgRS, but its C-terminal region corresponding to EMAPII does not (3). Similarly, p43 binds to p38 via its N-terminal domain. 2 Thus, we anticipated that the release of EMAPII would not destabilize the quaternary structure of the complex.
Examination of each of the components of the complex after caspase treatment revealed that, with the exception of ProRS, aminoacyl-tRNA synthetases remain associated. ProRS is carried by a multifunctional polypeptide (32) containing an Nterminal domain corresponding to GluRS, a linker region made of repeated units with nonspecific RNA binding properties (33), and a C-terminal ProRS domain. Cleavage of the GluProRS polypeptide by uncontrolled proteolysis led to the release of ProRS from the complex (30). Because of the size of the GluRS (105 kDa) and ProRS (60 kDa) polypeptides observed after cleavage by caspase, the finding that antibodies raised against the repeated units selectively recognized the N-terminal GluRS polypeptide of 105 kDa, and the known consensus sequence for caspase 7, the aspartate residue at position 857 from the sequence 854 DQVD 857 of human GluProRS is likely to correspond to the cleavage site (Fig. 8). AspRS and LysRS are also subjected to proteolysis, but their activities are not affected. The short polypeptides removed after cleavage by caspase 7 are therefore likely to be located at the N terminus of the two proteins, position corresponding to the eukaryotic-specific sequences that characterize the mammalian enzymes. According to the crystal structure of yeast AspRS (34) and of E. coli LysRS (35), the removal of a short polypeptide at the C-terminal extremity of the two proteins should result in their inactivation. Indeed, the conserved motif 3 of class IIb aminoacyl-tRNA synthetases is located at the very C terminus of these proteins. The potential cleavage sites for caspase 7 are indicated in Fig. 8.
Of particular interest is the finding that the immediate consequence of the cleavage of p43 into two equal moieties is the loss of its tRNA binding ability. The isolated domains have a weak RNA binding capacity (K D ϭ ϳ7.5-40 M), even though the N-terminal domain is very rich in basic residues and displays a calculated pI Ͼ 9. EMAPII-like polypeptides are recurrent domains associated with various proteins. In yeast, the Arc1p protein associates with MetRS and GluRS and acts as a cofactor for tRNA delivery to the synthetase (12). A polybasic sequence from the N-terminal region appended to the EMAPIIlike domain of Arc1p is also required for efficient RNA binding. In plants, an EMAPII-like domain is appended at the C terminus of MetRS and synergizes with the MetRS domain for binding tRNA (8). Thus, EMAPII domains always require additional sequences to potentiate their interaction with tRNA molecules.
It is worth noting that besides p43, the other targets of caspase 7 in the complex also concern peptide appendices that are eukaryote-specific and/or that are known to contribute RNA binding domains. The DQVD sequence that might be the caspase 7 recognition site in GluProRS is located in the third of the repeated units that form the linker polypeptide between the GluRS and ProRS domains. We recently determined that in human MetRS, an enzyme that possesses a single of these repeats, this polypeptide extension provides the MetRS core domain with a higher catalytic efficiency for tRNA aminoacy-lation. 3 Similarly, the N-terminal polypeptide extension of human LysRS contributes an RNA binding domain that facilitates tRNA aminoacylation. 4 Consequently, in vivo processing of components of the multisynthetase complex by caspase 7 would have as an immediate consequence to restrict the availability of aminoacylated tRNAs and should result in the inhibition of protein synthesis. Following onset of apoptosis, cleavage of p43 and of other components of the complex would be a means to arrest translation in cells engaged in programmed cell death. Although other factors of the translation machinery, including eIF2␣ (36) and eIF4G (37), are cleaved by caspase 3 during inhibition of translation in apoptotic cells, their proteolytic products do not possess cytokine activity. Secretion of EMAPII, the C-terminal domain of p43, results in the recruitment of macrophages (14,18,38) that engulf apoptotic cells, thus preventing inflammation caused by the release of their cellular content because of secondary necrosis of apoptotic cells. In this regard, p43 may compare with a molecular fuse. In its native pro-EMAPII form, it is an important cofactor for aminoacylation. After the fuse has blown, translation is irreversibly switched off, and EMAPII enters the cell death signaling pathway. In vitro assays showed that full-length recombinant p43 is also a potent cytokine (17,19). 5 This suggests that association of p43 within the multisynthetase complex inhibits its cytokine activity and/or sequesters proEMAPII in a cellular compartment and prevents its entry into apoptotic pathways until it is cleaved off the complex. Only a processed form of p43 is recovered in the supernatant of apoptotic cells. 6 The pathway of EMAPII secretion remains to be deciphered.