Purification and Characterization of Elongation Factor G from Bovine Liver Mitochondria”

The mitochondrial protein synthesis translocase elongation factor G,, (EF-G,t) from bovine liver has been purified to greater than 90% homogeneity by a combination of conventional gravity and high perform- ance liquid chromatography. The purification scheme results in an approximate overall 14,000-fold purifi- cation with 2% total recovery of EF-G,t activity. Gel filtration chromatography and sodium dodecyl sulfate- polyacrylamide gel electrophoresis indicate that the mitochondrial factor is a single polypeptide with a molecular weight of 80,000. EF-G,t displays similar levels of activity on its homologous mitochondrial ri- bosomes and on Escherichia coli ribosomes. The mitochondrial translocase is sensitive to temperatures above 37 “C, but the factor is partially protected from heat inactivation in the presence of GTP or GDP. The activity of EF-G,r is inhibited by treatment of the factor with N-ethylmaleimide. In contrast to all other translocases tested to date, EF-Gmt is completely re- sistant to the inhibiting effect of fusidic acid when tested on its homologous ribosomes. displays weak

The mitochondrial protein synthesis translocase elongation factor G,, (EF-G,t) from bovine liver has been purified to greater than 90% homogeneity by a combination of conventional gravity and high performance liquid chromatography.
The purification scheme results in an approximate overall 14,000-fold purification with 2% total recovery of EF-G,t activity. Gel filtration chromatography and sodium dodecyl sulfatepolyacrylamide gel electrophoresis indicate that the mitochondrial factor is a single polypeptide with a molecular weight of 80,000. EF-G,t displays similar levels of activity on its homologous mitochondrial ribosomes and on Escherichia coli ribosomes. The mitochondrial translocase is sensitive to temperatures above 37 "C, but the factor is partially protected from heat inactivation in the presence of GTP or GDP. The activity of EF-G,r is inhibited by treatment of the factor with N-ethylmaleimide.
In contrast to all other translocases tested to date, EF-Gmt is completely resistant to the inhibiting effect of fusidic acid when tested on its homologous ribosomes. It displays weak sensitivity to this antibiotic when assayed in the presence of heterologous E. coli ribosomes.
Mammalian mitochondria contain a protein synthesizing system distinct from that of the cell cytoplasm. The translational machinery of mammalian mitochondria remains to be fully explored, but recent progress indicates that this system has some characteristics which distinguish it from other protein synthesizing systems. For instance, animal mitochondrial ribosomes have a low ribosomal RNA to protein ratio resulting in a sedimentation value of about 55 S compared with the values of 80 S for cytoplasmic and 70 S for prokaryotic ribosomes (1). Bovine liver mitochondrial ribosomes can selectively interact with bacterial EF'-Tu, but not with the corresponding cytoplasmic factor EF-1 in aminoacyl-tRNA binding to the A site (2). Recently, an EF-Tu.EF-Ts complex has been purified from bovine liver mitochondria.
This complex is active on Escherichia coli ribosomes; however, unlike the corresponding prokaryotic and eukaryotic complexes, mitochondrial EF-Tu. EF-Ts cannot be dissociated by guanine nucleotides (3). Animal mitochondrial ribosomes function with prokaryotic EF-Tu, but they display no activity with either the prokaryotic or the eukaryotic translocases.
* This work was supported in part by National Institutes of Health Grant GM32734. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. preparations.
The EF-Gmt activity in these crude extracts remains quite stable when stored at -70 "C. However, due to the lability of EF-Tu/Ts,, , samples were generally processed within 1 week after preparation.
The total initial activity of EF-G,, varied with different livers, and the most active preparations were often obtained from livers of black male Angus cattle. In order to obtain extracts having maximal activity, the liver was always placed on ice within 30 min after the animal was slaughtered.
Purification of the Bovine Liver EF-G,,-The purification scheme of EF-G,, consisted of both conventional gravity and high pressure liquid chromatography.
In the first step of the isolation procedure, the crude mitochondrial extract (3) was applied to a gravity DEAE-Sepharose column. EF-G,, eluted immediately following EF-Tu/Ts,~ at 220 mM KC1 with about 71% recovery of activity and a 27-fold increase in specific activity (Table I). This step separated EF-G,t from EF-Tu/ Tsmt (data not shown). The sample at this stage is quite stable and can be stored for several weeks at -70 "C. EF-G,, was further purified on a gel filtration column (Sephacryl S-200). This step resulted in an extremely high recovery of EF-G,, activity with a 4-fold increase in specific activity (Table I). The native molecular weight of EF-G,, was determined to be 80,000 from a 50-ml S-200 column using yeast alcohol dehydrogenase, bovine serum albumin, and hen ovalbumin as standard molecular weight markers. The molecular weight of EF-Gmt is comparable to those observed for other organellar translocases and for E. coli EF-G (4-8).
The remainder of the purification scheme employed HPLC columns for effective purification of EF-Gmt. For the first HPLC purification step, the sample was applied to a TSKgel DEAE-5PW column. As indicated in Fig. 1, essentially all of the protein was retained by this resin. However, when the gradient was apphed, EF-G,, eluted toward the back of the main protein peak, and 75% of the enzyme activity could be recovered in fractions containing only 10% of the applied protein. This step allowed a B-fold increase in specific activity from the previous step (Table I). The sample was further purified after application onto a Progel-TSK heparin-sepharose-5PW column. As indicated in Fig. 2, the majority of the EF-G,, activity eluted with the front shoulder of the second protein peak. This step yielded 55% of the EF-G,t activity with lo-fold increase in specific activity from the previous purification step (Table I). EF-G,, activity at this stage appears to be labile upon prolonged storage at -70 "C due to low amounts of protein present (less than 0.5 mg).
For the final purification of EF-G,, the sample was applied to a TSKgel SP-5PW column. To preserve the maximum EF-G,t activity, this procedure was generally performed within 3 days of the previous step. The final purification step resulted in an elution of a symmetrical protein peak containing about 30-50% of input EF-G,, activity (Fig. 3). This step resulted in the removal of substantial amounts of contaminating protein (Fig. 4). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of each fraction from the TSKgel SP-5PW column containing EF-Gmt activity indicated that fractions from the front portion of the peak contained faint traces of contaminating proteins. Fractions corresponding to the center of the protein peak, which contained the highest EF-G,t activity, were purified essentially to homogeneity (Fig.  4B). These fractions, representing about 10% of the input EF-G,, activity, were selected for the following characterization experiments.
This sample is estimated to have a 2-fold column.
The column was developed as described under "Experimental Procedures." Aliquots of each fraction were diluted 5-fold, and 5-4 samples were assayed for EF-G,, activity (0). Absorbance at 280 nm was monitored (-) using a scale from 0 to 0.5. A series of linear salt gradients (---) were used to develop the column. I  I  I  I  I   I  I  I  I  10 20 A Progel-TSK purified sample was subjected to chromatography on TSKgel SP-5PW as described under "Experimental Procedures." Aliquots of each fraction were diluted lo-fold, and 5-J samples were assayed for EF-G,, activity (0). Absorbance at 280 nm was monitored (-) using a scale of 0 to 0.2. A three-part gradient (---) was used to develop the column. increase in specific activity from the previous step. The recovery of EF-Gmt activity from the TSKgel SP-5PW column was dependent on the amount of protein loaded. Samples containing higher amounts of protein generally gave higher recoveries of activity. For this reason, we recommend that two to three samples purified up to the TSKgel DEAE-5PW stage be combined and processed together through the remainder of the purification scheme. The protein concentration of the pure sample was difficult to determine precisely due to the very low amounts of protein present. The estimated amount of protein in the selected fractions containing the highly purified EF-G,, was about 10 Fg. Overall, this purifi- cation scheme results in approximately a 14,000-fold increase in the specific activity of EF-G,, with a total recovery of about 2% of the activity. The purified factor consists of a single polypeptide chain with a molecular weight of 80,000 (Fig. 4). This value corresponds to the molecular weight determined by gel filtration chromatography indicating that EF-G,, functions as a monomer. An estimate of the turn-over number indicated that 1 pmol of EF-G,, promotes the incorporation of 300 pmol of phenylalanine in the poly(U)-directed polymerization assay. The approximate specific activity of the pure mammalian mitochondrial translocase is about 4 x 10' units/mg protein. E. grucilis chloroplast and mitochondrial translocases have specific activities of about 1 x lo6 and 5 x lo", respectively (4,6). Pure cytoplasmic EF-2 from pig liver has a specific activity comparable to that of EF-G,, (9).

Interchangeability
of EF-G,, on Heterologous Ribosome-As indicated above, we routinely used E. coli ribosomes to assay the activity of EF-G,, in the poly(U)-directed polymerization assay. EF-G,, is equally active on heterologous E. coli ribosomes as on its homologous ribosomes (data not shown). However, it has been shown (2) that mitochondrial ribosomes do not reciprocate this interchangeability, and E. coli EF-G shows no polymerization activity on mitochondrial ribosomes. This stringent recognition possibly suggests that mitochondrial ribosomes have evolved a specific site on which only its homologous translocase can function. It has previously been reported that mammalian EF-G,, is not active on E. coli ribosomes (17). This observation was based on an effort to detect EF-G,, activity in poly(U)directed polymerization assay using E. coli tRNA and ['"Cl phenylalanine rather than precharged tRNA. We believe that the absence of activity in this system arose from a lack of tRNA charging rather than from a failure of EF-G,, to function on E. coli ribosomes. The above report also attempted to measure EF-G,, activity in crude mitochondrial extracts using an uncoupled GTPase assay in the presence of either mitochondrial or E. coli ribosomes. However, the amount of crude extract used probably contained too little EF-G,l to be detected in this assay. It is also conceivable that the mam-malian translocase may not possess an inherent uncoupled GTPase activity as does its prokaryotic counterpart.
EF-G,r Protection from Heat Inactivation by Guanine NUcleotides-Purified EF-G,, is quite stable at temperatures up to 37 "C, but its activity declines rapidly thereafter until complete inactivation is observed at 65 "C after 15 min of incubation (data not shown). We tested the ability of guanine nucleotides to protect EF-G,, from heat inactivation. As seen in Fig. 5, with 30 min of incubation at 44 "C, EF-G,, retained only 25% of its original activity. However, in the presence of GDP or GTP, EF-G,, retained 55 and 70% of its activity, respectively. All translocases which have been examined to date show an interaction with GTP (18,19) and many are protected from heat inactivation by the presence of guanine nucleotides (20). This GTP protection effect has also been observed with the translocase from archaebacterium Hulobacterium halobium (21). The protection observed is thought to be caused by the binding of GTP to the translocase, resulting in a conformational change of the protein.
Effect of Fusidic Acid on EF-G,,,! Activity-Fusidic acid is a steroidal antibiotic which inhibits translocation by stabilizing the ribosome-translocase-GDP complex. All translocases tested to date display sensitivity to this antibiotic. However, as seen in Fig. 6, the mammalian mitochondrial translocase is completely resistant to fusidic acid when tested with its homologous ribosomes. When tested on E. coli ribosomes, EF-G,, loses 50% of its polymerization activity in the presence of 1 mM fusidic acid. Concentrations of the antibiotic required to inhibit 50% of a variety of translocases have been determined (4). There is a variation of over 1000-fold in the antibiotic sensitivities by these factors. E. coli EF-G requires 35 pM for inhibition, while the wheat germ cytoplasmic EF-2 is inhibited by 9 pM fusidic acid. The organellar translocases EF-Gchl and EF-G,, from E. grucilis are inhibited by fusidic acid concentrations of 3 and 20 pM, respectively, when assayed on E. coli ribosomes, while the cytoplasmic factor EF-2 of this organism required 550 pM fusidic acid for inhibition when tested on wheat germ 80 S ribosomes. Yeast EF-G,, like the mammalian mitochondrial factor, requires up to 1 mM for 50% inhibition when assayed on E. coli ribosomes. In general, fusidic acid sensitivity appears to be dependent primarily on the translocase and not the ribosome used (4). Bovine EF-G,,, however, appears to diverge from this pattern since it is resistant to fusidic acid on its homologous ribosomes, but is