SRCHIVES OF BIOCHEMISTRY AND BIOPHYSICS 166, 463468 (1973)

A Novel Reaction of Reticuiocyte Peptide-Chain Elongation Factor,

EF2, with Guanosine Nucleotides

TERESA LEE, PHOEBE TSAI,’ AND ROGER HEINTZ

Department of Biochemistry and Biophysics, Iowa State University, Ames, Iowa 60010

Received November 13, 1972

The formation of phenylalanyl puromycin from phenylalanyl-tRNA, bound non- enaymically or enzymically to reticulocyte ribosomes, requires the peptide-chain elongation factor, EF22, and GTP. However the GTP analogue, GDPCP, may re- place GTP to a significant extent in this reaction. Other purine or pyrimidine nucleo- tides have little or no activity. Multistep experiments with either GTP or GDPCP indicate that binding of EF2 to the ribosome for subsequent peptide formation may be a portion of the activity of the EF2 (independent of the translocation reaction) during the elongation process. Neomycin inhibits the formation of phenylalanyl puromycin using either GTP or GDPCP in this system.

A concensus of opinion on the current understanding of the peptide chain elonga- tion phase of protein synthesis is that this process requires the hydrolysis of two guano- sine triphosphate molecules to guanosine diphosphate and inorganic phosphate for t’he formation of each peptide bond (l-3). This stoichiometry, GTP : peptide bond, has been established by studying partial reactions of the elongation process: aminoacyl-tRNA binding to a ribosome : messenger RNA com- plex and the translocation reaction.

In reticulocytes at approximately physio- logical magnesium concentrations the bind- ing of aminoacyl-tRNA t’o ribosomes rc- quires a protein fraction, EFl, GTP, and the appropriate messenger RNA. The existence of an EFl : aminoacyl-tRNA: GTP complex corresponding to that observed in prokaryo- tic systems has not been completely estab- lished but may provide the substrate for this

1 Present address : Department of Biochemistry, University of Iowa Medical School, Iowa City, Iowa 52240.

2 Abbreviations used in this paper are : GDPCP : Guanylyl methylene diphosphonate; EFl : Reticu- locyte aminoacyl transferase I ; EF2 : Reticulocyte aminoacyl transferase II; Rib. : Washed reticulo- cyte ribosomes.

binding reaction (4, 5). During or subse- quent to this binding process GTP is hydro- lyzed to GDP and Pi. Aminoacyl-tRNA in the presence of the appropriate messenger RNA may be bound to washed reticulocyte ribosomes nonenzymically. This process does not require EFl or GTP but does require a high magnesium: potassium ion ratio (1, G, 7).

The entirety of the translocation step is less well understood. Most evidence indi- cates that this process involves: a require- ment for GTP hydrolysis mediated by the EF2 protein, the movement of messenger RNA codons relative to reactive sites on the ribosome, and a change of aminoacyl- or peptidyl-tRNA from one reactive site to another. These sites have been termed the donor or peptide site and the acceptor or amino acid site (3,7). The primary definition of these sites still lies in the reactivity of aminoacyl-tRNA with puromycin. Amino- acyl-tRNA in the donor site reacts to form aminoacyl puromycin while that in the acceptor site will not. The actual peptide synthesis is mediated by an enzyme, pep- tidy1 transferase, which is tightly bound to the larger ribosomal subunit (3).

Our experiments indicate that, at least

463 Copyright @ 1973 by Academic Press, Inc. All rights of reproduction in any form reserved.

464 LEE, TSAI, AND HEINTZ

under certain conditions, the entire function of the translocase (EF2) may not be ex- plained by the simple model implied above. We describe a system in which the formation of phenylalanyl puromycin from phenylal- anyl-tRNA bound to washed reticulocyt’e ribosomes programmed with polyuridylic acid is dependent upon the EF2 protein fraction. The reaction requires GTP, how- ever the GTP analogue, GDPCP, may significantly substitut’e for GTP. This pep- tide synthesis with either GTP or GDPCP may reflect a portion of the physiological function of EF2 and in both cases the reac- tion is inhibited by the antibiotic, neomycin.

MATERIALS AND METHODS

Deoxycholate washed reticulocyte ribosomes and the EF2 protein fraction were prepared as described by Arlinghaus et al. (8). The binding enzyme (EFl) was purified by slight modifications of the procedure of Hardesty et al. (9). The EFZ fraction showed no aminoacyl-tRNA binding activity and the washed reticulocyte ribosome preparation by itself had little or no activity in either the aminoacyl-tRNA binding or transloca- tion assays. Phenylalanyl-tRNA, using yeast tRNA and [%]phenylalanine (sp act 50 rCi/ pmole), was prepared by established procedures (10, 7) and the GDPCP was acquired from Miles Laboratories, Inc. All other reagents were pur- chased from common commercial sources.

The assays for nonenzymic and enzymic binding of phenylalanyl-tRNA to reticulocyte ribosomes as well as for the formation of phenylalanyl puro- mycin were performed essentially as described by Heintz et al. (7). The enzymic binding and phenyl- alanyl puromycin formation reaction mixtures contained 33 lll~ Tris-HCl, pH 7.5, 6.7 mM MgCls and 67 mM KCI. In the nonenzymic binding reac- tion the magnesium and potassium ion concentra- tions were 13 mM and 6.7 mM, respectively (in- cluding 33 mM Tris-HCl, pH 7.5). An exception to the procedures described by Heinta et al. (7) was that dithiothreitol (1 mM) was included in the nonenzymic binding mixture. The ribosome con- centration in all initial aminoacyl-tRNA binding reactions was 1.0 mg/ml.

The procedure for the isolation of ribosomes from the initial binding reaction mixture include : dilution (4-lo-fold) with the appropriate buffer- salt solution at 4°C (for enzymic binding 6.7 mM MgC12, 67 m KCl, and 33 mM Tris-HCl, pH 7.5 and for nonenzymic binding 13.3 mM MgCL, 6.7 mM KCl, and 33 HIM Tris-HCl, pH 7.5)) centrifuga- tion at 105,OOOg for 60 min, several rinsings of the

ribosomal pellet with 2-3 ml of 0.25 M sucrose, gentle suspension of the ribosome pellet in 0.25 M

sucrose and centrifugation of the ribosome solu- tion at 10,OOOg for 10 min to remove aggregated material.

The basic reaction mixture for phenylalanyl puromycin formation contained in a total volume of 0.6 ml: 33 ells Tris-HCI, pH 7.5, 6.7 mM MgCln, 67 mM KCl, 67 m 1-phenylalanine, 10 mM reduced glutathione, 0.67 mM puromycin, 3.3 PM GTP (concentrations and subst,itutions as indicat’ed in various tables and figures), 116 fig/ml EF2, and 0.4-0.9 mg/ml ribosomes carrying 25-35 pmole of [14C]phenylalanine/mg of ribosomes (bound en- zymically or nonenzymically). The reaction mixture was routinely incubated for 20 min at 37°C and the phenylalanyl puromycin formation measured via the ethyl acetate ext,raction pro- cedure of Leder and Bursztyn (11). The specific activities of the two elongation factors, EFl and EF2, were 68 pmole of phenylalanyl-tRNA bound per minute per milligram of EFl and 11.4 pmole of phenylalanyl puromycin formed per minut,e per milligram of EF2 under the conditions of our standard assays.

Characterization of the peptide product, phenylalanyl puromycin, utilized paper chroma- tography on Whatman #I paper. In all cases the product (greater than SOY,) behaved upon such chromatography as expected for phenylalanyl puromycin. The product as chromatographed from the ethyl acetate phase of the Leder-Bursztyn assay (11) was stable to mild alkaline conditions (sufficient to hydrolyze adenosyl phenylalanine). Acid hydrolysis produced a radioactive component identical to L-phenylalanine. The solvents listed (A is n-butyl-alcohol:acetic acid:water; 78:5:17 and B is benzene:pyridine:water; 1O:lO:l) are representative of a number of different solvent mixtures used in such chromatographic analyses (see Table III).

RESULTS

The formation of phenylalanyl puromycin from phenylalanyl-tRNA, nonenzymically bound to washed reticulocyte ribosomes, is dependent upon the EF2 protein fraction and, apparently, GTP (Table I). However, significant peptide-bond synthesis, as mea- sured by phenylalanyl puromycin formation, occurred when GTP was replaced by its “nonhydrolyzable” analogue, GDPCP. This peptide bond forming reaction (phenylalanyl puromycin) absolutely requires the presence of the EF2 protein fract’ion (Table I, Ref. 7).

The concentrations of GTP and GDPCP

FUNCTION OF EUKARYOTIC ELONGATION FACTOR, EF2 465

TABLE I

THE FORMATION OF PHENYLALANYL PUROMYCIN FROM NONENZYMICALLY BOUND

PHI~Nk-L.4LaNYL-tR.NAa

Conditions Phenylalanyl puromycin

(%)

Complete 70 Minus puromycin 4 Minus EF2 8 Minus GTP 12 GDPCP replaces GTP 48

a The results are expressed as the y0 of non- enzymically bound phenylalanyl-tRNA released in the puromycin assay as described in Methods section. The quantity of phenylalanyl-tRNA bound was 25-35 pmolejmg of ribosomes. These results represent an average of more than ten ex- periments using different preparations of ribo- somes, RF2, phenylalanyl-tRNA, GTP, and GDPCP.

necessary for phenylalanyl puromycin for- mation are similar (Figs. 1 and 2). The ulti- mate extent of peptide bond formation, also illustrated in Table I, is different for GTP and GDPCP. This observation is not due to a difference in t’he rates of reaction compar- ing GTP with GDPCP as the time courses of phenylalanyl puromycin formation using saturating concentrations of either guanosine nucleotide are similar. The essential differ- ence is in the extent of t’he peptide bond formation. Figure 2 is included to indicate that the participation of GDPCP in EF2 dependent phenylalanyl puromycin forma- tion is seen using an enzymically formed ribosome : phenylalanyl-tRnTA complex. Both systems are somewhat artificial in that the initiation factors are probably not in- volved and the optimal magnesium ion con- cent’ration is relatively high. However this peptide bond forming system is dependent upon EF2 and must reflect some portion of the activity of this peptide elongation factor.

This activity of EF2 with the guanosine nucleotides (GTP and GDPCP) does not require the simultaneous presence of puro- mycin (Table II). In this experiment, we have first bound phenylalanyl-tRNA to ribosomes nonenzymically; a process which requires only a relatively high hIg+: Ii+ ratio

and the presence of poly U as a messenger RXA. After isolation (Methods) the ribo- somes are incubated with EF2 alone or EF2 plus either GTP or GDPCP and reisolated. These reisolated ribosomes are t,hen incu- bated with only puromycin or puromycin plus GTP. The extent of phenylalanyl puro- mycin formation w&h either GTP or GDPCP (2nd incubation) was similar to that seen in the usual two step reaction (Table I). EF2 by itself in the second reac-

GTP OR GOPCF (AM)

FIG. 1. The GTP and GDPCP concentration dependence of the phenylalanyl puromycin forma- tion from nonenzymically bound phenylalanyl- tRNA. The standard phenylalanyl puromycin formation assay was performed as described in the “Methods” section and the Table II with the indicated concentrations of GTP and GDPCP. The results are expressed as a percentage of the ribosome bound phenylalanyl-tKP\TA converted to phenylalanyl puromycin.

FIG. 2. The GTP and (;DPCP concentration dependence of phenylalanyl puromycin formation from enzymically bound phenylalanyl-tRNA. The standard phenylalanyl puromycin formation assay was performed as described in the Methods section and Table II with the indicated concen- trations of GTP and GDPCP. The results are expressed as a percentage of the ribosome bound phenylalanyl-tRNA converted to phenylalanyl puromycin.

466 LEE, TSAI, AND HEINTZ

TABLE II

REQUIREMENTS FOR A THREE-STEP REACTION LEADING TO PHENYLALANYL PUROMYCIN

FORMATIONS

Second incubation Third incubation Phenylalanyl

Puromycin, GTP, None 70 EF2

GTP, EF2 GDPCP, El% EF2 EF2

Puromycin Puromycin Puromycin Puromycin,

GTP

65 45 10 12

a The conditions for the experiment are the same as in Table I and “Methods” with the condi- tion that the ribosomes were isolated by centrifu- gation between the second and third incubations as described in the text. The concentration of ribosomes in the second and third incubations was approximately 0.5 mg/ml.

tion mixture led to very little phenylalanyl puromycin synthesis. Therefore, such a three-step experiment indicates that in the absence of GTP (or GDPCP) EF2 is not bound to the ribosomes, at least not in amounts so as to be effective for phenylal- any1 puromycin formation in the third incubation.

As the results shown above were unex- pected (“Introduction”), possible artifacts concerning the activity of EF2 need to be ruled out. The concentrations of the com- mercial preparations of GTP and GDPCP necessary for phenylalanyl puromycin for- mation in this system are similar (Figs. 1 and 2). This result is strong evidence against the GDPCP preparation being contaminated with significant amounts of GTP. Also the GDPCP preparations used will not support hemoglobin synthesis, polyphenylalanine synthesis and are only slightly active in the enzymic binding (EFl) of phenylalanyl- tRNA to the washed reticulocyte ribosomes (unpublished results).

The amount of phenylalanyl-tRNA bound enzymically to ribosomes with GDPCP is less than 20 % of that found with GTP after isolation of the ribosomes by centrifugation. The cellulose nitrate filtration assay of Nirenberg and Leder (26) gives a signifi-

cantly larger amount of phenylalanyl-tRNA binding, comparing GDPCP with GTP. Therefore, some difference exists between the phenylalanyl-tRNA : ribosome complexes formed with these two guanosine nucleo- tides. The complex formed using GDPCP is apparently less stable than that formed with GTP and is mostly lost during the relatively long centrifugation procedure under the conditions described for these experiments (“Methods”). The observations strongly in- dicate that the activity of GDPCP in phenylalanyl puromycin formation is due to the analogue not some contamination by GTP.

The product formed in this reaction appears to be phenylalanyl puromycin (Table III). As analyzed by paper chromatog- raphy (Methods) the product is distinct from phenylalanine, puromycin, diphenylalanine, and t,riphenylalanine. The separation of the product from oligophenylalanyl puromycin compounds has not been proven. However, the amount of synthesis of such peptides to be expected in our experiments is much too low to account for the observed synthesis of phenylalanyl puromycin (1,6). For example, Hardesty and his co-workers (1) have shown that diphenylalanine formation in similar reaction mixtures requires GTP which may not be replaced by GDPCP. Also the nucleo-

TABLE III

C~~R~MAT~~+RAPHI~ SEPARATION OF PHENYL- ALANYL PUROMYCIN FROM OLIGOPHENYL-

ALANINE" PEPTIDES

Compound Rf

Solvent A Solvent B

Phenylalanine 0.40 0.04 Diphenylalanine 0.75 0.40 Triphenylalanine 0.84 0.50 Puromycin 0.45 0.68 Phenylalanyl puro- 0.75 0.88

mycin Product after acid 0.42 0.04

hydrolysis

(1 Paper chromatographic separations of the above compounds were performed as described in the “Methods” section. In each case, approxi- mately90y0 of the radioactive product migrated as shown for phenylalanyl puromycin.

FUNCTION OF EUKARYOTIC ELONGATION FACTOR, EF2 467

tides, GDP, GMP, ATP, UTP, and CTP, show little or no activity in this peptide bond forming reaction.

Further evidence that phenylalanyl puro- mycin formation in the described system reflects a portion of the peptidc chain elonga- tion process on reticulocyto ribosomes is shown in Table IV. Neomycin, a strong inhibitor of hemoglobin synthesis on reticulo- cyte polyribosomes, also inhibits phenylal- any1 puromycin synthesis using either GTP or GDPCP. The mechanism of the inhibit’ion by this antibiotic is not complet’ely under- stood but seems similar in many properties to the inhibition by cycloheximide (12). Neo- mycin does not greatly effect the extent of enzymic or nonenzymic binding of aminoacyl- tRNA to the ribosomes but does inhibit poly- phenylalanine synthesis. The inhibition by neomycin of the extent of polyphenylalanine and phenylalanyl puromycin formation is dependent upon the concentrat’ion of GTP or GDPCP (25).

A summary of the results indicates t’hat under the described conditions an aspect of the activity of the EF2 protein in peptide chain elongation requires a guanosinenucleo- tide but that GTP may be replaced to a significant extent by its analogue, GDPCP.

DISCUSSION

The EF2 protein is generally believed (with a large amount of supportive evidence) to be the translocase enzyme in reticulocytes (I, 3). The translocation process requires the hydrolysis of GTP to GDP and inor- ganic phosphate. Therefore, the EF2 de- pendent formation of phenylalanyl puromy- tin in the above experiments in which the requirement for GTP may be significantly replaced by GDPCP (Table I) is difficult to explain in light of current thoughts about the function of EF2 (I, 3, 5). For example EF2 dependent formation of diphenylalanine on reticulocyte ribosomes required GTP which could not be replaced by GDPCP (1). The extent of diphenylalanine synthesis was only about 20% of the total phenyl- alanyl-tRNA bound to the ribosomes while the extent of phenylalanyl puromycin forma- tion in our system is 50-800/o of the bound phenylalanyl-tRNh depending upon whether

TABLE IV

INHIBITION OF PHENYLALANYL PUROMYCIN FORMATION BY NEOMYCIN

Experi- ment

System Phenylalanyl puromycin

(%)

1 Complete” (GTP) 80 2 Same as 1 except GTP re- 52

placed by GDPCP 3 Complete plus 0.67 mM 25

Neomycin 4 Same as 3 except GTP re- 25

placed by GDPCP

a The standard phenylalanyl puromycin for- mation assay from nonenzymically bound phenyl- alanyl-tRNA was performed as described in the “Methods” section and Table II with the follow- ing exceptions: in Exps. 1-4 the GTP or GDPCP concentrations were 6.7 PM.

GTP or GDPCP was used. The difference be- tween the two systems may be that in the sys- tem utilized by Hardesty’s laboratory the diphenylalanine product comes from ribo- somes carrying phenylalanyl-tRNA in both the donor and acceptor sites while we are pri- marily looking at peptide bond formation via puromycin with only one phenylalanyl-tRNA per ribosome. Since GDPCP replacement of GTP results in 40-50% of the bound phenylaIanyl-tRNA being converted to phenylalanyl puromycin one would conclude that approximately half of the bound phenylalanyl-tRNA is in the donor site. It should be stressed that essentially all phenyl- alanyl puromycin formation in this system using either GTP or GDPCP requires EF2.

The observed results indicate that the tot’al function of EF2 during peptide-chain elongation has not been adequately de- scribed. Perhaps the simplest explanation of our results may be that in the binding of phenylalanyl-tRiSA to the washed reticulo- cyte ribosomes in our somewhat artificial systems, enzymically or nonenzymically, a significant amount of the phenylalanyl- tRNA is bound to the donor site of the ribo- some. However, in some undefined fashion the EF2 is necessarily bound to the ribo- some in order for peptide bond synthesis via peptidyl transferase to ensue (13). Such binding of EF2 to the ribosome would

468 LEE, TSAI, AND HEINTZ

normally require GTP or some derivative thereof but the analogue, GDPCP, may partially substitute for GTP in our studies.

The binding of EF2 may be necessary for the proper ribosomal conformation for pep- tide synthesis. Ribosomal conformation has been implied to be an important factor of ribosomal activity during protein synthe- sis. Recently evidence supporting such an idea has been reported (14-17). Also illus- trative of recent evidence which may aid in bridging the void of physical-chemical under- standing of the peptide synthesis sites on ribosomes are a group of papers which show that the elongation factors, Tu and G, as their respective active complexes (3) are simultaneously incompatible upon the same prokaryotic ribosome (18-23). The under- standing of such reactive sites on ribosomes must await further physical and enzymolog- ical descriptions of the various steps in pro- tein biosynthesis.

SCKNOWLEDGMENTS

We appreciate the review of this manuscript by -- -- --- _.I --. several people, especially Dr. J. Horowitz and Wesley Tanaka. This investigation was supported by a National Institutes of Health Research

Grant (HE 12549).

REFERENCES

1. HARDESTY, B., CULP, W., AND MCKEEHAN, W. (1969) Cold Spring Harbor Symp. Quant. Biol. 34. 331.

2. MOLDAVE, K., GAL.~SIR~KI, W., RAO, P., AND

SILER, J. (1969) Cold Spring Harbor Symp. Quant. Biol. 34, 347.

3. LUCBS-LENBRD, J., AND LIPMANN, F. (1971) Annu. Rev. Biochem. 40,409.

4. HRADEC, J., Biochem. J. (1972) 126, 933.

5. SKOGERSON, L., AND MOLD~VIC, K. (1968) J. Biol. Chem. 243, 5354.

6. ARLINGHAUS, R., SCHAFFER, J., AND SCHWPET, It. (1964) Proc. Xat. Acad. Sci. USA 61, 1291.

7. HEINTZ, II., SAL.ZS, M., AND SCH~ISET, 12. (1968) Arch. Biochem. Biophys. 126,488.

8. ARLINGH~US, R., HEINTZ, R., AND SCHWEIST, R. (1968) Meth. Enzymol. 12,70.

9. HARDESTY, B., MCKEEHAN, W., AND CULP, W. (1971) Meth. Enzymol. 20, 316, 330.

10. HOSKINSON, It., AND KHORANA, H. (1965) J. Biol. Chews. 240, 2129.

11. LRDER, P., AND BURSZTYN, H. (1966) Biochem. Biophys. Res. Conrmun. 26, 233.

12. OBRIG, T., CULP, W., MCKEEHAN, W., AND HARDESTY, B. (1971) J. Biol. Chem. 246,174.

13. BALIGA, B., AND MUNRO, H. (1972) Biochim. Biophys. Acta 277, 368.

14. CHUAXG, D., AND SIMPSON, M. (1971) Proc. Nat. Acad. Sci. USA 66, 1474.

15. SCHREIER, M., AND NOLL, H. (1971) Proc. Nat. Acad. Sci. USA 68, 805.

16. SHERMAN, M. (1972) Eur. J. Biochewr. 26, 291.

17. COUTSOGEORGOPOULIS, C., FICO, R., AND MILLER, J. (1972) Biochem. Biophys. Res. Commun. 47, 1056.

18. MILLER, D. (1972) Proc. Nat. Acad. Sci. USA 69, 752.

19. HAMEL, E., KOKA, M., AND NAK.4MOT0, T. (1972) J. Biol. Chem. 247,805.

20. SANDER, G., MARSH, R., AND P~RMEGGIANI, A. (1972) Biochem. Biophys. Res. Commun.

47,866. 21. RICHMAN, N., AND BODLEY, J. (1972) Proc.

Nat. Acad. Sci. USA 69, 686. 22. CBBRER, B., VAZQUEZ, D., AND MODOLELL, J.

(1972) Proc. Nat. Acad. Sci. USA 69, 733. 23. RICHTER, D. (1972) Biochem. Biophys. Res.

Commun. 46, 1850. 24. NIRI<;NBERG, M., AND P. LEDER (1964) Science

146, 1399. 25. LEE, T., HEINTZ, J., AND HEINTZ, Ii. (1973),

in preparation.