Ann. Rev. Biochem. 1984. 53: 119-{)2 Copyright © 1984 by Annual Reviews 1 nco All rights reserved

STRUCTURE OF RIBOSOMAL RNA

Harry F. Noller

Thimann Laboratories, University of California, Santa Cruz, California, 95064

CONTENTS

PERSPECTIVES AND SUMMARy............................................................................................................. 119 rRNA Species-Nomenclature ... .... .... .... .... ..... ... ............................. .... .... ...... .. .... ..... .... ... .... ..... ..... . 120

PRIMARY STRUCTURE............................................................................................................................ 121

SECONDARY STRUCTURE....................................................................................................................... 124 General Approaches ........................ .... .. ... .... ... .... ... .. .. .. ... ......................... ..... .... ... ..... ... ..... .... .. ... .... .. 124 Description of Secondary-Structure M ode/s.......... .... ........ .... ............ .... .. .. ........ ...................... .. 129 Phylogenetic Comparison ............ .. ... . ..... .... ... .... ... . ..... .... ... ..................... .... .... .... .... ... . .... ..... ...... ... .. 136 Experimental Tests..... . .. .. ... ........ ......... .... ..... ... . ... .... ................. .... ... .... ......... ..... ... ..... ................ ... . ... 136

TERTIARY STRUCTURE........................................................................................................................... 143

QUATERNARY STRUCTURE .... ... ........ .... .............................. .... ......... ...... ............................................... 146 r-Protein-rRN A Interactions..... ..... ... ..... .... ... ................. .... ... ..... .. .. .... ..... . .. .... .............................. 146 rRNA-rRNA Interactions........................................ ......... ............ ................................... .. .. .. ........ 150

STRUCTURE AND FUNCTION ...•..................................•....••...•...•...................................••..••...•..•••..••.... 150 rRNA Genetics.. .... ............. .... .... ... .... ................ .. .... .... ... ......... ... ... ...................... ..... .... .... .... .... .... ..... 151 Specific Structure- Function Correlations.... .... .... .... .... .... ........ ............. ... ..... .... .... .... ..... .... ... ..... . 151

PROSPECTS ...•..••..••............ ;.... ... .... . .. .................. .... .... .... .... .... .... .... ........... ..... .. .. ...... ... .... ... ......... ............. . 157

PERSPECTIVES AND SUMMARY

Ribosomes are ribonucleoprotein particles that are responsible for trans­lation of the genetic code. Unlike other cellular polymerases, ribosomes typically contain 50 to 60 percent RNA as an integral part of their structures. It is an intriguing problem to understand why this is so, and its elucidation is likely to be inseparable from an understanding of the structure, function, assembly, and evolution ofthese particles. This review is an attempt to summarize recent progress in our understanding of ribo­somal RNA (rRNA) structure, and to point out some of the implica-

1 19 006(}.-41 54/84/0701-D1 19$02.00

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

120 NOLLER

tions of these findings as they relate to broader biological questions. The

reader is referred to previous reviews on the topics of ribosome structure and function (1-6) and ribosomal RNA (7- 14). For the most part, emphasis is on the large (I6S- and 23S-like) rRNAs ; consideration of 5S rRNA is confined to a few salient topics. Detailed discussions of 5S rRNA and 5.8S rRNA may be found in several recent review articles (1 5-20). A recent compilation of 5S rRNA sequences is in (21).

Many complete sequences are now available for the large rRNAs from organisms representing a wide phylogenetic spectrum. This has made possible elucidation of the secondary structures of the rRNAs by compara­tive sequence analysis and identification of highly conserved as well as phylogenetically variable regions of these molecules. The secondary structures begin to suggest details of the molecular organization of the rRNAs, outlining major structural domains and placing certain constraints on tertiary folding. Clues to the three-dimensional structure of rRNA in the ribosome are beginning to emerge, but we are not yet at the stage where a detailed model can be built.

The question of the direct involvement of rRNA in translation has become an increasingly interesting one in recent years. Affinity labeling experiments and specific protection of rRNA from chemical and enzymatic probes by functional ligands imply participation of RNA in a variety of ribosomal functions. Genetic studies, particldarly in the yeast mitochon­drial system, have shown that point mutations in rRNA genes can confer functional alterations upon ribosomes. The suggestion that ribosomal RNA plays a fundamental and direct role in translation is supported by the available experimental evidence and is entirely consistent with the very highly conserved nucleotide sequences and secondary structure elements that are found in these molecules.

r RNA Species-Nomenclature

In prokaryotic cells (in both the eubacterial and archaebacterial lines), ribosomes contain three ribosomal RNA molecules, usually called 5S, 1 6S, and 23S rRNA. These. molecules contain about 120, 1 540, and 2900 nucleotides, respectively (23-30). This is true for chloroplast ribosom(:s as well (3 1-32, 35-36), although mitochondria sQow a wide range of variability in the sizes of their rRNAs (37- 49), and appear to lack 5S rRNA (50). Many chloroplasts contain a 4.5S rRNA corresponding to the 3' terminal 96 nucleotides of 23S rRNA (5 1 , 52), and some contain a 7S rRNA, corre­sponding to the 5' 291 nucleotides of 23S rRNA (53). The 23S-like rRNA from Paramecium mitochondria also contains a 7S rRNA corresponding to 'the 5' '" 3 10 nucleotides of 23S rRNA (44). In eukaryotic cytoplasmic ribosomes, the three main classes ofrRNA are usually called 5S, 17S or I8S,

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 121

and 26S or 28S rRNA. Most eukaryotic cytoplasmic ribosomes also contain a 5.8S rRNA, which corresponds to the 160 nuclcotidcs at the 5'­terminus of prokaryotic 23S rRNA (29, 54-56); some insects also have a 2S rRNA corresponding to the 3' '" 25 nucleotides of 5.8S rRNA (57j:

For the sake of clarity, I often refer to rRNAs from small ribosomal subunits as "16S-like" rRNAs, or "small subUI;tit" rRNAs ; similarly, the large rRNAs from the large subunit are referred to as "23S-like" or "large subunit" rRNAs. To avoid confusion, I always refer to 5S rRNA as such. Other nomenclature (e.g., 12S, 1 8S, etc) is used where specifically required. In referring to specific positions in rRNA structures, I use here the numbering of Escherichia coli 16S, 23S, and 5S rRNA (22-23, 29, 58) unless otherwise noted.

PRIMARY STRUCTURE

At this writing, nucleotide sequences for over twenty 16S-like rRNAs, and over a dozen 23S-like rRNAs have been completed (Table 1 ). For the most part, these have been deduced by DNA sequencing of the corresponding cloned rRNA genes. The DNA sequencing approach has been favored for several reasons: Ribosomal RNA gene�are easily cloned, since it is usually relatively easy to obtain rRNA for use as a hybridization probe, and rRNA genes are present in multiple copies in most organisms ; restriction enzymes provide a wide variety of precise fragments for sequencing ; and both strands of the DNA can be sequenced, providing independent verification of the sequence. Drawbacks to DNA sequencing are that location of the ends of the mature rRNAs and information about post-transcriptionally modified nucleotides must be obtained independently. Also, the extent of sequence heterogeneity in the rRNA population is not apparent using this approach. In any case, the accuracy of the resulting sequence is by far the most crucial factor in any consideration of sequencing methodology, and in this regard DNA methods are presently superior.

Examination of Table 1 shows that the size of the 16S-like and 23S-like rRNAs varies over a two to three fold range. The smallest sequenced 16S­like rRNAs are the trypanosome mitochondrial 9S rRNA (597 residues) (37 ; Table 1) and the mammalian mitochondrial 12S rRNAs (�950 residues). The largest examples are the eukaryotic cytoplasmic 1 8S rRNAs (59-62 ; Table 1), which can comprise over 1800 nucleotides. This extreme variation in size is not to be taken as an absence of structural constraihts, however. As is discussed below, there is clear homology between the higher­order structures of r RN As from all sources. Furthermore, certain regions of rRNA show extremely high sequence conservation. It is convenient to think of "typical" 1 6S-like and 23S-like rRNAs as having about 1 500 and 2900

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

122 NOLLER

nucleotides, respectively, and the other types as being variants of these. Typical rRNAs would comprise those of eubacteria, archaebactt�ria, chloropiasts, and plant mitochondria; the variant types would include eukaryotic cytoplasmic and the remaining mitochondrial rRNA-s.

The content of post-transcriptionally modified nucleotides in rRNA is much less than that of tRNA, and is typically less than 1% in prbkaryotes. In the cases of E. coli I6S rRNA and Xenopus laevis I8S rRNA, all or most of the modified nucleotides have been identified (60, 67-69) and placed in the RNA chain (23-24,60). In eubacteria, the majority of modifications are base methylations (67-69). In E. coli 16S rRNA, these are clustered in six regions (Figure 1) : ':GS27; �G966 and �C967; �G1:207; ,!Cm1402 and

Table 1 Sequenced 16S-like and 23S-like rRNAs

Organism Residues' Reference

16S-like Eubacteria

E. coli 1542 23,24 B. brevis (1540) 25 P. vulgaris 1544 26 A. nidulans (1487) 27

Chloroplasts Maize chloroplast 1490 31 Tobacco chloroplast 1485 32 Euglena chloroplast 1491 33 Chlamydomonas chloroplast 1475 34

Archaebacteria H. volcanii 1469 28

Eukaryotes S. cerevisiae cytoplasmic (18S) 1799 59 X. laevis (18S) 1825 60 Rat liver (18S) 1874 61,62 Dictyostelium (17S) 1873 a Maize (18S) b

Mitochondria Trypanosome 5970 37 Human 954 38 Bovine 958 39 Mouse 956 40 Rat 953 41 Yeast 1 686 42, 43

Aspergillus 1437 44 Paramecium (1607) 45 Wheat 1957 d Maize ( 1962) e

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA

Table 1 (continued)

Organism

23S-like Eubacteria

E. coli B. stearothermophilus

Chloroplasts Maize chloroplast Tobacco chloroplast

Eukaryotes S. cerevisiae (26S) S. carlsbergensis (26S) Physarum (268) X. laevis (288) Rat (288)

Mitochondria Trypanosome· Human Bovine Mouse Rat Paramecium Yeast Aspergillus

• M. Sogin , personal communication. b J. Messing, personal communication . • P. Sloor, personal communication. d M. Gray, personal communication. 'R. Sederoff, personal communication. r S. Gerbi, personal communication.

Residues' Reference

2904 29 2931 30

2903 35 2904 36

3392 65 3393 64 3788 63

(4110) f 4718 66

1152 37 1559 38 1571 39 1582 40 1584 41

(2380+ ) 47 3273 48 2768 49

'In cases where there is some uncertainty about the actual length of the mature rRNA chain, the best estimate for the number of residues is given in parentheses.

123

�C1407; mU1498; �G1516' m�A1518 and m�A1519 (23-24). In E. coli 23S rRNA, pseudouridines and ribothymidines are found, in addition to base and ribose methylations (67-68). There is also evidence for the existence of other kinds of base modifications in E. coli 23S rRNA, as yet uncharacterized (S. Turner, unpublished). In X. laevis 18S rRNA, there are numerous ribose methylations in addition to base modifications (60), as appears to be generally true for eukaryotic cytoplasmic rRNAs (70-71). The modified bases tend to occur in positions in the sequence analogous to those in eubacteria, while the ribose methylations are more widely distributed. In most cases, the post-transcriptional modifications in E. coli 16S rRNA have been found to be located in regions that are accessible to

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

124 NOLLER

chemical probes in the intact 30S ribosomal subunit (72). These regions furthermore tend to be highly conserved in primary and secondary structure (10, 1 3) (see Structure and Function section).

Most intriguing are the regions of very high conservation of primary structure in the large rRNAs. In the 16S-like rRNAs, positions 322-·329, 5 1 5-533, 691-699, 1047- 1061 , 1 39� 1407, and 1492-1506 are nearly universal (13; Figure 1). Only the mitochondria show significant variation within these sequences. The three sequences around positions 530, 1400, and 1 500, each containing about twenty conserved nucleotides, are particularly striking. The loop around position 530 maintains its conserved sequence for the most part even in the most extreme mitochondrial examples, and is virtually unchanged across the three major phylogenetic lines. One can only conclude that these nucleotide sequences are among the most ancient of those presently in existence, and that their persistence in the face of evolutionary pressures is a measure of the crucial role of ribosomal RNA in the translation process. A detailed analysis of the phylogenetic conservation of each section of the 16S-like rRNAs may be found in (13). In the 23S-like rRNAs, sequences that are universal, or nearly so, include those at the approximate positions 671-677, 803-811, 1664--1677, 1833-1838,

190� 1905, 1927-1943, 1959-1973, 199� 1996, 2059-2065, 2445-2454, 2467- 2483, 250�2506, 2552- 2556, 2583-2589, and 2654-2665.

The high homology among the sequences of rRNAs from phylogeneti­cally diverse sources is by itself persuasive evidence that all ribosomes are related through a common ancestral ribosome. Based on this notion, Woese and colleagues have used rRNA sequences, or RNase T 1 oligomers thereof, to measure evolutionary distances between organisms (73). The power of this approach is documented by their d iscovery of the third major line of descent, the archaebacteria (74) ; sequences of 16S rRNAs from these organisms were shown to be no more closely related to those of eubacteria than to those of �l.lkaryotes. Still another application is the use of comparative sequence analysis in deducing secondary structure, as de­scribed in the following section.

SECONDARY STRUCTURE General Approaches

COMPARATIVE SEQUENCE ANALYSIS Determination of secondary structure, even in a relatively small RNA molecule, can be surprisingly difficult, as demonstrated in the case of E. coli 5S rRNA. Although only 120 nucleotides in length, eight years elapsed between elucidation of its primary (22) and secondary (75-76) structures. The large number of spurious secondary structures for this molecule is well documented (77). Because the com-

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 125

plexity of the problem of secondary-structure determinations increases approximately as the second power of the length of the RNA chain in question, one can readily sense the need for extremely reliable criteria with which to test the validity of each base-pairing possibility for the large rRNAs. As clearly demonstrated in the cases of tRNA and 5S rRNA, the most convincing proof for the existence of biologically significant pairing (short of X-ray crystallographic evidence, perhaps) comes from compara­tive sequence analysis.

Briefly, the comparative approach begins with the assumption that the molecules under comparison have essentially the same secondary structure. The primary structures are then aligned, using positions of high sequence homology as a guide. Success in this approach is crucially dependent upon the precision of the sequence alignment; if there is uncertainty at this stage, any conclusions about secondary structure can be taken to be equivocal. Next, base changes between the compared sequences are noted, and where compensating base changes maintain complementarity between two potential pairing regions, this is taken as evidence for the existence of a true helix at that position. Conversely, uncompensated changes, leading to mismatches (except in the case of G-C to G-U changes), are evidence against the helix. Clearly, the more independent examples of compensating changes that can be found for a given helix, the more likely it is to be correct. In studies on the large rRNAs, a useful criterion has been to consider two independent sets of compensating base changes within a helix as proof for its existence ( 13, 78). In principle, comparative sequence analysis amounts to assessing the data from a kind of preexisting genetic experiment in which all existing organisms can be considered pseudo-revertants of the various mutations in rRNA genes that have occurred during evolution.

Most investigators now use the comparative approach in studies on the secondary structures of the large rRNAs (10, 1 3, 78-82a). Discrepancies among structural models proposed by different laboratories can usually be attributed to the degree of emphasis placed on the comparative approach relative to other criteria, or to lack of rigor in its use, most commonly in the sequence-alignment step. The more sequences that become available, the more straightforward the alignment process becomes, as patterns of conservation vs variation become established. Furthermore, as in the case of the tRNA cloverleaf, the initial secondary-structure model becomes a prediction that must be fulfilled by a precise fit to each newly derived nucleotide sequence. That a 16S rRNA secondary structure derived on the basis of only three complete sequences and RNase T 1 oligomer catalogs has survived testing by all of the currently available 24 complete 16S-like rRNA sequences (which represent a very broad sampling of the phylogenetic spectrum ; Table 1) is in itself convincing evidence for its validity.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

126 NOLLER

COMPUTER ANALYSIS Because of the size of the large rRNA molecules, it is natural to employ computer methods at various levels of secondary­structure analysis. These approaches can be distinguished by the extent to which the problem is handled in the computer, and by which aspects of the problem are addressed by this method. At one extreme, the computer may be used simply to store sequence information; at the other extreme, the output may take the form of a fully developed secondary structure. Most of these methods center on free-energy minimization algorithms (83--88), using empirically determined values for stability parameters (89). Some approaches also allow incorporation of additional biological information, such as comparative sequence information (85), and positions that are susceptible to single or double strand-specific probes (88). Thus far, satisfactory secondary structures for the large rRNAs have not been obtained by use of purely computational approaches, although this appears to have been achieved in the case of 5S rRNA (85). A major difficulty in applying computer-based methods to the large rRNAs is that the requirement for computing time and/or working memory usually increases as the third or higher power of the length of the polynucleotide chain. This should become a less serious obstacle as faster processors with much larger memories become available and as more efficient algorithms are developed for making use of comparative sequence information.

FREE-ENERGY PREDICTION Another problem that continues to be explored is the estimation of free energies of potential RNA secondary structures. If the true biological structure(s) of an RNA molecule is that which has the lowest global free energy, then it is important to determine which aspects of the structure contribute significantly to this number. Values for nearest­neighbor stacking energies and costs of unpaired loops, bulges and mismatches have been empirically estimated (89) and refined (90--91 ; I. Tinoco, personal communication) for helical RNA oligomers. These are potentially pow�rful tools, but alone do not presently appear to suffice for prediction of secondary structure. For example, in certain regions of 16S rRNA, the biologically significant structure (as judged by all other criteria, including comparative sequence analysis) appears to comprise helices that are predicted to be significantly less stable than some that can be rigorously excluded by comparative al1alysis (79). It may well be that when one compares the total free energies of all helices in competing structures, the global free energy is in fact lower in the biologically significant structure, even though individual helices may not be the most stable ones on a helix­by-helix basis. A practical difficulty arises in attempting to test this possibility, since the total number of potential secondary models to be tested for a molecule the size of 16S rRNA is greater than 10100 (259) (for

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 127

comparison, the number of fundamental particles in the universe is estimated to be about 1080).

Another important consideration is that structure other than secondary may contribute significantly to the stability of the folded RNA molecule. There are indications that certain tertiary features oftRNA are in fact more sta�le than some secondary interactions (92). Tight binding of multivalent cations may also make important contributions. Finally, ribosomal RNAs exist as ribonucleoprotein particles, interacting strongly and specifically with many ribosomal proteins. Binding of r-pr:oteins to rRNA is known to alter the thermal stability of the higher-order structure of the RNA, usually decreasing or eliminating low-temperature "premelting" effects that are observed in melting curves of the naked rRNAs, and often increasing the cooperativity or sharpness of the melting transition (93). The influence of ribosomal proteins on the secondary structure of rRNA is probably limited, however, judging by the close agreement of biochemical probe experiments on naked rRNA with the phylogenetically derived secondary-structure models (see below). Very likely, all of these factors �ltimately play a part in determining and stabilizing the structure of rRNA, but quantitative estimation of their respective contributions to the conformational free energy is not presently within reach.

EXPERIMENTAL METHODS A wide variety of experimental approaches has been employed in the study ofrRNA structure. It is probably fair to say that experimental methods alone are not yet sufficient to deduce a correct secondary structure but are an extremely important test of the structural features derived by comparative sequence analysis. The most successful of these approaches include (a) the use of chemical and enzymatic probes, both of which are represented by single strand- and double strand-specific classes (72, 95-1 1 1); (b) two-dimensional gel systems, in which RNA fragments from partial nuclease digests, associated by base pairing in the first dimension, are resolved under dissociative conditions in the second dimension (25, 108, 141) ; (c) oligonucleotide probes, which presumably bind to free, unpaired regions of the rRNA ( 1 12-1 15) ; (d) direct observation of structural features in partially unfolded rRNA by electron microscopy (120, 260); and nuclear magnetic resonance techniques ( 1 1 6-119).

Among the more useful single strand-specific chemical probes are kethoxal (G-specific) (72, 95), diethyl pyrocarbonate (A- and G-specific) (96), bisulfite (C- and V-specific) (79), dimethyl sulfate (C-specific as a single­strand probe ; G-specific as a "major groove" or tertiary-structure probe at 7N) (96-97), and m-chloroperbenzoate (79) and monoperphthalate (98) (both A-specific). The reliability of chemical modification in obtaining detailed structural information has been verified in several cases by studies

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

128 NOLLER

on tRNA, where results can be compared with a known structure (96, 99). Examples of chemical reagents that have been used as double strand­specific probes are the psoralen derivatives aminomethyltrioxsalen (AMT) and hydroxymethyltrimethylpsoralen (HMT) ( 100-10S). Psoralens inter­calate between stacked base pairs in nucleic acids and undergo photo­chemically induced cycloaddition to pyrimidines on opposite strands (106). A number of ribonucleases have been used as structural probes, including RNase T 1 (24), RNase A (24), RNase T 2 ( 107), Sl nuclease (108) (all single strand specific) and cobra venom RNase ( 109-1 1 1) (double strand specific). A potential danger in the use of nucleases, particularly single strand­specific ones, as structural probes is that cleavage of the sugar-phosphate chain may lead to significant unfolding, or other rearrangement of the RNA structure. Thus, it is important to distinguish between primary and secondary cleavage events, and to rely only on primary cleavage sites as indicators of structure; in smaller RNAs, this has been achieved by an elegarit method utilizing alternate S' and 3' end labeling ( 1 1 1). Oligo­nucleotides complementary to specific sequences iii rRNA have been used to probe single-stranded regions ( 1 1 2- 1 1S). With the ready availability of synthetic DNA oligomers of predetermined sequence, this general ap­proach should become increasingly useful. A potential hazard is that the oligomer may itself perturb the RNA structure, in cases where the stability of an oligomer-rRNA duplex is significantly more stable than an existing structural feature of the rRNA. Also, care must be taken to ensure (e.g. by the use of RNase H) that the mode of binding of the oligomer is by conventional base pairing ( 1 14).

Nuclear magnetic resonance has been used with considerable success in probing the structure ofSS rRNA (1 16-1 1 8), and the colicin fragment of 16S rRNA ( 1 19), comprising the 3'-terminal 49 nucleotides of the latter molecule. Application of the Nuclear Overhauser Effect (NOE) (92) allows uriambiguous assignment of individual proton resonance to specific nucleotide residues. One is. then permitted to make measurements in solution to study the dynamic behavior of individual nucleotides, as well as to establish or confirm the existence of base pairing and, in principle, other higher-order structure. Unfortunately, this method appears to be limited to the study of small RNA molecules or RNA fragments. Another physical approach that has been employed with some success is electron microscopy ( 120, 260). RNA molecules are spread on grids under partial unfolding conditions, and the sizes of resulting loops and tails are measured. In this approach, choice of unfolding cpnditions is crucial, so that some features are maintained while others are disrupted, without at the same time introducing structural rearrangements. The resolution of this method is somewhat limited, not only by the resolution of the electron microscope,

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 129

but also because small hairpins or other local structure, too small to be resolved, nevertheless contribute to variable shortening of the apparent length of the RNA chain.

Description of Secondary-Structure Models

General secondary-structure models for 16S and 23S rRNAs have been proposed by three different groups of investigators (7, 10, 13, 78-82) and numerous proposals for the secondary structure of 5S rRNA have been made (1 5-20,75-76). Here, discussion will center on the models presented in Figures 1-3 (13 ; R. Gutell, H. Noller, C. Woese, unpublished) ; 16S rRNA models proposed by the other two groups (80-8 1), are in general agreement with the model shown in Figure 1 . A detailed discussion of specific differences between the three 16S rRNA models may be found elsewhere ( 13). The model for the secondary structure of23S rRNA shown in Figure 2 differs in several significant aspects from those proposed by the other two groups (82-82a), summarized below. As in the case of 16S rRNA, there is better agreement with the Strasbourg-Freiburg model (82a) than with the Berlin-Freiburg model (82).

16S rRNA The secondary-structure model for 16S rRNA presented in Figure 1 (13) contains a number of refinements that have been added since the last published descriptions (10). Significantly, every one of the helices that was considered proven by comparative analysis in the first published version of the structure (7) has survived to the present version, having been tested by some twenty additional complete 16S-like RNA sequences and an abundance of experimental information. Refinements have been made possible mainly by the number and phylogenetic variety of newly completed sequences (Table 1 ). The helices considered to be proven by comparative sequence analysis (13, 79) (two or more phylogenetically independent base-pair changes occurring in a given helix) are indicated in the figures by shading. Few possibilities for additional base pairing remain, and nearly all of the helices depicted in Figure 1 are supported by com­parative analysis. Helices not shown as shaded are either so constant in sequence that base changes have not yet been identified (e.g. 564--570/880-886) or are so variable that strictly analogous structures have not been found in two or more organisms for a given region (e.g. 1 435-1445/1457-1466). Nonshaded helices should therefore be taken as tentative.

The 16S rRNA molecule is subdivided into three major structural domains, and one minor domain, by three sets of long-range base-paired interactions (Figure 1). The 5' domain (residues � 26-557) is defined by the helix 27-37/547-556 ; the central domain (residues � 564--912) by the helix 564--570/880-886; and the 3' major (residues �926-1 39 1) and 3' minor

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

1 30 NOLLER

(residues ,..., 1392-1542) domains by the 926-933/1384-1391 helix. The extent to which these domains, defined here only in terms of secondary structure, correspond to true structural domains, is relevant to our understanding of ribosomal architecture. Fragments of 16S rRNA, cor­responding closely to the three major domains, have actually been isolated in connection with studies on ribosomal protein binding sites (see section on Quaternary Structure). The ability of certain r-proteins to remain bound to these fragments, or even to rebind to RNA fragments from which proteins have been removed (121-124), supports the suggestion that these domains are true structural entities. This is not to imply that there may not be extensive interactions among domains, however.

Within each domain, the structure is organized into series of simple and compound helices, separated by various interior loops and bulges. Many of these are structural types not found in tRNA. Single-bulged nucleotides are found at positions 31 , 55, 94, 397, 746, 1042, 1049, 1227, and 1441. The bulged adenosine at position 1227 is an extremely conserved feature, found in all 16S-like rRNAs thus far sequenced. The bulged nuc1eotides at positions 3 1, 397, and 746 are nearly always present, although their precise positions may shift, and in some cases they are replaced by mismatches or other irregularities. Multiple G · U pairs are almost always found in the 829-840/846-857 helix, although their positions are variable.

Frequent juxtaposition of A and G, most often at the ends of helices, but sometimes at internal positions (e.g. positions 148/174, 321/332, 663/742, 665/741 , 1413/1487, 1417/1483, and 1418/1482) suggested that actual anti­anti A-G pairing may occur in rRNA (7, 78-79). This possibility is sup­ported by several lines of evidence. First, comparative sequence analysis shows that in the archaebacteriuni H alobacterium volcanii, yeast mitochon­dria, and all eukaryotic cytoplasmic 18S rRNAs sequenced to date, the 321/332 A-G pair is replaced by a normal Watson-Crick pair ( 13). In the eukaryotes, an A-G juxtaposition appears instead at positions 3 18/335. Conversely, the two G' U pairs at positions 1425/1475 and 1426/1474 become A-G juxtapositions in Bacillus brevis. Residues involved in the above-mentioned A-G pairs are furthermore generally resistant to single strand-specific reagents, both chemical and enzymatic. Digestion of B. brevis 30S ribosomal subunits with RNase A under mild conditions releases pairs of fragments from the 1410-1430 and 1470-1490 regions, which remain stably associated under gel electrophoresis conditions in spite of no less than 5 A-G juxtapositions and several G • U pairs (25). Surprisingly, the double strand-specific cobra venom RNase cleaves between the 1417/1483 and 1418/1482 A-G pairs in E. coli and Bacillus stearothermophilis 16S rRNAs and in yeast 18S rRNA (126). A-G pairing has been observed in the

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 131

Figure 1 Secondary-structure model for eubacterial 16S rRNA (E. coli) (7, 10, 13, 79). Sequence and numbering are from (23, 58). Helices considered proven by comparative

sequence analysis (two or more pairs of compensating base changes in the helix proper) are

shaded.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

1 32 NOLLER

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 1 33

Figure 2 Secondary-structure model for eubacterial 23S rRNA (E. coli) (78; R. Gutell, H. Noller, C. Woese, unpublished). Helices considered proven (cf Figure 1) are shaded. Sequence and numbering are from (29). (Other half of figure on facing page.)

,

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

134 NOLLER

U 120- U G

A C C C

G U G G

A C

G C 10

C G 1 e C G ·. G G

u 30 U A 1 eCCe IIO-C GC GC CC UG A AU_40 A rii! IJr,ilij,1;l; G A

G C GAU G C C G C G U G G A C U C

C G G IAAA A A \ CAAG

G " / �70 60 50 A • U G

G ' A A • U

100-G G A G G U

C A G

u-eo G

u G

G

Figure 3 Secondary-structure model for eubacterial 5S rRNA (E. coli) (15, 75, 242). Helices considered proven (cf Figure 1) are shaded. Sequence and numbering are from (22).

structure of tRNAPhe, although it involves m;G (127-128). An early objection to anti-anti A-G pairing was that the exocyclic 2-amino group of guanine would not be fully able to hydrogen-bond with water (129). Thus, the A-G pair in tRNAPhe could be rationalized, in that methylation would obviate that theoretical objection. More recently, direct evidence for formation of A-G pairs has come from NMR studies of DNA oligomers (130 ; D. Patel, unpublished) ; duplexes containing A-G pairs appear to have stabilities comparable to those that have G . U pairs at the corresponding positions. Another group of investigators has noted the high proportion of A-G juxtapositions flanking helices in rRNA, and has suggested that anti­syn pairing may be involved ( 125). In conclusion, the existence of bona fide A-G pairs in rRNA seems to be a very real possibility, and they may even be relatively commonplace structural features, along with G' U pairs ; in retrospect, our understanding of base pairing in RNA appears to have been skewed by DNA pairing rules, which are turning out to be a special case.

Somewhat surprising is the absence of long, regular helices in rRNA. Instead, the structure is formed by joining of many short helices, the

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 135

junctions of which tend to create a variety of structural irregularities. The reason for this kind of architecture has been rationalized in two ways (10): Multiple small helices allow a much more complex three-dimensional structure, which could approach that of a globular protein. Another reason may be that long, stable helices could create thermodynamic "traps," pre­venting unfolding of nonproductive intermediates during assembly. Signi­ficantly, no "knots" ( 131 ) are found in the secondary structures of the rRNAs (apart from the 9-13/21-25 vs 1 7-20/915-918 helices), i.e. where the loop contained by a helix is involved in pairing with a sequence outside the helix. This may also be part of a strategy to avoid dead ends in assembly.

23S rRNA Figure 2 shows a currentJllodel for the secondary structure of 23S rRNA (78; R. Gutell, H. Noller, C. Woese, unpublished), again derived principally on the basis of comparative sequence analysis. Over one hundred individual helices are distributed among six domains, defined, as in 16S rRNA, by long-range base-paired interactions. These are referred to as domains I (residues 16-524), II (579-1261), III (1295-1645), IV (1648-2009), V (2043-2625) and VI (263�2882). Structural organization within domains is analogous to that of 16S rRNA ; short helices are connected by interior loops and bulges. Again, evidence for A-G pairing is found (78). A notable example of the latter is the tandem pairs 1039/ 1 1 16 and 1040/1 1 1 5, whi(;h are replaced by standard G-C and A-U pairs in B. Stearo­thermophilus and Zea mays chloroplast, respectively. Many single-base bulges are found, at about the same frequency as for 16S rRNA, as well as some double-base bulges (e.g. residues 1 321 and 1 322).

5S rRNA Numerous groups have presented detailed discussions of 5S rRNA secondary structure recently (15-20), and the reader is referred to these for a more complete treatment. The first models derived on the basis of comparative sequence analysis for eubacterial (75) and eukaryotic (76) 5S rRNA many years ago are currently accepted as generally correct, although some new elements have been added to the basic models by various authors. Figure 3 shows a model that incorporates most of the currently accepted refinements to the original Fox-Woese proposal for E. coli 5S rRNA. Its five helices exemplify the kinds oflocal features seen in the large rRNAs. There are examples of single- and double-base bulges, an A-G pair (in this case unsubstantiated by comparative evidence) and "helices connected by internal loops. The small size of 5S rRNA makes it amenable to computer analysis, and one such method (85) (which, significantly, incorporates a comparative algorithm) has produced a credible secondary structure for this molecule.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

136 NOLLER

Phylogenetic Comparison

Comparative. sequence analysis has allowed derivation of specific secondary-structure models for each of the rRNAs whose complete sequence is known. Again, models for the secondary structures of the various mitochondrial chloroplast, and eukaryotic cytoplasmic species have been presented by several groups. Here, the discussion centers on structures based on ( 13). Figure 4 shows some representative examples of 16S-like secondary structures ( 1 3). There is striking overall similarity between structures of 16S-like RNAs from the three major evolutionary lines, although the eukaryotic example shows considerable variation from the other two types. The central framework of the structure, including the long-range interactions that define the domains, is conserved even in mammalian mitochondria rRNAs, where 40% of the structure (relative to E. coli) is absent (37-40). Also readily identifiable in all cases are the parts of the structure containing the most highly conserved seque�ces (see also below). Similar conclusions can be drawn from comparison of 23S rRNA structures, which also share the same fundamental architecture.

Variation in size among the different 16S-like and 23S-like rRNAs is accounted for by insertions and deletions, which tend to be constrained to a few specific regions of these molecules ( 13). Thus, deletions in the 16S-like rRNAs tend to occur around positions 80, 200, 400-500, 590-650, 840, and 1450, which can be characterized as the most phylogenetically variable regions of the sequence in those organisms that retain these regions ( 1 3). Insertions in mitochondrial 16S-like rRNAs tend to occur around positions 1 140 and 1450 while insertions in the eukaryotic cytoplasmic rRNAs are in the 200, 395, 580, 840, 1 140, and 1450 regions (60-61, 1 32, 133); most of the increased size of the eukaryotic 18S rRNAs can be accounted for by the insertion of � 200 nucleotides around position 580. Similarly, the greater size of eukaryotic cytoplasmic 23S-like rRNAs is mainly due to insertions around positions 270, 550, 650, 1400, 1730, 2200, and 2800 (64-66, 1 34). The inserted sequences in both cases tend to be highly variable, and are much less conserved phylogenetically than the sequences in the parts of the molecule for which there is corresponding structure in the prokaryotic examples.

Experimental Tests.

SINGLE OR DOUBLE STRAND-SPECIFIC PROBES It was mentioned earlier that large fragments of the 16S and 23S rRNAs obtained by partial nuclease digestion correspond well to domains (or subdomains) generated by secondary structure, providing experimental support for the overall structural organization implicit in their secondary-structure models. At a

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 137

Figure 4 Phylogenetic comparison o f secondary structures of 16S-like rRNAs from (a) eubacteria (E. coli), (b) archaebacteria (H. volcanii), (c) eukaryote (S. cerevisiae), and (d) a "minimal" structure showing those features common to all sequenced 16S-like rRNAs, including mitochondrial types. Positions of insertion of additional nucleotides in eukaryotic rRNA with respect to bacterial rRNA are indicated by 25 and 226 in (c).

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

138 NOLLER

detailed level, each nucleotide can, in principle, be tested experimentally by the chemical and enzymatic probes listed above. In general, attack by a single strand-specific probe is taken as evidence for the residue at that position existing in single-stranded conformation (or conversely, for a double-strand probe). In some cases, resistance to attack has been well documented ; here, of course, it should be kept in mind that tertiary or quaternary structure as well as base pairing may account for protection. About 85%-90% of these kinds of probe data (summarized in 10, 1 3, 79) are in agreement with the secondary-structure models described here ; this level of accord is probably an underestimate, because the consistency of results between any two experimental approaches is also usually only on the order of 85%-90%. The lack of complete agreement between different experi­ments can possibly be attributed to whether the RNA molecule is probed as part of the ribosome or in the free state, differences in io'nic strength, magnesium ion concentration, methods of RNA isolation, and, most importantly, whether (and how) the RNA was ascertained to be in its "native" conformation. In the case of enzymatic probing, the problem of secondary cuts can be a significant one, as discussed above. Finally, correct identification of modified residues in a large RNA molecule is 9ften technically difficult, and the reliability of such data are often difficult to assess from literature accounts.

Oligonucleotides can, in principle, be used as probes of single-stranded regions of rRNA. An important question to be asked is, how much of the rRNA molecule exists as open, single-stranded structure, capable of base pairing with such oligomers? Detailed studies with chemical probes suggest that 16S rRNA, even in naked form is surprisingly highly structured, and has very few regions, if any, containing many consecutive unstructured nucleotides. In one type of approach, an oligodeoxynucleotide (AAGGAGGT) complementary to the 3'-terminal region of 16S rRNA (residues 1 534-- 1 541) was synthesized and used as a probe of the accessi­bility of its cognate sequence in naked RNA, 30S subunits and 70S ribosomes ( 1 15). It was found that naked 16S rRNA or 30S ribosomes in the salt-depleted "inactive" form do not bind the oligomer, whereas active 30S subunits are capable of binding it ; this finding in itself was unanticipated, since the RNA of the inactive form of 30S subunits is generally much more reactive toward kethoxal (1 35). Furthermore, protein S21 is required for oligomer binding. These results are presumably indicative of the exposure of residues 1534-- 1 541 to solvent, and dramatically demonstrate the care that must be taken in preparing ribosomes or rRNA for such probing experiments. Most significantly, the effect of a specific !"ibosomal protein on rRNA topography is clear, and the likely conclusion is that S21 induces a conformational change in the 30S subunit that in some way exposes the 3' -terminal sequence.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 139

In another approach ( 1 13-1 14), a random mixture of DNA hexamers is hybridized with 16S rRNA in 30S ribosomal subunits and the accessible positions are. then determined by cleavage with RNase H. In these experiments, nucleotides around positions 8-:-13, 996-998, 1408-1410, 1495, 1500-1 506, and 1 531-1 532 showed accessibility to the oligomers, con­sistent with the secondary-structure model. However, nucleotides around positions 80, 1044-1046, 1423, and 1484-1485 also bound oligomers; binding of a hexanucleotide to any of these regions requires disruption of helices considered to be proven by comparative sequence analysis. To \ account for this discrepancy, it can be argued that the regions of 16S rRNA in question can exist, at least transiently, in unpaired conformation while in a 305 ribosomal subunit. It must remain a possibility that the oligomers themselves inay help to convert the rRNA to the open form, by virtue of their ability to bind the single-stranded form.

The latter experiments are examples of a few specific instances where phylogenetically established helices are reproducibly found to be sus­ceptible to single strand-specific probes under certain conditions. These include the 9- 13/21-25 and 17-20/915-918 helices and the 1050-1067/1 189-1204 and 1410-1490 regions of the 16S-like rRNAs (72, 79, 97, 133, 135), and the 2063-2103 region of 23S-like rRNAs (78, 1 34, 136). Whether these results are mere reflections of the inherent instability of these structures under experimental conditions or clues to structural dynamics of rRNA is an open question at the present time.

DIRECT EVIDENCE FOR INTRAMOLECULAR INTERACTIONS Use of various psoralen derivatives has provided confirmation of base-paired features of rRNA (summarized in Table 2) and, in addition, has suggested RNA-RNA interactions that are not readily accounted for by secondary structure. Lack of agreement between early secondary-structure models ( 137) based on psoralen cross-linking (101-102) with those derived from comparative evidence can possibly be ascribed to a combination of factors. Some cross­links were obtained using naked rRNA at very low ionic strength (101), where the danger of disrupting conformation is significant. Identification of cross-linked regions was in some cases incorrect because of difficulties in determining the polarity of the RNA chain in electron micrographs; this problem has now been overcome (138). Another factor is the possibility that psoralen may cross-link nucleotides that are involved in tertiary, as well as secondary, interactions (see Tertiary Structure section).

Nitrogen mustards ( 139) and irradiation with ultraviolet light (140) have also been used to produce RNA-RNA cross-links. The former react readily with nucleophiles, primarily 7N of guanine, in the case of nucleic acids, and this reaction proceeds with base-paired as well as unpaired guanines. Cross-links induced by UV irradiation usually involve cycloaddition

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

140 NOLLER

Table 2 Experimental evidence for 168 rRNA secondary structure from isolation of non covalently associated or cross-linked pairs of RNA fragments

Noncovalent Psoralen UV Helix association' cross-1inkedb cross-linked"

17-20/915-918 + 39---42/394-403 + 73-82/87-97 + 122-128/233/239 + 136-142/221-227 + 240-245/281-286 + } 247-249/279-277 + + 252-259/267-274 + 28g...:.295/302-31 1 + 437---440/494-497 + 442---446/488---492 + 455---462/470---477 + + 564-570/880-886 + 576-580/761-765 + 584--587/754--757 + } + 588-595/644-651 + + } + 597-606/633--641 + 612-617/623--629 + 655-662/743-751 + 666-672/734--740 + 821-828/872-879 + + 829-840/846-857 + 926-933/1384--1391 + + 938-943/1340-1345 + 984-990/1215-1221 + (+)d 997-1003/1037-1044 + 1006-1012/1017-1023 + + 1046-1053/1205-1211 + (+) 1055-1060/1 197-1202 + 1063-1067/1189-1193 + 1068-1071/1104--1107 + 1113-�117/1183-1187 + + 1239-1247/1290-1298 + + 1308-13 14/1323-1329 + 1350-1356/1366-1372 + + + 1409-1430/1470-1491 + 1435-1445/1457-1466 +

• Sources: (25,108,141). b Sources: (101-103, lOS, 155). "Source: (140). d In cases where there is some uncertainty in the identification of cross-linked nuc1eotides, the

probable assignment is given in parentheses.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 141

between the 5 and 6 positions of stacked pyrimidines. Thus, cross-linking of residues that are nonadjacent in the RNA primary structure cannot occur if they are both involved in the same helix. This is consistent with the finding that UV-induced cross-links occur in regions of the structure containing internal loops (Tables 2 and 3) where bases on opposite chains may be able to stack on one another ; results of this kind may shed light on the details of base stacking in such areas of structural irregularity.

Another method that has been used to identify base-paired regions employs two-dimensional gel electrophoresis of partial nuclease digests of non-cross-linked rRNA (25, 108, 141). Pairs of fragments that remain associated in the nondenaturing first dimension are dissociated and resolved in the second dimension under denaturing conditions. This

Table 3 Experimental evidence for 23S rRNA secondary structure from RNA-RNA cross-linking and electron microscopy

Helix

16-25/515-524 281-293/347-359 295-297/341-343 588-601/656-669 812-817/1 190-1 195 1013-1019/1 143-1 149 868-870/907-909 1206-1209/1 237-1240 1 213-1222/1 227-1 236 1405-1415/1 587-1597 1435-1437/1 555-1557 1481-1492/1498-1508 1656-1663/1997-2004 1710-1750 region 1841-1846/1894-1899 2043-2050/261 8-2625 2063-2070/2441-2447 2630-2637/278 1-2788 2836-2843/2874-2882 2851-2856/2861-2866

a Source : (260). b Source: (104). o Source : (139). d Source : (158).

Cross-linkage with

Electron microscopy' Psoralenb

Nitrogen mustardC

+

+ +

+

+

+ + +

+

} + +

+

+

+

+

+

"( + )' ( + )

, In cases where there is some uncertainty in the identification of cross-linked nuc!eotides, the probable assignment is given in parentheses.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

1 42 NOLLER

method has the advantage that no chemical modification is employed, but anomalous reassociation of fragments following nuclease digestion may be a potential hazard. Evidence for many helices has emerged from these studies (summarized in Table 2), as well as some interactions that do not correspond to features ofthe secondary-structure models (discussed below).

There is striking agreement between results obtained by electron microscopy of partially unfolded rRNAs and the proposed secondary structures. Large loops seen in such preparations of 23S rRNA correspond closely to long-range base-paired interactions in the secondary-structure model (Table 3 ; Figure 2 ; 260). The critical factor required for reproducible observation of these features appears to be the presence· of adequate magnesium ion concentrations during preparation. The fact that the long­range interactions are detected at all suggests that they are more stable than many of the other interactions under these conditions. Inspection of the helices involved shows that they are, almost without exception, predicted to be quite stable from standard free-energy calculations. This may have important implications for ribosome assembly, possibly ensuring that each domain is maintained as a separate structural entity during early stages of folding.

Results of the above studies, in which direct experimental evidence has been obtained for intramolecular association of specific regions of 16S and 23S rRNA, are summarized in Tables 2 and 3. Data are much more abundant for 16S rRNA, as is the case for the structure-specific probe experiments. There is clear evidence for association or close juxtaposition of sequences making up 37 individual helices in 16S rRNA. The results listed in Tables 2 and 3 include only those that correspond to helices whose existence is compatible with comparative sequence analysis and represent most of the available data. Base pairing consistent with the comparative sequence criteria, outlined above, has so far not been found for the remaining examples of experimentally observed intramolecular associ­ation. This can be explained in one of three ways : (a) phylogenetically consistent helices exist for those interactions, but have not yet been found ; (b) experimental conditions have caused rearrangement of rRNA conforma­tion ; or (c) the interactions are tertiary in nature. None of these possibilities can yet be rigorously ruled out. Experimental conditions in some cases involve very low ionic strength (101) and/or chelation of magnesium (108), and even use of 2-M urea (141), all of which are known to perturb rRNA conformation. The noncovalent association experiments involve cleaving the RNA chain with nucleases, and the possibility of subsequent structural rearrangement must be considered. Results from electron microscopy suffer from low resolution, usually on the order of ± 20 nucleotides at best, and determination of the polarity of the RNA chain has often been a problem.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 1 43

Finally, many of these studies are performed on naked rRNA, and while it is becoming clear that its structure must be very much like that found in the ribosomes, significant differences may exist ; in any case, such studies depend crucially on the methods of isolation and treatment of the RNA before the actual experiment. Most intriguing, however, is the possibility that some of these interactions may be clues to tertiary folding of the rRNA, to be discussed below (see Tertiary Structure section).

Recently, detailed investigations of rRNA conformation using high­resolution NMR methods have directly confirmed some secondary­structure features, and have established a basis for future studies on rRNA such as solution dynamics and RNA-protein interaction. Analysis of 5S rRNA has directly identified features of the 1-10/ 1 1 0-1 1 9 and 79-86/92-97 helices, using the NOE to assign imino proton resonances to nucleotides in adjacent base pairs ( 1 1 6-1 1 8). Similarly, the 1 506-15 1 5/ 1520-1529 helix of 1 6S rRNA has been studied using the 3'-terminal colicin fragment ( 1 1 9). Unfortunately, this approach is limited to the study of relatively small rRNA molecules or fragments, up to about the size of 5S rRNA.

TERTIARY STRUCTURE

Most important for understanding the biological role of rRNA will be a description of its three-dimensional structure with its associated proteins. We are clearly far from the realization of this goal, which, as for other large biological structures, will likely require X-ray crystallographic analysis. Nevertheless, it is possible that a general scheme for the three-dimensional folding of the large rRNAs can be achieved in the absence of any crystallographic information. Relevant data are emerging from two general approaches. In the first approach, the electron microscope (e.m.) model for the structure of the ribosome and its subunits ( 142-143) is taken as a starting point, and attempts are made to assign specific features of rRNA to positions in the e.m. model. In tbe second kind of approach, the secondary­structure models for the rRNAs are taken as a starting point, and attempts are made to fold them into three dimensions. The two approaches are formally independent, and so there is the possibility to test or confirm the predictions of one approach using the other.

In order to assign a site in rRNA to a location in the e.m. model, there must be some recognizable feature in the electron micrograph that can be related, directly or indirectly, to the site in the RNA. Many individual ribosomal proteins have been localized by immuno-electron microscopy ( 142-143) and by neutron diffraction studies ( 144) ; since some of these have known binding sites on the rRNA, it can be inferred that the latter must be located reasonably near the positions of their cognate proteins. More

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

144 NOLLER

recently, antibodies have been raised to methylated nucleotides located at unique positions in the l6S rRNA chain, and to haptens coupled to the 5' or 3' termini, permitting direct immuno-e.m. localization of these sites (145-1 52). By these general approaches, the sites in 16S rRNA identified with proteins S7 (U 1 240 and C1378 or GI379), S8 (residues ", 590-650), S 15 (residues '" 650-750) and S20 (residues '" 250-280) can be located according to the positions of the corresponding proteins. Similarly, �G527 (146), m�Al5 u i and m�A15 19 (145),.and the 5 ' (1 50) and 3 ' ( 147-149) �ermini of 16S rRNA have been placed directly. Furthermore, the protems found associated with the 3' major domain (124) (residues '" 930-1390) are all located in the "head" of the e.m. model. Less direct assignments for proteins S4, S6, and S18 can also be made (10). These placements are summarized in Figure 5.

In 23S rRNA, similar references can be made although there is presently much less information available pertaining to its structure in the large subunits. Its central lobe is proximal to 5S rRNA, from studies using antibodies to end-labeled 5S rRNA ( 151-1 52), and from the location of its

�-------------------------- S 20------�1 ,------------ s 8 I I �-------------- S I 5--------��----�

S7. S9.S I3.SI9{

6 I I I �2A I I I L----3I

en�_1 11 1

L-______________ 5 en

�

L--_______ �: I Figure 5 Placement of specific features of 16S rRNA structure in the electron microscopy model (142) of the 30S ribosomal subunit. Positions of ribosomal proteins ( 142-144) and sites in the RNA (145-1 52) are indicated.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 145

cognate proteins, U 8 and L25 ( 15 1). The peptidyl transferase center (see below) has also been placed near the central lobe, from the positions of proteins affinity-labeled by various tRNA derivatives, and from the site of attachment of the antibiotic analog iodoamphenicol (1 53). The 2100-2200 region of 23S rRNA can be placed on the "left-hand" lobe of the 50S subunit, because of its association with protein U (142-143). Finally, the 5' and 3' termini have been placed on the lower rear of the particle, by end­labeling studies ( 154).

In the second approach, folding the molecule into three dimensions, is based on information establishing proximity between regions of the RNA that are non proximal in the secondary-structure model. As discussed above, some of these data may reflect tertiary contacts. Relevant findings include cross-links, induced by psoralen, nitrogen mustard, and UV irradiation, and noncovalent association of RNA fragments for which no helices supported by comparative analysis are known.

From studies employing psoralen, cross-links were found between the following regions of 16S rRNA (to be taken approximately, as the precise nucleotides invoived have not yet been determined) : 360/1330, 620/1420, and 960/1510 ( 105) (all determined by sequencing cross-linked fragments) ; and 1/680, 10/1 1 80, 490/980, 610/1320, and 930/1540 (1 55) (all determined by electron microscopy). All of these results imply interaction between domains, and several are found between the 5' and 3' major domains. Evidence for interaction between the latter two domains comes also from studies in which large fragments from these two regions co-migrate under conditions of gel electrophoresis in the absence of proteins (108, 141 , 1 56-1 57). Interaction between the 5' and 3'-major domains is also implied by the location ofS20, which binds to the 5' domain (8), in the "head" region ofthe e.m. model, surrounded by proteins that bind to the 3' major domain (143).

Cross-linked sites in 23S rRNA that represent potential tertiary contacts are 570/2030, 740/2610, and 1 780/2570 ( 158) (from uv irradiation) and 760/1570 (139) (from nitrogen mustard). These results imply contact between domains II and III, II and V, IV and V.

Singlet-singlet energy transfer has been used to measure inter- and intramolecular distances between the ends of the rRNAs ( 159). Distances between fluorophores attached to the 3' ends of 16S rRNA and 5S rRNA or 23S rRNA were estimated to be about 55 and 7 1 A, respectively. The corresponding distance between 5S rRNA and 23S rRNA was too large to be measured accurately with the available probes but was estimated to be greater than 65 A.

Even if all the results described here are taken as evidence for tertiary structure, far more information of this kind will be required to construct low resolution three-dimensional models for rRNA. Nevertheless, the apparent spatial proximity of sites that are distant from each other in the

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

146 NOLLER

primary and secondary structure, for example the 960 and 1 500 regions of 16S rRNA (both of which are implicated in ribosomal function :; see Structure and Function section), is valuable information that will be of use in formulating models of both structure and function.

Cross-linking of G41 and G72 in 5S rRNA (1 60) has stimulated three specific proposals for "tertiary" interactions in this molecule (16, 160-161). Two of thse involve loop-loop Watson-Crick base pairing, of residues 37-40/73-76 in one model (160), and 41-44/74-77 in another ( 161). Besides the high susceptibility of both V40 and G41 to single strand-specific probes (162-163), even in intact ribosomes ( 164-165), which prevents their involvement in any permanent base-paired interactions, comparative sequence analysis does not support proposed pairings. Several phylogeneti­cally independent disproofs can be found for both models. The third model proposes parallel base pairing of 38-40 with 76-78 (16). In this case, the sequences &re more highly conserved, but changes in Photobacterium and Aspergillus, for example, produce V-V and A-C mismatches, respectively. Thus, the available evidence, both experimental and phylogenetic, argues against these proposals. Furthermore, it remains to be ruled out that the cross-link itself, obtained in comparatively low yield ( 160), is not due to the B conformer of 5S rRNA, in which G41 and G72 are believed to be brought into juxtaposition (1 66).

QUATERNARY STRUCTURE r-Prote in-r RNA Interact io ns A complete discussion of the subject of interaction of ribosomal proteins with ribosomal RNA is beyond the scope of this article. The reader is refetred to other specific review articles on the subject (8). Here, the questions to be addressed are : What are the specific recognition signals embodied in the structure of rRNA by which ribosomal proteins discern their correct binding sites? And, do the proteins significantly alter rRNA conformation, either locally or globally?

RECOGNITION SIGNALS The answer to the first question depends on a detailed description of ribosomal protein-binding sites. Information bear­ing on this has been obtained principally by methods that include (a) the classic "bind and chew" approach (in which a protein is bound to rRNA, extraneous rRNA removed by RNase treatment, and the remaining bound RNA characterized), (b) chemical or VV-induced cross-linking, (c) pro­tection of the rRNA from chemical modification by the bound protein, (d) modification/selection experiments (in which the RNA is partially chemically modified, and the fraction of rRNA molecules retaining their

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 147

ability to bind the protein are analyzed to see which sites are not modified), and (e) electron microscopy of protein-rRNA complexes.

Most informative are those protein binding sites that have been localized to a relatively concise region of the rRNA. Some ribosomal proteins yield very large RNA fragments in nuclease protection experiments ; the "binding-site fragment" for protein S4, for example, contains about 500 nucleotides, about seven times the mass of the protein itself. It is unlikely that the protein recognizes more than a fraction of this fragment, but protein-RNA interactions may stabilize the structure and thereby enhance the RNA-RNA interactions. Unfortunately, it is often difficult to discern which features of these large fragments are the actual protein contact sites. Proteins that have been found to yield RNA binding-site fragments of more manageable size (mass on the order of that of the protein, or smaller) include S8, S15, Ll , L l 1 , Ll7, Ll8, and L25. Binding sites for these proteins have been localized approximately within 16S rRNA for S8 (residues 585-610, 630-654) and S15 (residues 655-675, 730-755) ( 122, 167), 23S rRNA for Ll (residues 2090-2200) ( 168), and L l 1 (residues 1050-1 1 10) (169) and 5S rRNA for Ll8 (residues 15-35, 45-69) (1 70-172) and L25 (residues 70-1 10) ( 1 1 , 1 72-73). Most of these fragments contain on the order of 40-60 nucleotides, and probably far fewer nucleotides actually make contact with protein in each case. It is important to keep in mind that these proteins may well bind RNA sites outside their respective "binding fragments," but that the other contact sites could escape detection if the binding constants are low, or if the RNA moiety is unusualy unstable to nuclease treatment.

It is reasonable to ask whether there are any structural features common to these various protein-binding sites. Interestingly, the majority would appear to contain two helices connected by an internal loop. Because of the abundance of such structures in rRNA, it is difficult to judge the significance of this. Binding-site fragments for S 15, Ll8, and Ll7 all contain single­bulged adenosine residues, and the L18 site appears to contain, in addition, a pair of bulged adenosines (Figures 1-3). The S8 and Ll binding sites tend to be unusually rich in G · U pairs (6 and 4, respectively). The best-studied protein-binding sites are those in 5S rRNA, largely due to the ease of preparation of this experimental system. For both L l 8 and L25, there is good evidence that the proteins bind to double-helical structures. In both cases helices are protected from attack by cobra venom RNase by the bound proteins (172), and L18 protects several base-paired guanines from reaction with dimethyl sulfate (171 ). Furthermore, spectroscopic studies with Ll8 and L25 (94, 174) and NMR studies with L25 (1 1 8) indicate that the RNA helices are not disrupted when the proteins bind. Single-stranded residues flanking the binding-site helices have also been shown to be protected from chemical and enzymatic probes by the proteins in the cases

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

148 NOLLER

of U 8 and SIS ( 122, 1 70--172). Protein S8 has been cross-linked to single­stranded nucleotides (approximate positions 593-597, 629-633, and 651-654) on either side of, and between, its two binding-site helices (175). These results seem to imply that the ribosomal proteins recognize both helical and nonhelical elements. The possible role of irregularities such as bulged nucleotides, multiple G · U pairs and A · G pairs is at present an open question. The existence of bulged nucleotides in protein-binding sites from systems other than ribosomal suggests at least some connection between them ; whether the bulge is a recognition signal for protein binding or is stabilized by the protein for some other purpose is not clear (171). What is clear is that the bulged nucleotides are often highly conserved phylogeneti­cally ( 17 1), and in at least two well-documented cases (SIS and U 8) are located well within protein-binding sites. It is not unreasonable to conclude that much of the architecture of the ribosomal RNAs has evolved to accommodate interactions with ribosomal proteins and, very likely, translational factors. The extent to which this is the case will not be clear without a much more complete description of the protein-binding sites in rRNA.

Studies on the translational regulation of ribosomal protein biosynthesis indicate that the structures of their mRNAs have similarly evolved to accommodate RNA-protein interactions. Certain ribosomal proteins that are known to bind directly to rRNA act as translational repressors by binding to polycistronic mRNAs, usually encoding several r-proteins, and in some cases other proteins crucial for transcriptional and translational processes ( 176). The regulatory sites on these mRNAs have been proposed to involve structures that mimic the rRNA binding sites for the regulatory r-proteins (177-180).

PROTEIN-INDUCED CONFORMATIONAL CHANGES Is the conformation of rRNA significantly affected by interaction with ribosomal proteins? There is controversy currently concerning the overall size and shape of naked 16S rRNA compared with 30S ribosomal subunits. There are studies using electron microscopy ( 181 ) and hydrodynamic methods (1 82) that show the naked 16S rRNA as having essentially the same size and shape as the 30S subunit. Other studies, however, also based on electron microscopy (183) and hydrodynamic approaches (1 84), have reported very different shapes and dimensions for the two cases, and describe the naked rRNA as unfolded or somewhat amorphous compared with the intact subunit. Here, the underlying question is whether the rRNA itself contains the requisite information for folding, which would merely be stabilized by the proteins, or whether some of the large-scale structuring of the molecule is completely dependent on the proteins. Thermodynamic studies on the process of in

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 149

vitro assembly of 30S (185) and 50S ( 186) subunits show that a conforma­tional rearrangement necessary for correct assembly [the "RI --+ RI*" conversion] depends on prior binding of several ribosomal proteins to the rRNA in both cases. The sedimentation constant of the 30S assembly intermediate particle changes from 21-22S to 25-26S as a result of the RI --+ RI* conversion (187), indicating a substantial change in global con­formation (i.e. most likely a compaction of the particle). Little is known, however, concerning the specific structural changes that occur during these events.

Studies on specific rRNA-protein interactions show that little change in rRNA conformation results in most cases. An exception is the binding of L 1 8 to 5S rRNA, which causes an increase of about 20% in the magnitude of the 263-nm circular dichroism band of the RNA (99, 1 74). Since this parameter is usually taken as a direct measure of he Ii city, the result has been interpreted as an increase in base pairing in 5S rRNA induced by the binding of Ll8 (94) ; alternatively, it has been interpreted as a change in the structural regularity of existing double-stranded segments of the molecule ( 174). It is interesting to consider that, if Ll8 were to stabilize the irregular extensions to its two binding-site helices beyond their respective bulged nucleotides (i.e. stabilize the pairing of 1 6-17/67-68 and 28-30/54-56) (Figure 3), this would amount to an increase in helicity of approximately the observed magnitude . .

High-resolution NMR studies on 5S rRNA, its L25 binding fragment (residues 1-1 1 , 71-120), and their respective interactions with L25 show effects of protein binding on the RNA structure ( 1 16-1 18). All of the assignable perturbations involve the 79-85/91-97 helix, and include the disappearance of the NOE linking base pairs Us2/A94 with GS3/C93• In addition, several new unidentified low field resonances appear ; a tantaliz­ing possibility is that some of these may represent hydrogen bonds involved in RNA-protein contacts. Clearly, this approach is a potentially powerful tool for elucidation of specific details of protein-RNA interaction.

In conclusion, perturbation of rRNA structure by r-proteins is poorly understood. A reasonable scenario for ribosome assembly is that proteins capable of binding independently to rRNA recognize, for the most part, preexisting structural features (most likely a combination of primary, secondary, and tertiary) or in some cases may trap conformers that exist only transiently in the absence of bound proteins ; this may then permit the binding of proteins that make protein-protein contacts, or whose RNA binding sites are stabilized by attachment of the primary binding proteins. Early steps may involve stabilization of structure within each domain, followed by domain-domain interactions. The lack of solid information pertinent to the structure of rRNA during assembly is a measure of how

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

1 50 NOLLER

little we understand of this process, in spite of the considerable information available regarding the assembly of ribosomal proteins.

r RNA -r RNA Intera ctions

Thus far, no convincing evidence for direct interactions between the ribosomal RNAs has come to light. Suggestive complementarities between 5S rRNA and the large rRNAs in E. coli ( 188-1 89) do not withstand comparative analysis. Nor have phylogenetically consistent base-pairing schemes for 16S-23S rRNA interactions been identified, although direct cross-linking between the large rRNAs in 70S ribosomes has been reported (1 30).

STRUCTURE AND FUNCTION

Early studies tended to emphasize ribosomal proteins as the components likely to be directly involved in ribosomal function. This can probably be attributed to a variety of circumstances. Enzymes are made of protein, so it was anticipated that ribosomal proteins would be responsible for the various translational functions. Ribosomal mutants, most notably those carrying antibiotic resistance alleles, were found to be due to altered proteins at the outset ( 190) ; this was probably because most genetic studies of this kind were carried out on E. coli and other eubacteria, which carry multiple copies of the rRNA genes in their genomes, masking potential rRNA-related phenotypes. In vitro reconstitution studies again drew attention to the function of proteins ( 191 ), although the requirement for 5S rRNA became apparent (186, 1 92). Finally, the greater inherent chemical reactivity of proteins tended to deemphasize rRNA in affinity labeling and chemical modification studies.

It can now be said that there is at least a comparable body of evidence to support the reverse paradigm, that the essence of ribosomal function is embodied in its RNA, and that ribosomal proteins serve to enhance its function. A first consideration is the evolutionary argument, the chicken­and-egg problem, that the original ribosome cannot have used proteins for any crucial purpose, since the ribosome had yet to make them. It is also more plausible to imagine a ribosome originally constructed from RNA, to which proteins have been added during the course of evolution, than the converse. Furthermore, both in vitro and in vivo studies have demonstrated that ribosomes lacking individual ribosomal proteins are functionally active (193-194). Finally, the concept of a "functional nucleic acid," one that is capable of directing the making and breaking of covalent bonds" for example, is no longer confined to the realm of speculation, since the discovery of a self-splicing RNA (195). The chemical nature of RNA would appear to have been greatly underestimated.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

rRNA Gene tics

STRUCTURE OF RIBOSOMAL RNA 1 5 1

Genetic studies have now amply documented the role of rRNA i n several functional contexts. Methylation of rRNA, or the lack of it, has been shown to affect resistance to the antibiotics kasugamycin ( 196), erythromycin (197), viomycin (199), and thiostrepton (169). Mitochondria usually have only a single set of rRNA genes, in contrast to most eubacteria. Strikingly, antibiotic resistance in mitochondrial ribosomes is almost always attrib­utable to mutations in the rRNA. Thus, resistance to chloramphenicol (201-204), ·erythromycin (205), and paromomycin (43) have been found to be due to point mutations in rRNA genes in yeast and mammalian mitochondria. Cold-sensitive mutations mapping in mitochondrial rRNA genes have also been found (206). More recently, erythrQ�ycin resistance in E. coli has also been attributed to a mutation in a plasmid-encoded rRNA gene (207). Use of site-directed mutagenesis of cloned E. coli rRNA genes has produced a number of mutants in 1 6S and 23S rRNA (208-209). Surprisingly, most of these mutations affect the growth rate of the organism significantly. A most interesting finding is an ochre suppressor in yeast mitochondria that maps in or near its 1 5S rRNA gene (21 0). In view of these results, it can be said that genetic evidence alone provides a compelling case for a direct role for rRNA in translation.

Dramatic examples of the effect of small alterations of the covalent structure of rRNA are provided by colicin E3 and IX-sarcin. These toxic proteins completely inactivate protein biosynthesis by cleaving single phosphodiester bonds in rRNA ; the A1439-G 1494 linkage of 16S rRNA by colicin E3 (21 1-212), and the G266cA2662 linkage of 23S-like rRNAs by IX-sarcin. Both scissions occur in universally conserved sequences (1 3).

Specific S truc ture-Func tion Correla tions

In the past few years, evidence from a wide variety of experimental approaches has emerged to create a first impression of many of the functional properties of ribosomal RNA, and in many cases specific sites in the RNA have been related to specific functions. It is useful to review these findings briefly. and to sense the extent to which the structures of the ribosomal RNAs are beginning to take on biological meaning.

mRNA INTERACTIONS AND INITIATION That mRNA initiation sites are selected by direct interaction with 16S rRNA as proposed by Shine & Dalgarno (2i5) is now well documented, and has been reviewed in detail by Steitz (216) and Gold et al ( 198). Comparative analysis of a large number of mRNA sequences shows a varying degree of complementarity between mRNA at a position about 10 nucleotides distal to the 5' side of the initiator codon and the conserved CCUCC sequence at position 1 535-1539 of

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

1 52 NOLLER

16S rRNA. This has now been amply substantiated by a number of elegant biochemical and genetic studies (21 7-219). It is likely that such a mech­anism operates in eubacteria and archaebacteria, but not in eukaryotic, cytoplasmic ribosomes nor in most mitochondria.

The initiation factor IF3, involved in promoting dissociation of vacant 70S ribosomes and binding mRNA during initiation of translation (:220), has long been suspected of binding to 16S rRNA. Cross-linking of IF3 to 16S rRNA has been achieved by photolysis using near-UV irradiation (221). At least one site of cross-linking is reported to be in the 3'-terminal region ; removal of the 3'-terminal colicin fragment decreases the binding of IF3 to 30S subunits. Recently, protection of specific residues in the 3'-terminal region of 16S rRNA has been demonstrated in studies of the interaction of IF3 with the 49 nucleotide colicin fragment (222). The 3'-terminal region is also implicated in subunit association (see below), consistent with the possibility that IF3 exerts its "dissociation factor" activity by binding to a region of 16S rRNA that contributes significantly to the strength of the 30S-50S interaction.

Kasugamycin inhibits binding of the initiator tRNA to ribosomes, and as such appears to be a specific inhibitor of translational initiation (223). Resistance to this antibiotic has been shown to be conferred by nonmethy­lation of the two adenosines (normally N-6-dimethylated) at positions 15 18 and 1 5 19 of 16S rRNA ( 196). Thus, by several independent lines of evidence, the 3'-terminal domain of 1 6S rRNA appears to be directly implicated in the process of translational initiation.

In an affinity labeling study in which a derivatized trinucleotide mRNA analog was used to probe the rRNA environment around the presumed mRNA binding site, residues 462 and 474 of 16S rRNA were shown to be Hi.beled (224). The possible generality ofthis result is diminished by the clear absence of the region encompassing residues 455-477 in 16S-like rRNAs from chloroplasts (31-34), in at least one archaebacterium (H. volcanii) (28), in mitochondria (37-44), and in eukaryotic cytoplasmic ribosomes (59--62) ; furthermore, the sequence in this region is phylogenetically variable among eubacteria (13). It may be relevant that the 455-477 region has been found to be a prominent site for: psoralen intercalation (103), in that the probe used in the above studies contained a phenylgyloxal moiety, itself a potential intercalator. Several groups have used other derivatized oHgo­nucleotides, as well as poly(U) and poly-(4-thio-U) as affinity probes of the mRNA binding site ; these have all been shown to react to some extent with 16S rRNA, but the sites of reaction have not yet been found (225-227).

ASSOCIATION OF RIBOSOMAL SUBUNITS When 30S and 50S ribosomal subunits associate to form 70S monosomes, a number of sites that are accessible to single and double strand-specific probes (both chemical and

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 153

enzymatic) become shielded (228-230). The protected regions of 16S rRNA are mainly, but not exclusively, in the central and 3' minor domains, and include positions 337, 674, 703, 705, 773, 791 , 803, 8 18, 1064, 1405, 1408, 1409, 1490, 1497, and 1 5 16. Protection results must be interpreted with the caveat that an induced conformational change can, in principle, give rise to protection that would otherwise be most simply interpreted as the result of intermolecular interaction. Following the simpler interpretation, these results place specific sites in 16S rRNA (mainly in the central and 3' minor domains) in the region of contact between the 30S and 50S ribosomal subunits, often referred to as the subunit interface.

Whether a particular site is crucial for intermolecular interaction can be tested by modification/selection experiments. Modification of 30S subunits with kethoxal destroys their ability to interact with 50S subunits (23 1). To test which kethoxal-reactive sites were responsible, researchers partially inactivated a population of 30S subunits by limited reaction with kethoxal. Association-competent 30S subunits were then selected from the modified population by virtue of their ability to form 70S monosomes. The modified sites in the association-competent subunits were compared with those of the total modified population. Several sites were found not to be modified, or much less extensively modified, in the competent population. These sites are residues 674, 703, 705, 791 , 8 18, 1064, 1497, and 15 16 (23 1). These results provide evidence that several sites in 16S rRNA that are protected from kethoxal, RNase T 1 , and cobra venom RNase in 70S ribosomes are also crucial for subunit association ; 50S subunits strongly select against 30S subunits in which these sites are kethoxal modified. Electron microscopic analysis of crystalline arrays of eukaryotic ribosomes, in which rRNA could be selectively visualized by contrast matching, show the subunit interface to be rich in RNA (232). Placement of rRNA in this region of the ribosome, commonly thought to be the center of the translational process, again points to a direct functional role for rRNA.

tRNA-RIBOSOME INTERACTIONS Certain tRNAs contain modified bases in the wobble position oftheir anticodons rendering them capable ofnear-UV photochemistry. When in the ribosomal P site, they react to form a cycloaddition product wi'th C1400 of 16S rRNA, in high yield (233-234). The significance of this finding is that it places the site of codon-anticodon interaction in extremely close proximity ('" 4 A) to a nucleotide in the middle of one of the most highly conserved sequences known to biologists, the 1390-1410 region of 16S rRNA. This is unlikely to be coincidental. This region of 16S rRNA has already been placed at the subunit interface by the studies described above.

Paromomycin-resistant mutants have been shown to have alterations at the base of the stem flanking this region of 16S rRNA. In yeast mito-

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

154 NOLLER

chondria C1409 is replaced by G (43), and in Tetrahymena 1 8S rRNA, G1491 becomes A (E. Blackburn, personal communication) ; in both cases the 1409/1491 base pair is disrupted (Figure 1 ). The proximity of the site of base changes conferring resistance to paromomycin, an antibiotic known to cause misreading, to the region of 16S rRNA cross-linked to the anticodons of several tRNAs is striking.

Several sites in 16S rRNA remain kethoxal-reactive in vacant 70S ribosomes. However, when polysomes ate formed, four of these sites, at positions 530, 966, 1 338, and 1 5 1 7, become substa�tially protected, in proportion to the tRNA occupancy ofthe polysomes (235). These protected sites may correspond to the sites of inactivation of tRNA binding by kethoxal, reported in earlier studies (236). The very highly conserved sequence in the loop at position 530 has already been noted. All four of the protected sites are in conserved sequences and are associated with secondary-structure features that are common to all 16S-like rRNAs. It is significant that three of the clusters of methylated bases in 16S rRNA are found among these four sites.

A great deal has been said concerning the postulated interaction of the GTJjJC sequence of tRNA with the GAAC sequence (residues 44-47) of 5S rRNA, both of which are quite conserved (77). Evidence for this is based on experiments using various oligonucleotide analogs of the GTJjJC sequence or tRNA fragments which inhibit aminoacyl tRNA binding, presumably by acting on the 50S subunit (237-239). Less convincing are studies placing the binding site for such fragments on 5S rRNA (240). Recent studies by Pace and colleagues (241 , 242) have tested this hypothesis directly. Residues 42-46 or 42-52 were excised from 5S rRNA, and 50S ribosomal subunits were reconstituted using the deleted RNA. Ribosomes utilizing the reconstituted 50S subunits were tested for their activity in protein synthesis, programmed either by poly U (241) or by phage MS2 RNA (242). In addition, activity of these ribosomes in f· Met-tRNA binding, EF-T-dependent aminoacyl tRNA binding, ppGpp synthesis, synthesis of phage proteins of the correct size and misreading under various conditions were studied. Ribosomes carrying deleted 5S rRNA were found to be active in all of these assays, although some diminution of activity was noted in AUG- [but not poly (A, U, G) or MS2 RNA] directed fMet-tRNA binding and ppGpp synthesis, especially in the case of the larger (residues 42-52) deletion. In particular, EF-Tu-depcndent aminoacyl-tRNA binding was affected very little by deletion of three of the four nucleotides in the GAAC sequence (242). Comparative analysis of the relevant sequences, in fact, predicts that this should be so. In most eukaryotic 5S rRNAs, the analog of the GAAC sequence is GAUC, yet the sequence in eukaryotic noninitiator tRNAs remains GTJjJC. Although eukaryotic initiator tRNAs have an altered

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 1 55

sequence, GAljJC, an apparent compensation to restore complementarity to 5S rRNA, plant 5S rRNAs retain the GAAC sequence in their 5S rRNAs, arguing against possible interaction of 5S rRNA with initiator tRNA in this fashion. Tn conclusion, the possibility of specific interaction(s) between the GTljJC sequence of tRNA and some site in the ribosome is still alive, but that any exists between tRNA and the GAAC sequence of 5S rRNA seems finally to have been ruled out.

Possible contact between the 3' terminus (acceptor end) of tRNA and rRNA has been explored extensively. However, because of its necessary proximity to the peptidyl transferase region, these studies are discussed in the following section.

PEPTIDYL TRANSFERASE The ribosomal activity responsible for catalysis of peptide bond formation is called peptidyl transferase. Affinity-labeling experiments as well as isolation of mutants resistant to peptidyl transferase­related antibiotics both indicate that the structure responsible for this activity is at, or near, a region in domain V of 23S rRNA.

An antibiotic closely identified with the peptidyl transferase function is chloramphenicol (243). Chloramphenicol-resistant mutations mapping in the 23S-like RNA genes have been found in several types of mitochondria. The sites of mutation have been found to be located at positions corre­sponding to residues 2447, 2451 , 2452, 2503, and 2504 (201-204). In the secondary-structure model (Figure 2) these sites are all clustered around the very highly conserved central loop in domain V. Erythromycin is also a peptidyl transferase inhibitor, although its mode of action is quite different from that of chloramphenicol (243). An erythromycin-resistant mutation in yeast mitochondria has been shown to result from a base substitution at a position corresponding to 2058 of 23S r RN A (205). This site is also found on the central loop in domain V (Figure 2). Erythromycin resistance in Staphylococcus ( 1 97) has been shown to be due to dimethylation of adenine in a GAAAG sequence somewhere in 23S rRNA. The presence of a GAAAG sequence in 23S rRNA at positions 2057-2061 , precisely at the site of the mitochondrial mutation discussed above, suggests that this is also a sequence crucial for sensitivity to erythromycin in bacteria.

A wide range of affinity probes of the peptidyl transferase site have been constructed, usually by attachment of reactive groups to the aminoacyl end of tRNA or to antibiotics implicated in this function (153). In several studies, these probes have been shown to react with 23S rRNA (244=-25 1). In a few cases, oligonucleotide sequences from the sites of reaction have been reported (252-253), but only in one case has the precise position of reaction with 23S rRNA been determined (A. Barta, personal communication). This elegant study utilizes a photoreactive benzophenone derivative of yeast

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

1 56 NOLLER

Phe-tRNA to label E. coli ribosomes. The position of the labeled site was first localized to the 2500-2600 region of23S rRNA by hybridization of23S rRNA containing the covalently bound radioactive benzophenone-Phe moiety with defined restriction fragments of the 23S rRNA gene. The precise position of attachment was identified by DNA-primed reverse transcription of the modified RNA, in which the covalently mod:ified residue causes the reverse transcriptase to "pause," generating a band on the gel pattern corresponding to cDNA chains terminating at that position. Remarkably, the main site of modification is U 2 5 84, with some modification ofU 2 5 8 5, once again placing the peptidyl transferase site in the proximity of the central loop in domain V.

ELONGATION-GTPase ACTIVITIES Thiostrepton appears to directly inhibit ribosomal GTPase, whether EFG, EFT or IF2 dependent (254). Thio­stepton resistance has been localized to methylation of A1 067 of 23S rRNA in Streptomyces azureus (200), implicating this region of domain II of 23S­like rRNAs in ribosomal GTPase functions. This is the same region of the 23S rRNA that binds protein L 1 1 , itself strongly implicated in GTPase­related functions (255). Recently EFG has been covalently cross-linked to 23S rRNA within the region 1055-108 1, possibly to A1067, further sub­stantiating the above hypothesis (256).

"SWITCHES" Ribosomes move along the mRNA chain during translation ; other kinds of possible inter- and intramolecular movement can be imagined to take place during translation, and raise the question of whether ribosomes have "moving parts." If this turns out to be the case, and their movement is fundamental to the process of translation, then very likely they will be found to be constructed, at least in part, from RNA (by the same evolutionary arguments summarized above).

It has been suggested previously that movement of r RN A structure could be generated by the making and breaking of helices (7, 12, 75). Early proposals (7) have been ruled out by comparative sequence analysis since the elucidation of many complete rRNA sequences, nor do more recent models ( 12) survive comparative analysis when rigorous alignment criteria are invoked. Only one pair of helices has been identified that could be considered to contain mutally exclusive structures. These are the 9-1 3/21-

25 and 1 7-20/91 5-9 18 helices of 16S rRNA (Figure 1), which could, in principle, coexist. They are, however, the only two helices thus far confirmed by comparative analysis where the loop contained by a helix is involved in pairing with a sequence outside the helix, and is thus formally analogous to a "knot." Biochemical probe experiments, summarized above, often show these sequences to be susceptible to single strand-specific agents. Whether or not the two helices actually coexist in vivo is presently an open question.

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 157

Another attractive hypothetical switch mechanism involves coaxial stacking and un stacking of helices. This possibility has been discussed previously in connection with 5S (257), 16S (1 3), and 23S (78) rRNA. Such a mechanism could bring about precise, large-scale conformational changes without necessarily involving disruption of base pairing or other hydrogen bonded interactions. To my knowledge, direct evidence for the existence of such a mechanism is also lacking.

Yet another possibility would be switches that involve tertiary structure. Direct evidence for a tertiary structural change in isolated 5S rRNA has been found (258). A conformational change triggered by increase in pH, temperature or cation concentration, all within physiological range, produces a structure that is more compact, has greater cation-binding capacity but less proton binding. It has been suggested that this conforma­tional switch may be biologically significant, and could provide the necessary movement for translocational events in the ribosome (258). The parts of the 5S rRNA involved in the observed transition are unknown.

PROSPECTS

Studies of ribosome structure and function seem to be about to enter a new phase. Structural studies are poised at the threshold of the identification of elements of tertiary structure. Evidence from biochemical and genetic studies is converging on the identification of functional sites in ribosomal RNA. Specific models for protein-RNA recognition seem imminent. Supporting all of this are the broad underpinnings of phylogenetic comparison, bringing the evolutionary perspective to bear not only on questions of evolution per se, but on the actual details of the structure and mechanism of this ancient machine.

ACKNOWLEDGMENTS

I wish to thank R. Gutell and B. Weiser for computer-generated drawings, and R. Gutell, J. Prince, K. Triman, S. Turner, and L. Zagorska for reading the manuscript. Secondary-structure models have been developed in collaboration with C. Woese. This work was supported by NIH grant No. GM-1 7129.

Literature Cited

1 . Chambliss, G., Craven, G. R, Davies, J., Davis, K., Kahan, L, Nomura, M. ed. 1979. Ribosomes : Structure, Function and Genetics. Baltimore, Md : Univ. Park.

2. Garrett, R. A. 1979. In Chemistry of Macromolecules lIB, 25 : 127-77. Balti­more, Md : Univ. Park

3. Brimacombe, R, Stiiffler, G., Wittmann,

H. G. 1978. Ann. Rev. Biochem. 47 : 2 17-49

4. Wittmann, H. G. 1982. Ann. Rev. Bio­chem. 5 1 : 155-83

5. Liljas, A. 1982. Prog. Biophys. Mol. Bioi. 40 : 16 1-228

6. Nierhaus, K. H. 1982. Curro Top. Micro­bioi. Immunol. 97 : 8 1-1 55

7. Noller, H. F. 1979. See Ref. 1 , pp. 3-22

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

158 NOLLER

8. Zimmermann, R. A. 1979. See Ref. 1 , pp. 135-70

9. Bogdanov, A. A., Kopylov, A. M., Shatsky, I. N. 1980. Sub-Cell. Biochem. 7 : 81-1 16

10. Noller, H. F., Woese, C. R. 1 98 1 . Science 2 12 : 403-1 1

1 1. Noller, H. F., van Knippenberg, P. H. 1983. Horizons in Biochemistry and Biophysics, ed. F. Palmieri, 7 : In press. London : Wiley

12. Brimacombe, R., Maly, P., Zwieb, C. 1983. Prog. Nucleic Acid Res. Mol. Bioi. 28 : 1-48

1 3. Woese, C. R., Gutell, R. R., Gupta, R., Noller, H. F. 1983. Microbiol. Rev. In press

14. Thompson, J. F., Hearst, J. E. 1983. Cell 33 : 19-24

15. Garrett, R. A., Douthwaite, S., Noller, H. F. 1981 . Trends Biochem. Sci. 6 : 137-39

16. Bohm, S., Fabian, H., WeiHe, H. 1982. Acta Bioi. Med. Ger. 41 : 1-16

1 7. de Wachter, R., Chen., M.-W., Vanden­berghe, A. 1982. Biochimie 64 : 3 1 1-29

18. Nazar, R. N. 1982. Cell Nucleus 10 : 1-28

19. MacKay, R. M., Spencer, D. F., Schnare, M. N., Doolittle, W. F., Gray, M. W. 1982. Can. J. Biochem. 60 : 480--89

20. Delihas, N., Andersen, 1. 1982. Nucleic Acids Res. 10 : 7323-44

21 . Erdmann, V. A., Huysmans, E., Vandenberghe, A., De Wachter, R. 1983. Nucleic Acids Res. 1 1 : r105-33

22. Brownlee, G. G., Sanger, F., Barrell, B. G. 1 967. Nature 215 : 735-36

23. Brosius, J., Palmer, M. L., Kennedy, P. J., Noller, H. F. 1978. Proc. Natl. Acad. Sci. USA 75 : 4801-5

24. Carbon, P., Ehresmann, C., Ehres­mann, B., Ebel, J.-P. 1979. Eur. J. Bio­chem. 100 : 399-410

25. Kop, J., Kopylov, A. M., Noller, H. F., Siegel, R., Gupta, R., Woese, C. R. 1983. Nucleic Acids Res. Submitted for publication

26. Carbon, P., Ebel, J.-P., Ehresmann, C. 1 98 1 . Nucleic Acids Res. 9 : 2325-33

27. Tomioka, N., Sugiura, M. 1983. Mol. Gen. Genet. 19 1 : 46-50

28. Gupta, R., Lanter, J., Woese, C. R. 1983. Science 221 : 656-59

29. Brosius, 1., Dull, T. J., Noller, H. F. 1980. Proc. Natl. Acad. Sci. USA 77 : 201-4

30. Kop,J., Wheaton, V., Gupta, R., Woese, C. R., Noller, H. F. 1983. Nucleic Acids Res. Submitted for publication

3 1 . Schwartz, Z., Kossel, H. 1 980. Nature 283 : 739-42

32. Tohdoh, N., Sugiura, M. 1982. Gene 1 7 : 2 1 3-18

33. Graf, L., Roux, E., Stutz, E., Kossel, H. 1982. Nucleic Acids Res. 10 : 6369- 81

34. Dron, M. , Rahire, M., Rochaix, J.-D. 1982. Nucleic Acids Res. 10 : 7609-·20

35. Edwards, 1., Kossel, H. 1981. Nucleic Acids Res. 9 : 2853-69

36. Takaiwa, F., Sugiura, M. 1982. Eur. J. Biochem. 124 : 13-19

37. Eperon, l. c., Janssen, J. W. G., Hoeijmakers, J. H. J., Borst, P. 1983. Nucleic Acids Res. 1 1 : 105-25

38. Eperon, I. c., Anderson, S., Nierlich, D. P. 1980. Nature 286 : 460--67

39. Anderson, S., de Bruijn, M. H. L., Coulson, A. R., Eperon, 1. C., Sanger, F., Young, I. G. 1 982. J: Mol. Bioi. 1 56 : 683-71 7

40. van Etten, R . A., Walberg, M . W., Clayton, D. A. 1980. Ce// 22 : 157-70

41. Kobayashi, M., Seki, T., Katsuyuki, Y., Koike, K. 1981 . Gene 1 6 : 297-307

42. Sor, F., Fukuhara, H. 1980. C.R. Acad. Sci. Ser. D 291 : 933-36

43. Li, M., TzagoiolT, A., Underbrink­Lyon, K., Martin, N. C. 1982. J. BioI. Chern. 257 : 5921-28

44. Kochel, H.-G., Kuntzel, H. 1981 . Nucleic Acids Res. 9 : 5689-96

45. Seilhammer, J. J., Olsen, G. J., Cummings, D. G. 1983. J. Bioi. Chern. Submitted for publication

46. Saccone, C., Cantatore, P., Gadaleta, G., Gallerani, R., Lanave, c., Pepe, G., Kroon, A. M. 1981 . Nucleic Acids Res. 9 : 4139-48

47. Seilhammer, J. J., Gutell, R. R., Cummings, D. 1. 1983. J. Bioi. Chem. Submitted for publication

48. Sor, F., Fukuhara, H. 1983. Nucleic Acids Res. 1 1 : 339-48

49. Kochel, H. G., Kuntzel, H. 1982. Nucleic Acids Res. 10 : 4795-801

50. Buetow, D. E., Wood, W. M. 1978. Sub­Cell. Biochem. 5 : 1-85

51. MacKay, R. M. 1981 . FEBS Lett. 123 : 1 7-18

52. Edwards, K., Bedbrook, J. , Dyer, T., Kassel, H. 1981. Biochem. Int. 2 : 533-38

53. Rochaix, J.-D., Darlix, 1.-L. 1982. J. Mol. Bioi. 159 : 383-95

54. Nazar, R. N. 1980. FEBS Lett. 1 19 : 212-14

55. Jacq, B. 1981. Nue/eic Acids Res. 9 : 2913-32

56. Walker, T. A., Johnson, K. D., Olsen, G. J., Peters, M. A., Pace, N. R. 1982. Biochemistry 21 : 2320--29

57. Jordan, B. R., Jourdan, R., Jacq, B. 1976. J. Mol. Bioi. 101 : 85-105

58. Brosius, J., Dull, T. J., Sleeter, D. D., Noller, H. F. 1981 . J. Mol. Bioi. 148 : 107-27

59. Rubtsov, P. M., Musakhanov, M. M., Zakharyev, V. M., Krayev, A. S.,

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 159

Skryabin, K. G., Bayev, A. A. 1980. Nucleic Acids Res. 8 : 5779-94

60. Salim, M., Maden, B. E. H. 1981. Nature 291 : 205-8

61. Chan, Y.-1., Gutell, R. R., Noller, H. F., Wool, I. G. 1984. 1. Bioi. Chern. In press

62. Torczynski, R., Bollon, A. P., Fuke, M. 1983. Nucleic Acids Res. 1 1 : 4879-90

63. Otsuka, T., Nomiyama, H., Yoshida, H., Kukita, T., Kuhara, S., Sakaki, Y. 1983. Proc. Natl. Acad. Sci. USA 80 : 3 163-67

64. Veldman, G. M., Klootwijk, J., de Regt, V., Planta, R., Branlant, C., et al. 1981 . Nucleic Acids Res. 9 : 6935-52

65. Georgiev, O. I., Nikolaev, N., Hadjiolov, A. A., Skryabin, K. G., Zakharyev, V. M., Bayev, A. A. 1981. Nucleic Acids Res. 9: 6953-58

66. Chan, Y.-1., Wool, I . G. 1983. Nucleic Acids Res. In press

67. Fellner, P., Sanger, F. 1968. Nature 219 : 236-38

68. Fellner, P. 1969. Eur. 1. Biochem. 1 1 : 12-27

69. Van Charldorp, R., Heus, H. A., Van Knippenberg, P. H. 1981. Nucleic Acids Res. 9 : 271 7-25

70. Chai, Y. e, Busch, H. 1978. Biochemistry 1 7 : 2551-60

71 . Klootwijk, J., Planta, R. J. 1973. Eur. 1. Biochem. 39 : 325-33

72. Noller, H. F. 1974. Biochemistry 1 3 : 4694-703

73. Fox, G. E., Stackebrandt, E., Hespell, R. B., Gibson, J., Woese, C R., et al. 1 980. Science 209 : 457-63

74. Woese, C R., Fox, G. E. 1977. Proc. Natl. Acad. Sci. USA 74 : 5088-90

75. Fox, G. E., Woese, e R. 1975. Nature 256 : 505-7

76. Nishikawa, K., Takemura, S. 1974. 1. Biochem. 76 : 935-47

77. Erdmann, V. A. 1976. Prog. Nucleic Acids Res. Mol. Bioi. 1 8 : 45-90

78. Noller, H. F., Kop, J., Wheaton, V., Brosius, J., Gutell, R. R., et al. 198 1 . Nucleic Acids Res. 9 : 6 167-89

79. Woese, C R., Magrum, L. J., Gupta, R., Siegel, R. B., Stahl, D. A., et al. 1 980. Nucleic Acids Res. 8 : 2275-93

80. Stiegler, P .. Carbon, P., Zuker, M., Ebel, J. P., Ehresmann, e 1981. Nucleic Acids Res. 9 : 21 53-72

81 . Zwieb, C, Glotz, C, Brimacombe, R. 1981. Nucleic Acids Res. 9 : 3621-40

82. Glotz, C, Zwieb, C, Brimacombe, R., Edwards, K., Kossel, H. 1981. Nucleic Acids Res. 9 : 3287-306

82a. Branlant, C., Krol, A., Machatt, M. A., Pouyet, J., Ebel, J. P., Edwards, K., Kossel, H. 1981 . Nucleic Acids Res. 9 : 4303-24

83. Tinoco, t, Uhlenbeck, O. C, Levine, M.

D. 197 1 . Nature 230 : 362-67 84. Pipas, J. M., McMahon, J. E. 1975.

Proc. Nat!. Acad. Sci. USA 72 : 2017-21 85. Studnicka, G. M., Rahn, G. M.,

Cummings, I . W., Salser, W. 1978. Nucleic Acids Res. 9 : 3365-87

86. Waterman, M. S., Smith, T. F. 1978. Math. Biosci. 42 : 257-66

87. Nussinov, R., Jacobson, A. 1980. Proc. Nat!. Acad. Sci. USA 77 : 6309-13

88. Zuker, M. , Stiegler, P . 1981 . Nucleic Acids Res. 9 : 133-47

89. Tinoco, I., Borer, P. N., Dengler, B., Levine, M. D., Uhlenbeck, O. C., et al. 1973. Nature New Bioi. 246 : 40-41

90. Salser, W. 1977. Cold Spring Harbor Symp. Quant. Bioi. 42 : 985-1002

91 . Ninio, J. 1979. Biochimie 61 : 1 1 33-50 92. Johnston, P. D., Redfield, A. G. 1979. In

Transfer RNA : Structure, Properties and Recognition, ed. P. R. Schimmel, D. Soli, J. N. Abelson, pp. 191-205. Cold Spring Harbor, NY : Cold Spring Harbor Lab.

93. Bear, D. G., Schleich, T., Noller, H. F., Douthwaite, S., Garrett, R. A. 1979. FEBS Lett. 1 00 : 99-102

94. Bear, D. G., Schleich, T., Noller, H. F., Garrett, R. A. 1977. Nucleic Acids Res. 4 : 25 1 1-26

95. Litt, M. 1969. Biochemistry 8 : 3249-53 96. Peattie, D. A., Gilbert, W. 1980. Proc.

Natl. Acad. Sci. USA 77 : 4679-82 97. Mankin, A. S., Kopylov, A. M., Bog­

danov, A. A. 1981. FEBS Lett. 1 34 : 1 1-14

98. Silberkang, M., RajBhandary, U. 1., Liick, A., Erdmann, V. A. 1983. Nucleic Acids Res. 1 1 : 605-17

99. Rhodes, D. 1975. 1. Mol. BioI. 94 : 449-60

100. Rabin, D., Crothers, D. M. 1979. Nucleic Acids Res. 7 : 689-703

101. Wollenzien, P. 1., Hearst, J. E., Thammana, P., Cantor, C R. 1979. J. Mol. BioI. 135 : 255-69

102. Thammana, P., Cantor, C. R., Wollen­zien, P. 1., Hearst, J. E. 1979. 1. Mol. Bioi. 1 35 : 271-83

103. Turner, S., Thompson, J. F., Hearst, J. E., Noller, H. F. 1982. Nucleic Acids Res. 1 0 : 2839-49

104. Turner, S., Noller, H. F. 1983, Bio­chemistry 22.: 4159-65

105. Thompson, J. F., Hearst,J. E. 1983. Cell 32 : 1 355-65

106. Hearst, J. E. 1981. Ann. Rev. Biophys. Bioeng. l0 : 69-86

107. Vigne, R., Jordan, B. R., Monier, R. 1973. J. Mol. BioI. 76 : 303-1 1

108. Ross, A., Brimacombe, R. 1979. Nature 28 1 : 27 1-76

1 09. Vassilenko, S. K., Ryte, V. C 1 975. Biokhimya 40 : 578-83

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

160 NOLLER

1 10. Vassilenko, S. K., Carbon, P., Ebel, J. P., Ehresmann, C. 1981. J. Mol. Bioi. 1 52 : 699-721

1 1 1. Douthwaite, S., Garrett, R A. 1981. Biochemistry 20 : 7301-7

1 12. Lewis, I. B., Doty, P. 1977. Biochemistry 16 : 5016-25

1 13. Kopylov, A. M., Chichkova, N. V., Bogdanov, A. A., Vasilenko, S. K. 1975. Mol, BioI. Rep. 2 : 95-100

1 14. Mankin, A. S., Skripkin, E. A., Chich­kova, N. V., Kopylov, A. M., Bogdanov, A. A. 1981. FEBS Lett. 131 : 253-56

1 1 5. Backendorf, c., Ravensbergen, C. J. C., Van der Plas, J., van Boom, J. H., Veeneman, G., Van Duin, J. 1981. Nucleic Acids Res. 9 : 1425-44

1 16. Kime, M. J., Moore, P. B. 1983. FEBS Lett. 153 : 199-203

1 17. Kime, M. J., Moore, P. B. 1983. Bio­chemistry 22 : 2615-22

1 18. Kime, M. J., Moore, P. B. 1983. Bio­chemistry 22 : 2622-29

1 19. Baan, R. A., Hilbers, C. W., van Charldorp, R, Van Leerdam, E., Van Knippenberg, P. H., Bosch, L. 1977. Proc. Natl. Acad. Sci. USA 74 : 1028-31

120. Edlind, T. D., Bassel, A. R. 1980. J. Bacterial. 141 : 365-73

121 . Ehresmann, c., Stiegler, P., Carbon, P., Ungewickell, E., Garrett, R. A. 1980. Eur. J. Biochem. 103 : 439-46

122. Miiller, R, Garrett, R. A., Noller, H. F. 1979. J. Bioi. Chem. 254 : 3873-78

123. Zimmermann, R. A., Mackie, G. A., Muto, A., Garrett, R. A., Ungewickell, E., et al. 1975. Nucleic Acids Res. 2 : 279-302

124. Yuki, A., Brimacombe, R. 1975. Eur. J. Biochem. 56 : 23-34

125. Traub, W., Sussman, J. L. 1982. Nucleic Acids Res. 10 : 2701-8

126. Douthwaite, S., Garrett, R. A. 1983. J. Mol. Bioi. In press

127. Rich, A., RajBhandary, U. L. 1976. Aim. Rev. Biochem. 45 : 805-60

128. Clark, B. F. C. 1977. Proc. Nucleic Acid Res. Mol. Bioi. 20 : 1-19

129. Crick, F. H. C. 1966. J. Mol. Bioi. 19 : 548-55

130. Kan, L.-S., Chandrasegaran, S., Pulford, S. M., Miller, P. S. 1983. Proc. N atl. Acad. Sci. USA 80 : 4263-265

131 . Cantor, C. R. 1979. See Ref. 1, pp. 23-49 132. Stiegler, P., Carbon, P., Ebel, J. P.,

Ehresmann, C. 1981. Eur. J. Biochem. 120 : 487-95.

133. Hogan, J. J., Gutell, R. R., Noller, H. F. 1983. Biochemistry. Submitted for publication

134. Hogan, J. J., Gutell, R. R, Noller, H. F. 1983. Biochemistry. Submitted for publication

135. Hogan, J. J., Noller, H. F. 1978. Bio­chemistry 17 : 587-93

136. Herr, W., Noller, H. F. 1978. Bio­chemistry 1 7 : 307-15

137. Cantor, C. R., Wollenzien, P. L., Hearst, J. E. 1980. Nucleic Acids Res. 8 : 1855-72

138. Wollenzien, P. L., Cantor, C. R. 1982. Proc. N atl. Acad. Sci. USA 79 : 3940-44

139. Stiege, W., Zwieb, C., Brimacombe, R 1982. Nucleic Acids Res. 10 : 721 1-99

140. Zwieb, c., Brimacombe, R. 1980. Nucleic Acids Res. 8 : 2397-1 1

141. Glotz, c., Brimacombe, R. 1980. Nucleic Acids Res. 8 : 2377-95

142. Lake, J. A. 1979. See Ref. 1, pp. 207-36 143. St(iffler, G., Bald, R, Kastner, B.,

Liihrmann, R, Stiiffler-Meilicke, M., Tischendorf, G. 1979. See Ref. 1, pp. 1 71-205

144. Moore, P. B. 1979. See Ref. 1, pp. 1 1 1-33

145. Politz, S. M., Glitz, D. G. 1977. Proc. Natl. Acad. Sci. USA 74 : 1468-72

146. Trempe, M. R, Ohgi, K., Glitz, D. G. 1982. J. Bioi. Chem. 257 : 9822-29

147. Olson, H. M., Glitz, D. G. 1979. Proc. N atl. Acad. Sci. USA 76 : 3769-73

148. Shatsky, I. N., Mochalova, L. V., Kojouharova, M. S., Bogdanov, A. A., Vasiliev, V. D. 1979. J. Mol. BioI. 133 : 501-15

149. Liihrmann, R., Stiiffler-Meilicke, M., StOffler, G. 1981. Mol. Gen. Genet. 182 : 369-76

150. Mochalova, L. V., Shatsky, I. N., Bogdanov, A. A., Vasiliev, V. D. 1982. J. Mol. Bioi. 159 : 637-50.

151. St6ffler-Meilicke, M., St6ffler, G., Odom, O. W., Zinn, A., Kramer, G., Hardesty, B. 1981 . Proc. Natl. Acad. Sci. USA 78 : 5538-42

152. Shatsky, I. N., Evstafieva, A. G., Bys­trova, T. F., Bogdanov, A. A., Vasiliev, V. D. 1980. FEBS Lett. 121 : 97-100

153. Cooperman, B. S. 1979. See Ref. 1, pp. 531-54

1 54. Shatsky, I. N., Evstafieva, A. G., Bys­trova, T. F., Bogdanov, A. A., Vasiliev, V. D. 1980. FEBS Lett. 122 : 251-55

1 55. Wollenzien, P. L., Cantor, C. R. 1982. J. Mol. BioI. 159 : 151-66

156. Spitnik-Elson, P., Elson, D., Avital, S., Abramowitz, R. 1982. Nucleic Acids Res. 10 : 1995-2006

1 57. Spitnik-Elson, P., Elson, D., Avital, S., Abramowitz, R. 1982. Nucleic Acids Res. 10 : 4483-92

158. Stiege, W., Glotz, c., Brimacombe, R. 1983. Nucleic Acids Res. 1 1 : 1687-1706

1 59. Odom, O. W., Robbins, D. J., Dottavia­Martin, D., Kramer, G., Hardesty, B. 1980. Biochemistry 1 9 : 5947-54

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

STRUCTURE OF RIBOSOMAL RNA 161

1 60. Hancock, J. , Wagner, R. 1 982. Nucleic Acids Res. 1 0 : 1257-69

161 . Pieler, T., Erdmann, V. A. 1982. Proc. Natl. Acad. Sci. USA 79 : 4599-603

162. Monier, R. 1 974. In Ribosomes, ed. M. Nomura, A. Tissieres, P. Lengyel, pp. 141-68. Cold Spring Harbor, NY : Cold Spring Harbor Lab.

163. Noller, H. F .. Garrett, R. A. 1979. J. Mol. Bioi. 1 3 2 : 637-48

1 64. Noller, H. F., Herr, W. 1974. J. Mol. Bioi. 90 : 181-184.

165. Delihas, N., Dunn, I. 1., Erdmann, V. A. 1975. FEBS Lett. 58 : 76-80

166. Weidner, H., Yuan, R., Crothers, D. M. 1977. Nature 266 : 193-94

167. Zimmermann, R. A., Singh-Bermann, K. 1979. Biochim. Biophys. Acta 563 : 422-31

168. Branlant, C., Krol, A., SriWidada, I., Ebel, I. P., Sioor, P., Garrett, R. A. 1976. Eur. J. Biochem. 70 : 457-69

169. Thompson, 1., Schmidt, F., Cundliffe, E. 1982. J. Bioi. Chem. 257 : 791 5--1 7

1 70. Garrett, R . A., Noller, H. F . 1979. J. Mol. Bioi. 1 32 : 637-48

17l. Peattie, D. A., Douthwaite, S., Garrett, R. A., Noller, H. F. 1981. Proc. Natl. Acad. Aci. USA 78 : 7331-35

172. Douthwaite, S., Garrett, R. A. 1982. Biochemistry 21 : 2313-20

173. Douthwaite, S., Garrett, R. A., Wagner, R., Feunteun, I. 1979. Nucleic Acids Res. 6 : 2453-70

174. Spierer, P., Bogdanov, A. A., Zimmer­mann, R. A. 1978. Biochemistry 1 7 : 5394-42

175. Wower, I., Brimacombe, R. 1983. Nucleic Acids Res. 1 1 : 141 9-37

176. Nomura, M., Dean, D., Yates, I. L. 1982. Trends Biochem. Sci. 7 : 92-95

177. Nomura, M., Yates, I. L., Dean, D., Post, L. E. 1 980. Proc. Nat!. Acad. Sci. USA 77 : 7084-88

178. Olins, P. O., Nomura, M. 1981. Nucleic Acids Res. 9 : 1 757-64

1 79. Gourse, R. L., Thurlow, D. L., Gerbi, S. A., Zimmermann, R. A. 1 98 1 . Proc. N atl. Acad. Sci. USA 78 : 2722-26

1 80. Branlant, C., Krol, A., Machatt, M. A., Ebel, I.-P. 1 98 1 . Nucleic Acids Res. 9 : 293-307

181 . Vasiliev, V. D., Selivanova, O. M., Koteliansky, V. E. 1978. FEBS Lett. 95 : 273-76

182. Allen, S. H., Wong, K.-P. 1978. Biochemisry 253 : 8759-66

183. Boublik, M., Robakis, N., Hellmann, W. 1982. Eur. J. Cell. Bioi. 27 : 177-84

1 84. Tam, M. F., Dodd, J. A., Hill, W. E. 1981. J. Bioi. Chem. 256 : 6430-34

185. Nomura, M., Held, W. A. 1974. See Ref. 1 62, pp. 193-223

1 86. Nierhaus, K. 1979. See Ref. 1, pp. 267-94

187. Held, W. A., Nomura, M. 1973. Bio­chemistry 1 2 : 3273-81

188. Herr, W., Noller, H. F. 1975. FEBS Lett. 53 : 248-52

189. Azad, A. A. 1979. Nucleic Acids Res. 7 : 1 9 1 3-29

190. Davies, J., Nomura, M. 1972. Ann. Rev. Genet. 6 : 203-34

191 . Nomura, M., Mizushima, S., Ozaki, M., Traub, P., Lowry, C. V. 1 969. Cold Spring Harbor Symp. Quant. BioI. 34 : 49-61

192. Erdmann, V. A., Fahnestock, S., Higo, K., Nomura, M. 1971. Proc. Natl. Acad. Sci. USA 68 : 2932-36

193. Dobbs, E. R. 1979. J. Bacterial. 140 : 734-37

194. Garrett, R. A. 1983. Trends Biochem. Sci. 8 : 75-76

195. Kruger, K., Grabowski, P. J., Zaug, A. I., Sands, I., Gottschling, D. E., Cech, T. R. 1982. Cell 3 1 : 1 47-57

196. Helser, T. L., Davies, J. E., Dahlberg, J. E. 1972. Nature New BioI. 235 : 6-8

197. Lai,.C.-I., Dahlberg, J. E., Weisblum, B. 1973. Biochemistry 12 : 457-60

198. Gold, L., Pribnow, D., Schneider, T., Shinedling, S., Singer, B. S., Stormo, G. 1981 . Ann. Rev. Microbial. 35 : 365-403

199. Yamada, T., Mizuguchi, Y., Nierhaus, K. H., Wittmann, H. G. 1978. Nature 275 : 460-61

200. Thompson, 1., Schmidt, F., Cundliffe, E. 1982. J. Bioi. Chem. 257 : 79 15--1 7

201. Dujon, B. 1980. Cell 20 : 1 85-97 202. Kearsey, S. E., Craig, I. W. 1981. Nature

290 : 607-8 203. Blanc, H., Adams, C. W., Wallace, D. C.

1981. Nucleic Acids Res. 9 : 5785-95 204. Blanc, H., Wright, C. T., Bibb, M. I.,

Wallace, D. c., Clayton, D. A. 1981. Proc. Natl. Acad. Sci. USA 78 : 3789-93

205. Sor, F., Fukuhara, H. 1982. Nucleic Acids Res. 1 0 : 6571-77

206. Bolotin-Fukuhara, M. 1979. Mol. Gen. Genet. 177 : 39-46

207. Sigmund, C. D., Morgan, E. A. 1982. Proc. Natl. Acad. Sci. USA 79 : 5602-6

208. Gourse, R. L., Stark, M. J. R., Dahlberg, A. E. 1982. J. Mol. Bioi. 1 59 : 397-416

209. Stark, M. I. R., Gourse, R. L., Dahlberg, A, E. 1982. J. Mol. BioI. 1 59 : 417-39

210. Fox, T. D., Staempfli, S. 1982. Proc. Nat!. Acad. Sci. USA 79 : 1583-87

2 1 1 . Bowman, C. M., Dahlberg, J. E., Ikemura, T., Konisky, 1., Nomura, M. 1971. Proc. Nat!. Acad. Sci. USA 68 : 964-68

2 12. Senior, B. W., Holland, I. B. 1971. Proc. Natl. Acad. Sci. USA 68 : 959-63

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.

162 NOLLER

213. Schindler, D. G., Davies, J. E. 1977. Nucleic Acids Res. 4: 1097-1 10

214. Endo, V., Wool, !. G. 1982. J. BioI. Chem. 257 : 9054--60

2 15. Shine, J., Dalgarno, L. 1 974. Proc. Natl. Acad. Sci. USA 7 1 : 1 342-46

216. Steitz, J. A. 1979. See Ref. 1, pp. 479-95 2 17. Steitz, J. A., Jakes, K. 1975. Proc. Natl.

Acad. Sci. USA 72 : 4734--38 2 18. Taniguchi, T., Weissman, C. 1978. J.

Mol. BioI. 1 1 8 : 533-65 219. Dunn, J. J., Buzash-Pollert, E., Studier,

F. W. 1978. Proc. Natl. Acad. Sci. USA 75 : 2741-45

220. Grunberg-Manago, M. 1979. See Ref. 1 , pp. 445-77

221. Cooperman, B. S., Dondon, J., Finelli, J., Grunberg-Manago, M., Michelson, A. M. 1977. FEBS Lett. 76 : 59-63

222. Wickstrom, E. 1983. Nucleic Acids Res. 1 1 : 2035-52

223. Okuyama, A., Machiyama, N., Kino­shita, T., Tanaka, N. 1971. Biochem. Biophys. Res. Commun. 43 : 196-99

224. Wagner, R., Gassen, H. G., Ehresmann, C, Stiegler, P., Ebel, J. P. 1976. FEBS Lett. 67 : 3 12-15

225. Fiser, I., Scheit, K. H., Kuechler, E. 1977. Eur. J. Biochem. 74 : 447-56

226. Budker, V. G., Girshovich, A. S., Grineva, N. I., Karpova, G. G., Knorre, D. G., Kobets, N. D. 1973. Dokl. Akad. Nauk SSSR 2 1 1 : 725-28

227. Towbin, H., Elson, D. 1978. Nucleic Acids Res. 5 : 3389-407

228. Santer, M., Shane, S. 1977. J. Bacterial. 130 : 900--10

229. Chapman, N, M., Noller, H. F. 1979. J. Mol. Bioi. 109 : 1 3 1-49

230. Vassilenko, S. K., Carbon, P., Ebel, J.-P., Ehresmann, C. 1981 . J. Mol. Bioi. 152 : 699-721

231. Herr, W., Chapman, N. M., Noller, H. F. 1979. J. Mol. Bioi. 1 30 : 433-49

232. Kiihlbrandt, W., Unwin, P. N. T. 1982. J. Mol. Bioi. 156 : 431-48

233. Schwartz, I., Of eng and, J. 1 978. Bio­chemistry 1 7 : 2524--30

234. Prince, J. B., Taylor, B. H., Thurlow, D. L., Of eng and, J., Zimmermann, R. A. 1982. Proc. Natl. Acad. Sci. USA 79 : 5450-54

235. Brow, D. A., Noller, H. F. 1983. J. Mol. Bioi. 163 : 27-46

236. Noller, H. F., Chaires, J. B. 1972. Proc. Natl. Acad. Sci. USA 69 : 3 1 1 5-1 18

237. Of eng and, J . , Henes, C 1 969. J. BioI. Chem. 244 : 6241-53

238. Shimizu, N., Hayashi, H., Miura, K.-I. 1970. J. Biochem. 67 : 373-87

239. Richter, D., Erdmann, V. A., Sprinzl, M. 1973. Nature New BioI. 246 : 1 32-35

240. Erdmann, V. A., Sprinzl, M., Pongs, O. 1973. Biochem. Biophys. Res. Commun. 54 : 942-48

241. Pace, B., Matthews, E. A., Johnson, K. D., Cantor, C R., Pace, N. R. 1982. Proc. N atl. Acad. Sci. USA 79 : 36-40

242. Zagorska, L., Van Duin, J., Noller, H. F., Pace, B., Johnson, K. D., Pace, N. R. 1983. J. Bioi. Chem. In press

243. Vazquez, D. 1979. Antibiotic Inhibitors of Protein Biosynthesis. Berlin Heidel­berg New York : Springer-Verlag

244. Bispink, L., Matthaei, H. 1973. FEBS Lett. 37 : 291-94

245. Girshovich, A. S., Bochkarcva, E. S., Kramarov, V. M., Ovchinnikov, Y. A. 1974. FEBS Lett. 45 : 213-17

246. Breitmeyer, J. B., Noller, H. F. 1 976. J. Mol. Bioi. 101 : 297-306

247. Barta, A., Kuechler, E., Branlant, C, SriWidada, J., Kro!, A., Ebel, J. P. 1975. FEBS Lett. 56 : 1 70-74

248. Sonenberg, N., Wilchek, M., Zamir, A . 1977. Eur. J. Biochem. 77 : 2 17-22

249. Yukioka, M., Hatayama, T., Morisawa, S. 1975. Biochim. Biophys. Acta 390 : 192-308

250. Johnson, A. 1979. See Ref. 92, pp. 487-99

251 . Greenwall, P., Harris, R. J., Symons, R. H. 1974. Eur. J. Biochem. g2 : 225-34

252. Eckermann, D. 1., Symons, R. H. 1978. Eur. J. Biochem. 82 : 225-34

253. Yukioka, M., Hatayama, T., Omori, K. 1977. Eur. J. Biochem. 73 : 449-59

254. Cundlilfe, E. 1979. See Ref. l, pp. 555-81 255. Moller, W. 1974. See Ref. 1 62, pp. 7 1 1-

3 1 256. Skold, S.-E. 1983. Nucleic Acids Res.

1 1 : 4923-32 257. Stahl, D. A., Luehrsen, K. R., Woese, C.

R., Pace, N. R. 1 981 . Nucleic Acids Res. 9 : 6129-37

258. Kao, T. H., Crothers, D. M. 1980. Proc. Natl. Acad. Sci. USA 77 : 3360-64

259. Stein, P. R., Waterman, M. S. 1978. Discrete Math. 26 : 261-72

260. Klein, B. K., King, T. c., Schlessinger, D. 1983. J. Mol. Bioi. 168 : 809-30

Ann

u. R

ev. B

ioch

em. 1

984.

53:1

19-1

62. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Yor

k on

03/

24/1

3. F

or p

erso

nal u

se o

nly.