Characterization of Elongation Factor-1A (eEF1A-1) and eEF1A-2/S1 Protein Expression in Normal and wasted Mice*

,

Peptide elongation factor-1A (eEF1A-1) 1 is an abundant protein, once thought to be ubiquitously expressed; its major role is to mediate the transfer of charged aminoacyl-tRNA to the A site of the ribosome during peptide elongation. We have previously shown that the mRNA expression of this gene declines in rat brain, heart, and skeletal muscle during development (1,2), to the extent that in adult skeletal muscle, there is an almost complete loss of eEF1A-1 mRNA expression; a sister gene, eEF1A-2/S1, is expressed instead (3). Furthermore, in contrast to the ubiquitous expression of eEF1A-1 in many cell types, eEF1A-2/S1 expression is limited to the terminally differentiated cells of the brain, heart, and skeletal muscle (3,4).
Recent evidence suggests that eEF1A-2/S1 is an alternative peptide elongation factor-1A for neurons and myocytes, because the amino acid homology between them is 92% (5,6), and eEF1A-2/S1 can perform peptide elongation in vitro (7). The absence of eEF1A-2/S1 mRNA, as a result of a deletion in the promoter and first exon of the eEF1A-2/S1 gene, was recently identified in the mutant mouse wasted (8). wasted mice exhibit severe muscle wasting and degeneration of motor neurons, beginning at 3 weeks after birth and die by the age of 28 days (8 -11). This coincides with the decline of eEF1A-1 expression in brain, heart, and skeletal muscle, without the normal concordant appearance of its sister gene; because neither form of elongation factor-1A is found, the consequential loss of protein synthesis ability is the likely cause of death.
Distinguishing between eEF1A-1 and eEF1A-2/S1 protein expression was long hampered by the absence of isoform-specific antibodies; for example, it was impossible to determine the proportion of eEF1A-1 versus eEF1-A2/S1 in any given skeletal muscle biopsy. This is specifically true with regard to a recent report, which concludes that there is no change in eEF1A in young and old muscle (12); this conclusion was based on a study using a pan-monoclonal antibody that recognizes both isoforms and thus cannot discriminate either qualitatively or quantitatively between the two sister forms of this gene family. To investigate the presence of eEF1A-2/S1 protein and differentiate its level from that of its sister, during mouse development, we generated a polyclonal antiserum that specifically recognizes the eEF1A-2/S1 isoform and does not cross-react with eEF1A-1. Using it along with another antiserum that specifically recognizes eEF1A-1, we characterized the protein presence of both peptide elongation factor-1A isoforms during development in normal and wasted mutant mice.

MATERIALS AND METHODS
Peptide Synthesis and Production of eEF1A-1-and eEF1A-2/S1specific Antibodies-A highly purified, sequence-specific peptide of eEF1A-2/S1 protein, CB5 (see Fig. 1A), was obtained by high performance liquid chromatography on a C 18 reverse-phase column, using an acetonitrile and trifluoroacetic acid gradient. CB5 protein was solubilized in 20 mM phosphate buffer and coupled to keyhole limpet hemocyanin carrier protein (Pierce) through amino-terminal cysteine residues. Antibodies were generated by multiple injections of 200 g of the peptide in Freund's incomplete adjuvant (Pierce), followed by two booster injections given at 3-week intervals with the same amount of antigen. One week after the final injection, the rabbits were sacrificed, and blood was collected. The eEF1A-1-specific antibody was raised against eEF1A-1 protein-specific peptide HT7 (see Fig. 1B), as previously described (13).
Tissue Preparation-Brain, heart, muscle, liver, and spleen samples were obtained from mouse, rat, and human autopsy tissues. Tissue extraction and preparation were performed as previously described (14), with minor modifications. All tissues were extracted by homogenization in a 2-ml tissue grinder (Wheaton, Millville, NJ) with 1 ml of homogenizing buffer (300 mM sucrose, 150 mM KCl, 30 mM Tris-HCl, 5 mM MgCl 2 , 1.5 mM dithiothreitol, 1 mM EDTA, pH 8.0, and 1.5% Triton X-100) containing protease inhibitors (10 g/ml aprotinin, 2 g/ml leupeptin, 10 g/ml pepstatin, and 2 mM phenylmethylsulfonyl fluoride) (ICN Biochemicals Canada, Montréal, Quebec). Protein extracts were then incubated with 150 g/ml DNase I (Roche Molecular Biochemicals) on ice for 1 h and sonicated for 1 min in an ice bath, followed by centrifugation at 3000 rpm for 15 min. The resulting pellet was discarded, and the supernatant was designated as total protein. Protein concentration was assayed by a Bio-Rad protein assay reagent (Bio-Rad), using bovine plasma globulin as a standard.
Expression and Purification of Recombinant Proteins-The construc-tion of recombinant plasmids and expression of eEF1A-2/S1 protein in the glutathione S-transferase (GST) system using the pGEX-2T plasmid vector (Amersham Pharmacia Biotech) were reported previously, and the recombinant plasmid was referred to as GST-eEF1A-2/S1 (13). We also amplified the mouse eEF1A-1 cDNA by polymerase chain reaction (PCR) and subcloned it into the pGEX-2T expression vector, resulting in the GST-eEF1A-1 construct. The recombinant vectors were used to transform Escherichia coli. Individual clones were isolated, and the constructs were analyzed by DNA sequencing to identify those carrying the desired fragment, without any mutations that might have FIG. 1. Peptide sequences used to derive eEF1A-1-and eEF1A-2/S1-specific polyclonal antibodies. Diagrammatic representation of the CB5 peptide used to make the eEF1A-2/S1-specific antibody (A) and the HT7 peptide used to generate the antibody specific to eEF1A-1 protein (B).

FIG. 2.
Expression and induction of eEF1A-1 and eEF1A-2/S1 in a GST system. A, Coomassie Brilliant Blue staining of total protein extracts (soluble fractions) of isopropyl-1-thio-␤-D-galactopyranoside-induced and non-induced GST, GST-eEF1A-2/S1, and GST-eEF1A-1 fusion proteins, as described under "Materials and Methods." Note the increase in quantity of the protein bands of 26 kDa for GST and 76 kDa for both GST-eEF1A-2/S1 and GST-eEF1A-1 upon induction with isopropyl-thio-␤-D-galactopyranoside, as indicated by the arrow. B, Coomassie Brilliant Blue staining of fusion proteins purified in non-denaturing conditions, using affinity chromatography with Sepharose 4B beads. As indicated, protein bands were detected at 26 kDa for GST and 76 kDa for GST-eEF1A-2/S1 and GST-eEF1A-1. C, immunoblotting analysis of purified fusion proteins using the eEF1A-1-specific polyclonal antibody. A protein band was detected at 76 kDa for the GST-eEF1A-1 fusion protein, with no evidence of cross-reaction with the GST-eEF1A-2/S1 fusion protein or with the GST protein alone. D, immunoblots of purified GST-eEF1A-2/S1 fusion protein probed with eEF1A-2/S1-specific antibody in the presence of increasing concentrations of either the HT7 or CB5 peptide. Soluble HT7 peptide inhibits binding of eEF1A-1-specific polyclonal antibody, whereas CB5 does not. E, immunoblots of purified GST-eEF1A-2/S1 fusion protein probed with eEF1A-2/S1-specific antibody in the presence of increasing concentrations of either the HT7 or CB5 peptide. Contrary to D, the soluble HT7 peptide does not compete with the eEF1A-2/S1-specific polyclonal antibody for binding with the fusion protein, whereas soluble CB5 inhibits binding. been introduced by PCR. Extraction and purification of GST and GST fusion proteins (GST-eEF1A-2/S1 and GST-eEF1A-1) using glutathione-Sepharose 4B beads were performed as previously described (15).
Electrophoresis and Immunoblotting-SDS-polyacrylamide gel electrophoresis was carried out on 10 or 12% gels using a discontinuous Laemmli system (16). For each assay of protein profile determination, 50 g of total protein extract was loaded onto the SDS-polyacrylamide gel. After electrophoresis, the proteins were either stained with Coomassie Brilliant Blue R-250 or transferred onto nitrocellulose paper. The nitrocellulose membranes were stained with Ponceau S (Sigma) to visualize the electrophoretic pattern of total protein and subsequently destained with Tris-buffered saline (50 mM Tris-HCl, 500 mM NaCl) containing 0.5% Tween 20 (TBST), followed by blocking with 5% milk in TBST for 1-2 h. The membranes were probed with the anti-eEF1A-1 or anti-eEF1A-2/S1 polyclonal antibodies for 2 h at room temperature in TBST at a 1:2000 dilution. The membranes were then washed four times with TBST for 10 min each, incubated with goat anti-rabbit immunoglobulin conjugated to horseradish peroxidase (Cappel, Durham, NC) for 1 h at room temperature in a 1:20,000 dilution, and washed four times as described above. Nitrocellulose blots incubated with horseradish peroxidase-conjugated secondary antibodies were developed using enhanced chemiluminescence detection (ECL) (Amersham Pharmacia Biotech) according to the manufacturer's protocols and exposed to x-ray film.
Immunoprecipitation-eEF1A-1 or eEF1A-2/S1 was immunoprecipitated from total protein extracts of adult rat brain, liver, and skeletal muscle, using a monoclonal anti-EF1A antibody (17) (Upstate Biotechnology, Lake Placid, NY) that recognizes both proteins. One mg of total protein extracts was pre-cleared with 50 l of protein G-agarose (Sigma) for 30 min at 4°C. Of the pre-cleared tissue extract, 720 g was incubated with 3 g of anti-EF1A monoclonal antibody for 17 h at 4°C. Immune complexes were collected with 100 l of protein G-agarose (Sigma) for 3 h at 4°C. The protein G-agarose/immune complex was washed four times with 1 ml of ice-cold lysis buffer containing 1 mM phenylmethylsulfonyl fluoride, resuspended in 100 l of SDS-loading buffer, and heated at 95°C for 10 min. Twenty-five l of the immunoprecipitated mixture was used for Western blotting analysis using the monoclonal anti-eEF1A-1, polyclonal anti-eEF1A-1, or anti-EF1A-2/S1 antibodies.
Densitometry and Quantification-Western blot analysis of protein patterns from wild-type, heterozygous, and mutant mice from embryonic day 18 (E18) and postnatal days 1 (P1), P7, P14, P20, and P28 were scanned and quantified on a Molecular Dynamics Densitometer SI scanner (Sunnyvale, CA). The band intensities were quantified by Im-ageQuaNT TM software 4.1. For each sample, means and standard deviations were derived from specimens isolated from three different animals.
Immunofluorescence Labeling and Confocal Microscopy-All mouse tissues processed for immunofluorescence analysis were immediately embedded in Optimum Cutting Temperature (OCT) compound (VWR, Montréal, Quebec) at Ϫ80°C after dissection. Frozen sections (10 m) were obtained using a Leica microtome type 2800 Frigocut. The tissue sections were incubated with either the anti-eEF1A-1 antibody (1:1000) or the anti-eEF1A-2/S1 antibody (1:100). The secondary antibody used was goat anti-rabbit fluorescein isothiocyanate-conjugated (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). For double labeling, serial sections of samples previously stained for eEF1A-1 or eEF1〈-2/S1 immunoreactivity were incubated with a mouse antibody against a 68-kDa neurofilament neuronal marker (NF-68) (Sigma), followed by incubation with rhodamine-conjugated goat anti-mouse antibody (Cappel). The histochemical preparations were examined using a Bio-Rad MRC-600 laser scanning confocal microscope, equipped with a ϫ 60 oil immersion objective. Images were scanned on two channels (red and green) and merged to produce a signal profile. In this model, all regions exhibiting colocalization of red and green emitters produce yellow fluorescence. The images were collected using the COMOS software package (Bio-Rad) and stored on an optical disc.
To verify the specificity of these polyclonal antibodies against eEF1A-1 or eEF1A-2/S1 proteins, we subcloned the eEF1A-1 and eEF1A-2/S1 cDNAs in the pGEX-2t vector, resulting in a GST fusion protein, and produced recombinant proteins (referred to as GST-eEF1A-1 and GST-eEF1A-2/S1, respectively). After induction with isopropyl-1-thio-␤-D-galactopyranoside, FIG. 3. Immunoblotting analysis of polyclonal antibodies to eEF1A-1 and eEF1A-2/S1 with monoclonal anti-eEF1A-1 immunoprecipitates from adult rat protein extracts. Total protein extracts from 3-month-old rat tissues (liver (L), skeletal muscle (M), and brain (B)) were immunoprecipitated using a commercially available eEF1A monoclonal antibody. Western blot (WB) analyses are shown for both the total protein extracts and the immunoprecipitated mixtures, using a monoclonal antibody to eEF1A-1 (A); a secondary antibody alone (goat anti-mouse, control for monoclonal antibody) (B), detecting only Ig in the immunoprecipitation panel (IP); and eEF1A-2/S1-and eEF1A-1-specific antibodies detecting the former protein in muscle and brain but the latter in liver and sparsely in brain in both total protein and immunoprecipitation (IP) panels (C and D, respectively).
We used a monoclonal antibody that recognizes both eEF1A-1 and eEF1A-2/S1 in an immunoprecipitation reaction against total protein extracts from adult rat tissues (liver, skeletal muscle, and brain) and determined the presence of eEF1A-1 and eEF1A-2/S1 in the immunoprecipitated products by Western blot analysis, using either anti-eEF1A-1 or anti-eEF1A-2/S1 antibodies. Fig. 3 shows the total and immunoprecipitated protein extracts probed with the anti-eEF1A-1 monoclonal antibody (A), secondary antibody alone as a control (Panel B), and the anti-eEF1A-1 and anti-eEF1A-2/S1 antibodies (C and D, respectively). Evidently the monoclonal antibody for eEF1A-1 immunoprecipitates both eEF1A-1 and eEF1A-2/S1 proteins; probing the immunoprecipitates with this antibody detects the presence of both eEF1A isoforms, eEF1A-1 in liver and eEF1A-2/S1 in skeletal muscle and brain. The presence of eEF1A-1 (which has a slightly smaller molecular weight) was not detected in muscle and brain, because we used adult tissues, where the level of eEF1A-1 protein is drastically diminished in later postdevelopmental stages (described in this report; Figs. 4 and 7). However, eEF1A-1 was detected in liver, an eEF1A-2/S1-negative tissue (Fig. 3A). To verify this finding, we probed the immunoprecipitates with anti-eEF1A-1 and anti-eEF1A-2/S1 antibodies; we found eEF1A-1 only in the liver (and a small amount in the brain (Fig. 3D)) and eEF1A-2/S1 only in skeletal muscle, and the latter as a major band in brain (Fig. 3C). These results verify the specificity of the anti-eEF1A-1 and anti-eEF1A-2/S1 polyclonal antibodies in tissue extracts and demonstrate that the commercially available monoclonal antibody to eEF1A-1 recognizes both isoforms of eEF1A indiscriminately.
Tissue-specific Presence of eEF1A-2/S1 Protein in Mice, Rats, and Humans-Our laboratory reported previously that eEF1A-2/S1 mRNA expression is restricted to three tissues: brain, heart, and skeletal muscle. The appearance of eEF1A-FIG. 4. eEF1A-1 and eEF1A-2/S1 proteins in adult mouse, rat, and human tissues. Immunoblotting analysis of eEF1A-1 and eEF1A-2 protein presence in mouse (3-months-old), rat (3-months-old), and human autopsy (68-year-old woman) tissues. Extraction was performed as described under "Materials and Methods." IA, various mouse tissues (brain (B), heart (H), skeletal muscle (M), liver (L), spleen (S)) were probed with eEF1A-1-specific antibody, revealing a 50-kDa protein band in abundance in liver and spleen. The 50-kDa protein is also present in moderation in brain and muscle but absent in heart (arrow). IB, on the other hand, probing the same mouse tissues with eEF1A-2/S1-specific antibody reveals a differentially expressed 50-kDa protein band in brain, heart, and muscle alone (arrow), in addition to two prominent nonspecific 66-and 45-kDa bands (*). IC, immunoblotting with rabbit preimmune serum verifies that the additional bands seen in IB (*) are the result of some pre-immune reaction in the rabbits used for immunization. IIA and IIB, in rat tissues, eEF1A-1 and eEF1A-2/S1 proteins exhibit the same pattern of expression; i.e. eEF1A-2/S1 protein is restricted to the same three tissues, brain, heart, and muscle, whereas eEF1A-1 protein is prominent in liver and spleen (arrow) but reduced in abundance in brain and skeletal muscle and absent in heart. The same nonspecific reactions associated with the eEF1A-2/S1-specific antibody are also observed (*). IIIA and IIIB, human autopsy tissues (68-year-old woman), also probed with the same antibodies, reveal a similar tissue-specific pattern of eEF1A-1 and eEF1A-2/S1 protein expression (arrow). Again, the same nonspecific antibody reaction is seen as with the preimmune serum (*).
2/S1 mRNA is a late differentiation event, occurring after myotubes are formed and neuronal terminal differentiation is completed (1). Using anti-eEF1A-1 and anti-eEF1A-2/S1 antibodies, we investigated the presence of eEF1A-1 and eEF1A-2/S1 proteins in various tissues previously characterized for positive or negative expressions of eEF1A-2/S1 mRNA. Immunoblotting analyses were performed on various tissues isolated from adult mice, rats, and humans. A major band of 50 kDa in mouse liver and spleen was detected using the anti-eEF1A-1 antibody (Fig. 4IA). The 50-kDa eEF1A-1 protein was also present at a very low level in heart and reduced in abundance in brain and skeletal muscle (Fig. 4IA). By probing with anti-eEF1A-2/S1, a 50-kDa protein was detected only in brain, heart, and muscle (Fig. 4IB). Anti-eEF1A-2/S1 also detected two other bands at 66 and 45 kDa; to determine whether these two additional bands were specific, a reaction with a rabbit antiserum obtained prior to immunization was compared with the anti-eEF1A-2/S1 reaction, in which the preimmune serum also yielded 66-and 45-kDa protein bands but lacked the specific reaction with the 50-kDa eEF1A-2/S1 protein (Fig. 4IC). We conclude that the 66-and 45-kDa bands detected using anti-eEF1A-2/S1 are the result of nonspecific reactions of the preimmune serum.
Similar studies in adult rat (3 months) and human autopsy tissues (68-years-old) revealed the same pattern of expression found in mice. In rat and human tissue, eEF1A-2/S1 protein is restricted to brain, heart, and muscle (Fig. 4, IIA, IIB, IIIA, and IIIB). However, eEF1A-1 protein levels in these three eEF1A-2/S1-positive tissues vary among the three tested species, humans, mice, and rats, ranging from non-detectable to modest levels. This variation may be due to species differences. Nevertheless, results of these studies at the protein level are in overall agreement with the specific gene expression patterns examined at the message level, because eEF1A-2/S1 mRNA is only expressed in brain, heart, and skeletal muscle (3,6). We also investigated rat smooth muscle (bladder and stomach) for both eEF1A-1 and eEF1A-2/S1 proteins (Fig. 5); lung and spleen were also studied, along with liver and skeletal muscle, which we know are negative and positive, respectively, for eEF1A-2/S1 protein. Fig. 5A shows that eEF1A-1 protein is present in all tested tissues, whereas eEF1A-2/S1 is only present in skeletal muscle but not in smooth muscle of aorta, bladder, or stomach or any other eEF1A-2/S1-negative tissues (including liver, lung, and spleen) (Fig. 5B).
Genotyping of Wild-type and Mutant Mice-The wasted mouse is a spontaneous mutant strain from the Jackson Laboratories (9). The mutant phenotype was recently shown to correlate with a 15.8-kb deletion, including the promoter and first non-coding exon of the eEF1A-2/S1 gene (8) (Fig. 6A). To detect the presence of the mutant allele, we designed oligonucleotide primers to amplify around the boundary of the deletion (the P1 fragment), with the forward primer upstream and the reverse primer downstream, of the 15.8-kb deleted region (Fig.  6A). PCR using these two primers amplifies a 2.9-kb DNA fragment from both homozygous (Ϫ/Ϫ) and heterozygous (ϩ/Ϫ) mice (Fig. 6B, lanes 2 and 4). This PCR product is not observed with the wild-type (ϩ/ϩ) mouse DNA, because the expected PCR reaction yields a product larger than 18 kb (Fig. 6B, lane  3). To detect a wild-type allele, oligonucleotide primers were designed to amplify a 456-bp fragment within the 1.5-kb deletion (the P2 fragment). The absence of the fragment was used to identify the homozygous mutants (Fig. 6B, lane 6), whereas the PCR product is readily obtained in the parallel PCR reactions of the wild-type and heterozygous animals (Fig. 6B, lanes  7 and 8).
Presence of eEF1A-1 and eEF1A-2/S1 Proteins in Wild-type, Heterozygous, and wasted Mice-Using anti-eEF1A-1 and anti-eEF1A-2/S1 antibodies, we expanded the previous findings of FIG. 5. eEF1A-1 and eEF1A-2/S1 proteins in smooth muscle. To investigate the protein presence of eEF1A-1 and eEF1A-2/S1 in smooth muscle, immunoblotting analysis was performed on adult rat bladder and stomach tissues, as well as skeletal muscle, liver, lung, and spleen. A, various rat tissues were probed with eEF1A-1-specific antibody. eEF1A-1 is very abundantly expressed in smooth muscle tissues from bladder and stomach as well as in eEF1A-2/S1-negative tissues such as liver and spleen. B, the same rat tissues were probed with the eEF1A-2/S1-specific antibody. eEF1A-2/S1 is detected only in skeletal muscle, not in smooth muscle or eEF1A-2/S1-negative controls, including liver, lung, and spleen. the mRNA levels of both isoforms in wasted mutant mice to a study of their protein presence, detailing the temporal pattern of eEF1A-1 and eEF1A-2/S1 protein levels. In wild-type mice, there is a switch from a predominant eEF1A-1 presence to eEF1A-2/S1 protein in brain, heart, and skeletal muscle during development. eEF1A-1 protein levels are high in embryonic life but gradually decline with postnatal age (Fig. 7). In comparison, the level of eEF1A-2/S1 protein is low in embryonic tissues and increases in abundance up to P28 after birth (Fig. 7, A-C). Interestingly, some residual eEF1A-1 protein is still detectable in eEF1A-2/S1-positive tissues at P20 and P28. As expected, liver does not express eEF1A-2/S1 protein at any developmental time point (Fig. 7D), consistent with previous reports where eEF1A-2/S1 mRNA was not detected.
The patterns of eEF1A-1 protein presence in heterozygous and wild-type mice are quite similar (Fig. 7). However, in experiments directly comparing eEF1A-2/S1 protein levels in wild-type and heterozygous animals, we detected a decrease in the protein presence of eEF1A-2/S1 in heterozygous animals (data not shown). This was expected, because heterozygous animals lack one eEF1A-2/S1 allele. In homozygous mutant mice we detected no eEF1A-2/S1 protein throughout the entire period of development, up to and including postnatal day 28 (Fig. 7). The absence of eEF1A-2/S1 protein has no effect on eEF1A-1 protein abundance during development in all three tissues that are normally eEF1A-2/S1-positive; the scheduled reduction in eEF1A-1 protein quantity is indistinguishable from that in wild-type mice. Quantitation of eEF1A-1 and eEF1A-2/S1 expression levels in brain, heart, and muscle of wild-type and heterozygous mice reveals an inverse relationship between the expression levels of eEF1A-1 and eEF1A-2/ S1. In other words, when eEF1A-1 protein is very abundant, during early embryonic development, eEF1A-2/S1 levels are very low and vice versa later. Table I summarizes elongation factor abundance, expressed as a ratio of eEF1A-1 to eEF1A-2/S1 Western blot band intensities (eEF1A-1/eEF1A-2 ratio). In all eEF1A-2/S1-positive tissues, the eEF1A-1/eEF1A-2 ratio is high during embryonic life and gradually decreases during postnatal development.
Immunofluorescence Microscopy-Western blot analysis TABLE I Relative eEF1A-1/eEF1A-2/S1 protein ratio in wild-type, heterozygous, and mutant mice ϩ/ϩ ϩ /Ϫ Ϫ /Ϫ eEF1A-1 eEF1A-2/S1 eEF1A-1/ eEF1A-2/S1 eEF1A-1 eEF1A-2/S1 eEF1A-1/ eEF1A-2/S1 eEF1A-1 eEF1A-2/S1 eEF1A-1/ eEF1A-2/S1  7. Differential expression of eEF1A-1 and eEF1A-2/S1 proteins during development of wild-type, heterozygous, and mutant mice. Mouse tissues from brain (A), heart (B), skeletal muscle (C), and liver (D) obtained at ages of embryonic day 18 (E18) and postnatal (P) days 1-, 7-, 14-, 20-, and 28-days-old were used for immunoblotting assays from wild-type (ϩ/ϩ), heterozygous (ϩ/Ϫ), and mutant (Ϫ/Ϫ) mice. Protein extracts of the four selected tissues were processed for immunoblot analysis with eEF1A-1-and eEF1A-2/S1-specific antibodies, as described under "Materials and Methods." demonstrates that reactivity against eEF1A-1 in the brain of wild-type mice (20-day-old, P-20) is weak, whereas eEF1A-2/S1 reactivity is intense (Fig. 7A). However, some eEF1A-1 protein is still detectable in the mouse brain, making it unclear whether both isoforms are coexpressed in the same cells. To resolve this matter, we analyzed eEF1A-1 expression by immunohistochemical studies in sections of wild-type mouse brain at postnatal day 20. To assess whether eEF1A-1 expression in brain is contributed by neurons or by non-neuronal cells, we double stained with either anti-eEF1A-1 or anti-eEF1A-2/S1 antibodies, anti-glial fibrillary acidic protein antisera for detection of non-neuronal astrocytes, and an antibody against neurofilament (NF-68). Neurofilament is specifically expressed in neurons, serving as a marker for neuronal localization in brain sections (Fig. 8B, H, K, and N, red). Double staining of wildtype brain sections with anti-eEF1A-2/S1 (Fig. 8A) and anti-NF-68 (Fig. 8B) revealed prominent colocalization of both proteins; merging of the red and green images produced a yellow fluorescence (Fig. 8C), indicating that eEF1A-2/S1 protein is present in the neuronal cells of the brain. We labeled nonneuronal GFAP-positive cells (Fig. 8D) to demonstrate the presence of eEF1A-1 immunoreactivity (Fig. 8E) in the nonneuronal population of the brain sections and found GFAP and eEF-1A-1 coimmunoreactivity (Fig. 8F), indicating that both GFAP and eEF1A-1 proteins are localized in the non-neuronal population. As expected, eEF1A-2/S1 protein is not detected in mutant brain (Fig. 8G), although neurons show positive staining with anti-NF-68 in the cell body (Fig. 8, H and I, solid arrows) and dendrites (Fig. 8I, open arrows). On the other hand, brain sections of postnatal day 20 mutant mice reveal similar low level expression patterns of eEF1A-1 in a few cells (Fig. 8J) but strong NF-68 (Fig. 8K), with no obvious colocalization between cells positive for NF-68 (Fig. 8L, open arrows) and those staining weakly for eEF1A-1 (Fig. 8L, solid arrows). Similar to results from mutant mice, double staining wild-type samples with both anti-eEF1A-1 (Fig. 8M, green) and anti-NF-68 (Fig. 8N, red) shows little colocalization (with the exception of two or three neurons with weak eEF1A-1 presence), demonstrating that most eEF1A-1-positive cells have lost this gene expression by this stage of development (Fig. 8O, arrows). These results show that the residual eEF1A-1 presence following the onset of eEF1A-2/S1 appearance is contributed by the non-neuronal cells in the brain or by at most low level presence in neurons. It is likely that analogous situations exist in the heart and muscle, where fibroblasts and satellite cells, respectively, may house most of the eEF1A-1 protein, whereas eEF1A-1 is absent in cardiomyocytes and skeletal myotubes, as in neurons. DISCUSSION We have generated an eEF1A-2/S1-specific polyclonal antibody, which recognizes the eEF1A-2/S1 protein with high affinity. This success, combined with the existing peptide-derived anti-eEF1A-1 (13), provided us two antibodies that distinguish between eEF1A-2/S1 and eEF1A-1 proteins. These antibodies allowed us to demonstrate the following. (i) eEF1A-1 protein presence declines at postnatal day 7, with the concomitant increase of its sister protein, eEF1A-2/S1, in brain, heart, and skeletal muscle. (ii) wasted mutant mice, as expected, do not express eEF1A-2/S1 protein. protein abundance occurs on schedule in wasted mutant mice, as in their wild-type counterparts, indicating that the decline of eEF1A-1 protein abundance is regulated independently of eEF1A-2/S1 protein presence. Therefore, it appears that by postnatal day 28 wasted mutant mice are deficient in both isoforms of peptide elongation factor-1A, which are essential for normal protein synthesis.
During the course of development of the wasted mutant mouse, we failed to detect eEF1A-2/S1 protein at any postnatal stage, which is in agreement with the absence of eEF1A-2/S1 mRNA in mutant mice (8). It is important to note that the absence of eEF1A-2/S1 protein is not compensated for by any increase in eEF1A-1, in agreement with Chambers et al. (8), who detected no substantial level of eEF1A-1 mRNA in the brain, heart, or muscle of wasted mice. Therefore, these tissues in the wasted mutant do not express any known peptide elongation factor-1A; so it is to be expected that there is a deficiency of protein synthesis in these tissues. However, no reports so far describe protein synthesis in wasted mice. In the mouse brain, a residual amount of eEF1A-1 remains even in adulthood, in contrast with heart and muscle, where its expression is almost undetectable. It is well known that the brain has a very heterogeneous cellular composition; thus, to determine whether eEF1A-1 expression is found in neurons, where eEF1A-2/S1 is expressed, immunohistochemical studies were performed. We demonstrate that the residual expression of eEF1A-1 in the adult brain is attributed mostly to non-neuronal cells, suggesting that only neurons express eEF1A-2/S1 as a peptide elongation factor-1A, which is consistent with the previous report of Lee et al. (19).
Taken together, our findings favor the possibility that the cause of death in wasted mutant mice may be a drastic reduction or absence of peptide elongation factor-1A. Despite the absence of eEF1A-2/S1 in the wasted mutant mice, the possibility that other mutations coexist with the eEF1A-2/S1 deletion cannot be excluded, as argued by Chambers et al. (8). Therefore, the phenotype of the wasted mutant may result from additional mutations. In addition to muscle wasting and motor neuron loss, the wasted phenotype is also associated with immune system abnormalities and defects in mechanisms of DNA repair (20), where the presence and role of eEF1A-2/S1 would be obscure.
There are many possible functional implications for a shift in expression from eEF1A-1 to eEF1A-2/S1 in brain, heart, and skeletal muscle during development. It is reported that eEF1A-2/S1 performs peptide elongation in vitro, with the only known difference being a slower rate of GDP dissociation (7), which may result in increased translational fidelity. Indeed, a correlation between translational accuracy and GTP hydrolysis has been reported (21)(22)(23)(24). It is also possible that eEF1A-2/S1 retains the protein elongation role of eEF1A-1 but not all of its other functions, most importantly the bundling of actin filaments (25) and severing of microtubules (26). Investigation into the peptide domains responsible for the shared protein elongation function and for the dissimilar non-canonical functions between the two sister eEF1As will definitively explain the development-dependent shift in expression of these two genes. Nonetheless, our results show that in neurons, myotube and cardiomyocytes, eEF1A-1 is the embryonic form of peptide elongation factor, whereas its sister, eEF1A-2/S1, is the adult form.