Eur. J. Biochem. 171, 293-299 (1988) 0 FEBS 1988

Structural and functional studies of the interaction of the eukaryotic elongation factor EF-2 with GTP and ribosomes Lars NILSSON and Odd NYGARD Department of Cell Biology, The Wenner-Gren Institute, University of Stockholm

(Received September 24, 1987) - EJB 87 1082

The structure of the guanosine nucleotide binding site of EF-2 was studied by affinity labelling with the GTP analogue, oxidized GTP (oGTP), and by amino acid sequencing of polypeptides generated after partial degradation with trypsin and N-chlorosuccinimide. Native EF-2 contains two exposed trypsin-sensitive cleavage sites. One site is at Arg66 with a second site at L ~ s ’ ~ l / L y s ~ ~ ~ . oGTP was covalently bound to the factor between Arg66 and L Y S ~ ~ ’ . After further cleavage of this fragment with the tryptophan-specific cleavage reagent N-chlorosuccinimide, oGTP was found associated with a polypeptide fragment originating from a cleavage at Trp261 and Trp343. The covalent oGTP . EF-2 complex was capable of forming a high-affinity complex with ribosomes, indicating that oGTP, in this respect, induced a conformation in EF-2 indistinguishable from that produced by GTP. Although GTP could be substituted by non-covalently linked oGTP in the factor and ribosome-dependent GTPase reaction, the factor was unable to utilize the covalently bound oGTP as a substrate. This indicates that the conformational flexibility in EF-2 required for the ribosomal activation of the GTPase was inhibited by the covalent attachment of the nucleotide to the factor. EF-2 cleaved at Arg66 were unable to form the high-affinity complex with ribosomes while retaining the ability to form the low-affinity complex and to hydrolyse GTP. The second cleavage at Lys571/ Lys572 was accompanied by a total loss of both the low-affinity binding and the GTPase activity.

The eukaryotic elongation factor EF-2 is a single polypep- tide chain with a molcular mass of 95 kDa [l]. The factor has at least three distinguishable functional properties. Guanosine nucleotides, GTP and GDP, form binary complexes with EF-2 [2], resulting in a conformational change in the factor, thereby adapting EF-2 for interaction with the ribosome [3, 41. The EF-2 . GTP complex associates with high affinity to ribo- somes, having peptidyl-tRNA in the A-site, thereby promot- ing the translocation of the growing peptide chain from the A-site to the P-site [5 - 71. The translocation is coupled to a hydrolysis of GTP to GDP and inorganic phosphate [8, 91. The catalytic domain, presumably located on the factor [lo], is activated by the interaction of the binary EF-2 . GTP complex with the ribosome [l 11. The post-translocation ribosomes formed have a reduced affinity for the factor [7], resulting in a facilitated release of the EF-2 . GDP complex during the ribosomal interaction with the ternary EF-1 . GTP . amino- acyl-tRNA complex.

Mild trypsin treatment of EF-2 results in the formation of four major fragmentation products, T1, T2, T3 and T4 with molecular masses of 82, 48, 34/33 and < 10 kDa respectively [4]. The cleavage products T3 and T4 contain the two terminal parts of the EF-2 polypeptide chain with the T2 fragment

Correspondence to 0. NygBrd, Cellbiologiska Avdelningen, Bio- logi E5, Wenner-Gren Institut, Stockholms Universitet, S-106 91 Stockholm, Sweden

Abbreviations. EF, elongation factor; oGTP, oxidized GTP; GuoPP[CH2]P, guanosine 5’-[P-y-methyleneltriphosphate.

Enzyme. Trypsin (EC 3.4.21.4).

originating from an internal part 141. Limited trypsin treat- ment of EF-Tu results in an analogous degradation at the N-terminal without affecting the function of the factor in poly(phenyla1anine) synthesis [12].

Oxidized GTP (oGTP) has been used as a reactive GTP analogue for affinity labelling of the GTP-binding domain of translational factors [4, 131. The reagent contains two reactive aldehyde groups in the ribose ring, capable of reacting with neighbouring amino groups [14]. By combining trypsin treat- ment with specific chemical cleavage at the tryptophan resi- dues, the GTP-binding domain was localized to an ap- proximately 5-kDa fragment positioned in the center of the T2 fragment [4].

In this report, we have studied the kinetics of the tryptic degradation of EF-2 and the effects of the cleavage on the functional properties of the factor. The factor is rapidly cleaved at Arg66, with a second cleavage at after completion of the first reaction. The high-affinity bind- ing of EF-2 to the ribosome was lost after the cleavage at Arg66. However, the ability of the factor to hydrolyse GTP and to form a low-affinity ribosome complex was unaffected until the factor was cleaved at The results indicated that the EF-2- and ribosome-dependent GTPase was associated with the low-affinity complex. oGTP was bound to EF-2 in the region between G1y262 and Trp343. The covalent EF-2 . oGTP complex could form the high-affinity complex with ribosomes, but the factor was unable to use the covalently bound oGTP as a substrate in the GTPase reaction. The localization and structure of the GTP-binding domain on EF-2 is discussed in view of the data from the prokaryotic elongation factor Tu.

294

MATERIALS AND METHODS

Chemicals

Trypsin and soybean trypsin inhibitor were obtained from Sigma Chemical Co. (St Louis, MO, USA). Sodium dodecyl sulphate, acrylamide, bisacrylamide and Amberlite mixed- bed ion-exchanger MB-1 were from Serva Feinbiochemica (Heidelberg, FRG). The sodium dodecyl sulphate was re- crystallized according to Hunkapiller et al. [15] and the acrylamide and bisacrylamide were purified with Amberlite prior to use. lZ5I, [Y-~'P]GTP and [3H]GTP were from Amersham International, UK. Oxidized GTP was synthesized from [p3'P]GTP or l3H]GTP as described in [14]. Sephadex G-50 F and G-75 SF were from Pharmacia Fine Chemicals (Uppsala, Sweden). Glass-fibre filters GF/C were from Whatman Biochemicals Ltd (Maidstone, Kent, UK). The glass-fibre filters were treated with trifluoroacetic acid and 3-aminopropyltriethoxysilane as described by Aebersold et al. [16]. All chemicals for the gas-phase sequencing were from Applied Biosystems (Foster City, CA, USA).

[3H]oGTP-containing tryptic polypeptide was performed ac- cording to Lischwe and Ochs [20]. The generated fragments were separated on SDS/polyacrylamide gradient gels contain- ing 10-20% (w/v) acrylamide [4] and transferred to glass- fiber filters as above. The stained filters were exposed to autoradiography for 3 weeks at -8O"C, and the fragmen- tation products were sequenced as above.

GTPase activity

The ribosome- and factor-dependent GTPase activity was determined in reaction mixtures containing 22 pmol4OS and 60s ribosomal subunits, 22 pmol trypsin-treated EF-2 and 100 mM KC1, 20 mM Tris/HCl (pH 7.6), 5 mM MgC12, 7 mM 2-mercaptoethanol. The released [32P]P04 was deter- mined as previously described [lo].

The GTPase activity of the covalent EF-2 . [y-32P]oGTP complexes was determined in the presence of 31 pmol4OS and 60s ribosomal subunits and 40 pM [y-32P]-labelled nucle- otides as indicated.

Isolation of EF-2 and ribosomes

EF-2 was isolated from rat liver as previously described [lo] and labelled with '''1 as indicated [17]. Rat-liver ribosomal subunits were isolated as previously described [18].

Separation of the tryptic polypeptides by gel filtration

'251-labelled EF-2 (100 pg) was treated with trypsin as described above. The samples were applied onto columns (5 x 100 mm) containing Sephadex G-75 superfine [21], equi- librated in buffer as indicated. and consecutive 2 2 5 4 frac-

Formation ofthe covalent EF-2 . oGTP complex

EF-2 (1 mg/ml) was incubated for 30 min at 30°C in reac- tion mixtures containing 100 mM KCI, 20 mM Tris/HCI (pH 7.6), 0.1 mM EDTA, 10% (v/v) glycerol, 15 mM 2-mercaptoethanol and 400 pM radioactive oGTP. Non- specific labelling with [3H]oGTP was minimized by including an exess of free amino groups in the reaction mixture. The EF-2 . GTP complexes were separated from non-bound oGTP by gel filtration on columns (100 x 6 mm) containing Sephadex G-50 fine equilibrated in 100 mM KCl, 20 mM Tris/ HCl (pH 7.6), 0.1 mM EDTA, 10% (v/v) glycerol and 15 mM 2-mercaptoethanol.

Trypsin-cleaved EF-2

EF-2 was treated with trypsin (enzyme/substrate ratio 1 : 100, w/w) at 37°C as indicated. Aliquots were withdrawn after 0,2,4, 8, 16 and 32 min and the proteolysis was stopped by adding trypsin inhibitor (inhibitor/trypsin ratio 1 : 1, w/w).

Isolation of peptides f o r sequencing

The covalent EF-2 . [3H]oGTP complex was treated with trypsin for 1 h as described above and the reaction stopped by heating for 1 min at 100°C in the presence of an equal volume of SDS-electrophoresis sample buffer [19]. The poly- peptides were separated by SDS/polyacrylamide gel elec- trophoresis according to Laemmli [I91 using 7- 15% (w/v) linear gradient gels 141 and transferred to glass-fibre filters as described by Aebersold et al. [16]. The filters were stained for 2min with 0.1% (w/v) Coomassie brilliant blue in 40% (v/v) methanol, 10% (v/v) acetic acid, destained in 40% (v/v) methanol, 10% (v/v) acetic acid for 2 min and dried. The stained bands were excised and used for sequencing in a 470 A gas-phase sequenator (Applied Biosystems). In order to avoid blocking of the N-terminus, all chemicals used were of highest purity available [16]. N-Chlorosuccinimide cleavage of the

tions were collected. The radioactivity was determined in a gamma counter and 100-p1 samples were analysed on SDS/ polyacrylamide gradient gels containing 7 - 15% (w/v) acrylamide 141. The gels were stained with 0.1% (w/v) Coomassie brilliant blue in 40% (v/v) methanol, 10% (v/v) acetic acid.

Ribosomal binding of the GTP . EF-2 complex

Ribosomal complexes containing EF-2 were formed by incubating 22 pmol '25T-labelled EF-2 in final volumes of 100 pl, containing 100 mM KC1,4.5 mM magnesium acetate, 30 mM Hepes/KOH (pH 7.6), 7.5 mM 2-mercaptoethanol, 100 mM sucrose, 0.5 mM GuoPP[CH2]P and 40s and 60s ribosomal subunits as indicated, for 10min at 37°C. The ribosomal binding was analysed by sucrose gradient cen- trifugation as previously described [3]. Alternatively, the com- plexes were fixed by the addition of 0.1 vol. 5% (v/v) glutar- aldehyde in 100 mM KC1 and 100 mM Tris/HCl (pH 7.6) prior to centrifugation.

For determination of the binding of the EF-2 . oGTP complex, isolated 251-labelled EF-2 . oGTP complexes (184 pmol) were incubated in the presence of 26 pmol6OS and 40s ribosome and 0.5 mM GuoPP[CH2]P as indicated.

RESULTS

Identification of the tryptic cleavage sites in EF-2

Trypsin cleavage of elongation factor EF-2 resulted in the formation of four major cleavage products, T1, T2, T3 and T4 with apparent molecular masses of 82, 48, 34/33 and < 10 kDa respectively [4]. The kinetics of the degradation was analysed by SDS gel electrophoresis of EF-2 treated with trypsin for various incubation times. The content of the indi- vidual tryptic polypeptides was quantified by use of a laser densitometer (Table 1). As seen in Table 1, EF-2 was rapidly cleaved to the T1 and T4 fragments. This reaction was com-

29 5

Fig. 1. Alignment of the tryptic fragments of EF-2. (A) The fragments. (B) Coomassie staining of the tryptic fragments transferred to 3-aminopropyltriethoxysilane-treated glass-fibre filter as described in Materials and Methods. (C) Amino acid sequence obtained from the tryptic fragments in B. (D) Cleavage point

Table 1. Quantification of the different tryptic polypeptides by laser densifometer scannings of SDSlpolyacrylamide gels EF-2 was treated with trypsin as described in Materials and Methods at 37°C as indicated. Values in parentheses refer to the molar ratio of the polypeptides

Trypsin EF-2 T1 T2 T3 T4 treat- ment

min Yo

- - - 0 100 -

4 - 82 (0.89) 7 (0.12) 2 (0.06) 9 (1.00) 8 - 73 (0.79) 14 (0.24) 6 (0.18) 7 (0.78)

16 - 45 (0.49) 29 (0.49) 18 (0.54) 8 (0.89) 32 - 30 (0.33) 35 (0.60) 26 (0.78) 9 (1.00)

2 - 91 (1.00) - - 9 (1.00)

pleted within 2 min (Table 1). Thereafter, the T1 polypeptide was slowly degraded into the T2 and T3 fragments, with a half-life of approximately 15 min (Table 1). The T4 fragment was further degraded, resulting in a slightly smaller fragment. This reaction was completed within a 16-min incubation.

We have previously shown that the T2 and T3 products derive from a cleavage of the T1 polypeptide and that the T3 and T4 fragments contain the two terminal parts [4]. The orientation of the fragments with respect to the amino ter- minus of EF-2 was determined by amino acid sequencing of the separated fragments. As seen in Fig. 1, both the T1 and the T2 fragments had the same amino-terminal sequence. Thus, the T3 fragment contained the carboxy terminus of the intact EF-2 polypeptide. The N-terminal sequence of EF-2 was different from that of T1/T2 (Fig. l), showing that none of these fragments were derived from the amino terminus of EF-2. In addition, the results suggest that the T4 polypeptide contained the N-terminus of the intact EF-2. The alignment of the tryptic fragments is summarized in Fig. 1.

Recently, the amino acid sequence of hamster EF-2 was determined by DNA sequencing [l]. Thus, the available infor- mation allows localization of the tryptic cleavage sites with respect to the primary sequence of the factor. The N-terminal 13 amino acids of the T1/T2 polypeptides aligned perfectly

with the hamster sequence starting from C Y S ~ ~ . Thus, the T1 polypeptide derived from a tryptic cleavage at Arg66 (Fig. 1). T3 is composed of two closely migrating polypeptides [4]. After sequencing this polypeptide mixture, the 34-kDa and 33-kDa fragments were shown to derive from a proteolytic cleavage at Lys571 and Lys572 respectively (Fig. 1).

Affinity labelling of the GTP-binding site

The reactive GTP analogue oGTP labels EF-2 at a site located near the center of the T2 polypeptide [4]. In order to obtain a more detailed positioning of the GTP-binding site, the N-chlorosuccinimide-derived fragments were sequenced. The T2 fragment was prepared from EF-2 labelled with [3H]oGTP and further cleaved with N-chlorosuccinimide. The generated peptides were separated by SDS/polyacrylamide gel electrophoresis and blotted on to glass-fiber filters. The filters were stained and the labelled fragments were identified by autoradiography (Fig. 2B). By comparison of the hamster EF-2 sequence [l] with the sequence data obtained, the N-terminal amino acids of fragments 3, 5 , 8 and 9 was found to be Ala222 (fragment 8), G1y262 (fragments 3 and 9) and

(fragment 5). Determination of the N-terminal se- quence of the T2 polypeptide (fragment 1) was difficult after N-chlorosuccinimide cleavage, indicating that the treatment resulted in a blocked N-terminus (compared Figs 1 and 2). Furthermore, all fragments expected to have an identical N-terminal sequence to that of the T2 fragment contained blocked N-termini (Fig. 2C). Thus, the origins of the non- sequencable polypeptides 2,4,6 and 7 were deduced from their molecular masses (Fig. 2B). The smallest peptide (Ala222 - Trp26 ') expected from the tryptophan cleavage, was lost dur- ing separation of the generated fragments.

The smallest labelled poypeptide (no. 9) originated from a cleavage at Trp261 and Trp343, indicating that oGTP was covalently linked to an amino group in the G1y262-Trp343 region. According to the proposed alignment of the N-chloro- succinimide fragments (Fig. 2A), all labelled polypeptides (1 -4, 8 and 9) contained this region. Furthermore, the poly- peptides lacking the Gly262-Trp343 sequence (no. 5 , 6 and 7) were not labelled by t3H]oGTP (Fig. 2B), indicating that the labelling was highly specific.

296

Fig. 2. Alignment of the N-chlorosuccinimide-generated fragments from the trypticpolypeptide T2. (A) The polypeptides. (B) Coomassie staining of the N-chlorosuccinimide-generated fragments from [3H]oGTP-labelled EF-2, transferred to 3-aminopropyltriethoxysilane-treated glass- fibre filter as described in Materials and Methods (lane 1) and autoradiogram of lane 1 (lane 2). (C) Amino acid sequence obtained from the fragments in B. (D) N-terminal residue

Table 2. GTPase function of the covalent complex between oGTP and

The covalent [y-32P]oGTP . EF-2 complex was isolated by gel fil- tration and the GTPase activity determined as described in Materials and Methods

EF-2 Table 3. Binding of the EF-2 ' oGTP complex to empty reconstituted ribosomes The covalent complex between 12SI-labelled EF-2 and oGTP was isolated by gel filtration and the ribosomal binding determined as described in Materials and Methods

Additions [32P]P04 released

EF-2 EF-2 ["PIoGTP [32P]GTP ribo- actual above . oGTP somes back-

ground

Additions EF-2 bound to ribosomes

EF-2 El'-2 GuoPP- GTP . oGTP [CHzIP without with

fixation fixation

pmol pmol mo1/100 mol

80 116 80

116 - - - 80

116

80 116

-

- -

1 1 3 2 5 4

22 25 0 92 67 27 28 80 0

168 88 212 132 760 680 810 730

Functional properties of the EF-2 . oGTP complex

In order to assure that the oGTP-binding site in EF-2 was identical to the nucleotide-binding site involved in the binding of GTP and GDP, the functional properties of the EF-2 . oGTP complex were determined. After labelling EF-2 with ly-32P]oGTP and removal of the non-bound nucleotide, the ability of the modified factor to utilize the covalently bound oGTP as a substrate for the factor- and ribosome-dependent GTPase was determined. Approximately 90% of the factor was covalently labelled with oGTP after a 30-min incubation (not illustrated). However, only approximately 4% of the factor-bound oGTP was hydrolysed after a 10-min incubation in the presence of ribosomes (Table 2). The low hydrolysis

+ - - - 32.8 73.2 + - + - 45.1 87.8 - + - + 0.0 44.5 - + + - 69.0 95.0

rate was not due to an inability of oGTP to serve as substrate for the GTPase, as the analogue was hydrolysed by native EF-2. However, the efficiency of this reaction was consider- ably lower than that observed with GTP (Table 2), indicating a continuous inactivation of the native EF-2 by covalent bind- ing of the reactive analogue. The oGTP-labelled factor showed a reduced hydrolysis of supplied [y-32P]GTP. By determining the rate of hydrolysis at increasing concentrations of EF-2 [ll], the remaining activity was found to correspond to the hydrolysis catalysed by the fraction of EF-2 not covalently linked to oGTP (Table 2). Thus, covalent attach- ment of oGTP to EF-2 in the region of G1y262 - Trp343 result- ed in a loss of the GTPase activity even in the presence of added GTP.

Guanosine nucleotides are absolutely required for the in- teraction of EF-2 with the ribosome and the factor forms two types of ribosomal complexes in the presence of guanosine nucleotides. The high-affinity complex allows isolation of ribosome-bound EF-2 by centrifugation while the low-affinity complex is demonstrable only after glutaraldehyde fixation [3]. Since EF-2 formed a stable complex with oGTP this com- plex was purified and used for the formation of ribosome . EF-2 complexes. As seen in Table 3, the covalent EF-2

297

Fig. 3. Gel filtration of trypsinated EF-2. 1251-labelled EF-2 was treated with trypsin as described in Materials and Methods at 37°C for 10 min. The reaction was stopped with trypsin inhibitor and the sample was applied onto a column of Sephadex G-75 SF as described in Materials and Methods

. oGTP formed ribosomal complexes with both high and low affinity [3]. The affinity was comparable to that observed with native factor in the presence of the non-cleavable GTP analogue GuoPPICHz]P (Table 3). Due to the inability of the EF-2 . oGTP complex to hydrolyse GTP, the ribosomal binding was considerable more efficient than that observed in the presence of GTP [3]. As seen in Table 3, addition of GuoPPICHz]P increased the binding of EF-2 . oGTP to the ribosomes. This increase corresponded to the proportion of EF-2 not covalently modified by oGTP. Thus, the attachment of oGTP to EF-2 prevented additional binding of guanosine nucleotides. The results show that oGTP in the covalent EF-2 . oGTP induced the GTP-specific effect on the EF-2-ribosome interaction, suggesting that the nucleotide analogue binds to the GTP-binding site of the factor.

Stability of trypsinated EF-2

In order to determine whether the N-terminal T4 polypep- tide was dislocated from T1 after trypsin cleavage, lz5I-

labelled EF-2 was treated with trypsin and applied to a Sephadex G-75 column equilibrated in 100 mM KC1,20 mM Tris/HCl (pH 7.6), 0.1 mM EDTA, 10% (v/v) glycerol and 15 mM 2-mercaptoethanol. Fractions (225 pl) were analysed on SDS/polyacrylamide gels. As seen in Fig. 3, the T1 and T4 fragments were co-eluted from the column clearly ahead of the trypsin inhibitor, indicating that the polypeptides remained associated even after degradation with trypsin. The two poly- peptides remained associated even in the presence of 1 M KCl (data not shown), but were separated after equilibration of

Fig. 4. Separation ofthe tryptic polypeptides by gelfiltration at acidic pH. '251-labelled EF-2 was trypsinated as in Fig. 2 and the sample was applied onto a column of Sephadex G-75 SF as described in Materials and Methods

the column with 10% (v/v) formic acid and 20mM 2- mercaptoethanol (Fig. 4).

Functional properties of trypsin-treated EF-2 The functional properties of trypsinated EF-2 were ana-

lysed by determining the ability of the factor to associate with ribosomes and catalyse the hydrolysis of GTP. In order to avoid variations in the proportion of the tryptic polypeptides between the functional experiments, aliquots from the same trypsination reaction were used for both functional assays and determination of the content of the polypeptides. Further- more, the functional assays were performed with limiting amounts of EF-2, assuring a response proportional to the factor still active. As seen in Fig. 5, tryptic cleavage of EF-2 resulted in a rapid loss of the high-affinity complex formation. Only approximately 5% of the original activity remained after 2 min of trypsin treatment. Thus, the disappearance of the ability of the factor to form the high-affinity complex coincid- ed with the cleavage of EF-2 at Arg66 (Fig. 5). However, both the EF-Zcatalysed hydrolysis of GTP and the low-affinity binding was only slightly decreased after the 2-min incubation. These activities disappeared at a much slower rate than the high-affinity binding and more than 30% of the activity re- mained after a 32-min incubation. Thus, the loss of these activities followed approximately the degradation of T1 at Lyss71/Lys572 (Fig. 6).

DISCUSSION Elongation factor EF-2 is composed of a single polypep-

tide chain, with a molecular mass of 95 kDa [l]. Despite

298

100

75

- ae

.= 50

Y

>I

.- I

0 Q

25

0 2 4 8 16 min

32

Fig. 5 . Correlation of the high-aflinity ribosomal binding of trypsin- treated EF-2 with the content of’non-cleaved EF-2. For the formation of high-affinity complexes, 22 pmol 1251-labelled EF-2, treated with trypsin, were incubated with 22 pmol4OS and 22 pmol6OS ribosomal subunits as described in Materials and Methods; 100% corresponds to 21.4 pmo1/100 pmol ribosomes. High-affinity binding of EF-2 (A); content of non-cleaved EF-2 ( A )

100

75

0 ae

.= 50

v

>I

._ I

0 a

25

”\

0 2 4 8 16 min

32

Fig. 6. Correlation of’ the low-affinity ribosomal binding of trypsin- treated EF-2 and the EF-2- and ribosome-dependent GTPase activity with the content of EF-2 and TIIT4 complexes. EF-2 was treated with trypsin as described in Materials and Methods. Formation of the low- affinity ribosomal complex and the determination of the GTPase activity was as described in Materials and Methods; 100% corre- sponds to 38.5 pmol EF-2 bound/100 pmol ribosomes and 697 pmol GTP hydrolysed. Low-affinity binding (0) ; EF-2- and ribosome- dependent GTPase activity (m); content of non-cleaved EF-2 and T1/T4 complexes (0)

the large size, proteolytic degradation only resulted in the formation of four major polypeptide fragments, with a prima- ry cleavage at Arg66 and a second cleavage site at Lys571/ L Y S ~ ~ * . In the native factor, rapid cleavage at Arg66 resulted in the formation of two fragments, T1 and T4, with the molec- ular masses of 82 and < 10 kDa, respectively. The trypsin- sensitive site is highly conserved and similar degradation pat- terns have been observed with other elongation factors (EF- l a at Arg69 [21], EF-Tu [22] and EF-G [23, 241 at ArgS8). Fragmentation at Lys 571/Lys572 resulted in the formation of

two new polypeptides, T2 and T3, and a concomitant loss of the T1 fragment. This site was less exposed and the degrada- tion became detectable only after completion of the cleavage at Arg66. Binding of GTP to the factor leads to an exposure of the second cleavage site and a reduction in the ability of diphtheria toxin to ADP ribosylate the diphthamide residue in the C-terminal region [4, 25, 261. Thus, EF-2 shows a guanosine-nucleotide-dependent rearrangement of the struc- ture of the C-terminal domain, similar to that observed with

The T1 and T4 polypeptides, formed by cleavage of EF-2 at Arg66, remained associated in a complex that could only be separated after partial denaturation. Also in this respect EF-2 showed similarity with other elongation factors [21]. In trypsin-treated EF-Tu, the N-terminal 44 amino acids were found associated in a complex with the protein starting from amino acid 59. However, the region containing amino acids 45 - 58 could be released from the complex without affecting the function in poly(phenyla1anine) synthesis, indicating that the cleavage does not affect the factor-ribosome interaction [22]. In contrast, the T1/T4 complex generated by trypsin treatment of EF-2 was not capable of forming a high-affinity complex with the ribosome, suggesting that the structure of the N-terminal portion of the molecule was crucial for the formation of a stable EF-2 . ribosome complex. Interestingly, the cleaved factor was still able to form low-affinity ribosomal complexes and to catalyze hydrolysis of GTP, although with a slightly reduced efficiency.

The reactive GTP analogue oGTP was used to label the GTP-binding site on EF-2 [4, 131. The analogue was found covalently linked to the factor in the region between Gly262 and Trp343 (Fig. 2). Binding of the reactive analogue resulted in a shift of the EF-2 conformation characteristic for guanosine nucleotides [4]. Furthermore, oGTP bound covalently to EF-2 was able to substitute for GTP in the EF-2 . GTP . ribosome complex. Thus, also in this respect, the interaction of oGTP with EF-2 resulted in a complex with functional properties indistinguishable from that obtained with GTP, suggesting that the analogue was bound to the factor at the guanosine-nucleotide-binding site.

During the translocation process, the factor-bound GTP is hydrolysed to GDP and inorganic phosphate [8,9]. Although oGTP not covalently linked to EF-2 could serve as substrate for this reaction, EF-2 was not capable of hydrolysing covalently bound oGTP. Furthermore, the covalent EF-2 . oGTP complex could not hydrolyse added GTP. The reac- tion sequence GTP binding, ribosome attachment and GTP hydrolysis suggests that the GTPase active center of EF-2 is activated only after interaction with the ribosome. Thus, the results indicate that the conformational flexibility of the guanosine-nucleotide-binding domain required for activation of the GTPase was lost after covalent attachment of the reac- tive GTP analogue.

The N-terminal 161 amino acids of EF-2 contain se- quences homologous to the amino acids of the prokaryotic elongation factor Tu [l]. In EF-Tu these homologous regions have been identified as parts of the guanosine-nucleotide- binding site by high-resolution X-ray crystallography [27 - 291. In EF-2, oGTP labelled a region located a further 100- 180 amino acids towards the center of the polypeptide chain. Stereochemical analysis of the EF-Tu . GDP complex showed that the 2’,3‘-hydroxyl groups of the ribose ring are exposed to the solvent [29]. As seen by the formation of a covalent EF-2 . oGTP complex, the 2’,3‘-aldehyde groups of oGTP were located in close contact with the region containing

EF-G [24].

299

G1y262 -Trp343. Interestingly, both EF-1 and EF-G bind oGTP but none of the factors form a covalent complex with the analogue [13,30,31]. The oGTP-labelled region, G1y262 - Trp343, has no counterpart in EF-G [l], indicating that the region in EF-2 covalently bound to oGTP and the positioning of this region as a part of the guanosine-nucleotide-binding domain is unique to EF-2.

In analogy with the ras proteins, the amino acid Gly31, located within the T4 polypeptide, has been suggested to be vital for the GTPase activity [l , 321. Separation of T1 and T4 was associated with a loss of the GTPase activity and attempts to renaturate the T1/T4 complex were unsuccessful. Part of the T4 polypeptide, amino acids 20 - 28, is homologous to the 8-strand 1 of EF-Tu and parts of the N-terminus of T1 is homologous to the P-strands 2 -4 [28]. The interaction of these regions in the same 8-pleated sheet may be responsible for the difficulties in separating and the failure to reconstitute the T1/T4 complex.

The skilful technical assistance of Mrs Birgit Lundberg is grate- fully acknowledged. This work was supported by the Swedish Natural Science Research Council (B-BU 8463-102) and by the Carl Trygger and the Magn. Bergvall foundations. The help of Dr Ib Svendsen with the sequencing of the tryptic polypeptides is gratefully acknowledged.

REFERENCES 1. Kohno, K., Uchida, T., Ohkubo, H., Nakanishi, S., Nakanishi,

T., Fukui, T., Ohtsuka, E., Ikehara, M. & Okada, Y . (1986) Proc. Natl Acad. Sci. USA 83,4978 -4982.

2. Henriksen, O., Robinson, E. A. & Maxwell, E. S. (1975) J . Biol. Chem. 250, 720 - 724.

3. Nygird, 0. & Nilsson, L. (1984) Eur. J. Biochem. 140, 93-96. 4. Nilsson, L. & Nygird, 0. (1985) Eur. J . Biochem. 148,299 - 304. 5. Nolan, R. D., Grasmuk, H. & Drews, J. (1975) Eur. J. Biochem.

6. Nombela, C. & Ochoa, S. (1973) Proc. Natl Acad. Sci. USA 70,

7. Nygird, 0. & Nilsson, L. (1985) Biochim. Biophys Acta 824,

8. Chuang, D.-M. & Weissbach, H. (1972) Arch. Biochem. Biophys.

50,391 -402.

3556-3560.

152 - 162.

152, 114-124.

9. Taira, H., Ejiri, S. & Shimura, K. (1972) J . Biochem. (Tokyo) 72,

10. Nilsson, L. & Nygbrd, 0. (1984) Biochim. Biophys. Acta 782,49 -

11. Nilsson, L. & Nygird, 0. (1986) Eur. J . Biochem. 161, 111 - 11 7. 12. Jacobson, G. R. & Rosenbusch, J. P. (1976) Biochemistry 15,

13. Nurten, R. & Bermek, E. (1980) Eur. J . Biochem. 103, 551 -555. 14. Loewe, P. N. & Beechey, R. B. (1982) Bioorg. Chem. 11, 55-71. 15. Hunkapiller, M. W., Lujan, E., Ostrander, F. & Hood, L. E.

16. Aebersold, R. H., Teplow, D. B., Hood, L. E. &Kent, S. B. H.

17. Bolton, A. E. & Hunter, W. M. (1973) Biochem. J . 133, 529-

18. Nygird, 0. &Nika, H. (1982) EMBO J . 1, 357-362. 19. Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685. 20. Lischwe, M. A. & Ochs, D. (1982) Anal. Biochem. 127,453 -457. 21. Amons, R., Pluijms, W., Roobol, K. & Moller, W. (1983) FEBS

22. Wittinghofer, A,, Frank, R. & Leberman, R. (1Y80) Eur. J . Bio-

23. Skar, D. C., Rohrbach, M. S. & Bodley, J. W. (1975) Biochemistry

24. Kashparov, I . A., Semisotnov, G. V. & Alakhov, Y . B. (1981)

25. Sperti, S., Montanaro, L. & Mattioli, A. (1971) Chem. Biol.

26. Montanaro, L., Sperti, S. & Mattioli, A. (1971) Biochim. Biophys.

27. Morikawa, K., LaCour, T. F. M., Nyborg, J., Rasmussen, K. M., Miller, D. L. & Clark, B. F. C. (1978) J . Mol. Bid. 125, 325- 338.

1527- 1535.

54.

5105- 51 10.

(1983) Methods Enzymol. 91, 227-236.

(1986) J. Biol. Chem. 261,4229-4238.

538.

Lett. 153, 37 - 42.

chem. 108,423-431.

14, 3922 - 3926.

Eur. J . Biochem. 118, 417-421.

Interact. 3, 141 - 148.

Acta 238, 493 -497.

28. Jurnak, F. (1985) Science (Wash. DC) 230, 32-36. 29. LaCour, T. F. M., Nyborg, J., Thirup, S . & Clark, B. F. C. (1985)

30. Bodley, J. W. & Gordon, J . (1974) Biochemistry 13, 3401 -3405. 31. Bermek, E., Nurten, R., Bilgin-Aktar, N., Odabas, U. G. &

Sayhan, 0. 2. (1985) in Mechanism of protein synthesis (Bermek, E., ed.) pp. 180- 191, Springer-Verlag, Berlin.

32. Seeburg, P. H., Colby, W. W., Capon, D. J., Goeddel, D. V. & Levinson, A. D. (1984) Nature (Lond.) 310,644-649.

EMBO J. 4,2385 - 2388.