Folding on the ribosome of Escherichia coli tryptophan synthase β subunit nascent chains probed with a conformation-dependent monoclonal antibody

J. Mol. Biol. (1992) 228, 351-358

Folding on the Ribosome of Escherichia coli Tryptophan Synthase p Subunit Nascent Chains Probed with a Conformation-dependent Monoclonal Antibody Alexey N. Fedorov’, Bertrand Friguet’, Lisa Djavadi-Ohaniance’ Yuli B. Alakhov3 and Michel E. Goldberg’ 1Institute of Protein Research Academy of Sciences of Russia 142292 Pushchino, Moscow Region, Russia 2Unite’ de Biochimie Cellulaire (CNRS URA 1129) Institut Pasteur, 28 rue du Docteur Roux 75724 Paris Cedex 15, France 31nstitute of Bioorganic Chemistry Academy of Sciences of Russia 142292 Pushchino, Moscow Region, Russia (Received 22 April

1992, accepted 28 July

1992)

Experimental analysis of protein folding during protein synthesis on the ribosome is of nascent polypeptides and the rendered very difficult by the low concentration heterogeneity of the translation mixture. In this study, an original approach is developed for analysing nascent polypeptide structures still carried by the ribosome. Folding on the ribosome of nascent chains of the /3 subunit of Escherichia coli tryptophan synthase was investigated using a monoclonal antibody (mAb 19) recognizing a conformation-dependent antigenic determinant. Upon synthesis of fi subunits in an E. coli coupled transcriptiontranslation system, it is shown that ribosome-bound nascent polypeptides can react with the monoclonal antibody provided their size is above 115 kDa, which is smaller than that’ of both the N-terminal proteolytic and cristallographic domains (29 and 21 kDa, respectively). The gene fragments coding only for the 115 kDa polypeptide, with and without stop codon at the end of the corresponding mRNAs, were constructed and expressed in a cell-free wheat germ translation system. It is shown that antibody 19 reacts with this polypeptide either bound to the ribosome or free in solution. That the 11.5 kDa polypeptide acquires a condensed structure is shown by gel filtration in native conditions and by urea gradient gel electrophoresis. Moreover, it is demonstrated that this condensed structure resembles that of native BZ in the vicinity of the epitope for antibody 19. Indeed, the affinity of antibody 19 for the 11.5 kDa fragment, either free or bound to the ribosome, was measured (6 x lO*m-‘) and shown to be close to that for native b2 . It is therefore proposed that the polypeptide chain may start to fold during its biosynthesis and that, even before the appeara,nce of an entire domain, a folded intermediate is formed that already exhibits some local structural features of the native state and of an immunoreactive intermediate previously detected during the in vitro refolding of denatured complete /I chains. Keywords: Protein folding on ribosomes; monoclonal

351 $08.00/0

conformational

probes.

beginning to become clear (Goldberg, 1985; Ptitsyn, 1987; Kim & Baldwin, 1990). However, the folding of mature proteins during in vitro renaturation may greatly differ from the folding in viwo in the course of synthesis on ribosomes. On the ribosome, a protein is synthesized from the N to the C-terminal end;

The main experimental approach for the study of protein folding is to investigate the renaturation process of mature proteins from the denatured (fully unfolded) state (Anfinsen & Scheraga, 1975: Creighton, 1984; Jaenicke, 1987). A general pattern for the renaturation process of globular proteins is oozz-2836/92/2203514,8

antibodies;

0

1992 Academic

Press Limited

352

A. N. Fedorov et al.

therefore folding of nascent polypeptide chains may already proceed from the N-terminal part during t’ranslation. Moreover. the attachment of the nascent chain by the C-terminal end to the tRNA moiety in the ribosome may affect its ability t,o fold. The aim of this work was to investigate the folding on ribosomes of the /? subunit of tryptophan synthase by means of a conformation-specific monoclonal antihod?. A similar approach had been attempted previously for fl-galactosidase (Hamlin & Zabin, 1972), but the immunological probe used was a polyclonal immune serum, of which the specificity for the native conformation was difficult to ascertain. On t’he contrary. the monoclonal antihod! (mAb 19) used here has been thoroughly charac%erized in previous st,udirs. It recognizes an antigenic determinant carried by the X-terminal part, of the native protein (Friguet rt al.. 1986). The kinetics of appearance of this determinant during the &L vitro refolding have been investigated (Blond & Goldberg, 1987; Murry-Hrelier & Goldberg, 198X: Blond-Elguindi & Goldberg. 1990). Tt was shown that the binding of mAh 19 on hot,h the entire 1 subunit (44 kDa) and its N-terminal proteolgtic fragment (29 kDa) occurs quite early during their ire vitro refolding, but only after several folding steps have been achieved. It was also shown that the early immunoreact,ivc intermediate has the samct rate c*onstant, of association with mAb 19 as native for tnAb 19 is at least P and that its affinity a value only fivefold lower than that 2~1oY. found for the folded Fl domain. This strongI!, suggests either that the population of immunoreactive molecules essentially contains molecules wit,h an epitope approximating t,he native one, or that it) contains a significant amount (at. least 20(:,,) of molecules with the native epitope. Thus. as a result of the folding steps leading to t8hc early immunoreactive species. some structural features resembling those of the na,tive state appear on the polypeptide chain. In order to test whether or not these folding steps can occur on ribosome-bound nascent chains, we tried! in this study, t’o detect and charac%erize [35S]methionine labelled nascent polypeptides still bound to the ribosomes that react with the conformation-dependent ant,ibody 19. For this purpose. synthesis of the p suhunit was done in an E8cherichia coli coupled transcript,ion/translation cell-free system supplemented with a plasmid carrying genes coding for both the r and the /3 subunits of 14:. coli t,ryptophan synthasci. The distribution of complete and nascent chains was analysed af%er a sucrose gradient centrifugation (for details SW legend to Fig. 1). Figure 1A shows that nndcr the conditions used only few final or abortive products are released from the ribosomes and that most of the radioactivity is observed wit,hin the mono- and polyribosome fractions. We then tried to specifically trap. with mAb 19. immunoreactive chains still carried by the ribosomes. This required two conditions to he satisfied. First. all the polyribosomes had to be transformed int,o monoribosomes (i.e. 1 ribosome carrying I

Figure 1. IXstribution of synthesized l)olypeptidc+ among components of thv c~~upled trarls~rif)tion-translation system befhrr. A. and after. B. rrlicWx~c~c~cxJ nu&asc: digestion. Thr 6. voli c~)upletl tr.ansc+)tiot,, translation system was pyart4 a.s tlrscvihtvl ),rt~viousl? (Zubay. 1973) with some minor modifications. Thr systrrn was supplemented with a plasmid coding for thr r acid /i subunits of E. coli tryptophan sgnthast, (a gift from 1)~ Zetina) and [3SS]m&hioninr as radirjaotivr labrl for synthesized (Lhains. A total of I.? ml of t,ransc.riptioll, translation mixture was incubated for 45 ruin at :li’(’ and diluted in 300 ~1 of buffer .A (20 rnM-Tris-H(‘1. pH 7%. 100 mM-K-acetate. IO mM-Mg-acetate). Thrn (‘a(ll, to a final concentration of fi rnM and 10 units of’ ~.alciundependent micrococc~al nuclease (Koehrinaer-hIalltIhrirt1 J were added. Samples were incubated IO min at 20°C’ an{1 EGTA was added to a tinal concentration of 12 I~IYI. Samples were loaded on top of a 1% ~1 isokinetic, 11) to S9?;, sucrose gradient (van der Zeijst Xr Blormers. I!)%) prepared with buffer A. (‘entrif’ugation in a Beckman S\z’ 4LTi swinging bucket rotor was done at 235.000 g :mcl 4°C for 2 h. Fract.ions (0-S ml) were collected, the ahsol+)ante at 260 nm was measured (continuous line) anti trichloroacetic acid-precipitated radioactivity after heat hydrolysis was count,ed (broken line). The positions 111 monorlbosomes are shown by arrows. rpm. count,s!min.

nascent. chain) to prevent binding of the antibottj. to polyribosomes (tarrying nascent polypeptides 01‘ different sizes. including those unable to interac.1 with the antibody. Second. all free nbortivv polypeptides as well as mature polyprptides relrasctl from the ribosomrn had t’o be removed. To 0t)tain only monoribosomrs. digest ion of polyribosomes by micrococcal nuclease was achieved. After nuclease digestion, the radioariivc, labelled polypeptides were revealed predominantjly among monoribosomen (Fig. 1 H). However. somr radioactivity was also found on the top of the gradient. indicating t,he presence of released poly-

Communications

peptides. Thus, to separate monoribosomes from all other components of the translation mixture, a centrifugation step through a glycerol cushion was carried out after the RNase treatment and the pellet containing monoribosomes was collected (for details see legend to Fig. 2). To reveal immunoreactive /? subunit nascent chains, the monoribosome fraction was subjected to immunoadsorbtion with the conformation-dependent monoclonal antibody 19 and two control antibodies, coupled to Sepharose beads. This was done in conditions preventing disruption of the connection between peptidyl-tRNA and the ribosome. The size range of the immunoreactive nascent chains was then estimated by electrophoresis of the immunoadsorbed material on an SDS/l0 to 25 ‘$6 polyacrylamide gel, followed by autoradiography of the radiolabelled peptides. Figure 2 shows the result of SDS/polyacrylamide gel electrophoresis of the translation products before and after immunoadsorbtion with the different antibodies. Before immunoadsorbtion, a set of radioactive labelled polypeptides, the larger of which was similar in molecular mass to the p chain, was observed. Some nascent chains were found to accumulate and give discrete bands, indicating that synthesis of a large number of polypeptide chains is initiated but few of them could reach completion. An explanation for pausing in translation could be either rare codon usage or the high amount of plasmid available and the limiting concentration of [3sS]methionine as compared to the concentration of the other amino acids in the translation mixture. The results presented in Figure 2 show that nascent chains of molecular masses ranging from about 14 to 44 kDa have been trapped by antibody 19. As expected, with an antibody (mAb 164) recognizing a continuous epitope localized in the /? chain between residues 270 and 290 (Friguet et al., 1989), only high molecular mass nascent chains, down to about 30 kDa. were trapped. With a control antibody that is not directed against fi2, no significant immunoprecipitation was observed. To determine the minimal length of nascent’ chains recognized by antibody 19, a similar experiment was carried out, but a urea/acrylamide gel (which is more adapted to the analysis of small polypeptide chains than the gel system of Fig. 2) was used for the electrophoretic analysis of the immunoprecipitated nascent polypeptides. The results (Fig. 3A) show that the minimal mass of the reacting peptides was 11.5 kDa (a minor component of 8 kDa is also observed), which is smaller than the mass of the N-terminal proteolytic domain (29 kDa) and N-terminal crystallographic domain (21 kDa). To investigate in more detail the structural properties of the 11.5 kDa N-terminal polypeptide either bound on the ribosome or free in solution, two DNA fragments coding for a N-terminal polypeptide close to 11.5 kDa were obtained using Polymerase Chain Reaction technology. A DR;A fragment coding for the first 103 amino acid residues and a DNA fragment with a stop codon

Ig G added

353

-

19

164

C

kDa

-

46

-

30

-

21.5

-

a

b

c

14.3

d

Figure 2. SDS/polyacrylamide gel electrophoresis of nascent polypeptide chains carried by the ribosomes and immunoadsorbed with different monoclonal antibodies. The E. coli coupled transcription/translation system was prepared in a total volume of 190 ~1. After 45 iin incubation at 37”C, 300 ~1 of buffer A, CaCI, to 5 IIIM and 100 units of micrococcal nuclease were added. Bfter 20 min incubation at 20°C. EGTA was added to a final concentration of 10 mM and the sample was layered onto 0.5 ml of 40% glycerol prepared in buffer A, and centrifuged in a Beckman Ti 50 fixed angle rotor at 225,000 g and at 4°C for 25 h. The pellet of monoribosomes was resuspended in 280 ~1 buffer A and divided into 3 equal aliquots. Different antibody-coupled Sepharose beads, previously washed twice with @2 ml of buffer A, were added in each aliquot and incubated for 15 min at 4°C. After immunoadsorbtion the beads were washed 3 times with @5 ml of the same buffer with @1 O/ONonidet P40. The adsorbed material was eluted by the sample buffer for electrophoresis. The mixtures were subjected to SDS/polyacrylamide gel electrophoresis (Anderson et al., 1973) on a 10 to 25% polyacrylamide gel and the [35S]methionine radioactivity was revealed after autoradiography. (a) Nascent chains after stopping biosynthesis on the ribosomes. Nascent chains immunoadsorbed onto Sepharose beads coupled with (b) mAb 19, (c) mAb 164 and (d) a monoclonal antibody not specific for the /3 subunit. Molecular mass marker positions are indicated in kilodaltons (kDa).

after the first 102 amino acid residues were constructed. The calculated molecular masses of the corresponding polypeptides are 11,332 and 11,186 kDa, respectively. The DNA fragments were cloned in the pTZ 19 plasmid under bacteriophage T7 phage promotor. mRNAs ending at codon 103 (either stop or LyslO3) were prepared from the

354

A. N. Fedorov et al.

Ig G added

Ig G added

19

-

19

-

19

kDa

i

-12.5

-6.5

a

b

c

a

b

C

d

Figure 3. Immunoreactivity

of nascent polypeptide chains of tryptophan synthase fl subunit with the c,onformat,iondependent monoclonal antibody 19. A, Pattern of immunoreactive nascent polypeptide chains carried by the ribosomes. The E. co& coupled transcription/translation system was prepared in a total volume of 150 ~1 and ribosome-linked nascent chains were obtained as described in Fig. 2. Urea-SDS/polyacrylamide gel electrophoresis (Shtigger K: von ,Jagow. 1987) was used for separation of immunoadsorbed peptides. a. Control beads. b, Beads with coupled mAb 19. (‘. Radioactive labelled marker proteins (Amersham). B, Translation and immunoreactivitv of the K-terminal 11.5 kL)a polypeptide carried by the ribosome and free in solution. Tryp B gene fragments coding ior the 1st 102 and 103 amino acid residues of the tryptophan synthase fl subunit were amplified as described by Saiki ot al. (1988) and cloned in pTZ 19 plasmid behind the T7 phage promotor using the general procedures described by Rambrook et al. (1989). In the case of the 1st fragment? a TAA stop codon was introduced instead of Lysine 103. mRNAs were prepared as described earlier (Sambrook et al., 1989). Translation of both mRNAs was done during 1 h at 25°C in a cell-free wheat germ translation system in a total volume of 25 ~1 according to Erickson & Blobel (1987). The mixtures, before and after 10 min immunoadsorbtion at 4°C with antibody 19 coupled to Sepharose beads, were subjected to SDS/polyacrylamid~ gchl electrophoresis and autoradiography (cf. Fig. 2). Total radioactivity incorporat,ion after translation of mRl”jA with and without, stop codon (a,c. respectively); radioactivity obtained by immunoadsorbtion onto I& beads of the translat,ion product of mRh’As with, b, and without, d, the stop codon.

DKA templates by T7 RNA phage polymerase (see legend to Fig. 3B) and translated in a cell-free wheat germ translation system. Translation was done in conditions of a large excess of mRNA over ribosomes to prevent polyribosome formation and therefore to promote synthesis of the polypeptides essentially on monoribosomes. Figure 3B shows the synthesis of the corresponding polypeptides and demonstrates that they interact with the conformation-dependent

antibody 19. Sucrose gradient sedimentation. similar to those of Figure 1, showed that the polypeptides synt,hesized from the RNA without a stop codon remained linked to the ribosomrs while those synthesized from the RKA with the stop codon were released from the ribosomes. Yet. in both cases thev interact with mAb 19 in a similar way. This indicates that the exposed X-terminal part of’ the polypeptide on t>he ribosome has an immunoreactivr conformation similar to that of the corresponding

Communications

355

A

used here for structure probing, X’

X

a

bcdefghi

B

6000

(I) E

6000

ii 0

4000

2000

0

abcdetghi C

4

1

The validity of this test relies on the following considerations. In the case of ent’ire /I chains and of the N-terminal 29 kDa fragments. it has been shown that the early immunoreactive conformation must be formed before the antibody can bind, thus ruling out that this conformation would be induced by the antibody (Murry-Brelier & Goldberg, 1988; Blond-Elguindi & Goldberg, 1990). Furthermore, as discussed above, studies on the rate and equilibrium constants of the association of mAb 19 with the immunoreactive intermediat’e have clearly demonmolecules represtrated that the immunoreactive sent a significant fraction (and perhaps the

majority) of the molecule population, thus ruling out that the binding of mAb 19 to a very minor fraction of the molecules would result in a drastic shift of an equilibrium between reactive and nonreactive molecules. To ascertain this conclusion, we measured the equilibrium association constants of mAb 19 for the 11.5 kDa fragment (free or ribosome bound) and compared it with that reported for native p2. To determine the affinity of mAb 19 for the ribosome-bound fragment, we used the basic principle of the previously described ISLISA competition method (Friguet et al., 1985) in the following way. [ 35S]methionine-labelled nascent 11.5 kDa fragments were synthesized in the wheat germ translation system supplemented with mRNA without a stop codon (see legend to Fig. 3). Portions of the translation mixture were incubated with various known concent’rat,ions of mAb 19 to let

.

3

-

I> .

0

5

10

(l/a,)

x lo-',

15

20

M-'

Figure 4. Determination of the affinity of mAb 19 for the nasoent 11.5 kDa fragment. Portions (3 ~1) of the translation mixture containing the [35S]methionine radiolabelled ribosome-linked fragment (about 100,000 counts/ min obtained with mRNA without the stop codon) were incubated for 90 min at 4°C with 150 ~1 of buffer A containing different’ concentrations of mAb 19. In all, 50 ~1 of Sepharose beads coupled with mAb 19 were then added to each sample and the mixture was subjected to gentle shaking for 30 s. A total of 1 ml of buffer A supplemented with @l o/, Nonidet P40 was added and the beads were separated by centrifugation for 15 s at 7500g in an Rppendorf centrifuge. The pelleted beads were washed twice more in the same way with buffer A, and then resuspended in 20 ~1 of the SDS sample buffer for 1, 1 .rnl 1 , 1 I rnnon r- n .1--

described in the legend to Fig. 2. The gel was then exposed for 18 h in a /?-imager 1000 (RioSpace Instruments, Paris, France). A, total radioactivity pattern: from left to right, the samples resulted from incubation in. a, the absence of mAb 19, b to h, the presence of mAb 19 at a concentration (expressed in binding sites) of 6 x lo- lo 2x 10-9. 6 x 10-9, 2 x 10-s, 6x lOmE, 2 x 10d7 and IO’” M respectively. and, i, the absence of mAb 19 as in a, but trapping the antigen with Sepharose beads coupled to a monoclonal antibody that does not, recognize the bZ subunit of tryptophan synthase. This sample served as a control for the specificity of t’he immunotrapping. B, Profile obtained by pro,jecting, along line XX’ (see pattern in A), the radioactivity counted in t~he band corresponding to the Il.5 kDa fragment: the total number of counts in each peak corresponds to the radioactivity detected in each slot for that fragment. C. Klotz plot for the binding of mAb 19 to the nascent fragments: the fraction of bound fragments was equal to Y = (R, - R,)/(R, - Ri) where R, was the radioactivity in slot n. R, the radioactivity in slot a and Ri the background radioactivity corresponding to non-specific binding to the beads obtained from slot i. The free mAb 19 concentration, a,, was considered as equal to the total mAb 19 concentration since the latter was always much larger than the total fragment’ concentration. The OPT ., ,‘,I I PI1 II I,?

OM

Urea

+

8M

B Figure 5. Evidence for a compact structure of the immunoreactive 115 kI)a X-terminal polypeptide. X, (:rl tiltration mixt’ure containing the newly synthesized analysis of the 11.5 kDa N-terminal fragment. The translation [“Slmethionine radiolabelled fragment (400,000 counts/min in a total volume of 150 ~1) was applied onto a Sephadex G75 Superfine column (1 x 35 cm) equilibrated in a 20 mM-Hepes. 200 mm-potassium acetate buffer, pH 7.6. The flow rate was 2.5 ml/h and 0.4 ml fractions were collected. The radioactivity incorporated in the polypeptide was count.cad after precipitation with trichloroacetic acid and plotted versus the fraction number. The inset shows a representation of K=, = ( V, - V,)/( V, - V,) as a function of the Stokes radius of the marker protein. I’, is the elution volume of t.he protein. V, the total column volume and V, the void volume of the column. B. Urea gradient gel elect’rophoresis of the I I.5 ki)a K-terminal fragment. An acrylamide-urea gradient gel was prepared as described by Creighton (1979). The translation system containing 400,000 counts/min of the synthesized [ 35~ S] methionine radiolabelled fragment was applied onto thr gel. After migration at, 4°C for 3 h at 400 1’ in 50 mlvr-Tris-acetate buffer, pH 72. t,he radioactivity of the polypeptidts was revealed by autoradiography. cpm. counts/min

equilibrium be reached. Then, a small volume of mAb 19 bound to Sepharose beads was added to each portion. The mixtures were incubated for 30 seconds, at which time further reaction of mAb 19 with the nascent chains was interrupted by dilution and immediate centrifugation of the beads. This resulted in the trapping onto the beads of a constant fraction (about 5%) of the nascent fragments that had not reacted with mAb 19 at equilithe concentration of unreacted brium. Thus, could be determined by fragments nascent t.he radioactivity of t.he fraction measuring immunoadsorbed ont,o the beads. To do that, the submitted to was immunoadsorbed material electrophoresis on a polyacrylamide gel in the presence of SDS. The gel was dried and scanned for 1000 (Biospace radioactivity on a fi-imager Instruments, Paris, France), a high sensitivity and low background radiation scanner built according to the principle described by Charpak et al. (1989), and specially adapted to the detection of 35S. The result of such a scan is shown on Figure 4A. The amount of radioactivity contained in each sample was then determined by integrating the number of counts in each peak of the profile shown in Figure 4B. Analysing the results as indicated by Friguet et al. (1985) gave a value of 6 x lo8 M--~ for the affinity constant. The same value was obtained when this

mAb 19 for the 11.5 kDa fragment dissociated from the ribosomes. Moreover. this value was close to that, ( lo9 K ‘) reported previouslv by Friguet et nl. (1986) for native &. Since the eq;ilibrium constant of the antigen/antibody interaction reflects the structural complementarity bet,ween the antihod) and its epitope, these observations rule out that the observed immunoreactivity would reflect a conformational equilibrium between a minority of’ immunoreactive and a majority of’ non-immunoreactive molecules. This would result in a reduced affinity for the antibody (Furie it rrl.. 197.5). Similarly, it is unlikely that the immunoreactive conformation would be induced by t’he antibody since this would reduce the affinity by the energy involved in the conformational change (Friguet, et al., 1989). Therefore, like /? chains as well as their proteolyt,ic N-terminal domain, the 1 I.5 kDa fragment seems to spontaneously adopt a folded COW formation sharing with fiZ a local structural pattern in the vicinity of the epitope recognized by mAb 19. This was confirmed by two hydrodynamic methods utilizing nanogram amounts of radiolabelled polypeptides synthesized in the cell-free translation system. First, gel filtration in native conditions was done. After synthesis of the polypeptide, directed bl the mRNA with a stop codon to produce free polypeptide, the total translation mixture was analysed

357

Communications

the polypeptide coincided with that of cytochrome c (Fig. 5A). The calibration of the column with three marker proteins yielded a satisfactory linear plot of K,, as a function of Stokes radii (inset to Fig. 5A). From this calibration and the elution volume, one could estimate that the Stokes radius of the 11.5 kDa fragment in solution was close to 1.8 nm, a value compatible with a condensed globular state for a polypeptide chain of this molecular mass. That the peptide is in a condensed state was confirmed by a second method. urea/gradient gel electrophoresis according to Creighton (1979). After translation of the mRNA having the stop codon, the total system was applied onto the gel. A control experiment was performed with lysozyme. which does not undergo unfolding even at 8 M urea (Creighton, 1979). The protein band observed for lysozyme after the migration did not show the S-shaped aspect typical of a denaturationlrenaturation transition (data not shown). On the contrary, the electrophoretic patt’ern of the 11.5 kDa polypeptide showed a transition between a rapidly migrating condensed folded state and a more slowly migrating unfolded state (Fig. 5R) with a transItIon midpoint near 4 M urea (at high urea concentration 2 peptide bands are observed; t.his micro-heterogeneity could be due to cis-trans proline isomerization or to some S-S bridge Therefore, according to the two format’ion). the 11.5 kDa N-terminal fragment methods, behaved as a condensed globular polypeptide. These results, taken together with those showing that the conformation-dependent ant’ibody 19 binds t,o the 1I.5 kl)a fragment with a high affinit?. indicate that this fragment indeed folds mto a cbondensed globular structure that is native-like at least in the region of the epitope to mAb 19. That mAb 19 also binds. with the same high-affinity, to ribosome-hound nascent chains comprising at least the 11.5 kl)a N-terminal polypeptide strongly suggests that the protein may start to fold during its synt,hesis on the rihosome and that local structural patterns closely resembling native ones may be formed before completion of synthesis of the entire protein or even the N-terminal domain. Tn the light of the three dimensional structure of the Salmonella

t~phinrurium

tryptophan

synthase

fl

subunit (Hyde rf al.. 1988) the 11.5 kl)a polypeptide in t,he native protein corresponds to the first three cc-helixes and two b-strands of the b chains. The appearance of the epitope to mAb 19 is therefore likely t.o reflect the formation. on the rihosomes. of a secbondary/tertiary structure in this region of the molecule that much resembles that of t.he correspending region in t,he native protein.

l)at’a on the kinetics of refolding of the fl subunit have indicated that the native-like epitopr is formed within a “molten globule” (Goldberg et al.. 1990). The half-t.ime of appearance of this immunorracative intermediate (10 s at, 12°C) is shorter than the time for the protein synthesis in viao (about 0.5 min at 37Y’) according to t’he average rates of mRNA translation in V~VO (Gouy 8 Grantham, 1980: Sorensen et al.. 1989). Thus the folded nascent

polypeptides we identified on the ribosomes bring solid support to the hypothesis of cotranslational folding. To further test this hypothesis, immunochemical pulse labelling experiments. to compare the rate of appearance of the folded structure to the rate of chain elongation, are in progress. We thank I. Morosov for the E. coli coupled system and V. Chemeris for help in some experiment,s. This work was supported by funds from the Tnstit,ut’ Pasteur. the Universite Paris 7, the Centre National de la Recherche Scientifique (URA 1129), the French Association against Myopathies (AFM) and the Academy of Sciences of Russia. One of us (A.N.F.) benefited from an EMBO Fellowship.

References Anderson. C. W., Baum. P. R. & (iesteland. R. F. (1973). Processing of adenovirus 2-induced probeins. .I. Viral. 12. 241-254. Anfinsen, C. B. $ Scheraga. H. A. (1975). Experimental and t’heoretical aspects of protein folding. ddvan. Protein

(‘hem. 29. 205-300.

Blond. S. & Goldberg. M. E. (1987). Partly native epitoprs are already present on early intermediates in the folding of trpptophane synt’hasr. I’roc. ~Vnt. dcad. Sci., 1’S.A. 84, 1147~1151. Blond-Elguindi. S. B (A~ldberg. M. E. (1990). Kinetic characterization of early immunoreactive intermediat,es during the refolding of guanidine-unfolded Eschrrichia coli tryptophan synthase f12 subunits. Biochemixtry. 29. 2409-2412. Charpak, G.. Dominik. W. & Zaganidis. S. (1989). Optical imaging of the spatial distribution of p-particles emerging from surfaers. Proc. Sot. drnd. Sci.. I’.S.A. 86. 1741m1745. C’rrighton. T. E. (1979). Electrophorrtic analysis of the unfolding of proteins by urea. ,I. Mol. Rio/. 129. 235~264. C’reighton. T. E. (1984). Pathways and mechanisms of protein folding. Advan. Biophys. 18. l-20. Erickson. A. H. R: Blobel. G. (1987). (“ell-free translation of messenger RXA in a wheat germ system. Methods Enmymol. 96, 38-50. Friguet. B.. (‘haffotte. A. F., Djavadi-Ohaniance. L. & (ioldbrrg. M. E. (1985). Measurements of the true affinity cbonst.ant in solution of antigen-antibody ~omplrxes by enzyme-linked immunosorbent assay. J. fmmunol. Methods, 77, 305-319. Friguet. B.. Djavadi-Ohanianct. T,. 8: (Goldberg. M. E. (1986). (‘onformational changes induced by domain assrmblg wit)hin the f12 subunit of Ir:schurichia coli tryptophan synthase analysed with tnonoclonal ant,ibodies. EILT. ,J. Biochmr. 160, 593 597. Friguet, B.. ljjaradi-Ohaniance. L. & (:oldbrrg. M. E. (1989). Polypeptide-antibody binding mechanism: conformational adapt,ation tnvest,ipat,rd by equilibrium and kinetic analysis. Rpg. I,mmuno2. 140. 355-376. Furie. B.. Schecht,er, A. N., Sachs. I). II. & .4nfinsen. C. B. (I 975). An immunological approach to the conformational equilibrium of staphylococral nrrclease. J. Mol. Biol. 92. 497-506. Coldberg. M. E. (1985). The second t,ranslation of the genetic message: protein folding and assembly. Trends Biochem. Sci. 10, 388-391. Goldberg, M. E., Semisotnor. (:. I’.. Friguet, B..

358

A. N. Fedorov et al.

Kuwajima, K., Ptitsyn, 0. B. & Sugai, 8. (1990). An early immunoreactive folding intermediate of the t,ryptophan synthase /12 subunit is a ‘molten globule’. FEBS Letters, 263, 51-56. Gouy. M. & Grantham, R. (1980). Polypeptide elongation and tRNA cycling in Escherichia coli: a dynamic approach. FEBS Letters, 115, 151-155. Hamlin, J. & Zabin, I. (1972). P-Galactosidase: Immunological activity of ribosome-bound growing polypeptide chains. Proc. Na.t. Acad. Sci., U.S.A. 69, 412416. Hyde, C. C., Ahmed, S. A., Padlan, E. A., Miles. E. U’. & Davies, D. R. (1988). Three-dimensional structure of the tryptophan synthase L& multienzyme complex from Salmonella typhimurium. .I. Biol. Chem. 263. 17857r17871. Jaenicke. R. (1987). Folding and association of proteins. Progr. Biophya. Mol. Biol. 49, 117-237. Kim, P. 8. & Baldwin, R. L. (1990). Intermediates in the folding reactions of small proteins. Annu. Rec. Biochem. 59, 631-660. Murry-Brelier, ii. & Goldberg. AM. E. (1988). Kinetics of appearance of an early immunoreactive species during the refolding of acid-denatured Escherichia coli tryptophan synthase /12 subunit. Biochemistry. 27, 7633-7640.

Edited

Ptitsyn. 0. B. (1987). Prot’ein folding: hypotheses and experiments. J. Protein C’hem. 6, 273-293. Saiki, R. K.. Gelfand, I). H.. Stoffel, S., Schorf. S. J ., Higuchi, R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988). Primer-directed enzymatic amplification of DEA with a thermostable DNA polymerase. Science, 239. 487491.

Sambrook, .J.. Fritsch, E. F. & Maniatis. T. (1!#89). Molecular Cloning: A Laboratory Manua,l. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. NY. Shagger. H. & von Jagow, G:. (1987). Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 t,o 100 kDa. ilnal. Biochem. 166, 368-379. Sorensen. M. A.. Kurland, c’. G. & Pedersen. S. (1989). Codon usage determines translation rate in Escherichia coli: physical and rhemical dat,a. J. No!. Riol. 207, 365-377. van der Zeijst. B. & Bloemers, H. (1976). 4698 isokinrtica glycerol and sucrose gradients for density gradient centrifugation. In Handbook of Biochrm. nnd Mol. Biol.: Physical and Chemical Data (Fasman, G rd.) 3rd edit.. vol. 1, pp. 426464. CRC Press, Cleveland. Zubay. G. (1973). In Vitro synthesis of proteins in microbial systems. Annu. Rez!. (Tenet. 7. 267-287.

by B. Matthews