Proc. Nat. Acad. Sci. USAVol. 69, No. 12, pp. 3769-3773, December 1972

Structural and Functional Properties of Ribosomes Crosslinked withDimethylsuberimidate

(E. coli ribosomes/initiation complex/protein synthesis/60S ribosomes)

LAWRENCE I. SLOBIN

Division of Biological Sciences, Cornell University, Ithaca, New York 14850

Communicated by Leon A. Heppel, October 10, 1972

ABSTRACT To test whether a 30S ribosomal subunit-formylmethionyl-tRNA-mRNA complex is an obligatoryintermediate in protein synthesis, 70S ribosomes fromEscherichia coli were crosslinked with the bifunctionalimidoester, dimethylsuberimidate. Crosslinked ribosomescontained covalently joined 30S and 50S subunits, asjudged by their inability to dissociate at low Mg2+ concen-trations. Treatment of 70S ribosomes with high salt (1 MNH4Cl), either before or after reaction with the crosslink-ing reagent, produced two different crosslinked ribosomalparticles, one of "60 S" and the other "70 S." Preliminaryevidence indicates that both particles can bind N-acetyl-phenylalanyl-tRNA at low Mg2+ concentrations and areactive for polyphenylalanine syntheses.Crosslinked ribosomes were functional when tested with

poly(U) as an mRNA in systems requiring initiation fac-tors and N-acetylphenylalanyl-tRNA for activity. Underoptimal crosslinking conditions, they retained 80% of theactivity of unmodified ribosomes for polyphenylalaninesynthesis. Despite the maintenance of these functionalcapacities, such ribosomes had a sharply reduced abilityto bind fMet-tRNA and were completely inactive in proteinsynthesis with bacteriophage f2 RNA as a messenger. Weconclude that 70S ribosomes must dissociate into sub-units to initiate protein synthesis with natural mRNAs.

The formation of an initiation complex involving a 30S ribo-somal subunit, fMet-tRNAf, and mRNA has generally beenaccepted as the first step in the initiation of protein synthesisin bacteria. Much of the evidence for this model has beenreviewed by Guthrie and Nomura (1), who attempted todemonstrate that such a 30S initiation complex is an obliga-tory step in protein synthesis. Their evidence rested on experi-ments involving ribosomes labeled with heavy isotopes:fMet-tRNAf was found only on hybrid (heavy-light) ribo-somes produced during incubation with light ribosomal sub-units, and not on intact heavy ribosomes supplied at the start.However, as has recently been shown (2), free 70S ribosomesrapidly exchange subunits in bacterial extracts. Therefore, it ispossible that hybrid ribosomes could have been formed byexchange of subunits with free ribosomes. In view of recentevidence that free (runoff) 70S ribosomes can accumulate inEscherichia coli under certain growth conditions (3, 4), itseemed worthwhile to reinvestigate the problem of whether aninitiation complex involving a 30S ribosomal subunit is anobligatory step in protein synthesis.The major experimental obstacle to resolving this question

is the fact that 70S ribosomes are always potentially capable ofdissociation and reassociation, particularly at physiologicalMg2+ ion concentrations. This is particularly true of ribosomes

Abbreviations: AMe2Sub, dimethylsuberimidate; CLR, cross-

linked ribosomes, i.e., 30S-50S complexes covalently joined bytreatment with Me2Sub. Ribosomal subunits treated separatelywith Me2Sub, as well as the various forms of CLR, are indicatedby the S value of the particle; thus, 30S CLR, 60S CLR, etc.

3769

washed in buffers of high-salt concentration that are free ofextraneous nonribosomal components (mRNA, peptidyl-tRNA, etc.). For example, Nomura et al. (5) observed that70S ribosomes washed with high-salt concentrations couldform an initiation complex in the presence of initiation factors.The binding of fMet-tRNA to 70S ribosomes was inhibited byan excess of 50S particles. This finding led Nomura and hiscollaborators to suggest that initiation obtained with 70Sribosomes was due to a small amount of free 30S particles inequilibrium with the 70S ribosomes; 50S particles were in-hibitory because they reduced the fraction of 30S subunits.To prevent the dissociation of 70S particles, I have suc-

ceeded in crosslinking the ribosomal subunits to one anotherusing dimethylsuberimidate, a bifunctional imidoester. Someof the structural and functional properties of these ribosomesare reported here. I found that crosslinked ribosomes retainmany of their functional properties, except the ability to per-form protein synthesis with natural mRNA. In addition,treatment with crosslinking reagent leads to the formation,under certain conditions, of a 60S ribosome.

MATERIALS AND METHODS

E. coli strain Q13 cells were obtained from General Biochemi-cals. ["C ]Phenylalanine (492 Ci/mol) and [14C ]valine (50Ci/mol) were obtained from Schwarz/Mann; and [3H Imeth-ionine (5 Ci/mmol) was from Amersham/Searle. The follow-ing materials were also purchased: poly(uridylic acid) andE. coli B tRNA (General Biochemicals); pyruvate kinase(10 mg/ml) and puromycin (Sigma).Ribosomes (70 S) were prepared from cells ground with

alumina as described by Traub et al. (6). Salt-washed ribo-somes were prepared by two successive treatments of ribo-somal solutions with 1 M NH4Cl, followed by centrifugationthrough a cushion of 1.1 M sucrose containing 1 M NH4Cl.Only salt-washed ribosomes were used as unmodified controlribosomes in the assays described below.30S and 50S subunits were obtained by zonal centrifugation

of salt-washed 70S particles on a sucrose gradient (6). E. colisupernatant enzymes and crude initiation factors were pre-pared according to Traub et al. (6). ['4C]Phenylalanyl-tRNA(7), N-acetyl-[14C]phenylalanyl-tRNA (8), and formyl-[3H]-methionyl-tRNA (9) were synthesized by establishedprocedures from unfractionated E. coli tRNA. Only fMet-tRNAf (and not Met-tRNAm) is bound to ribosomes in thepresence of f2 RNA (10). Bacteriophage f2 RNA was pre-pared by phenol extraction (11). Dimethylsuberimidate dihy-drochloride was synthesized as described (12).

Assays. The poly (U)-directed binding of N-acetyl-[14C]Phe-tRNA to ribosomes by the nitrocellulose filter technique (13),as well as reactivity of bound N-acetyl-Phe-tRNA with puro-mycin and polyphenylalanine synthesis, were determinedaccording to Lucas-Lenard and Lipmann (14). Bacteriophage

Dow

nloa

ded

by g

uest

on

Dec

embe

r 9,

202

0

Proc. Nat. Acad. Sci. USA 69 (1972)

E

M0.8(B) 30 S 50 S (D) 50 S 68 S q

0.6 - 0.2

0.4-

0.1

0.2-

top boytm top bofom

FIG. 1. Sedimentation patterns of crosslinked ribosomes.Samples were centrifuged into 17-ml 10-30% linear sucrose

gradients in Buffer 1 containing 1 mM Mg(OAc)2 with a SW27.1rotor. (A) 30S subunits crosslinked for 8 hr and centrifuged for17 hr at 17,000 rpm. (B) 50S subunits, crosslinked for 8 hr andcentrifuged for 17 hr at 17,000 rpm. As can be seen, the SOSparticles were slightly contaminated with 30S subunits. (C)70S ribosomes (see text) crosslinked for 4 hr and centrifuged for4.5 hr at 27,000 rpm. (D) Same CLR as in (C), after purificationby zone centrifugation, centrifuged for 4.5 at 27,000 rpm.

f2 RNA-directed incorporation of ['4C valine into protein(6) and binding of f- [3H ]Met-tRNA to ribosomes (10) were

measured by the procedures of Nomura and his collaborators.

Crosslinking of Ribosomes. Ribosomal solutions (5 mg/ml of70S, 2.5 mg/ml of 30S and 50S subunits) were dialyzed over-

FIG. 2. Effect of p-hydroxymercuribenzoate on the stability ofCLR. CLR and conditions of centrifugation were the same as

described in Fig. 1D. Unmodified 70S ribosomes were not high-salt washed. The gradients contained 10 mM Mg(OAc)2. Ribo-somes (2.4 A260 units, in 0.2 ml) in Trist HCl [20 mM Trist HCl(pH 7.5)-50 mM KCl-10 mM MgCl2] were treated with 20 u1 of a

10 mM solution of the mercurial in Tris- HCl. After incubationfor 1 hr at 37°, 150-,ul aliquots were layered onto the sucrose

gradients. Control ribosome solutions (incubated in the absenceof the mercurial) were also analyzed. (A) Unmodified ribosomesand (B) treated ribosomes; (Al) CLR (same as in Fig. 1D) and

(Bl) the same ribosomes after treatment.

1 0.2

ew

o.iI

top boaffo top bottom

FIG. 3. Sedimentation patterns of high-salt-washed CLRI. (A)Ribosomes, after high-salt washing were crosslinked for 8 hr andcentrifuged into a 10-30% sucrose gradient (see legend to Fig. 1)containing 0.1 mM Mg(OAc)2. (B) CLR (same as in Fig. 1C)were subjected to high-salt washing and then analyzed as in (A).The gradients contained 1 mM Mg(OAc)2.

night at 4° against 500 volumes of a buffer containing 20mM NN-bis(2-hydroxyethyl)glycine (Bicine)-10mM MgC12-10 mM KCI-6 mM 2-mercaptoethanol (pH8.5). Solid Me2-Sub was neutralized with an equivalent amount of 1 NKOH, and a solution of ribosomes at 00 in Bicine bufferwas added immediately thereafter. The final Me2Sub con-

centration was 10 mM. The reaction mixture was kept at 00.

At appropriate times, aliquots of the reaction mixture were

removed and 0.1 volume of 1 M Tris- HCl buffer (pH 6.7) was

added in order to stop the reaction. After dialysis at 40 againsta buffer containing 10 mM Tris-HCl-30 mM NH4Cl-10 mMMgCl2-6 mM 2-mercaptoethanol (pH 7.5) (Buffer I), CLRwere recovered by centrifugation. The ribosomal pellets were

suspended in Buffer I and stored as a concentrated solution at00. The activity of ribosome preparations remained unalteredfor at least 1 month under these conditions.CLR that had not previously been subjected to a high-salt

wash were washed twice with Buffer I containing 1 MNH4Cl. CLR were recovered by centrifugation for 90 min at40,000 rpm in a no. 40 rotor in a Beckman L3-50 centrifuge.This centrifugation procedure left essentially all the free 30Ssubunits and most of the free 50S subunits in the supernatantfluid. CLR were further purified by zone centrifugation for4.5-5 hr at 27,000 rpm on 10-30% sucrose gradients in BufferI containing 1 mM Mg(OAc)2 in a SW27 rotor.

RESULTSCrosslinking of 70S ribosomes with Me2Sub

Preliminary experiments with Me2Sub-treated ribosomes fromE. coli indicated that, despite extensive reaction of ribosomalamino groups, the modified particles retained their ability tosynthesize polyphenylalanine with poly(U) as a message (12).These results suggested that it might be possible to covalentlyjoin ribosomal subunits to give crosslinked particles thatwould retain at least some of their functional properties. Fig.1 illustrates that this expectation was realized. After treatmentwith Me2Sub for 8 hr, about 70% of the starting 70S ribosomesappear to be covalently joined. This conclusion is based on theinability of these particles to separate into subunits when theMg2+ concentration is lowered to 1 mM. At this Mg2+ con-

centration, CLR sediment as 68S particles, and can readily bepurified by zone centrifugation (Fig. 1D). As can be seen (Fig.IA and B), both 30S and 50S subunits remain unaffected insedimentation behavior after Me2Sub treatment, and there isno tendency for the subunits, when crosslinked separately, toassociate at 1 mM Mg2+ or below.

(A) 50 6077SI4 4

(B) 50 61 74 SA 4 I

0.4

0.2

3770 Biochemistrv: Slobin

I

N1

Dow

nloa

ded

by g

uest

on

Dec

embe

r 9,

202

0

Proc. Nat. Acad. Sci. USA 69 (1972)

TABLE 1. Binding of N-acetylphenylalanyl-tRNA to CLR atJow M1g2+ concentrations: effect of initiation factors

Stimulation by

Initi- tRNA initiation factorsRibo- ation Poly- bound (+factors)/ %somes factors (U) (pmol) (-factors) Activity

70S + - 0.29 -70S - + 0.57 1270S + + 2.70 8.6 100CLR* - + 0.78 20CLR* + + 1.51 2.5 51

The reaction mixture (total volume 0.1 ml) contained the fol-lowing components: ribosomes (1.4 A260 units), Tris HCl buffer(26 mM, pH 7.4), NH4C1 (0.12 M), Mg(OAc)2 (5 mM), 2-mercaptoethanol (4 mM) GTP (0.3 mM), poly(U) (25 jig), crudeinitiation factors, when added (34,g), and N-acetyl-[14C]Phe-tRNA (9 pmol). After incubation for 10 min at 300, 2 ml of ice-cold binding buffer [20 mM Tris- HCl-100 mM NH4Cl-5 mMMg(OAc)2] was added to each tube and the contents werefiltered immediately through a nitrocellulose membrane. Themembranes were washed with five 3-ml portions of bindingbuffer, dried, and counted in a Packard Tri-Carb series 3000liquid scintillation spectrometer in a scintillation cocktail con-sisting of 4 g of Omnifluor (New England Nuclear Corp.) perliter of toluene.

* Crosslinked particles were prepared by treatment of 70Sribosomes with Me2Sub for 8 hr, followed by washing with 1 MNH4Cl and purification by zone centrifugation (see also Fig. 3B).

To further substantiate the conclusion that the ribosomalsubunits are covalently crosslinked to one another, bothunmodified and CLR were treated with p-hydroxymercuri-benzoate. Treatment with this mercurial agent causes ribo-somes to dissociate into subunits (15). Unmodified 70S ribo-somes are almost completely dissociated into 30S and 50Ssubunits, whereas purified CLR are unaffected by such treat-ment (Fig. 2).Production of 60S ribosomes by treatment of high-salt-washed particles with Me2Sub

Treatment of high-salt washed ribosomes with Me2Sub led toproduction of two particles that were readily distinguishableon sucrose gradients (Fig. 3A), a rapidly sedimenting com-ponent with an apparent S value of 74-77 S and a more slowlysedimenting component characterized by an S value of 59-61.These sedimentation coefficients were calculated accordingto Martin and Ames (16), and were based on assumed constantsedimentation rates for the ribosomal subunits at Mg2+ ionconcentrations between 0.1 and 10 mM. More precise deter-minations of sedimentation rates will be reported elsewhere.*The two forms of crosslinked 30-50S complexes will be re-ferred to here as 60S and 70S CLR.Both 60S CLR and 70S CLR particles bind AcPhe-tRNA

at low Mg2+ concentrations to about the same extent, andboth particles are active in polyphenylalanine synthesis.It has proven possible to prepare these two ribosomal particlesby first crosslinking 70S ribosomes (low-salt washed), fol-lowed by subsequent washing with 1 M NH4Cl (Fig. 3B).In the functional studies reported below, the CLR tested wereall prepared by high-salt washes subsequent to Me2Sub treat-ment. Therefore, they consist of a mixture of both 60S CLRand 70S CLR (Fig. 3B). Some CLR preparations also con-tained a small amount of lOOS particles (Fig. 2A and B), whichprobably represent crosslinked 70S ribosome dimers.

Functional properties of CLR programmed by poly(U)

Poly(U)-directed polyphenylalanine synthesis has many

features in common with protein synthesis dependent on

natural mRNAs. In particular, it was first shown by Lucas-Lenard and Lipmann (14) that the binding of AcPhe-tRNA toa ribosome-poly(U) complex at low Mg2+ concentrations was

dependent upon at least two initiation factors and was stimu-lated by GTP. In these features -this reaction appears tomimic natural initiation, which is dependent on fMet-tRNAf.Table 1 demonstrates that after treatment with Me2Sub for 8hr, CLR retain about 50% of the capacity of unmodifiedribosomes to bind AcPhe-tRNA in the presence of poly(U) at5 mM Mg2+. Binding to CLR was affected by crude initiationfactors, although the degree of stimulation (a maximum of 2.5-fold) was less than that observed with unmodified high-salt-washed ribosomes. It is possible that some or all of the initia-tion factors are covalently joined to some of the ribosomesafter Me2Sub treatment and, therefore, remain attached andfunctional even after high-salt washing. In this regard, pre-

liminary studies show that crosslinking of ribosomes after a

high-salt wash (see Fig. 3A) enhances the stimulatory effect ofinitiation factors on AcPhe-tRNA binding.To demonstrate that AcPhe-tRNA occupies the same func-

tional site (the P or donor site) in both CLR and unmodifiedribosomes, the sensitivity of the binding reaction to puro-mycin was tested. As expected (Table 2), puromycin releasedabout the same percentage of bound counts with both types ofribosomes.Having established that CLR could apparently form an

initiation complex with AcPhe-tRNA, I next examined theability of these ribosomes to act in a polyphenylalanine syn-thesizing system. The results, presented in Table 3, demon-strate that polyphenylalanine synthesis is effectively medi-ated by CLR. At the Mg2+ concentration used (5.5 mM), thesynthetic reaction was absolutely dependent on the presenceof AcPhe-tRNA. The data also show that treatment withMe2Sub for shorter times (2 or 4 hr) results in more activeCLR preparations. Fortunately, crosslinking of the tworibosomal subunits is essentially complete after 4 hr, as judgedby the fraction of CLR present after zone centrifugation atlow Mg2+ concentrations (unpublished observations). Thereason for the 20% reduction in polyphenylalanine synthesiz-ing activity after 2-4 hr of treatment with Me2Sub is notknown. Preliminary investigations indicate that CLR have a

somewhat higher activation energy in this reaction than dounmodified ribosomes.

TABLE 2. Effect of puromycin on the binding of N-acetyl-phenylalanyl-tRNA to CLR at low Mg2+ concentrations

Inhibition

Ribo-Ino. tRNA bound (pmol) byRibo- Initiation puromycinsomes factors +puromycin - puromycin (%)

70S - 0.56 0.57 070S + 0.55 2.42 77CLR - 0.36 1.0 64CLR + 0.59 1.50 61

The reaction mixture (see Table 1) was incubated for 10 min at370. Then, 10 1A of a 10 mM solution of puromycin in 10 mMTris* HC1 (pH 7.5) was added to one set of tubes and 10 ,ul ofTris HCl was added to another. Incubation was continued for5 min at 300, whereupon the tubes were processed as in Table 1.

Cross-Linked Ribosomes 3771

Dow

nloa

ded

by g

uest

on

Dec

embe

r 9,

202

0

Proc. Nat. Acad. Sci. USA 69 (1972)

TABLE 3. Poly(U)-directed polyphenylalanine synthesis byCLR at low Mg2+ concentration

pmol ofHr of polyphenyl-

Ribosomes crosslinking alanine Activity (%)

70S 0 8.9 100COR 1 5.3 60CLR 2 7.1 80CLR 4 6.8 77CLR 8 3.0 3430S + 50S 0 7.4 10030S CLR + 50S CLR 1 6.3 8530SCLR + 50SCLR 2 3.3 45

The reaction mixture (total volume 0.1 ml) contained thefollowing components: Ribosomes (70S and CLR, 1.7 A260 units;30 and 30S CLR, 0.5 A260 units; 50S and 50S CLR, 1.0 A260 units),Tris* HCI (10 mM, pH 7.5), NH4Cl (22 mM) crude initiationfactors (34 jg), supernatant enzymes (150,gg), pyruvate kinase(1 Ml), ATP (1 mM), GTP (0.03 mM), dithiothreitol (1 mM),2-mercaptoethanol (6 mM), phosphoenolpyruvate (5 mM),Mg(OAc)2 (5.5 mM), poly(U) (25 ,gg), N-acetyl-[14C]Phe-tRNA(9 pmol), and [14C]Phe-tRNA (13 pmol). After incubation for 30min at 25', 3 ml of 5% C13CCOOH was added to stop the reaction.The precipitates were heated for 15 min in a boiling-water bath,brought to room temperature, and filtered onto glass-fiber filters(Reeve Angel, 984 H). The filters were dried in a 600 oven for 15min and counted as described in Table 1. 0.18 pmol of phenyl-alanine was incorporated into acid-precipitable material in theabsence of poly(U).

Functional properties of CLR programmed by f2 RNAFrom the results presented thus far, it might be expected thatCLR would be highly active in functions mediated by a na-tural mRNA. This turned out not to be the case. It may beseen (Table 4) that f2 RNA-directed binding of fMet-tRNAwas reduced to 10-15% of the control values and that allCLR preparations tested have lost the capacity to incorporate

TABLE 4. Binding offMet-tRNA to CLR directedby phage f2 RNA

Hr of fMet-tRNAcross- Initiation bound %

Ribosornes linking factors (pmol) Activity

70S 0 - 0.059 470S 0 + 0.81 100CLR 1 - 0.063 4CLR 1 + 0.15 15CLR 2 - 0.052 3CLR 2 + 0.14 14CLR 4 _ 0.055 3CLR 4 + 0.143 14CLR 8 - 0.043 1CLR 8 + 0.110 10

The reaction mixture (total volume of 0.1 ml) contained thefollowing components: Ribosomes (1.7 A260 units), Tris- HC1(100 mM, pH 7.5), NH4C1 (100 mM), initiation factors whenadded (68 Ag), GTP (0.3 mM), Mg(OAc)2 (10 mM), f2 RNA (1.1A260 units), and f-[3HI Met-tRNA (18 pmol). After incubation for10 min at 37°, 3 ml of ice-cold binding buffer [100 mM Tris- HC1(pH 7.5)-100 mM NH4CI-10 mM Mg(OAc)2] was added toeach tube, and the contents were filtered immediately through anitrocellulose membrane. The membranes were washed with five3-ml portions of binding buffer, dried, and counted. 0.03 pmol off-[3H]Met-tRNA was bound in the absence of f2 RNA.

TABLE 5. Incorporation of ['4C]valine directed by phage f2RNA: activity of CLR

ValineHr of incorporated %

Ribosomes crosslinking (pmol) Activity

70S 0 67 100CLR 1 0.30 0CLR 2 0.31 0CLR 4 0.28 0CLR 8 0.30 030S + 50S 0 56 10030SCLR + 50SCLR 1 50 8930SCLR + 50SCLR 2 26 5030S + 50SCLR 8 23 4130SCLR + 50SCLR 8 14 25

The reaction mixture (total volume 0.1 ml) contained thefollowing components: ribosomes (30S and 30S CLR, 0.74 A260units; 50S and 50 S CLR, 1.6 A260 units; 70S and 70S CLR, 1.7A260 units), Tris HCO (10 mM, pH 7.5), NH4Cl (22 mM), dithio-threitol (1 mM), 2-mercaptoethanol (6 mM), pyruvate kinase(1 Ml), ATP (1 mM), GTP (0.03 mM), phosphoenolpyruvate (5mM), Mg(OAc)2 (10 mM), initiation factors (34 Mlg), calciumleucovorin (1.5 Mg), supernatant enzymes (150 Mg), and [14C]-valine (500 pmol). After 10 min at 370, 1.1 A260 units of f2 RNAwas added and the incubation was continued for an additional 30min at the same temperature. C13CCOOH (3 ml of a 5% solution)was added to stop the reaction. The precipitates were filtered andcounted. 0.28 pmol of [14C]valine was incorporated into acid-precipitable material in the absence of f2 RNA.

valine into f2 coat protein (Table 5). This lack of incorporationability does not appear to be due to inactivation of the indi-vidual ribosomal subunits; although there is a progressiveinactivation of both subunits when crosslinked separately,they still retain about 90% of their starting activity after a1-hr reaction with Me2Sub when mixed together in a cell-freeextract directed by f2 mRNA, and are still active to someextent even after 8 hr of reaction (Table 5). A similar decline inactivity was also observed for polyphenylalanine synthesis(Table 3), suggesting that the inactivation was not specific forprotein synthesis directed by a natural mRNA.

DISCUSSION

To answer the question whether the 30S subunit initiationcomplex is an obligatory intermediate in protein synthesis, itseemed necessary to avoid the possibility of ribosome associa-tion-dissociation that has proved an obstacle to a satisfyingresolution of this problem. To overcome this difficulty, ribo-somes were crosslinked with the bifunctional imidoester,dimethylsuberimidate. The crosslinked particles appear tocontain covalently joined 30S and 50S subunits as judged byseveral criteria: (i) CLR do not dissociate into ribosomalsubunits, even at 1 mM Mg2+ concentration or below (Figs.1 and 3). (ii) p-hydroxymercuribenzoate treatment, whichalmost completely dissociates unmodified 70S ribosomes, hasno effect on CLR (Fig. 2). (iii) Ribosomal subunits cross-linked separately do not associate at a Mg2+ concentration of1 mM, nor is there any evidence of subunit aggregation as aresult of Me2Sub treatment (Fig. 1).Ribosomes treated with Me2Sub for 8 hr and subsequently

purified by zone centrifugation contain about 55 crosslinkedlysine residues per mol (about 17% of the total lysine residuesare crosslinked), about 10 more than are contained in the sum

3772 Biochemistry: Slo.in

Dow

nloa

ded

by g

uest

on

Dec

embe

r 9,

202

0

Proc. Nat. Acad. Sci. USA 69 (1972)

of the individual subunits crosslinked separately (unpublishedobservations). Whether all of these. 10 extra crosslinks jointhe two subunits remains to be established.A surprising result of this investigation is the observation

that CLR can exist in at least two forms, a 60S and 70S CLR(Fig. 3), and that these two forms are functional in poly(U)-mediated test systems. A 60S form of E. coli ribosomes hasalso been reported by several laboratories (17, 18). It is per-haps significant that Schreier and Noll (17, 19) obtained 60Sribosomes only after high-salt washing of ribosomes; similartreatment is necessary to observe 60S CLR (compare Figs. 1and 3). Schreier and Noll (19), as well as others (20), havedescribed a number of pre- and post-translocational steps thatare characterized by the interconversion of these two ribo-somal forms. However, recent work on pressure-induced dis-sociation of eukaryotic (21), as well as E. coli, ribosomes (2,22) casts some doubt as to the actual existence of a 60S par-ticle. The relationship, if any, between 60S CLR and the 60Sribosomes reported by others is under active investigation.

All of the properties of a ribosome necessary for proteinsynthesis would seem to be largely unaffected by crosslinking.Thus, CLR can bind AcPhe-tRNA at low Mg2+ concentra-tions in a reaction stimulated by initiation factors (Table 1),the bound AcPhe is released by puromycin (Table 2) as ex-pected if it were occupying the P site, and apparently normalelongation processes mediated by elongation factors EF-Tand EF-G, and by peptidyl transferase, can occur (Table 3).

Nevertheless, fMet-tRNAf binding directed by f2 RNA wasgreatly reduced with CLR (Table 4), and no incorporation ofvaline into protein was observed (Table 5). It is very unlikelythat this inactivation was due to damage to the individualribosomal subunits, since subunits crosslinked individuallyand mixed together retained up to 90% of their starting activ-ity under conditions in which CLR are completely inactive(Table 5). In addition, 70S particles appear to be more re-sistant to inactivation by Me2Sub in polyphenylalanine syn-thesis as compared with the separated subunits (Table 3 andunpublished observations), suggesting that-as might beexpected-the functional properties of ribosomes are protectedby the association of subunits. These findings would appear toestablish conclusively that 30S ribosomes are an obligatoryintermediate in protein synthesis.The apparent differences between fMet-tRNAf and AcPhe-

tRNA as initiators deserves some comment. It has been ob-served that a stable initiation complex analogous to 30S sub-unit-mRNA-fMet-tRNAf cannot be isolated with AcPhe-tRNA as initiator (23, 24). At low Mg2+ concentrations,AcPhe-tRNA binding to ribosomes, although dependent oninitiation factors and GTP, is much less stable than that offMet-tRNAf. On the basis of this observation, Rudland et al.(25) suggested that there is enough similarity in tertiary struc-ture between fMet-tRNAf and AcPhe-tRNA to permit a weakcomplex between the phenylalanyl analogue and ribosomesthat is stable enough to be retained on nitrocellulose filters.

It appears likely that fMet-tRNAf (complexed with f2 andGTP) binds to a ribosome at a locus that formally resemblesthe P site (26), but which may be somewhat different than theP site that is active during peptide chain elongation (27-29).The diminished binding capacity of CLR for fMet-tRNAfcompared with AcPhe-tRNA suggest that these two RNAsoccupy different, or at least distinct, regions on a ribosome.

Perhaps AcPhe-tRNA binds initially to the true translationalP site, which remains intact in CLR.

Finally, the ability to rapidly fix ribosomes so they cannotdissociate into subunits, while they maintain many of theirfunctional properties, should be of some value in studies on theribosome cycle. The utility of bifunctional imidoesters in thestudy of ribosome topography has been discussed (12).

*NOTE ADDED IN PROOF

The S value of the more rapidly sedimenting CLR has beenfound to be 68 i 2 S rather than 74-77 S as mentiond in thetext. The S value of slower-sedimenting CLR is 61 ± 2 S. Itis possible to completely convert 68 S CLR to the 60 S form byfour successive washes in buffer I containing 1 M NH4Cl.

I thank Dr. E. B. Keller for advice on the preparation ofcharged tRNAs and for helpful discussions. This work was sup-ported by a grant from the National Science Foundation.

1. Guthrie, C. & Nomura, M. (1968) Nature 219, 232-235.2. Subramanian, A. R. & Davis, B. D. (1971) Proc. Nat. Acad.

Sci. USA 68, 2453-2457.3. Kohler, R. E., Ron, E. Z. & Davis, B. D. (1968) J. Mol.

Biol. 36, 71-82.4. Algranati, I. D., Bonzalez, N. S. & Bade, E. G. (1969) Proc.

Nat. Acad. Sci. USA 62, 574-580.5. Nomura, M., Lowry, C. V. & Guthrie, C. (1967) Proc. Nat.

Acad. Sci. USA 58, 1487-1493.6. Traub, P., Mizushima, S., Lowry, C. V. & Nomura, M.

(1971) Methods Enzymol. 20, 391-407.7. Ravel, J. M. & Shorey, R. -(1971) Methods Enzymol. 20,

309-310.8. Haenni, A. L. & Chapeville, F. (1966) Biochim. Biophys.

Acta 114, 135-148.9. Dubnoff, J. S. & Maitra, U. (1971) Methods Enzymol. 20,

248-251.10. Nomura, M. & Lowry, C. V. (1967) Proc. Nat. Acad. Sci.

USA 58, 946-953.11. Roberts, J. W. & Steitz, J. E. A. (1967) Proc. Nat. Acad.

Sci. USA 58, 1416-1421.12. Slobin, L. I. (1972) J. Mol. Biol. 64, 297-303.13. Nirenberg, M. & Leder, P. (1964) Science 145, 1399-1407.14. Lucas-Lenard, J. & Lipmann, F. (1967) Proc. Nat. Acad.

Sci. USA 57, 1050-1057.15. Tamaoki, T. & Miyazawa, F. (1967) J. Mol. Biol. 23, 35-46.16. Martin, R. G. & Ames, B. N. (1961) J. Biol. Chem. 236,

1372-1379.17. Schreier, M. H. & Noll, H. (1970) Nature 227, 128-133.18. vanDiggelen, 0. P., Heinsius, H. L., Kalousek, F. Bosch,

L. (1971) J. Mol. Biol. 55, 277-281.19. Schreier, M. H. & Noll, H. (1971) Proc. Nat. Acad. Sci.

USA 68, 805-809.20. Gupta, S. L., Waterson, J. Sopori, M. L., Weissman, S. M.

& Lengyel, P. (1971) Biochemistry 10, 4410-4421.21. Infante, A. A. & Baierlein, R. (1971) Proc. Nat. Acad. Sci.

USA 68, 1780-1785.22. vanDiggelen, 0. P., Oostrom, H. & Bosch, L. (1971) FEBS

Lett. 19, 115-119.23. Lucas-Lenard, J. & Haenni, A. L. (1968) Proc. Nat. Acad.

Sci. USA 59, 554-560.24. Grunberg-Manago, M., Clark, B. F. C., Revel, M., Rud-

land, P. S. & Dondon, J. (1969) J. Mol. Biol. 40, 33-44.25. Rudland, P. S., Whybrow, W. A., Marcker, K. A. & Clark,

B. F. C. (1969) Nature 222, 750-753.26. Benne, R. & Voorma, H. 0. (1972) FEBS Lett. 20, 347-351.27. Zagorska, L., Dondon, J., Lelong, J. C., Gros, F. & Grun-

berg-Manago, Al. (1971) Biochemie 53, 63-70.28. vanDuin, J., vanKnippenberg, P. H., Dieben, M. & Kurland,

C. G. (1972) Mol. Gen. Genet. 116, 181-191.29. Thach, S. 8. & Thach, R. E. (1971) Proc. Nat. Acad. Sci.

USA 68, 1791-1795.

Cross-Linked Ribosomes 3773

Dow

nloa

ded

by g

uest

on

Dec

embe

r 9,

202

0