Assembly of proteins and 5 S rRNA to transcripts of the major structural domains of 23 S rRNA

Article No. mb982185 J. Mol. Biol. (1998) 284, 227±240

Assembly of Proteins and 5 S rRNA to Transcripts ofthe Major Structural Domains of 23 S rRNA

Peter éstergaard1, Hien Phan1, Lise B. Johansen1,2, Jan Egebjerg2

Laust éstergaard1, Bo T. Porse1 and Roger A. Garrett1*

1RNA Regulation CentreInstitute of Molecular BiologyCopenhagen UniversitySùlvgade 83H, DK-1307Copenhagen K, Denmark2RNA Regulation CentreInstitute of Molecular andStructural Biology, AarhusUniversity, DK-8000Aarhus C, Denmark

E-mail address of the [email protected]

Abbreviations used: DMS, dimet1 cyclohexyl-3-(2-morpholinoethyl)-p-toluene sulphonate; TP50, total 501D, one-dimensional; 2D, two-dime

0022±2836/98/470227±14 $30.00/0

The six major structural domains of 23 S rRNA from Escherichia coli, andall combinations thereof, were synthesized as separate T7 transcripts andreconstituted with total 50 S subunit proteins. Analysis by one and two-dimensional gel electrophoresis demonstrated the presence of at least oneprimary binding protein associated with each RNA domain andadditional proteins assembled to domains I, II, V and VI. For all the com-binations of two to ®ve domains, enhanced assembly yields and/or newproteins were observed primarily to those transcripts containing eitherdomains I�II or domains V�VI. This indicates that there are two majorprotein assembly centres located at the ends of the 23 S rRNA, which isconsistent with an earlier view that in vitro protein assembly nucleatesaround proteins L24 and L3. Although similar protein assembly patternswere observed over a range of temperature and magnesium concen-trations, protein L2 assembled strongly with domains II and IV at4-8 mM Mg2� (the ®rst step of the two-step reconstitution procedure)and with domain IV alone at higher Mg2� concentrations (the secondstep). It is proposed that this change in protein-RNA binding provides abasis for the two-step reconstitution in vitro. A chemical footprintingapproach was employed on the reconstituted protein-domain complexesto localize a putative L4 binding region within domain I to a region thatis partially co-structural with the site on the L4-mRNA where L4 bindsand inhibits its own translation. A similar approach was used to map theputative binding regions on domain V of protein L9 and the 5 S RNA-L5-L18 complex.

# 1998 Academic Press

Keywords: 50 S subunit assembly; 23 S RNA domains; 5 S RNA-assemblysite; L4-23 S RNA site; L9-23 S RNA site
*Corresponding author
Introduction

The assembly process of the large ribosomal sub-unit is still only partially understood. The 23 S-likerRNA generates six major structural domains(Egebjerg et al., 1990a) that fold, undergo chemicalmodi®cation, and assemble proteins during theirtranscription in vivo. Although considerable efforthas gone into (a) identifying the proteins that bindstrongly to the RNA, (b) localizing their RNA bind-ing sites (reviewed by Draper, 1996), and (c) estab-

ing author:

hylsulphate; CMCT,carbodiimide metho-S subunit proteins;

nsional.

lishing cooperative protein-protein interactionsduring assembly (Dohme & Nierhaus, 1976;Herold & Nierhaus, 1987; Franceschi & Nierhaus,1990), only limited progress has been made indetermining the cooperative interplay of proteinswith 23 S RNA during 50 S subunit assemblyin vitro (Chen-Schmeisser & Garrett, 1976; Spiereret al., 1979) and in vivo (Liiv et al., 1996). One infer-ence from the early work was that protein assem-bly nucleates around the 23 S RNA bindingproteins L24 and L3 (Nowotny & Nierhaus, 1982).

This limited progress partly re¯ects the require-ment for a two-step reconstitution procedurein vitro, at least for the Escherichia coli ribosome,which involves increasing both the magnesiumconcentration and the temperature during a secondincubation step (Nierhaus & Dohme, 1974; Amilset al., 1978). In order to gain greater insight into the

# 1998 Academic Press

mailto:[email protected]

228 Protein Assembly to 23 S rRNA Domains

protein-RNA assembly process, the individualstructural domains of E. coli 23 S RNA, and combi-nations thereof, were transcribed and reconstitutedwith total 50 S subunit proteins over a range of sol-ution conditions, and the assembled proteins wereidenti®ed. In preliminary experiments, domain VIwas shown to assemble proteins L3 and L6, andprotein L2 bound to domain IV (Leffers et al.,1988). Here, we extend the approach to each of thesix domains and to all of their combinations.

Although many of the 23 S RNA binding sites ofthe primary binding proteins have been localizedand probed in some detail (reviewed by Draper,1996), those of protein L4 and the 5 S RNA-L18-L5complex have not yet been and the location of theL9 site has been mapped only approximately(Adamski et al., 1996). Moreover, L4 is a transla-tional repressor that has a binding site within theleader sequence of the mRNA for the S10 operon(Zengel & Lindahl, 1996) which is likely to bepartly co-structural with its rRNA binding site(Zengel & Lindahl, 1993). A footprinting approachwas employed to probe their binding sites ondomain transcripts.

Results

Characterization of constructs

The six main structural domains of 23 S RNA,and combinations thereof, were transcribed fromthe T7 promoter-containing vector, essentially asdescribed by Leffers et al. (1988). The main restric-tion sites that were used within the 23 S rDNA, or

Table 1. Details of the domain constructs

Domain Plasmid Co

I pJE101 DraI± 113(Bal31)±SstI363 � SstI36

II pJE105 SstI363(Bal31)±EcoRI845 � EcoRIIII pJE102 SphI1230(Bal31)±AvaI1658

IV pDIV HaeII1644±SacII2046

V pDVH AvaI2001±HaeII2626

VI pJE130 HaeII2626±HaeII65 in 5 S(Bal31)I-II pJE106 DraI± 114(Bal31)±SphI1230 � SphI-II-III pJE103 DraI± 114(Bal31)±SstI363 � SstI36

I-II-III-IV pJE123 DraI± 114(Bal31)±SalI1345 � SalI1

I-II-III-IV-V pJE143 DraI± 114(Bal31)±SphI1230 � SphI-II-III-IV-V-VI pJE145 DraI± 114(Bal31)±HaeII65 in 5 S(BII-III pJE104 SstI363(Bal31)±SphI1230 � SphI12

II-III-IV pJE124 SstI363(Bal31)±SalI1345 � SalI1345

II-III-IV-V pJE144 SstI363(Bal31)±SphI1230 � SphI12

II-III-IV-V-VI pPE104 SstI363(Bal31)±SacII2046 � SacII2

III-IV pJE122 SphI1230(Bal31)±SalI1345 � SalI13

III-IV-V pJE142 SphI1230(Bal31)±SalI1345 � SalI13

III-IV-V-VI pPE103 SphI1230(Bal31)±SacII2046 � SacIIV-V pJE125 HaeII1644±HaeII2626

IV-V-VI pPE102 HaeII1644±SacII2046 � SacII2046±V-VI pPE101 SacII2046±HaeII69 in 5 S(Bal31)VI-I pPE105 EcoRI2949 in pJE130±PstI2914 in pJE

VI-I-II pPE107 SstI363±HindIII50 to T7 promoter in

The constructs are listed that were generated for producing singlinformation on the restriction sites and fragments of 23 S rDNA thscript are indicated in column 4 (see also Figure 1A). Numbers indifrom the sequence numbering system of E. coli 23 S RNA (Brosius50-U2904ACAACGCCGAUCCCCGAG1-3

0 between domains VI and I.

vector, are listed in Table 1 and illustrated inFigure 1A. The terminal sequence of each singledomain transcript is shown in Figure 1B and,within one to four nucleotides these are also validfor the multi-domain transcripts. A ``clamp''sequence was introduced bordering the domain IIconstruct to constrain the ends of the transcript.The sequences of the DNA constructs were allchecked (see Materials and Methods). Each domaintranscript was subjected to gel ®ltration, whichremoved low levels of prematurely terminated ordegraded transcripts although very small amountsof degraded template DNA remained. The singledomains were electrophoresed in denaturing poly-acrylamide gels and yielded single bandsmigrating with the expected relative mobilities(Figure 2).

Reconstitution of domain transcripts

The domain transcripts were incubated withtotal 50 S subunit proteins (TP50). In general,lithium chloride-extracted proteins were employedbecause they are more readily renatured thanacetic acid-extracted proteins (Dijk & Littlechild,1979). They suffered the disadvantage, however,that limited protein degradation occurs duringextraction and some protein fragments assemblewith 23 S RNA, thereby rendering interpretation ofthe 2D gels more dif®cult. Moreover, extraction ofprotein L4 was always partial under these con-ditions. Initially, 5 S RNA was added to all recon-stitution mixtures containing single domain

nstruct23 S RNAsequence

3±EcoRI845(Bal31) 1±533

845±SalI1345(Bal31) 577±12621273±16611647±20452000±26242628±2904

I1230±SalI1345(Bal31) 1±1262

3±AvaI1658 1±1661

345±SacII2046 1±2045I1230±HaeII2626 1±2624al31) 1±2904

30±AvaI1658 577±1661±SacII2046 577±2045

30±HaeII2626 577±2624

046±HaeII65 in 5 S(Bal31) 577±2904

45±SacII2046 1273±2045

45±HaeII2626 1273±2624I2046±HaeII69 in 5 S(Bal31) 1273±2904

1647±2624HaeII69 in 5 S(Bal31) 1647±2904

2048±2904

130 � BamHI-7 in pJE101±EcoRI548 in pJE101 2628±533

pJE106 � HindIII50 to T7 promoter in pPE105±SstI363 2628±1262

e and multi-domain transcripts. Column 3 contains the essentialat were used in preparing the constructs. The ends of each tran-cated as subscripts in column 3 and numbers in column 4 deriveet al., 1981). Constructs pPE105 and pPE107 contained the linker

Figure 1. A, A map of the rrnB operon of E. coli on plasmid pJE145 showing the restriction sites that were used inthe formation of the various single and multi-domain constructs (see also Table 1). B, Terminal sequences of each ofthe single domain transcripts. The terminal nucleotide of each sequence corresponding to 23 S RNA is shown inbold-face and numbered. The extensions derive from cloning steps and the pUT719 vector (Leffers et al., 1988).Nucleotides designed to generate a ``clamp'' helix on domain II are underlined.

Protein Assembly to 23 S rRNA Domains 229

transcripts. However, since it affected only proteinassembly to domain V (data not shown), it wassubsequently added only to the transcripts thatcontained domain V. The RNA-protein complexeswere prepared and separated from unbound pro-teins by gel ®ltration (Figure 3A). Samples from allthe column fractions were examined in one-dimen-sional (1D) SDS/polyacrylamide gels, as is illus-trated for the L23-domain III complex (Figure 3B),

Figure 2. Electrophoresis of transcripts of the singledomains I to VI in a 6% polyacrylamide gel containing88 mM Tris-borate (pH 8.3), 25 mM Na2EDTA and 7 Murea after fractionation on gel-®ltration columns (seeMaterials and Methods). RNA bands were stained withtoluidine blue and the gel origin is indicated by a line.

to ensure that the complexes were not contami-nated with free protein. The largest free proteincomponent, and most likely contaminant of thesingle domain complexes, was the L10.(L12)4 com-plex (Figure 3B), which assembled only withdomain II (see below). Complexes fractionated byagarose gel electrophoresis or sucrose density-gra-dient centrifugation exhibited essentially the sameprotein composition (data not shown).

Reconstitution experiments were performed reg-ularly under conditions based on the two-step pro-cedure that yields active 50 S subunits (Nierhaus &Dohme, 1974). The temperature conditions of thetwo steps were maintained unchanged at 44�C and50�C, respectively (Nierhaus & Dohme, 1974), butthe magnesium conditions used were either4 mM/20 mM or 20 mM/20 mM Mg2�. Proteincontents were identi®ed unambiguously for all thelarger proteins by two-dimensional (2D) gel elec-trophoresis. The small proteins L28 and L31-L35were not reproducibly resolved under the exper-imental conditions used, and no attempt was madeto assign them to domains. The results for thesingle domain transcripts reconstituted under the20 mM/20 mM Mg2� conditions are illustrated inFigure 4A and listed in Table 2. They demonstratethat at least 21 of the 25 proteins examinedassembled directly to the single domains, and theresults were essentially the same under the twosets of magnesium conditions tested (4 mM/20 mM and 20 mM/20 mM Mg2�). These includeall the known primary binding proteins and sev-

Figure 3. A, Fractionation of the L23-domain III com-plex from total 50 S subunit proteins on a Superose 12column. B, The relative locations of L23 and the com-ponents of the pentameric L10.(L12)4 complex in the col-umn fractions analysed are indicated on the 1D SDS/polyacrylamide gel. Gels were calibrated with TP50(Dijk & Littlechild, 1979). Experimental details are givenin Materials and Methods.


eral others. When the double domain constructswere examined, additional protein assembly,including enhanced binding of L6, L15 and L29and new binding of L14, L17 and L19, wasobserved to the domain combinations I�II, V�VIand VI�I (Table 3). The results are illustrated forthe former two transcripts in Figure 4B. Again theresults were essentially the same for the 4 mM/20 mM Mg2� and 20 mM/20 mM Mg2� conditions.The results were also reproducibly observed in allthe larger, three, four and ®ve domain constructsthat contained the domain combinations I�II andV�VI. Surprisingly, no additional assembly effectwas reproducibly observed to any of the othermulti-domain constructs.

The effect of ionic conditions and temperatureon protein assembly

In order to gain some insight into the structuralsigni®cance of the low magnesium conditions used

in the ®rst assembly step, the patterns of proteinassembly were examined by simultaneously vary-ing the magnesium concentration of both steps, by2 mM increments, over the range 4 mM/4 mM to20 mM/20 mM. Proteins and RNA were incubatedin reconstitution buffer at the given Mg2� concen-tration, at the two temperatures, and fractionatedby gel-®ltration at the same Mg2� concentration. Ingeneral, the 2D gel results were similar to thosepresented above (Figure 4A) for the single andmulti-domain transcripts throughout this Mg2�

range, with one striking exception. L2 boundstrongly to both domains II and IV in the lowerMg2� range (4±8 mM; Figure 4C), whereas itbound exclusively to domain IV in the higherrange (10±20 mM Mg2�; Figure 4A). Moreover,weak assembly of L16, which is dependent on thepresence of L2 for its assembly (Romero et al.,1990), was also observed on domain II at 4 mMMg2�. These effects are summarized in Table 4.

Reconstitutions were performed at 42, 44 and50�C, and at 250 and 400 mM NH4Cl when noreproducible difference in protein identity andonly minor differences in the protein yields wereobserved.

Variations in assembly yields were observed forproteins L4 and L20. For L4, the extraction yieldsobtained with the LiCl-urea method varied and,moreover, aggregation occurred frequently in thesecond dimension of the 2D gels, such that lowyields were observed (e.g. domain I, Figure 4A).For L20, and L21, which is dependent on L20 forits assembly (Spierer & Zimmermann, 1976), weakspots were sometimes seen on domain II con-structs. For example, both spots are strong ondomain II in Figure 4A but very weak spots occurin the experiment performed at low Mg2� (4 mM;Figure 4C). However, this result was not generallyreproducible and samples prepared at higher mag-nesium concentrations also sometimes lacked theseproteins. This variability probably re¯ects differentef®ciencies of protein extraction rather than alteredassembly, although the highly basic L20 proteindoes tend to bind anomalously to 23 S RNA(Garrett et al., 1974).

Probing additional protein binding sites andthe 5 S RNA-L18-L5 complex on the domains

Proteins other than the primary 23 S RNA bind-ing proteins assembled to transcripts of domainsI�II and V�VI (Table 3). Therefore, complexes con-taining these domains (except for domain VI,which was described earlier; Leffers et al., 1988)were subjected to footprinting analyses. Complexeswere prepared as described in Materials andMethods, and divided into two parts. One partwas examined by 2D gel electrophoresis and theother was subjected to chemical modi®cation usingnucleotide-speci®c probes. The modi®cation pat-terns of the transcripts and complexes were com-pared by primer extension analyses (Leffers et al.,1988). These data were also compared with the cor-

Figure 4. Identi®cation of assembled proteins by electrophoresis in 2D polyacrylamide gels together with a smallamount of total 50 S subunit proteins. A, Proteins assembled to single domain constructs at 20 mM/20 mM Mg2�.B, Proteins assembled to the double domain constructs I�II and V�VI at 20 mM/20 mM Mg2�. C, Proteinsassembled to domain II at 4 mM Mg2�. Lf indicates ribosomal protein fragments that bind to rRNA. The relativestrong background of L10 indicates a small overlap with the L10.(L12)4 complex (see Figure 3). Gels contained 4%(w/v) acrylamide, bis-Tris acetate (pH 5), and 6 M urea in the ®rst dimension and 18% (w/v) acrylamide, KOH/acet-ate (pH 4.5), in the second dimension (modi®ed from Geyl et al., 1981): 0.01% Triton X-100 was added to the samplebuffer and pyronin G was used as tracking dye. Proteins were stained with Coomassie brilliant blue R250.


responding results for modi®ed 23 S RNA and50 S subunits (Egebjerg et al., 1990a and unpub-lished results). For domains I�II (see Figure 4B for

Table 2. Proteins that assemble to the individualdomains

Domain Proteins assembled at 20 mM Mg2�

DI L4, L24, L22, L29DII L10.[L12]4, L11, L13, (L15), L20, L21DIII L23DIV L2DV � 5 S RNA L1, L9, L27�L5, L18, (L25)DVI L3, (L6)

Proteins that assembled to each of the single domains underthe conditions of the two-step reconstitution procedure wherethe magnesium concentration was maintained at either 4 mM/20 mM or 20 mM/20 mM in the consecutive steps. The pro-teins were identi®ed by 2D polyacrylamide gel analysis and theresults are averaged from about ®ve independent experimentsfor each domain under each set of conditions. Proteins indi-cated in bold face have been designated as binding strongly to23 S RNA or to 5 S RNA (L18, L25). Parentheses denote weaklyassembled proteins.

protein content), only one protected region wasdetected outside of the previously characterizedL24 site (domain I) and the L11 and L10.(L12)4

sites (domain II; see below). It occurred withindomain I, where L24 produces a very complexfootprint centred on two regions (site A, 320±350and site B, 470±510; Egebjerg et al., 1987). The L24footprint showing altered reactivities to dimethyl-sulphate (DMS), kethoxal and 1 cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-p-toluenesulphonate (CMCT) was reproducible here, withthe minor revision that some kethoxal protectioneffects were stronger than observed earlier andnew protections were observed at G258, G498 andG506. The additional footprint that we observed(Figure 5) occurs on a structure lying between thetwo L24 regions, where the altered reactivities areindicated (Figure 5B). This region was marginallymore reactive in the domain transcripts than in23 S RNA (Egebjerg et al., 1987), suggesting thatthe domain structure is more ¯exible; in particularA265, A279, A371, A401 and A402 were more reac-

Table 3. Additional proteins that assembled to multi-domain constructs

Domains (2) Further assembly

DI-DII L15" , L29"

DII-DIII ±DIII-DIV ±DIV-DV ±DV-DVI L6", L14

L17, L19DVI-DI (L17)a

Proteins that assembled to all the multi-domain transcriptsduring two-step reconstitution at 4 mM/20 mM or 20 mM/20 mM Mg2� were identi®ed by 2D gel electrophoresis in atleast three independent experiments for each construct. Theprotein yields were estimated approximately from spot intensi-ties. Assembly additional to that observed for the singledomains (Table 2) was detected only for multi-domain tran-scripts that contained the double domains I�II, V�VI andVI � I and no additional assembly was detected for the three,four and ®ve-domain constructs. Proteins that were present inhigher yields than for the single domains are indicated by ".

a Weak binding protein.

Table 4. Additional assembly effects to single domainsobserved at low Mg2� concentrations

Protein Binds 10-20 mM Mg2�Additional binding at

4-8 mM Mg2�

L2 DIV DIIL15 DII DIIL16 ± DIIL17 ± DII/DVIL25 5 S RNA-DV 5 S RNA-DV

Proteins were reconstituted with single domain transcriptsunder the conditions of the two-step reconstitution procedure,where the indicated Mg2� concentration was maintained inboth reconstitution steps and only the temperature was altered.


tive. The footprint is probably produced, at leastpartly, by L4 because it is the only primary bindingprotein, apart from L24, that binds to domain I(Table 2), where it stimulates the assembly of L22and L29 (Spierer & Zimmermann, 1976; RoÈhl &Nierhaus, 1982).

L4 is also a translational repressor protein thatbinds to the leader sequence of the mRNAexpressed from the S10 operon at a site that hasbeen localized, indirectly, to a hairpin structure(Zengel & Lindahl, 1996). This interaction can becompeted out by a 23 S RNA fragment extendingfrom positions 1±368 of domain I, but not by asmaller fragment (positions 1±317; Zengel &Lindahl, 1993). Thus, the former fragment must beat least partially co-structural with the mRNAbinding site of L4. A comparison of the structuresof the putative L4 sites on 23 S RNA and mRNAreveals common primary and secondary structureat a helix-bulge junction (Figure 5C) that may con-stitute a common attachment site for L4.

For domain V, strongly altered reactivities,mostly protection effects, were observed on proteinassembly in two main regions. One region (pos-itions 2090±2220) includes the L1 binding site. Thissite is relatively unreactive in domain V and in thedomain V-protein complex, as it is in 23 S RNAand the 50 S subunit (Egebjerg et al., 1990a), whichis consistent with it generating a very compactRNA structure (Sloof et al., 1976). DMS protectioneffects were observed at A2173, which is inducedby L1 (Egebjerg et al., 1991), and at A2147, whichparticipates in a tertiary interaction in the L1 site(Figure 6A,B). Additional protection effects thatwere independent of the presence of 5 S RNAoccur in an internal loop structure located immedi-ately below the L1 site (Figure 6B) at A2198, A2199and A2225 Figure 6A) and they are likely to resultfrom L9 binding (see Discussion). Each of thesesites was accessible in 23 S RNA, although the for-mer two were not reported earlier, and they are

inert in 50 S subunits (Egebjerg et al., 1990a andunpublished results). The only differencesobserved between domain V and 23 S RNA in thisregion lie in the terminal loop 2210±2214, whichwas more reactive in the domain; C2214 was theonly site in this region that was weakly protectedin the presence of 5 S RNA (Figure 6A).

The second affected region extends from pos-itions 2250±2410, and exhibits many altered reac-tivities with DMS and kethoxal, including severalstrongly protected guanine residues (Figure 7A,B).Most of the effects were produced only in the pre-sence of 5 S RNA and are considered to derive pri-marily from the interaction of the 5 S-L18-L5complex with the rRNA region (Newberry &Garrett, 1980), although some of the effects mayresult from L27 binding, which does not assemblein the absence of the 5 S rRNA complex (unpub-lished data). The nucleotide reactivities observedfor free domain V were closely similar to those of23 S RNA (Egebjerg et al., 1990a and unpublishedresults). Only A2328, which is involved in a puta-tive tertiary interaction (Figure 7C), was exclu-sively modi®ed in 23 S rRNA. Each of the ®veprotection effects produced by proteins, in the pre-sence and absence of 5 S RNA, occurred at nucleo-tides G2330, G2357, G2360, G2409 and G2410,which are involved in, or adjacent to, tertiary inter-actions in the RNA (Figure 7C; see Discussion) andare inert in 50 S subunits (Egebjerg et al., 1990a).

With the exception of A2327 and A2352, all ofthe sites that are reactive in the domain V-5 SRNA-protein complex, including G2251, G2286,C2306, A2311 and G2398, which exhibit enhancedreactivity (Figure 7C), are accessible in 50 S sub-units (Egebjerg et al., 1990a).

Discussion

Proteins assembling to the individual domainsinclude the strongly binding proteins L1, L2, L3,L10.(L12)4, L11, L23 and L24 (Garrett et al., 1974),whose RNA sites have been localized using ribonu-clease digestion and chemical modi®cationapproaches (Figure 8). These proteins have beenshown in a kinetic study, performed at 0�C and inthe presence of an assembly inhibitor dye, to bindearly to 23 S RNA in vitro (Datta et al., 1986) and it

Figure 5. Protein assembly site in domain I. A, Footprint showing altered nucleotide reactivities with DMS,kethoxal and CMCT within the central region of domain I, with (�) and without (ÿ) assembled proteins. Con rep-resents control samples not treated with modifying chemicals. A, C and G represent sequencing tracks. B, The reactiv-ity changes that are additional to those produced by protein L24 alone (Egebjerg et al., 1987) are indicated on thesecondary structure of domain I by purple nucleotides (chemically reactive) and red circles (protection effects). TherRNA fragment that inhibits L4 binding to the leader region of its mRNA extends from positions 1 to 368 (markedwith arrow). The site of a putative tertiary interaction with domain V (Leffers et al., 1987) is indicated in green.C, Secondary structural similarities observed between the leader sequence of the L4-mRNA and the protected regionof domain I that constitutes the putative binding site of protein L4 (see B); red circles in the latter denote protein pro-tection sites.


Figure 6. Protein L9 assembly site on domain V. A, DMS protection effects that are independent of the presence of5 S RNA within and bordering the L1 site. Samples 1 and 4 were reconstituted with 5 S RNA and proteins, andsample 4 was treated with DMS. Samples 2 and 5 were reconstituted with proteins but no 5 S RNA, and sample 5was treated with DMS. Samples 3 and 6 were incubated alone and DMS was added to sample 6. B, Secondary struc-ture of the site where protein L1 has been footprinted (Egebjerg et al., 1991) and where some L1-23 S RNA cross-linksoccur (Osswald et al., 1990) is boxed. Chemically reactive nucleotides are violet. Protection effects observed in the pre-sence of proteins, with and without 5 S RNA, are indicated by blue and red circles, respectively. A weak protectioneffect at C2214 in the presence of 5 S rRNA (see A). Green broken lines represent putative tertiary interactions (Gutellet al., 1994). The lower boxed region probably corresponds to the L9 binding site (see the text).


is likely, since one or more of these proteins bindsper RNA domain (Figure 8), that they ``tune'' theRNA structure and facilitate subsequent proteinassembly (Bear et al., 1979).

Apart from proteins L23 and L2, which bound todomains III and IV, respectively, the remainder ofthe other 25 proteins probed assembled to domainsI, II, V and VI (Table 2). This suggests that thereare two main assembly centres, one withindomains I�II, where at least 13 proteins bind, anda second within domains V�VI, where at least 11proteins bind. This result provides a rationale forthe earlier designation of L24 and L3 as ``assemblyinitiator'' proteins on the basis of in vitro reconsti-tution experiments (Nowotny & Nierhaus, 1982),since they bind within domains I and VI, respect-ively (Figure 8), where both produce complex foot-prints that are indicative of substantial modulation

of the local structures (Egebjerg et al., 1987; Lefferset al., 1988). The fact that we did not observeadditional assembly effects within the transcriptsof domains VI�I or VI�I�II (Table 3) also suggeststhat the two centres assemble independently of oneanother.

Although the rRNA transcripts lacked post-tran-scriptional modi®cations, it is unlikely that thiswould have impaired protein assembly. It has beenshown that at least 17 of the 23 known modi®-cations in E. coli 23 S RNA are dispensable forboth assembly and function (Green & Noller, 1996)and, moreover, the remaining six (between pos-itions 2445 and 2523), which were not tested indi-vidually, are not conserved phylogenetically anddo not lie in a known protein assembly region ofdomain V. Furthermore, the assembly data comp-lement the results of protein-23 S RNA cross-link-


ing within the 50 S subunit. Several cross-linkswere induced by different bifunctional reagentsand localized, at least partially, on the 23 S RNAwhere up to four cross-linked sites per proteinwere observed, some within different RNAdomains (Osswald et al., 1990). The cross-linkedrRNA sites re¯ect either the binding site of the pro-tein or a close RNA neighbourhood, and the resultssummarized in Table 5 show that about half of theprimary binding proteins (L1, L2, L5, L11, L18)cross-link within their putative RNA bindingregions while all of the others, except L23, produceat least one cross-link within their assemblydomain (L3, L4, and L24).

Figure 7. Footprint of proteins and 5 S RNA assembled to(A) kethoxal and (B) DMS. Domain V modi®ed alone (ÿ). Gture of the 5 S RNA assembly region on domain V. Putativ(Leffers et al., 1987; Gutell et al., 1994). Nucleotide reactivitireactivities are denoted by arrowheads.

Our data also correlate well with earlier in vitroassembly studies of the endogenously produced23 S RNA fragments, 13 S RNA (�domains I-II),18 S RNA (�domains III to VI) and 11 S RNA(�domains V-VI; Chen-Schmeisser & Garrett, 1976;Spierer et al., 1979). A comparison is also possiblewith the protein content of de®ciently assembledparticles, in vivo, deriving from deletion mutants of23 S rDNA (Liiv et al., 1996). For those complexeswhere the deletions correspond approximately towhole domains or multi-domains, comparativedata are presented (Table 6). Proteins that areindirectly linked with a deleted RNA region in vivo,and are underlined, assemble directly with that

domain V (�) showing the main altered reactivities with, C and A denote sequencing tracks. C, Secondary struc-

e tertiary interactions are indicated by green broken lineses are indicated as described for Figure 6 and enhanced


RNA region in vitro (Figure 8). The correlation isclosest for the large in vivo deletions, which includemost of domains I-III and most of domains IV-VI(Table 6).

The RNA region attributed to binding L4(Figure 5B) yielded the only footprint in the wholeof domains I and II that has not previously beenassigned to a binding protein. It shows limitedstructural similarity to the putative L4 site on themRNA of the S10 operon (Figure 5C) and suchsimilarities are a common feature of the bindingsites of other rRNA binding proteins that act astranslational repressors (reviewed by Draper,1996). Earlier, a small seven-base deletion wasmade in the corresponding 23 S rDNA site (pos-itions 359±365; Figure 5B) and, although the resul-tant rRNA could assemble into 50 S subunits, thestructure of domain I was substantially altered andthe assembled 50 S subunits could not form 70 S

Figure 8. A representation of the secondary structure of Eto VI are boxed by thin lines. The footprint produced by theassembly data from the present study are included and pret al., 1991), L2 (Beauclerk & Cundliffe, 1988; Egebjerg et(Schmidt et al., 1981; Beauclerk et al., 1984; Egebjerg et al., 11987). Nucleotides are indicated by dots. Base-pairs that reby lines; if the evidence is con¯icting, the bases are juxtapo®ftieth nucleotide is marked by a dash.

ribosomes (Skinner et al., 1985), both of whichcould re¯ect the non-assembly of protein L4. U360is also hyper-reactive with a monovalent psoralenadduct in 23 S RNA but resistant in 50 S subunits,a further indicator of a protein binding site(Egebjerg et al., 1991; Egebjerg, 1993).

Earlier, two protein-RNA fragments were iso-lated from ribosomes after limited RNase T1 diges-tion, one containing proteins L5, L18, L25 and 5 SRNA and the other proteins L1 and L9. The partialoligonucleotide sequences of the RNA componentscorrespond to the domain V sequence (Branlantet al., 1976) and the assembly results for domain Vcorrelate with these protein contents (Table 2 andFigure 4A). For the former fragment, L9 is likely toinduce the reactivity changes that occur directlybelow the L1 site (Figure 6A). Indeed, this infer-ence is reinforced by the results of a toeprintingstudy on an L9-23 S RNA fragment complex

. coli 23 S RNA (Egebjerg et al., 1990a), where domains I5 S RNA and proteins is enclosed by a thick line. Proteinimary protein binding sites are shaded for L1 (Egebjergal., 1991), L3 (Leffers et al., 1988), L11 and L10.(L12)4

990b), L23 (Egebjerg et al., 1991) and L24 (Egebjerg et al.,ceive support from compensating changes are indicatedsitioned. Every 250th nucleotide is numbered and every

Table 6. Comparison with in vivo assembly dataobtained with deletion mutants of E. coli 23 S RNA

DeletionDomain

(approximate) Non-assembled proteins

40±608 DI L4, L14, L15, L20608±1350 DII L11, L13, L21, L23, L251350±1908 DIII-IV L232650±(�31) DVI L3, L14, L19

250±1656 DI-IIIL4, L10, L11, L13, L15, L20,L21, L22, L23, L29

1656±2773 DIV-DVIL1, L3, L5, L6, L9, L17, L18,L19, L25, L27

The assembly data are from Liiv et al. (1996), taken from 2Dgels when presented, where the deletions correspond approxi-mately to the domains indicated (see Figure 8 for more preciseassignments). Proteins are given that did not assemble, prob-ably owing to the given deletion. Underlined proteinsassembled directly to domains in the present study (summar-ized in Figure 8). The smaller proteins, as in this study, werenot assigned.


(Adamski et al., 1996), which suggests that theN-terminal domain of the protein interactsimmediately 50 to position 2231, close to the protec-tion effects observed at positions 2225, 2198 and2199 (Figure 6). The binding limit of the C-terminaldomain was assigned, using a similar approach, 50to position 2179 (Figure 6B), suggesting that theelongated protein extends into the L1 bindingregion.

The 5 S RNA assembly to 23 S RNA requiresproteins L18 and L5 but not L25, which binds to5 S RNA (Newberry & Garrett, 1980 and referencestherein). The low yield of L25 assembly comparedto L18 and L5 (Figure 4A, Tables 2 and 3; RoÈhl &Nierhaus, 1982) may re¯ect its anomalous optimalMg2� dependence for 5 S RNA binding (Spierer &Zimmermann, 1978); possibly, in vivo, it assemblesto the de novo transcript of 5 S RNA. Most of thereactivity changes observed in the region 2250±2410 were dependent on the presence of 5 S rRNAand are assigned to the binding of 5 S RNA-L18-L5complex together with protein L27 (Figure 7). Onlyvery few effects occurred in the absence of 5 SRNA and these all lie in regions implicated in ter-tiary interactions (Figure 7C), suggesting that somerenaturation of the RNA structure occurs duringreconsitution. The complex pattern of reactivitychanges in this region suggest that extensive pro-tein-RNA interactions occur, possibly concurrentlywith conformational changes in the 23 S RNA. Thelatter region has been cross-linked within the 50 Ssubunit to proteins L5 (G2307), L18 (G2307-U2320)and L27 (to fragments 2272-2283, 2320-2323 and2332-2337) using different chemical reagents(Osswald et al., 1990). Moreover, its functionalimportance is underlined by its extending fromG2252, which has been implicated in the bindingof C74 of P-site bound tRNA (Samaha et al., 1995),to near C2394, which has been implicated in E-sitetRNA binding (Moazed & Noller, 1989).

The requirement for a two-step procedure forreconstitution of the 50 S subunit of E. coli in vitrobut not in vivo (Nierhaus & Dohme, 1974) is notfully understood. Earlier results suggested that dis-crete groups of protein assembled during the tworeconstitution steps in vitro; 22 proteins in the ®rst

Table 5. Assignments of chemically cross-linked pro-teins within the 23 S RNA

Domain Proteins cross-linked within domains

DI L4, L23(2), L24, L28DII L4, L11*, L13, L17#, L21(2), L30DIII L9DIV L1(2), L2*, L14, L19DV � 5 S RNA L1(2)*, L5*, L6, L18*, L27(3), L33DVI L3, L32

The data are taken from Osswald et al. (1990). Proteins listedare those that were cross-linked within the individual domainsof 50 S subunits. Proteins that both assembled and were cross-linked within the same domain are underlined. Asterisks indi-cate that cross-linking occurred within the putative RNA bind-ing site (see Figure 8). L17#, cross-linked at the border ofdomains II and III.

and ten in the second (Dohme & Nierhaus, 1976),and that the assembly steps were followed byrate-limiting conformational changes (Sieber &Nierhaus, 1978). However, both kinetic studies ofassembly performed at 0�C and 4 mM Mg2� (Dattaet al., 1986), and the present results obtained underdifferent Mg2� and temperature conditions, showsome assembly of the second group of proteins L6,L14, L16, L19 and L27 (Figure 4 and Table 2). Thismay re¯ect the fact that the 50 S subunits can fol-low different assembly pathways in vitro, as wassuggested by Franceschi & Nierhaus (1990). How-ever, a conformational change has been detectedduring E. coli 50 S subunit assembly in vitro(Spillmann et al., 1977) and, moreover, forhaloarchaeal 50 S subunits, there is strong evidencethat a major conformational change of a ribonu-cleoprotein occurs on raising the salt concentrationfrom 1 M to 2 M (NH4)2SO4, which is crucial forfunctional subunit assembly (Sanchez & Amils,1995).

Given our data, where only small differencesare seen in protein assembly under the conditionsof the two Mg2�/temperature steps (Table 4), aconformational change is a likely rationale for theoccurrence of two-step assembly. Moreover, sincethe conformational change observed during E. coli50 S subunit assembly requires the presence ofproteins L4, L13, L20, L22 and L24 (domainsI�II), and is stimulated by L3 (domains V�VI;Table 2; Spillmann et al., 1977), it may result inthe juxtapositioning of the two protein assemblycentres within these RNA regions. Since L2 bindsto domains II and IV only under the conditionsof the ®rst assembly step (Figure 4C), it mayfacilitate bridging of the two RNA assemblyregions via the formation of tertiary interactionsincluding, for example, that between positions413±416 and 2407±2410 of 23 S RNA (Lefferset al., 1987), and subsequently dissociate fromdomain II during the second reconstitution step.


Materials and Methods

Construction of vectors for in vitro transcription ofrRNA domains

The strategy for cloning and expressing the domainswas similar to that described for domain VI (Leffers et al.,1988). Appropriate restriction fragments from the rrnBoperon of E. coli were prepared from the plasmidpKK3535 (Brosius et al., 1981) and cloned into a pUT719,a pUC19-derived T7 transcription vector that exhibits aunique StuI site preceded by the T7 promoter. Whenrestriction sites coincided with domain borders (Figure 1),they were used. Otherwise, larger fragments were clonedand their ends were subjected to a controlled Bal31digestion (Leffers et al., 1988). The Bal31 digestion strat-egy was employed to generate the 50-ends of domains I,II and III and the 30-ends of domains I, II and VI. Restric-tion sites were used to de®ne the 50-ends of domains IV,V and VI, and the 30-ends of domains III, IV and V. Theessential data on the restriction fragments and sites thatwere used in constructing single and multi-domain vec-tors, including those of domains VI-I and domains VI-I-II, are listed in Table 1. The vector for domain VI hasbeen described by Leffers et al. (1988). For single rRNAdomain vectors, the regions encoding the rRNA werecompletely sequenced and shown to coincide with therrnB sequence (Brosius et al., 1981). Vectors encodingmultiple rRNA domains were checked by restrictionmap analysis, and the sequences of the major junctionswere veri®ed by primer-directed DNA sequencing. Allvectors were transformed into the recAÿ strain E. coliXL1-Blue.

Preparation of the rRNA domains

The RNA domains were synthesized as run-off tran-scripts from linearized pUT719-derived plasmids. CsCl-puri®ed vectors (25 pmol) were linearized with BamHI,and with EcoRI for vectors encoding domains I and VI-I.Complete linearization was checked by agarose gel elec-trophoresis. The digestion mixture was extracted twicewith one volume of phenol and once with one volume ofchloroform, and the DNA was precipitated with threevolumes of ethanol, washed with 80% (v/v) ethanol,dried and dissolved in 1.2 ml of transcription buffer(0.75 mM rNTPs, 40 mM Tris-HCl (pH 8.0), 6 mMMgCl2, 1 mM spermidine, 5 mM DTT). T7 RNA poly-merase was added to generate run-off transcripts andthe reaction mixture was extracted twice with onevolume of phenol, precipitated with ethanol, dissolvedin 300 mM sodium acetate (pH 6.0), 5 mM MgCl2,1 mM EDTA and applied to a 1 cm � 30 cm Superose6 gel ®ltration column (Pharmacia) equilibrated in thesame buffer. The RNA peak was precipitated withethanol and washed with 80% ethanol. After drying,the RNA was dissolved in 10 mM Tris-HCl (pH 7.8),1 mM MgCl2 at 2 mg/ml and stored in aliquots atÿ80�C. Puri®ed RNA transcripts were checked ondenaturing polyacrylamide gels (see Figure 2) and con-centrations were estimated from A260 measurements.RNA domains (1 mg/ml) were renatured inT40M20A400E0.2D1, pH 7.8 (40 mM Tris-HCl (pH 7.8),20 mM MgCl2, 400 mM NH4Cl, 0.2 mM EDTA, 1 mMDTT) by heating at 56�C for ten minutes and coolingslowly to room temperature.

Isolation of 50 S subunit proteins

Total 50 S subunit proteins were isolated from E. colistrain MRE600 using a LiCl extraction procedure modi-®ed from that described by Dijk & Littlechild (1979). The50 S subunits (30 mg/ml), from a sucrose density gradi-ent in a zonal rotor in T10M1K100 (pH 7.4) (10 mM Tris-HCl, pH 7.4, 1 mM MgCl2, 100 mM KCl) were diluted to12.5 mg/ml with buffer containing 20 mM Hepes-NaOH(pH 7.0), 20 mM EDTA, 14 mM b-mercaptoethanol,100 mM PMSF, 200 mM benzamidine hydrochloride. Onevolume of 14 M LiCl was added with vigorous stirringand the mixture was left overnight at 4�C. After centrifu-ging at 100,000 g for two hours at 5�C, the supernatantwas dialysed (Spectrapor tubing; cut-off 6000 Da) against50 volumes of ice-cold 70 mM LiCl, 5 mM Hepes-NaOH(pH 7.0), 7 mM 2-mercaptoethanol, 50 mM PMSF,100 mM benzamidine hydrochloride, with one bufferchange and then against 50 volumes of ice-coldT40M20A400E0.2D1 (pH 7.8), 0.01% (v/v) Triton X-100,50 mM PMSF, 100 mM benzamidine-HCl with two bufferchanges and no protease inhibitors in the last dialysisstep. The TP50 concentration was estimated by UVabsorption (E230 nm (1 mg/ml) � 4.5) and diluted to0.5 mg/ml in T40M20A400E0.2D1 (pH 7.8), 0.01% TritonX-100, before freezing aliquots in liquid nitrogen andstoring at ÿ80�C.

Preparation and protein content ofprotein-domain complexes

Each RNA transcript was mixed with TP50 at a molarratio of 1:2. A 1.0 ml aliquot of TP50 (0.5 mg/ml inT40M20A400E0.2D1 (pH 7.8), 0.01% Triton X-100) wasthawed quickly at 42�C and placed on ice. For transcriptscontaining domain V, renatured 5 S RNA from E. coliMRE600 (Monier & Feunteun, 1971) was mixed at 1 mg/ml in T40M20A400E0.2D1, pH 7.8, with the proteins at anequimolar ratio and kept on ice for two minutes beforetransferring to a 42�C water bath, where the renaturedRNA domain transcript (1 mg/ml in T40M20A400E0.2D1,pH 7.8) was added, with mixing, and maintained for30 minutes before placing on ice. Reconstitutions werealso performed under the conditions of the ®rst(T40M4A400E0.2D1 (pH 7.8) at 44�C) and second steps(T40M20A400E0.2D1 (pH 7.8) at 50�C) of 50 S subunitreconstitution. Unbound TP50 was removed on a gel-®l-tration column equilibrated in T40M20A250E0.2D1 (pH 7.8),0.01% Triton X-100. A Superose 12 (1.6 cm � 40 cm) col-umn (Pharmacia) was used for domains I, III, IV or VIand a Superose 6 (1 cm � 30 cm) column (Pharmacia)was employed for the larger domain constructs, at ¯ow-rates of 1 ml/(min cm2). The column pro®les were ana-lysed on 1D SDS/polyacrylamide gels (Laemmli, 1970),where samples were precipitated with 10% (v/v) tri-chloroacetic acid and the pellets washed with ice-coldacetone. The peak containing the RNA-protein complexwas also examined on 2D gels (Geyl et al., 1981) afteradding 10 mg/ml bovine serum albumin, precipitatingwith 10% trichloracetic acid, washing with ice-coldacetone and drying. At this stage, a small amount of pur-i®ed 50 S subunits was sometimes added to produce aweak background of TP50 spots in the 2D gel. The RNA-protein complex was then dissolved (1 mg RNA/ml) in200 mM Tris-base, 10 mM EDTA, 5 mM DTT, and onevolume of 14 M LiCl was added with rapid mixing. Thesample was kept overnight at ÿ20�C before centrifugingand diluting the supernatant with one volume ofdistilled water. Proteins were precipitated with 10%


trichloroacetic acid, washed with acetone and driedthoroughly before loading onto the 2D gel.

Footprinting of proteins L4, L9 and the 5 S RNA-L5,L18 complex on single RNA domains

Transcripts containing domains I, II, I�II, I�II�III,and V, V�VI and IV�V�VI, were incubated inT40M10A250E0.1D0.5 (pH 7.8), in the absence and presenceof proteins, and 5 S RNA for those containing domain V,and subjected to chemical modi®cation, using the nucleo-tide-speci®c reagents kethoxal (guanosine; N1, N2), DMS(adenosine N1 > cytidine N3) and CMCT (uridineN3 > guanosine N1). The modi®ed sites were analysedby reverse transcription as described by Leffers et al.(1988). Potential protein sites were mapped in domains Iand II using primers with 30-ends at positions 228, 516,617, 770, 887, 1158 and 1309. For sites in domain V, theoligonucleotides were extended from positions 2102,2237, 2424, 2500, 2607, 2654 and 2888.

Acknowledgements

This research was supported by the Danish Centre forRNA Regulation and by grants from the Danish NaturalScienti®c and Medical Research Councils and the NovoNordisk and Carlsberg Foundations. L.B.J. and L.é.received PhD stipends from the Danish Center of Micro-biology. Ilia Leviev is thanked for help with some pre-liminary experiments.

References

Adamski, F. M., Atkins, J. F. & Gesteland, R. F. (1996).Ribosomal protein L9 interactions with 23 S rRNA:the use of a translational bypass assay to study theeffect of amino acid substitutions. J. Mol. Biol. 261,357±371.

Amils, R., Matthews, E. A. & Cantor, C. R. (1978). Anef®cient in vitro total reconstitution of the Escherichiacoli 50 S subunit. Nucl. Acids Res. 5, 2455±2470.

Bear, D., Schleich, T., Noller, H. F., Douthwaite, S. &Garrett, R. A. (1979). The ribosomal S4-RNA frag-ment melts cooperatively when complexed withprotein S4. FEBS Letters, 100, 99±102.

Beauclerk, A. A. D. & Cundliffe, E. (1988). The bindingsite for ribosomal protein L2 within 23 S rRNA ofEscherichia coli. EMBO J. 7, 3589±3594.

Beauclerk, A. A. D., Cundliffe, E. & Dijk, J. (1984). Thebinding site for ribosomal protein complex L8within 23 S rRNA of Escherichia coli. J. Biol. Chem.259, 6559±6563.

Branlant, C., Krol, A., Sriwidada, J. & Brimacombe, R.(1976). RNA sequences associated with proteins L1,L9, and L5, L18, L25 in ribonucleoprotein fragmentsisolated from the 50 S subunit of Escherichia coliribosomes. Eur. J. Biochem. 70, 483±492.

Brosius, J., Ullrich, A., Raker, M. A., Gray, A., Dull, T. J.,Gutell, R. R. & Noller, H. F. (1981). Constructionand ®ne mapping of recombinant plasmids contain-ing the rrnB rRNA operon of E. coli. Plasmid, 6,112±118.

Chen-Schmeisser, U. & Garrett, R. A. (1976). Distri-bution of protein assembly sites along 23 S rRNAof Escherichia coli. Eur. J. Biochem. 69, 401±410.

Datta, D., Changchien, L.-M. & Craven, G. (1986).Studies on the kinetic sequence of in vitro assemblyusing cibacron F3GA as a general assembly inhibi-tor. Nucl. Acids Res. 14, 4095±4111.

Dijk, J. & Littlechild, J. (1979). Puri®cation of ribosomalproteins from E. coli under nondenaturing con-ditions. Methods Enzymol. 59, 481±502.

Dohme, F. & Nierhaus, K. H. (1976). Total reconstitutionand assembly of the 50S subunit from Escherichiacoli. J. Mol. Biol. 107, 585±599.

Draper, D. (1996). Ribosomal protein-RNA interactions.In Ribosomal RNA Structure, Evolution, Processing andFunction in Protein Biosynthesis (Zimmermann, R. A.& Dahlberg, A., eds), pp. 171±198, CRC Press, BocaRaton, FL.

Egebjerg, J. (1993). Protein-RNA interactions in Escheri-chia coli ribosomes. PhD thesis, Aarhus University,Denmark.

Egebjerg, J., Leffers, H., Christensen, A., Andersen, H. &Garrett, R. A. (1987). Structure and accessibility ofdomain 1 of E. coli 23 S rRNA in free RNA, in theL24 RNA complex and in 50 S subunits. Impli-cations for ribosomal assembly. J. Mol. Biol. 196,125±136.

Egebjerg, J., Larsen, N. & Garrett, R. A. (1990a). Struc-tural map of 23 S rRNA. In The Ribosome Structure,Function and Evolution (Hill, W., Dahlberg, A.,Garrett, R. A., Moore, P. B., Schlessinger, D. &Warner, J. R., eds), pp. 168±197, A. S. M. Press,Washington, DC.

Egebjerg, J., Douthwaite, S. R., Liljas, A. & Garrett, R. A.(1990b). Characterization of the binding sites of pro-tein L11 and the L10.(L12)4 pentameric complex inthe GTPase domain of 23 S rRNA from Escherichiacoli. J. Mol. Biol. 213, 275±288.

Egebjerg, J., Christiansen, J. & Garrett, R. A. (1991).Attachment sites of primary binding proteins L1, L2and L23 on 23 S rRNA of Escherichia coli. J. Mol.Biol. 222, 251±264.

Franceschi, F. J. & Nierhaus, K. H. (1990). Ribosomalproteins L15 and L16 are mere late assembly pro-teins of the large ribosomal subunit. J. Biol. Chem.265, 16676±16682.

Garrett, R. A., MuÈ ller, S., Spierer, P. & Zimmermann,R. A. (1974). Binding of 50 S ribosomal subunit pro-teins to 23 S RNA of Escherichia coli. J. Mol. Biol. 88,553±557.

Geyl, D., BoÈck, A. & Isono, K. (1981). An improvedmethod for 2D gel electrophoresis: analysis of muta-tionally altered ribosomes of Escherichia coli. Mol.Gen. Genet. 181, 309±312.

Green, R. & Noller, H. F. (1996). In vitro complementa-tion analysis localizes 23 S rRNA posttranscriptionalmodi®cations that are required for Escherichia coli50 S ribosomal subunit assembly and function.RNA, 2, 1011±1021.

Gutell, R. R., Larsen, N. & Woese, C. R. (1994). Lessonsfrom an evolving rRNA: 16 S and 23 S rRNA struc-tures from a comparative perspective. Microbiol.Rev. 58, 10±26.

Herold, M. & Nierhaus, K. H. (1987). Incorporation ofsix additional proteins to complete the assemblymap of the 50 S subunit from Escherichia coli ribo-somes. J. Biol. Chem. 262, 8826±8833.

Laemmli, U. (1970). Cleavage of structural proteinsduring the assembly of the head of bacteriophageT4. Nature, 227, 680±685.

Leffers, H., Kjems, J., éstergaard, L., Larsen, N. &Garrett, R. A. (1987). Evolutionary relationships


amongst archaebacteria. A comparative study of23 S rRNA of a sulphur-dependent extreme thermo-phile, an extreme halophile and a thermophilicmethanogen. J. Mol. Biol. 195, 43±61.

Leffers, H., Egebjerg, J., Andersen, A., Christensen, T. &Garrett, R. A. (1988). Domain VI of E. coli 23 SrRNA. Structure, assembly and function. J. Mol.Biol. 204, 507±522.

Liiv, A., Tenson, T. & Remme, J. (1996). Analysis of theribosome large subunit assembly and 23 S rRNAstability in vivo. J. Mol. Biol. 263, 396±410.

Moazed, D. & Noller, H. F. (1989). Interaction of tRNAwith the 23 S rRNA in the ribosomal A, P and Esites. Cell, 57, 585±597.

Monier, R. & Feunteun, J. (1971). Preparation of 5 SRNA. Methods Enzymol. 20, 494±502.

Newberry, V. & Garrett, R. A. (1980). The role of thebasic N-terminal region of protein L18 in 5 S RNA-23 S RNA complex formation. Nucl. Acids Res. 8,4131±4142.

Nierhaus, K. H. & Dohme, F. (1974). Total reconstitutionof functionally active 50 S ribosomal particles fromEscherichia coli. Proc. Natl Acad. Sci. USA, 71, 4713±4717.

Nowotny, V. & Nierhaus, K. H. (1982). Initiator proteinsfor the assembly of the 50 S subunit of Escherichiacoli within ribosomes. Proc. Natl Acad. Sci. USA, 79,7238±7242.

Osswald, M., Greuer, B. & Brimacombe, R. (1990). Local-ization of a series of RNA-protein-cross-links sitesin the 23 S and 5 S rRNA from Escherichia coli,induced by treatment of 50 S subunits with threedifferent bi-functional reagents. Nucl. Acids Res. 18,6755±6760.

RoÈhl, R. & Nierhaus, K. H. (1982). Assembly map of thethe large subunit (50 S) of Escherichia coli. Proc. NatlAcad. Sci. USA, 79, 729±733.

Romero, D. P., Arredondo, J. A. & Traut, R. R. (1990).Identi®cation of a region of Escherichia coli riboso-mal protein L2 required for assembly of L16 intothe 50 S subunit. J. Biol. Chem. 265, 18185±18191.

Samaha, R. R., Green, R. & Noller, H. F. (1995). A base-pair between tRNA and 23 S rRNA in the peptidyltransferase centre of ribosomes. Nature, 377, 309±314.

Sanchez, M. E. & Amils, R. (1995). Absolute requirementfor ammonium sulphate for reconstitution of active70 S ribosomes from the extremely halophilicarchaeon Haloferax mediterranei. Eur. J. Biochem. 233,809±814.

Schmidt, F. J., Thompson, J., Lee, K., Dijk, J. &Cundliffe, E. (1981). The binding site for ribosomalprotein L11 within 23 S RNA of E. coli. J. Biol.Chem. 256, 12301±12305.

Sieber, G. & Nierhaus, K. H. (1978). Kinetic and thermo-dynamic parameters of the assembly in vitro of thelarge subunit from Escherichia coli ribosomes. Bio-chemistry, 17, 3505±3511.

Skinner, R. H., Stark, M. J. R. & Dahlberg, A. E. (1985).Mutations within the 23 S rRNA coding sequenceof E. coli which block ribosomal assembly. EMBO J.4, 1605±1608.

Sloof, P., Garrett, R. A., Krol, A. & Branlant, C. (1976).The binding site of L1 on 23 S rRNA of Escherichiacoli. Eur. J. Biochem. 70, 447±456.

Spierer, P. & Zimmermann, R. A. (1976). Cooperativeinteractions on the 50 S subunit of E. coli. J. Mol.Biol. 103, 647±653.

Spierer, P. & Zimmermann, R. A. (1978). Stoichiometry,cooperativity and stability of interactions between5 S RNA and proteins L5, L18 and L25 from the50 S ribosomal subunit of Escherichia coli. Biochemis-try, 17, 2474±2479.

Spierer, P., Wang, C. C., Marsh, T. L. & Zimmermann,R. A. (1979). Cooperative interactions among pro-tein-RNA components of the 50 S ribosomal subunitof E. coli. Nucl. Acids Res. 6, 1669±1682.

Spillmann, S., Dohme, F. & Nierhaus, K. H. (1977).Assembly in vitro of the 50 S subunit of Escherichiacoli ribosomes: proteins essential for the ®rst heat-dependent conformational change. J. Mol. Biol. 115,513±523.

Zengel, J. M. & Lindahl, L. (1993). Domain I of 23 SrRNA competes with a paused transcription com-plex for ribosomal protein L4 of Escherichia coli.Nucl. Acids Res. 21, 2429±2435.

Zengel, J. M. & Lindahl, L. (1996). A hairpin structure ofthe terminator hairpin required for ribosomal pro-tein L4-mediated attenuation control of the S10operon of Escherichia coli. J. Bacteriol. 178, 2383±2387.

Edited by D. E. Draper

(Received 21 May 1998; received in revised form 17 August 1998; accepted 19 August 1998)

Assembly of proteins and 5 S rRNA to transcripts of the major structural domains of 23 S rRNA

Documents

Transcript of Assembly of proteins and 5 S rRNA to transcripts of the major structural domains of 23 S rRNA