The N Terminus of Pro-endothelial Monocyte-activating Polypeptide II (EMAP II) Regulates Its Binding with the C Terminus, Arginyl-tRNA Synthetase, and Neurofilament Light Protein*

Background: Functional domains of pro-EMAP II play an important role in multi-aminoacyl tRNA synthetase (MSC) complexes. Results: The N terminus of Pro-EMAP II binds to its C terminus, arginyl-tRNA synthetase, and the neurofilament light subunit and contains a putative leucine zipper. Conclusion: The N terminus of pro-EMAP II facilitates its interaction with MSC complexes. Significance: Understanding these binding domains may provide important insights into transcriptional regulation. Pro-endothelial monocyte-activating polypeptide II (EMAP II), one component of the multi-aminoacyl tRNA synthetase complex, plays multiple roles in physiological and pathological processes of protein translation, signal transduction, immunity, lung development, and tumor growth. Recent studies have determined that pro-EMAP II has an essential role in maintaining axon integrity in central and peripheral neural systems where deletion of the C terminus of pro-EMAP II has been reported in a consanguineous Israeli Bedouin kindred suffering from Pelizaeus-Merzbacher-like disease. We hypothesized that the N terminus of pro-EMAP II has an important role in the regulation of protein-protein interactions. Using a GFP reporter system, we defined a putative leucine zipper in the N terminus of human pro-EMAP II protein (amino acid residues 1–70) that can form specific strip-like punctate structures. Through GFP punctum analysis, we uncovered that the pro-EMAP II C terminus (amino acids 147–312) can repress GFP punctum formation. Pulldown assays confirmed that the binding between the pro-EMAP II N terminus and its C terminus is mediated by a putative leucine zipper. Furthermore, the pro-EMAP II 1–70 amino acid region was identified as the binding partner of arginyl-tRNA synthetase, a polypeptide of the multi-aminoacyl tRNA synthetase complex. We also determined that the punctate GFP pro-EMAP II 1–70 amino acid aggregate colocalizes and binds to the neurofilament light subunit protein that is associated with pathologic neurofilament network disorganization and degeneration of motor neurons. These findings indicate the structure and binding interaction of pro-EMAP II protein and suggest a role of this protein in pathological neurodegenerative diseases.

Endothelial monocyte-activating polypeptide II (EMAP II) 2 was initially isolated and identified as a 22-kD proinflammatory cytokine secreted from a murine methylcholanthrene A-induced fibrosarcoma (1)(2)(3)(4). Later studies confirmed that EMAP II was the C terminus (aa 147-312) of pro-EMAP II cleaved under apoptosis and by protease inhibitors (5)(6)(7). Pro-EMAP II, also known as p43, AIMP1, and Scye I, encodes one component of the multi-aminoacyl tRNA synthetase complex (MSC) and is involved in protein translation. Mechanistically, the C terminus of EMAP II has been determined to induce apoptosis in endothelial cells and inhibit angiogenesis and tumor growth through inhibition of VEGF-mediated signaling and ␣5␤1 integrin-mediated deposition of the extracellular matrix protein fibronectin (8 -10). Expression and secretion of EMAP II is relevant to the prognosis of some cancer patients and found to be involved in diverse biological processes, such as lung development, immune responses, and glucose metabolism in the pancreas (11)(12)(13)(14)(15). These imply the importance of EMAP II in the diagnosis and treatment of cancer and other diseases (16 -20).
Intracellularly, pro-EMAP II directly interacts with other components of the MSC, such as p38, arginyl-tRNA synthetase (RARS), and glutaminyl-tRNA synthetase (QARS) (21)(22)(23)(24). Such an interaction plays an important role in the assembly and function of the MSC. In addition, through C terminus binding with the E3 ubiquitin ligase Smurf2, pro-EMAP II stabilizes Smurf2 and down-regulates the TGF-␤ signaling pathway (25). Additional studies unveiled functional mechanisms and domains of pro-EMAP II protein (26,27). Park et al. (28) iden-tified the N terminus 4 -46 aa region of pro-EMAP II to be important in promoting fibroblast cell proliferation and wound repair. Pro-EMAP II has also been determined to bind with heat shock protein HSP90B1/gp96, which plays a critical role in innate and adaptive immunity (16,29). On the basis of sequence analysis, Guo et al. (30) introduced a novel concept postulating that, during evolution, the N terminus of pro-EMAP II was preserved because of evolutionary pressure, and the C terminus gained cytokine functions through the accumulation of mutations, and that this may be the reason for the diverse functions of EMAP II.
Recently, studies have determined some biological roles for pro-EMAP II. Zhu et al. (31) found that pro-EMAP II interacts directly with the neurofilament light (NFL) chain and downregulates neurofilament phosphorylation. Depletion of pro-EMAP II resulted in neurofilament network disorganization and degeneration of motor neurons. In 2010, pro-EMAP II truncation at the C terminus was identified in consanguineous Israeli Bedouin kindred who suffered from Pelizaeus-Merzbacher-like disease (32)(33)(34), indicating the requirement of further evaluation of the structure and functionality of pro-EMAP II in different biological processes. The role of pro-EMAP II in MSC, regulation of gene expression and modification, the interaction with other proteins, and the functionality in different cell types require further and detailed exploration. Here we established an approach to study the structure and function of pro-EMAP II using GFP punctum analysis and pulldown assays. We found that the N terminus of pro-EMAP II can form a specific strip-like punctate structure. We characterized this structure as a putative leucine zipper of pro-EMAP II. We also determined that the C terminus of pro-EMAP II could bind with its N terminus to repress punctate structure formation and that it colocalizes to the neurofilament light subunit protein. Furthermore, our studies indicate that the 1-70 aa region in the N terminus of pro-EMAP II is responsible for its binding to the MSC subunits.

EXPERIMENTAL PROCEDURES
Cells, Reagents, and Antibodies-Human lung adenocarcinoma A549, human embryonic kidney 293, and human neuroblastoma SH-SY5Y cells were utilized in studies for neuronal function and differentiation. Cells were maintained in DMEM containing 10% fetal bovine serum and 5 mg/ml L-glutamine at 37°C in 10% CO 2 . Human fetal kidney HEK293 cells were maintained in DMEM containing 5% fetal bovine serum at 37°C in 5% CO 2 .
PCR Cloning and Detection-For cloning, PCR reactions were carried out using a high-fidelity DNA amplification kit (Sigma-Aldrich) containing 10 pmol of each primer and 10 ng of DNA template. For RT-PCR and real-time PCR, total RNA was isolated from cells using TRIzol reagent. Reverse transcription was carried out using the Superscipt III RT-PCR system according to the instructions of the manufacturer (Invitrogen). Oligonucleotides were synthesized by Invitrogen and Sigma-Aldrich. The N terminus 1-102 aa of the RARS gene were amplified using oligos 5Ј-AACGAATTCAACCATGGACG-TACTGGTGTCTGAGT-3Ј and 5Ј-CTTGTCGACCACTAG-CAGAGGAGGATTTTCCA-3Ј. The N terminus 1-108 aa of the human QARS gene were amplified using oligos 5Ј-AAC-GAATTCAACCATGGCGGCTCTAGACTCCCTGT-3Ј and 5Ј-CTTGTCGACCTCGAAGTCCACAGTGTCGAT-3Ј.
Plasmid Construction-The EMAP II plasmid was used as a template for the amplification of the N terminus and C terminus and truncation of human pro-EMAP II using oligonucleotides. PCR fragments were digested and cloned into vector pEGFP-N3 for overexpression of GFP fusion protein. Site-directed mutation was performed using overlapped oligonucleotides for PCR. In brief, N terminus and C terminus fragments containing mutations were amplified. The PCR fragments were combined and used for the PCR template. Site-directed mutated fragments were inserted into the pEGFP-N3 vector. PCR-cloned fragments were also cloned into vector pET28a(ϩ) and pGEX-4T-3 for overexpression of His-or GST-tagged protein in Escherichia coli. DNA sequencing and/or restriction digestion analysis was used to confirm the plasmid sequence. The reporter plasmids EYFP-GalT, Lamp1-YFP, pEGFP-C1-wtVHL, EGFP-supervillin, and DsRed-rab7 WT were purchased from Addgene (Cambridge, MA).
Transfection and Stable Clones-Cells were seeded in 24-well plates or in chamber slides 24 h prior to transfection. A549 cells were transfected using reagent TransIT LT1 or GeneExpresso TM 8000. HEK-293 cells were transfected using Lipofectamine 2000 according to the protocol of the manufacturer. Analysis for GFP punctum formation was performed on live cells. Cells were observed and photographed after 20 h under a fluorescence microscopy. 100 -200 GFP-positive cells were analyzed per experiment, and these were confirmed using three to six replicates from independent experiments performed on different occasions. For drug treatment experiments, a final concentration of 5 mg/ml BFA, 50 nM Bafilomycin A1, 1 mM DTT, 10 nM PS-341, and 50 mM chloroquine or PBS was applied to cells 1 h after transfection. For immunofluorescence analysis, cells were treated with 10 mg/ml of BFA or PBS for 4 h or with 10 nM of PS-341 or PBS for 8 h prior to further analysis. For stable clones of A549 or SH-SY5Y, cells were selected in complete medium containing 800 g/ml or 600 g/ml of G418 for 2 weeks. Single colonies were picked up for expanding cultures. The expression of GFP-fusion protein was confirmed using fluorescence microscopy and Western blot analysis.
Cross-linking Using DSS-Transfected cells were harvested and washed three times with PBS. Cells were then suspended in PBS at 1 ϫ 10 7 cells/ml and incubated with 1 mM DSS for 20 min. Reactions were quenched by adding Tris-HCl buffer (pH 7.5) to a concentration of 20 mM and incubated for 15 min. Cells were then pelleted, rinsed, suspended in 1ϫ SDS loading buffer, and sonicated to disrupt sticky genomic DNA prior to being subjected to SDS-PAGE analysis.
Immunofluorescence-Cells were rinsed with PBS, fixed in 4% paraformaldehyde, and permeabilized with 0.2% Triton X-100. After blocking in CAS solution, cells were incubated with the specific primary antibody, followed by specific secondary FITC-or Cy-3-conjugated secondary antibodies. Cells were washed with PBS four times and mounted with SlowFade Gold antifade reagent with DAPI (Invitrogen). A fluorescence microscope was used for imaging analysis.
Immunoblot Analysis-Cells were lysed at 4°C in radioimmune precipitation assay buffer or 1ϫ lysis buffer (50 mM Tris-Cl (pH 7.4), 10 mM MgCl 2 , 100 mM NaCl, 1% Triton X-100, 10% glycerol, 1 mM DTT, and 1ϫ protease inhibitor mixture). Lysate was cleared by centrifugation at 4°C at 13,200 rpm for 20 min. The protein concentration was measured using Bio-Rad protein assay reagents according to the manual of the manufacturer. Equal amounts of protein were subjected to SDS-polyacrylamide gel and blotted to a PVDF membrane. The blots were blocked in 5% nonfat milk and incubated with primary antibody, followed by the appropriate secondary antibody. Specific signals were measured using a detection kit from Amersham Biosciences.
Pulldown Assay-Pulldown assays were performed using the ProFound TM pulldown PolyHis protein:protein interaction kit (Pierce). Briefly, BL21 DE3-competent E. coli was transformed with plasmids constructed in vectors pET28a(ϩ) and pGEX-4T-3, and fusion proteins were induced with 2 g/ml isopropyl 1-thio-␤-D-galactopyranoside. Following rinsing with TBS at 4°C, pellets were suspended in Profound lysis buffer and incubated for 30 min on ice. The crude E. coli lysate was clarified through 12,000 ϫ g centrifugation, and the supernatant was collected and used as bait or prey protein for the pulldown assay. When PolyHis-tagged proteins were used as bait protein, a 40 mM final concentration of imidazole was used in binding and washing buffers, and proteins were eluted with buffer containing 290 mM imidazole.
Flow Cytometry Sorting and Mass Spectrometry-SH-SY5Y stable clone cells overexpressing aa 1-70 of pro-EMAPII fused with GFP were trypsinized for harvesting. After washing with PBS, cells were suspended in lysis buffer containing 1% Triton X-100 and rotated at 4°C for 30 min, followed by centrifugation at 1000 rpm for 10 min. Pellets were suspended in lysis buffer and washed again. Pellets were then resuspended in lysis buffer and sonicated. The supernatant was transferred to new tubes after centrifugation at 1500 rpm for 10 min and sorted by flow cytometry as described previously (35). The enrichment of GFP puncta was confirmed by microscopic observation. GFP puncta were pelleted down after centrifugation at 5000 rpm for 10 min. The pellet was suspended in 1ϫ SDS loading buffer, boiled, subjected to 12% SDS-PAGE, and run into separation gel for about 1 cm. The gel was stained and destained according to the Coomassie blue staining protocol. The gel was cut out and analyzed by mass spectrometry.
Statistical Analysis-Prism software was used to perform all statistical analyses. All results are expressed as mean Ϯ S.E. The significance of differences between two sample means was determined by unpaired two-tailed Student's t tests using 95% confidence intervals. p Ͻ 0.05 was considered significant. Water solubility was determined using peptide property calculator software. Coiled-coil structures were identified using Lupas software.

Overexpression of Truncated Isoforms of the Pro-EMAP II Protein Uncovered an N Terminus 1-70 aa Region Responsible for the Aggregation of EMAP II Punctum Formation-Although
much is known regarding the activities and functions of mature C terminus EMAP II, little is understood about the functions of its intracellular precursor form, pro-EMAP II protein. A549 cells overexpressing the full-length pro-EMAP II (aa 1-312) fused with GFP were noted to have a dot-like GFP punctate structure in Ͻ1% of GFP-positive cells (Fig. 1B). Multiple truncations of the C and N terminus regions of EMAP II were utilized to identify the region of pro-EMAP II responsible for the punctum formation. We determined that the 1-70 amino acid region in the N terminus promoted GFP punctum formation because 45.2% of GFP-positive cells contained GFP puncta (p Ͻ 0.01 compared with full-length pro-EMAP II, Fig. 1, D and J) and compared with 0% GFP punctum formation in truncated aa 99 -312 (  Fig. 1, H and J) of the N terminus of pro-EMAP II were truncated, the GFP punctum ratio decreased dramatically (p Ͻ 0.01 compared with aa 1-70 truncation). These results suggest that the entire 1-70 a region of pro-EMAP II is essential for GFP punctum formation.
In addition to the quantitative amount of GFP puncta, a profound morphological change of GFP puncta was also observed in transfection with 1-70 aa truncations. In transfection with full-length pro-EMAP II, GFP puncta were dot-like in structure and perinuclear, as shown in Fig. 1K. In contrast, the pro-EMAP II 1-70 aa truncation expressed a marked variation of punctum formation that was much stronger and strip-like, as shown in Fig. 1, L-S. However, no branches in the strip-like GFP punctate pattern were observed. Similar results were observed in the transfection of HEK293 (Fig. 2, A-C and G) and SH-SY5Y neuroblastoma cells (Fig. 2, D-G) using different transfection reagents, such as Lipofectamine 2000, TransIT LT1, FuGENE6, and GeneExpresso TM 8000 (data not shown). No punctate lesions were noted in the C terminus EMAP II 147-312 aa cells (data not shown).

Analysis of GFP Punctum Formation Uncovered a Role of Leucine-like Residues in the 1-70 aa Region of Pro-EMAP II-
Across species, the 1-70 aa region of pro-EMAP II (AIMP1) is highly conserved during evolution (Fig. 3A), implying the importance of the N terminus of pro-EMAP II. To investigate the function of the putative acetylation sites of the 1-70 aa region of pro-EMAP II, ten lysine residues were identified, and stripe replacement was performed to replace lysine residues with alanine residues found between amino acid 8 -11 (Fig. 4D) and 46 -63 (Fig. 4C). Transfection analysis showed no significant differences in GFP punctum formation between alanine replacements (Fig. 4, C and D) and the wild type (Fig. 4, B and Q).
We next turned our attention to the hydrophobic amino acids leucine and isoleucine within this region that have frequently been identified in protein-protein interactions and binding recognition sites for hydrophobic ligands, respectively. Site-directed replacement of leucine or isoleucine residues with alanine residue at amino acid Ile-213 Ala, Leu-253 Ala, Leu-313 Ala, Leu-323 Ala, Ile-373 Ala, Leu-383 Ala, Leu-423 Ala, Leu-553 Ala, Ile-593 Ala, and Leu-623 Ala abolished GFP punctum formation (Fig. 4, E, G-O, and Q). However, although replacement of isoleucine residues with an alanine residue at amino acid Ile-223 Ala showed no significant difference (Fig. 4F), exchange of a leucine for an alanine at amino acid Leu-663 Ala showed a significant decrease in GFP punctum formation compared with the wild type ( Fig. 4, P and Q). These results suggest that leucine and isoleucine residues are important in GFP punctum formation of 1-70 aa pro-EMAP II.
Regions within the C Terminus of Pro-EMAP II Were Determined to Induce Repression of GFP Punctum Formation and Induce Morphologic GFP Punctum Alterations-Overexpression of full-length pro-EMAP II markedly decreased GFP punc-  tum formation compared with the 1-70 aa isoform. To define the region within the pro-EMAP II protein that repressed its GFP punctum formation, serial internal truncations were performed. Internal truncation of pro-EMAP II between amino acids 71-253 dramatically increased the amount of GFP punctum formation, whereas truncation of pro-EMAP II between amino acids 71-157, 71-203, and 71-94 did not significantly impact GFP punctum formation (Fig. 5, A and B). Further internal truncations between amino acids 204 -253 were performed to identify the region within the amino acid span of 204 -253 that impacted the GFP punctum formation of pro-EMAP II. We determined that the internal truncation of 71-233 aa FIGURE 3. Region 1-70 aa of pro-EMAP II (AIMP1) is highly conserved across species and is predicted to have a coiled-coil structure. A, evaluation of the N terminus of pro-EMAP II determined that the region is highly conserved in evolution. B, using Lupas software, three potential coiled-coil structures in pro-EMAP II protein were identified, 1 M-80FI, 108G-148K, and 199L-219R, with these sequences extending for more than 35 residues, suggesting a probability of Ͼ80 -90% that they will assume a coiled-coil structure. resulted in a 1.8-fold increase in GFP punctum formation (p Ͻ 0.01) compared with the 71-223 aa truncation, indicating that the amino acid region between 203-233 of pro-EMAP II represses GFP punctum formation (Fig. 5, C and D).
To further narrow the repressing region, fragments of C terminus of pro-EMAP II were linked with the 1-70 aa region through a 15-aa flexible peptide linker ((GGGGS) 3 , Fig. 5F) (36). Optimization of codon pair use within the (GGGGS) 3 linker sequence is known to enhance protein expression. Transfection experiments containing C terminus pro-EMAP II fragments 147-170aa, 181-210 aa, or 200 -230 aa resulted in a respective marked 3.2-, 4.5-, or 7.0-fold decrease of GFP punctum formation compared with 1-70 aa alone, implying that multiple regions within the C terminus have repressive effects on GFP punctum formation (Fig. 5F). In addition to the amount of punctum protein formation, a change in morphology of GFP puncta was also observed in A549 cells transfected with fulllength pro-EMAP II, whereas internal truncation at 71-203 aa resulted in dot-like GFP punctation (Fig. 5E). However, internal deletion of the 71-212 aa span induced a small strip-like change to the GFP punctate structure, whereas further internal deletions of 71-223 aa, 71-233 aa, and 71-253 aa resulted in a strong and strip-like GFP punctate structure. Analysis of pro-EMAP II identified three potential coiled-coil structures, 1 M-80FI, 108G-148K, and 199 -219R, with these sequences, extending more than 35 residues, suggesting a probability of Ͼ80 -90% that they will assume a coiled-coil structure (Fig. 3B).
The 1-70 aa Region of Pro-EMAP II Binds with Its C Terminus-To examine whether the 1-70 aa region of pro-EMAP II binds with itself, HEK293 cells were transfected with GFP control and 1-70 aa fused with GFP. Cells were crosslinked using DSS. Western blot analysis determined that multiple bands of 35-kD increments were observed in the 1-70 aa transfection treated with DSS but not in GFP control and untreated cells (Fig. 6A). This suggests that the 1-70 aa region of pro-EMAP II fused with GFP might form a polymer structure within cells. Using His-tagged 1-70 aa as bait and GST-tagged pro-EMAP II fragments recombinant protein as prey (Fig. 6B), a His tag pulldown assay was performed to determine binding between the N and C terminus regions of pro-EMAP II. As shown in Fig. 6C, His-tagged 1-70 aa strongly pulled down GST-tagged 1-70 aa and the GST-tagged C terminus of pro-EMAP II (146 -312 aa) but not GST control, GST tagged 74 -146 aa, or the 1-70 aa mutant. These experiments revealed the binding of the 1-70 aa N terminus region to the C terminus of pro-EMAP II.

The N terminus Region 1-70 aa of EMAP II Is a Binding Partner of the MSC Complex RARS but Not of QARS-Previous
reports support the N terminus region of pro-EMAP II as a component of the MSC, where it has been identified as binding to itself as well as to p38, RARS, QARS, and the leucine zipper of AIMP2. To establish the binding domain of two of the key MSC proteins, RARS and QARS, we examined their ability to bind to the pro-EMAP II N terminus region extending from 1-70 aa. Using His-tagged 1-70 aa of pro-EMAP II as bait and the recombinant GST-tagged RARS and QARS N terminus as prey, we determined that pro-EMAP II pulled down GST-tagged RARS (1-102 aa) but not GST control or GST-QARS (1-108 aa) (Fig. 7A). Point-mutations of pro-EMAP II at Ile-21-3 Ala, Ile-37-3 Ala, Leu-38-3 Ala, Ile-21-3 Ala Ile-22-3 Ala Leu-25-3 Ala (Fig. 7, B and C) defined the RARS binding region to be dependent on the leucine at amino acid 42 and 55 as pro-EMAP II mutant Leu-42-3 Ala Leu-55-3 Ala did not immunoprecipitate with RARS (Fig. 7, B and C). Furthermore, cotransfection of 1-70 aa of pro-EMAP II fused with GFP or the DsRed reporter resulted in a two-colored colocalized punctate structure (Fig. 7E), whereas cotransfection of GFP empty vector control and DsRed fused with 1-70aa did not colocalize (Fig.   staining of transfected cells with the nucleus stain Hoechst 33343 indicated that 1-70 aa GFP puncta are not nuclear (Fig.  8, A-C). Although the addition of an NLS fused to the 1-70 aa and GFP partitioned the 1-70 aa GFP fusion protein to the nucleus, no GFP punctate lesions were observed in nuclei (Fig.  8D). Because of the cytoplasmic partitioning of the 1-70 aa pro-EMAP II GFP punctate lesions, we examined whether the Golgi apparatus, lysosomes, aggresomes, or endosomes markers colocalized with these punctate lesions. Because of a significant lack of effective organelle antibodies, cotransfection experiments were performed. Cotransfection of 1-70 aa GFP with the late endosome marker Rab7 fused with DsRed (Fig.  8E), with the Golgi marker GalT (Fig. 8F), or the lysosomal marker lamp1 fused with YFP showed no punctate colocalization (Fig. 8G). These studies were validated by examination of purified punctate lesions and subsequent mass spectrometry analysis of binding partners isolated and bound to the punctate lesions. Cross-linking assays revealed no interaction between the GFP punctate pro-EMAP II 1-70 aa formations and known organelle markers (data not shown). Although treatment of cells with BFA to induce the disassembly of the Golgi apparatus does not alter 1-70 aa GFP punctum formation (Fig. 8, H-M), consistent with the above observation that 1-70 aa GFP puncta do not colocalized with the Golgi apparatus or nuclei. Similarly, the 1-70 aa GFP puncta did not colocalize with the aggresome markers vimentin and ubiquitin even in the presence of the proteasome inhibitor PS-341 (data not shown). Furthermore, treatment of cells transfected with the mutant Leu-663 Ala, which demonstrates reduced punctum formation using PS-341, did not increase GFP punctum formation (data not shown).

Insoluble 1-70 aa Pro-EMAP II GFP Punctum Formation
Binds to NFL-To determine the binding partners associated with 1-70 aa GFP punctum formation, GFP punctate lesions were isolated following transfection of 1-70 aa pro-EMAP II into neuroblastoma SH-SY5Y cells. Our initial observations determined that these punctate formations were predominately insoluble because 1-70 aa GFP was detected in the insoluble fraction, whereas the GFP control was detected only in the supernatant by Western blot analysis (Fig. 9A). This was further supported by microscopic fluorescence analysis showing that the 1-70 aa GFP punctum formations were observed in cell debris after lysis in buffer containing 1% Triton X-100 and an intensive wash (Fig. 9B). Following brief sonication, 1-70 aa GFP puncta were released from debris into the supernatant and pelleted and underwent flow cytometry enrichment. Mass spectrometry analysis of the enriched 1-70 aa GFP punctate formations and their binding partners determined that they bound NFL polypeptide with 11 peptide spectrum matches. Immunoprecipitation from cells transiently transfected with the 1-70 aa GFP EMAP II N terminus confirms binding of NFL with the pro-EMAP II N terminus (Fig. 10E). Because mass spectrometry analysis and previous reports indicate colocalization of pro-EMAPII and the neurofilaments found in neurons and neuroblastoma cells (SH-SY5Y), cotransfection with 1-70 aa GFP and DsRed-NFL were undertaken. The human NFL subunit gene was cloned and fused with DsRed. Transfection of DsRed-NFL in A549 cells and SH-SY5Y cells induced punctum formation (Fig. 10, A-D). Furthermore, cotransfection of DsRed-NFL with either 1-70 aa (Fig. 10, A and C) or full-length pro-EMAP II fused (Fig. 10, B and D) with GFP resulted in colocalization and copunctate lesions supporting an interac- tion between the 1-70 aa region of pro-EMAP II and NFL fibers (Fig. 10, A-D). The C terminus 146 -312aa EMAP II C terminus does not immunoprecipitate with NFL (Fig. 10E).

DISCUSSION
Pro-EMAP II is abundantly expressed in all cell types and is a component of the multi-enzyme assemblage of 11 polypeptides that form the essential link to the structural organization of the mammalian cell cytoplasmic translation apparatus, the MSC complex. On the basis of bioinformatical analysis, pro-EMAP II protein has been predicted to have of an N terminus and C terminus (37). The N terminus contains a coiled-coil structure, which may play an important role in protein-protein interaction. Furthermore, the N terminus of pro-EMAP II has been reported to interact with its own N terminus, p38, RARS, QARS, and the leucine zipper of AIMP2 (21,23,24,38), suggesting that the N terminus may mediate the association of pro-EMAP II to MSC. However, the structure and function of N terminus of pro-EMAP II remains obscure. Through GFP punctum analysis and pulldown assays, we showed that the truncation of pro-EMAP II to 1-70 aa resulted in a strong and specific strip-like punctate structure that was mediated by the leucine-like residues within the 1-70 aa region of pro-EMAP II. In addition to the self-binding of the 1-70 aa region, we determined that it also binds the C terminus of pro-EMAP II through a putative leucine zipper motif and is the region that is the binding partner of the MSC complex polypeptide RARS but not QARS. We also determined that the punctate GFP pro-EMAP II 1-70 aa aggregate colocalizes and binds to the NFL subunit protein. These findings show that the 1-70 aa region of the pro-EMAP II N terminus is responsible for the binding of MSC subunits and may be the region of the N terminus of pro-EMAP II that facilitates the interaction of the MSC complex with RNA binding.
Transcriptional regulation is dependent on the flexibility and interaction of the participating protein complexes. Leucine zippers play an important role in this process by mediating protein-protein interactions and transcriptional regulation, where they serve as folding triggering sequences in coiled-coil structures (39,40). Pro-EMAP II, a component of the MSC complex, functions as a cofactor of aminoacyl-tRNA synthetases, whose involvement is as a general RNA binding domain (41). We hypothesized that pro-EMAP II contained a putative leucine zipper that would allow it greater flexibility in the mediation of protein-protein interactions. Using mutation analysis, we showed that specific leucine or isoleucine residues are critical for GFP punctum formation but that abundant lysine residues have little effect. Our cross-linking and pulldown assay results showed that the 1-70 aa region of pro-EMAP II interacts with itself, the C terminus of pro-EMAP II and N terminus RARS. Importantly, some leucine mutants abolished the binding with RARS, suggesting the relevance of the leucine binding regions. Therefore, our data provide detailed information to support that the N terminus of pro-EMAP II functions as a putative leucine zipper, which is consistent with previous predictions and other reports (38).
In contrast to the GFP punctate lesions seen following overexpression of the 1-70 aa pro-EMAP II, not all leucine zipperlike domains of the MSC proteins have a similar expression pattern. For example, although RARS is one component of MSC and contains a leucine zipper-like domain at the N terminus, overexpression of the RARS leucine zipper domain showed no GFP punctate structure. This implies the specific characteristics of the putative leucine zipper of pro-EMAP II. The underlying mechanism of GFP punctum formation remains unclear. Cross-linking analysis and pulldown analysis suggest that protein-protein interaction and protein polymerization may play FIGURE 9. Isolation of insoluble 1-70 aa pro-EMAP II GFP punctum formation using flow cytometry enrichment. GFP punctate lesions were isolated from neuroblastoma SH-SY5Y cells transfected with 1-70 aa pro-EMAP II. A, Western blot analysis of the supernatant and pellet determined that the 1-70 aa GFP was predominately found in the insoluble fraction (pellet) in contrast to the GFP vector control that was found only in the supernatant. B, systematic lysis with 1% Triton X-100, sonication, pelleting, and flow cytometry enrichment of the GFP punctate formations was performed and evaluated by light microscopy. Magnification, ϫ100. FAOS, flow-assisted organelle sorting. some role in forming a strip-like structure. Cotransfection experiments suggest that the punctate structure is not colocalized with the markers of the aggresome, Golgi apparatus, lysosome, or endosome. BFA treatment experiments also suggest that GFP punctum formation is not relevant to the endoplasmic reticulum pathway. Our data show that 1-70 aa GFP colocalizes with NFL. This is consistent with the finding of colocalization and interaction of pro-EMAP II with NFL (31), which implies that GFP punctum formation may be related to neurofilament structures.
Internal interaction of proteins is critical for the proper and effective folding and maintenance of the functional structure of a protein (42). Protein misfolding and aggregation may play important roles in the pathological process of disease processes such as neurodegenerative diseases. For example, in transfection experiments with full-length pro-EMAP II, only rare GFP puncta was observed compared with transfection with the N terminus 1-70 aa GFP. However, the presence of the C terminus of pro-EMAP II had a repressive effect on GFP punctum formation. These findings suggested that the C terminus of pro-EMAP II might fold back and bind with its N terminus region to cover up the active moiety of the putative leucine zipper. Through truncation of the C terminus, we were able to expose the putative leucine zipper, resulting in GFP punctum formation. On the other hand, folding up between N terminus and C terminus may also prevent the exposure of an active C terminus moiety. In our experiments to establish stable clones, we could easily obtain clones overexpressing 1-70 aa GFP, but it was difficult to obtain stable clones overexpressing C terminus EMAP II fused with GFP (data not shown). Previous studies in our laboratory also showed that the overexpression of the C terminus of pro-EMAP II inhibits cell proliferation (43). These data suggest that the N terminus putative leucine zipper plays an important role in the maintenance of proper soluble structure status and function of full-length pro-EMAP II, which are the results of natural selection during evolution.
Although the mechanism of GFP punctum formation is complicated, it is the result of specific or nonspecific protein internal interaction or protein-protein interactions. As we know, most native proteins are soluble in vivo. When proteins are truncated or mutated, spatial hindrance will be removed, and aggregates and other specific punctate structures may form. As shown in our experiments, overexpression of pro-EMAP II resulted in small dot-like punctate structures in a few cells. Gradual truncation of the C terminus of pro-EMAP II not only increases the ratio of GFP punctum formation but also resulted in a strong and long strip-like punctate structure, mimicking the process of gradual exposure of a putative leucine zipper of pro-EMAP II to the environment following gradual removal of its C terminus. Our experiments suggest that GFP punctum analysis can be used as a powerful tool to dissect protein structural information.
Feinstein et al. (33) found truncation of C terminus of pro-EMAP II in consanguineous Israeli Bedouin kindred who suffer from a severe axonal disease. Loss of the C terminus may be the cause of the disease. To our knowledge, the 1-97 aa of the N terminus of pro-EMAP II remain intact in patients with a C terminus truncation. In the brain, pro-EMAP II is highly expressed. In our experiments, truncation of the C terminus resulted in strong strip-like punctum formation. It is possible that such a punctate structure can be found in the neurons of these patients. Our data in SH-SY5Y human neuronal cells that have been studied widely as in vitro models of neuronal function and differentiation (44 -46) and in Parkinson disease (47)(48)(49), we show colocalization of 1-70 aa GFP with NFL. This is consistent with reports showing that pro-EMAP II colocalizes and binds with neurofilaments. Knockout mice of pro-EMAP II show axon development defects in motor neurons (31). These findings suggest that pro-EMAP II may play an important role in axon assembly. It is well known that protein aggregation plays an important role in many neurodegenerative diseases, such as Alzheimer disease. If punctate formation or exposure of FIGURE 10. Punctum formations of 1-70 aa and 1-312aa GFP pro-EMAP II colocalize and immunoprecipitates with NFL. Cells were cotransfected with either 1-70 aa or 1-312aa GFP pro-EMAP II and the DsRed-human NFL gene. Colocalization and copunctate lesions were identified in cells cotransfected with DsRed-NFL and either the 1-70 aa GFP (A and B) and 1-312 aa GFP pro-EMAP II (B and D). Cell lysates from using transient transfected with 1-70 aa GFP EMAP II, 1-312 aa GFP pro-EMAP II, 146 -312 aa GFP EMAP II, or empty vector were immunoprecipitated with GFP antibody. Western blot analysis using an NFL antibody confirmed binding of the NFL with the EMAP II N terminus domain and pro-EMAP II but not the C terminus (E) (immunoprecipitation with empty vector and control cell lysate was negative, data not shown). Magnification, ϫ100.
the putative leucine zipper occurs within these disease process, then one may consider their potential role in disease progression because our experiments show that even one amino acid mutation abolishes GFP punctum formation. This implies the possibility to alter protein structure through targeting the key amino acid residues, leading to the prevention and treatment of neurodegenerative diseases.
In conclusion, we established an approach to study the structure and function of pro-EMAP II using GFP punctum analysis and pulldown assays. Utilizing this technique, we determined that the N terminus of pro-EMAP II can bind with its C terminus to repress punctate structure formation, that it has a putative leucine zipper, that it is the region responsible for the binding of MSC subunits, and that it colocalizes to the NFL subunit protein. These findings indicate the structure and binding interaction of pro-EMAP II protein and suggest a role of this protein in pathological neurodegenerative diseases.