Structure Synthesis the Ribosomal Ribonucleic Prokaryotes · amongthe prokaryotes. Similarities among the ribosomes isolated from prokaryotes of different taxonomic orders andtheirstructural

BACTERIOLOGICAL REVIEWS, Dec. 1973, p. 562-603 Vol. 37, No. 4Copyright 0 1973 American Society for Microbiology Printed in U.S.A.

Structure and Synthesis of the RibosomalRibonucleic Acid of Prokaryotes

NORMAN R. PACENational Jewish Hospital and Research Center, and Department of Biophysics and Genetics, University of

Colorado Medical Center, Denver, Colorado 80206

INTRODUCTION ........................................................... 562GENERAL FEATURES OF PROKARYOTIC RIBOSOMES ....... ............... 563

Diversity in Bacterial Ribosomes .................... ............................ 563STRUCTURES OF THE PROKARYOTIC rRNA MOLECULES ...... ............ 564Prokaryotic rRNA Molecules................................................... 564rRNA Integrity and Ribosome Function ............ ............................. 565Eukaryotic Analogy ........................................................... 565Base Compositions ofrRNA from Prokaryotes .......... ......................... 565Primary and Secondary Structures of the rRNA ......... ........................ 568Question of Functional Heterogeneity in the rRNA Molecules ...... .............. 569Phylogenetic Differences in the rRNA of Prokaryotes ........ .................... 570Prokaryotic Origin for the Eukaryotic Organelles? ........ ...................... 572

GENES SPECIFYING rRNA ..................................................... 573Gene Dosage for rRNA .......................................................... 573Mapping the rRNA Genes in E. coli ...... ........ .............................. 573Mapping the rRNA Genes in B. subtilis ......................................... 575Isolation of the rDNA .......... ............................... 575General Features of the Prokaryotic Ribosomes ................... .............. 591Structures of the Prokaryotic rRNA Molecules .................................. 591Genes Specifying rRNA ........... .............................. 591Transcription of the rRNA Genes .......................................... 592Post-Transcriptional Processing of rRNA ....................................... 592

TRANSCRIPTION OF THE rRNA GENES ....................................... 576rRNA Transcriptional Unit ............... .......................... 576Interaction of the RNA Polymerase with the rDNA .............. ................ 577Initiation of rDNA Transcription .......................................... 578Regulation ofrRNA Synthesis During Balanced Growth ....... ................. 579Regulation of rRNA Synthesis During Perturbation of Balanced Growth ........ 581Regulation ofrRNA Synthesis During Morphogenesis ....... .................... 582

POST-TRANSCRIPTIONAL PROCESSING OF rRNA ....... .................... 583Hypothetical Tandem Transcript of the rDNA .......... ......................... 583Immediate Precursors of the rRNA Molecules .......... ......................... 584Function of the Precursor-Specific Sequences .......... ......................... 587Post-Transcriptional Modification of Nucleosides ......... ...................... 588Interaction of the rRNA with the Ribosomal Proteins ............................ 589

CONCLUDING COMMENTS ................... ........................ 590SUMMARY ........................................... 591ACKNOWLEDGMENTS .............. ............................. 592ADDENDUM IN PROOF.593LITERATURE CITED ............................................................ 593

INTRODUCTION mer of amino acids, is accomplished by acomplex interaction of numerous macromolecu-

Gene expression is generally considered in lar components which are collectively referredterms of two separable events, transcription and to as the translation apparatus. This collectiontranslation. The process of the transcription of includes the various aminoacyl transfer RNAdeoxyribonucleic acid (DNA) into informa- (tRNA) species; dissociable protein factors in-tional ribonucleic acid (RNA) molecules is car- volved in polypeptide chain initiation, elonga-ried out by RNA polymerases and regulated by tion, and termination; and, finally, the ribo-the various factors governing initiation or termi- some, which is the hub of the translationnation of RNA chains. The second event, trans- apparatus. The ribosome furnishes the surfacelation of informational RNA into a linear poly- upon which protein synthesis takes place, and it

562

on May 17, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

http://mmbr.asm.org/

rRNA OF PROKARYOTES

probably also coordinates and provides cata-lytic assistance to the numerous interactingcomponents of the translation apparatus.Because of its obvious importance and the

continuing progress in exploring its extremecomplexity, the prokaryotic ribosome, and moreparticularly that of Escherichia coli, has be-come a perennial favorite for review. This essayalso concerns the prokaryotic ribosome, but itwill be limited in large part to a consideration ofonly one of its ingredients, the ribosomal RNA(rRNA). A few review articles have been di-rected toward ribosomal RNA. Particularly val-uable recent efforts are those of Attardi andAmaldi (4), which are concerned with bothprokaryotic and eukaryotic rRNA, and of Bur-don (33), which consider only eukaryotic rRNA.Most reviewers of protein or RNA synthesis alsoconsider, at least in passing, rRNA. Many ofthese reviews will be referred to during thecourse of this essay. For the benefit of thosereaders who have only passing interest in therRNA of prokaryotes, I have included a sum-mary of what I consider to be the most note-worthy points (see Summary). The literaturesurvey for this article was completed in thespring of 1973.

GENERAL FEATURES OFPROKARYOTIC RIBOSOMES

Diversity in Bacterial RibosomesThe detailed structure of bacterial ribosomes

(as exemplified by those of E. coli) and theirfunction during protein synthesis have beenreviewed extensively (60, 146, 166, 208, 215,264). In brief, active ribosomes at physiologicalMg2+ concentrations have the sedimentationproperties of approximately 70S particles (304).The ribosomes of prokaryotes are distinctlysmaller than those of eukaryotes, which sedi-ment as 80S particles (301). At low Mg2+concentrations (305) or when not engaged inprotein synthesis (134, 172, 286), the 70S bacte-rial ribosomes dissociate into two components,one about 50S and the second approximately30S in size. The anhydrous molecular weight ofthe 50S component of E. coli is 1.55 x 106, andthat of the 30S particle is 0.9 x 101 (121). Theseparticle weights represent the total of numerouscomponents, both RNA and protein. The 50Ssubunit is constructed of 30 or more proteins(146), which comprise about 25% of its mass;the remainder is RNA. The 30S ribosomalcomponent is approximately 60% RNA and 40%protein, including at least 21 identifiable poly-peptide chains (146). When mixed, the RNAand protein moieties of the 30S ribosomalsubunits spontaneously recombine to yield par-

ticles which are functional in protein synthesis(208, 209). Functional 50S subunits of Bacillusstearothermophilus (83, 210) and E. coli (179)also have been reconstituted from their compo-nent proteins and RNA, but the in vitro assem-bly process appears rather more complex thanthat of 30S subunits.Ribosomes from prokaryotes other than E.

coli are similar in their general properties.Taylor and Storck (301) have carefully exam-ined and compared the sedimentation proper-ties of ribosomes purified from 25 microorga-nisms, including members of the taxonomicorders Pseudomonadales, Eubacteriales, andActinomycetales as well as representatives ofthe blue-green algae. All possess characteristic"70S" prokaryotic ribosomes, although the pre-cise sedimentation values vary slightly amongthe different organisms. Less precise measure-ments have been made on ribosomes and theirsubunits from members of other taxonomicorders, including Myxobacterales (266), Myco-plasmatales (129) and even Rickettsiales (255,302). All possess the 70S ribosomes characteris-tic of prokaryotes, which dissociate at low Mg2+concentrations into 50 and 30S subunits. Sincethe ribosomes of such an evolutionarily diversecollection of organisms exhibit these properties,it seems unlikely that exceptions will be foundamong the prokaryotes.

Similarities among the ribosomes isolatedfrom prokaryotes of different taxonomic ordersand their structural differences from eukaryoticribosomes are not limited to gross particle size.The RNA components of all prokaryotic ribo-somes thus far examined are similar in dimen-sions, although they differ considerably in de-tailed structure (see below), and these RNAsize classes are quite distinct from those ofeukaryotic cells (160, 300). Furthermore, theindividual proteins associated with the prokary-otic ribosomes display certain common features.For example, Traut and his colleagues (10, 287)examined the ribosomal proteins from variousrepresentatives of the orders Pseudomonadales,Actinomycetales, and Eubacteriales by polyac-rylamide gel electrophoresis in the presence ofsodium dodecyl sulfate and found extensivesimilarities regarding the quantities and molec-ular weights of the ribosomal proteins. Thesefeatures were not common with proteins puri-fied from eukaryotic ribosomes (10). The similarsizes of the classes of prokaryotic ribosomalproteins do not reflect similar primary struc-tures, however, since in the absence of sodiumdodecyl sulfate the ribosomal proteins isolatedfrom the various orders do not have similarelectrophoretic properties (321), and they gener-ally display little or no serological cross-reac-

VOL. 37, 1973 563


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

tion. As might be anticipated, ribosomal pro-teins purified from organisms within a singletaxonomic family (Enterobacteriaceae) showconsiderable antigenic similarity (236).

It is clear, then, that prokaryotes have di-verged greatly in the detailed structures of thecomponents of their ribosomes. Nevertheless,functional homology has been retained at leastamong diverse members of the order Eubac-teriales. This is evident in the observation (43,298) that 50S ribosomal subunits from thegenus Bacillus, including the thermophile B.stearothermophilus (2), can combine with E.coli 30S subunits to form hybrid ribosomeswhich are active in protein synthesis. Further-more, Nomura and his colleagues (211) havereconstituted functional, hybrid 30S ribosomalsubunits from the component proteins andRNA from different eubacteriales, including E.coli, Micrococcus lysodeikticus, Azotobactervinelandii, and B. stearothermophilus. This ispossible even though the ribosomal proteins(321) and ribosomal RNA (211, 308) of theseorganisms have little resemblance in their de-tailed primary structures. The interaction be-tween the RNA and protein components leadingto functional aggregation is not a nonspecificphenomenon, since rRNA isolated from rats oryeasts is incapable even of particle formationwith the proteins of 30S subunits from E. coli(308). Therefore, even though the RNA andprotein components of ribosomes from diverseprokaryotes have deviated significantly in theirstructures during evolution, they have retainedthe details which permit their functional in-teraction. This suggests that comparative stud-ies of the components of ribosomes from varioustaxonomically diverse prokaryotes should be ofconsiderable value in identifying those func-tional details.

STRUCTURES OF THE PROKARYOTICrRNA MOLECULES

Prokaryotic rRNA MoleculesDeproteinization of purified, bacterial ribo-

somes generally yields three distinct RNA com-ponents. Two of these, classified by their ap-proximate sedimentation velocities as 23S and5S rRNA's, are derived from the 50S ribosomalsubunit (145, 247). The third RNA component,16S rRNA, originates from the 30S ribosomalsubunit (145). The molecular weights of 23, 16,and 5S rRNA's are, respectively, about 1.1million, 0.55 million, and 0.04 million (145, 247,280). These weights correspond to polynucleo-tide chain lengths of about 3,300 nucleotides for23S rRNA, 1,650 nucleotides for 16S rRNA, and

120 for 5S rRNA. All of the rRNA componentspresent in actively growing E. coli are generallyconsidered to be metabolically stable (91), anddensity-labeling experiments have shown thatthere is no exchange of rRNA among the availa-ble pool of ribosomes in the cell (133, 181).Ribosomal RNA molecules of approximately

23, 16, and 5S in size have been isolated from allprokaryotes examined, including members ofthe order Pseudomonadales (76, 160, 177, 300),numerous members of Eubacteriales (76, 145,160, 190, 247, 300; S. Sogin, Ph.D Thesis, Univ.of Ill., 1971), Myxobacteriales (7), Ac-tinomycetales (23), Rickettsiales (107), and var-ious representatives of the blue-green algae(123, 160, 300; W. F. Doolittle, personal com-munication). Detailed comparisons have re-vealed, however, that 23S and 16S rRNA mole-cules isolated from various genera differ slightlyin their sedimentation properties (300) and intheir electrophoretic mobility through polyac-rylamide gels (160). Evidence has even beenpresented that the 23S and 16S rRNA's of E.coli may each be separated into several mobilityclasses by polyacrylamide gel electrophoresis(59, 257), but this heterogeneity probably re-flects minor diversity in the secondary struc-tures of the rRNA molecules in solution (59).Similarly, variation in the physical properties ofthe rRNA molecules from diverse genera maynot reflect chain length but rather conforma-tional differences. A decision between thesealternatives could be achieved by comparingthe electrophoretic mobilities of the rRNA com-ponents after heating in the presence of formal-dehyde so as to irreversibly denature any sec-ondary structure (16). Certainly the chainlengths of 5S rRNA from different organismsvary slightly. The 5S molecules ofPseudomonasfluorescens (76) and E. coli (32) are 120 nucleo-tides in length, whereas those of Bacillus sub-tilis are composed of only about 115 to 116nucleotides (222; S. Sogin, Ph.D. thesis, Univ.of Ill., Urbana, 1971).An interesting anomaly regarding the struc-

ture of 23S rRNA is emerging from studies ofcertain photosynthetic organisms. Lessie firstreported (155) that ribosomes isolated fromRhodopseudomonas spheroides and Rhodop-seudomonas capsulata do not possess a 23S-sized rRNA component, whereas other athio-rhodaceae, including Rhodospirillum rubrumand the Rhodopseudomonas species palustrisand gelatinosa, contain normal complements of23 and 16S rRNA's. Instead of 23S rRNA,Rhodopseudomonas spheroides 50S ribosomalsubunits contain two RNA molecules, one about16S in size (0.53 x 106 daltons) and a secor'

564 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

approximately 15S in size (0.42 x 10' daltons)(177, 244). It has been claimed that thesemolecules might be fragments of 23S rRNAgenerated during purification, and indeed 23S-sized molecules may be isolated from R. sphe-roides (19, 294), but these are dissociable byheat or formamide treatment into the 16 and15S "halves" (177, 244). Moreover, Marrs andKaplan (177) have shown that an RNA moleculeof intact, 23S size in fact transiently exists in R.spheroides 50S ribosome subunits, but that it iscleaved during growth to the smaller compo-nents. Lack of an intact 23S rRNA moleculedoes not affect the function of the R. spheroidesribosome (177), so it has been suggested that thecleavage products remain noncovalently associ-ated.The blue-green alga Anacystis nidulans also

cleaves its 23S rRNA into two fragments (70,293). The larger of these is about 0.88 x 106daltons, and the smaller is 0.17 x 10. daltons. Itis quite clear in this case that the fragments arenot an artifact of isolation. The cleavage processis very slow relative to ribosome maturationtimes; the half-life of intact 23S rRNA is severalhours (70). Furthermore, scission of the 23SrRNA is retarded by inhibitors of energy metab-olism, which implies that it is closely coupled togrowth rather than being the result of randomnuclease action (70). Ribosomal RNA labilityalso has been reported for species ofAgrobacter-ium (103) and for numerous eukaryotes (119,143, 283). The utility to these organisms ofcleaving the rRNA remains obscure.

rRNA Integrity and Ribosome FunctionThe existence of organisms which cleave their

23S rRNA molecules, for whatever reason, sug-gests that rRNA intactness is not necessary forribosome function. Apparently it is not. Severalinvestigators (34, 61, 92) have examined theeffects of ribonuclease treatment on the capac-ity of ribosomes from E. coli to carry outpolypeptide synthesis, with the conclusion thatas long as Mg2+ concentrations are sufficientlyhigh to stabilize the digested ribosomes, theyremain functional (34). Probably the divalentcations are chelated by adjacent polynucleotidefragments, and hence the structural integrity ofthe rRNA-protein aggregates is maintained(34). Unfortunately, all of these investigationsexamined only the capacity of digested ribo-somes to participate in polyuridylate-directedpolyphenylalanine synthesis. It would be ofinterest to test their ability to translate faith-fully a heteropolymer.

Eukaryotic AnalogyIt is instructive while considering the prokar-

yotic rRNA components to consider briefly theirevolutionary homologues among the eukaryotes.80S ribosomes from animal cells are com-

posed of two subunits, one about 60S and asecond about 40S in size (301). Deproteinizationof the smaller, 40S subunit yields one RNAmolecule, about 18S in size and 0.7 x 106 inmolecular weight (160, 300). The 60S subunityields three distinct RNA molecules. Thelargest of these is about 28S (300). This corre-sponds to a molecular weight of about 1.7 x106, although the exact molecular weight mayvary somewhat according to the animal source(160). The larger rRNA component of loweranimals and plants may be substantially smallerthan that of mammals, with molecular weightsas low as 1.4 x 106 recorded (160). The 60Sribosomal subunits, like the prokaryotic 50Ssubunits, also contain 5S rRNA (0.04 x 106daltons) as well as another low-molecular-weight RNA component which originally wastermed "7S" rRNA (228). "7S" rRNA, which isactually 5.5S and about 0.05 x 106 daltons, isintimately associated with 28S rRNA; it is re-leased by treatments with disrupt hydrogenbonding, for example by high concentrations ofurea or by heating in solutions of low ionicstrength (228). Consequently, this RNA specieshas been referred to as "28S-associated" (28S-A) rRNA (317).The ribosomes of eukaryotes, therefore, have

certain fundamental physical properties whichdistinguish them from the prokaryotic ribo-somes. The ribosomes derived from eukaryotesare significantly larger than those of prokary-otes; the same is true of the two high-molecular-weight rRNA components, 28 and 18S rRNA's.Finally, there are two low-molecular-weightrRNA components of eukaryotic ribosomes, ascontrasted with only one in prokaryotes. It istempting because of their equivalent sizes toassert that the eukaryotic 5S rRNA is the func-tional homologue of the prokaryotic 5S rRNA,and that 28S-A rRNA is a relative newcomer onthe evolutionary scene, but this may be incor-rect (see below). When knowledge of the de-tailed function of the rRNA components of bothprokaryotes and eukaryotes becomes available,it will be possible to decide just how profoundthe differences are between their ribosomes.

Base Compositions of rRNA fromProkaryotes

The nucleotide compositions of the high-molecular-weight rRNA components isolated

565VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

from a wide variety of prokaryotes have beendetermined. For comparative purposes, some ofthese are listed in Table 1, along with the DNAguanosine plus cytidine (G+C) content of theorganism indicated. The listed values are thoseof the combined 16 and 23S rRNA components;the nucleotide compositions of the individualmolecules are nearly the same, but carefulanalyses have revealed minor differences (119,280, 322). The noteworthy feature of the availa-ble rRNA base compositions is their remarkablesimilarity, even though the genome base com-positions, as reflected in the G+C content of theDNAs, are extremely variable. Of course, therRNA molecules are defined by a very smallfraction of the cellular DNA, and so their genesdo not significantly contribute to the observedG+C values of the DNA. The implication of thedisparate G+C values for the genome DNA, ascompared with the rather constant G+C con-tent of rRNA, is that the requirements of thetranslation apparatus for a particular rRNAstructure are rather fastidious. That is, what-ever selective pressures have driven the ge-nomes of the organisms toward their presentG+C compositions have affected very little thenucleotide composition of the rRNA cistrons;the rRNA molecules are evolutionarily highlyconservative in that regard. Nevertheless, thenucleotide sequences of the rRNA moleculesfrom diverse microorganisms differ extremely(see below). Therefore, the conservative fea-tures of the rRNA molecules which are reflectedin their base compositions are not chiefly theprimary structures, but rather their secondaryor tertiary structures. Evidence that these latterfeatures probably largely define the specificityof the interactions of the rRNA molecules withthe ribosomal proteins is provided by the find-ings of Nomura et al. (211) that functional,hybrid 30S ribosomes may be reconstituted

from the 16S rRNA and the 30S subunit pro-

teins of organisms whose rRNA componentspossess relatively little nucleotide sequence ho-mology. The possible importance of RNA se-

quences not associated with helical regionsmust not be ignored, however. Modification ofbases not involved in stable base pairing inpurified 16S rRNA in solution abolishes theability of the molecule to combine functionallywith ribosomal proteins (208).The 16 and 23S rRNA components each

contain, in addition to the four "natural" bases,minor quantities of a variety of modified nucleo-sides. All of the modified nucleosides known tobe associated with the bacterial rRNA mole-cules, except one (5-ribosyluridine), are methyl-substituted. This distinguishes the minor com-ponents of rRNA from those to tRNA, whichalso include hypermodified and thiol-sub-stituted nucleosides. Ribosomal RNA from E.coli also has been reported to contain thiol-sub-stituted nucleosides (51), but this observationproved to be a consequence of contaminatingtRNA (42, 81). The chemistry and biology of themodified nucleosides, with particular referenceto tRNA, have been reviewed recently andconcisely by Hall (108).The majority of detailed studies of methyl-

ated nucleosides in rRNA from prokaryoteshave focused on E. coli. The methyl-substitutednucleosides in E. coli rRNA fall into two classes,the first containing methylated bases and thesecond possessing a methyl substitution at the2'-O-position on the ribose moiety of the nucleo-side. About 1% of the total nucleotides of E. colirRNA are methylated (73, 86). It has beenestimated (73, 86) that the 16S and 23S rRNAmolecules contain approximately 22 and 23methyl groups, respectively, although some-what lower values have been reported (113).About 90% of the methylated nucleosides in the

TABLE 1. Base compositions of rRNA and total DNA from various bacteriaa

Organism A G C U rG+C Reference NGA(%) Reference

Mycoplasma hominis 28.8 26.9 19.5 24.7 46.4 129 27-29 129Proteus vulgaris 26.2 31.4 21.7 20.7 53.1 183 -38 120Bacillus megaterium 25.2 31.1 22.0 21.6 53.1 217 38 120Streptococcus pneumoniae 25.4 31.0 19.1 24.2 50.1 322 39 120Vibrio marinus 15381 26.9 30.9 21.2 20.9 51.1 217 40 49Bacillus subtilis 25.9 31.0 22.3 20.8 53.3 66, 183 43 120Escherichia coli 25.1 32.3 21.4 21.2 53.7 280 52 120Bacillus stearothermophilus 22.7 35.5 23.7 18.0 5P.2 217 52 217Spirillum intersonii 23.4 33.2 23.8 19.5 57.0 217 55 120Aerobacter aerogenes 25.5 31.5 21.9 21.1 53.4 183 -56 120Pseudomonas aeruginosa 25.7 31.6 21.7 21.0 53.3 114,183 -65 120Micrococcus lysodeikticus 22.7 32.7 21.2 23.1 53.9 322 72 120

a Abbreviations: A, arginine; G, guanosine; C, cytidine; U, uridine.

566 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

rRNA of E. coli are base substituted; theremainder are 2'-O methylated (73, 86, 203).This contrasts with the rRNA of the highereukaryotes, in which most of the methyl groupsare contained on the ribose moiety of thenucleosides (28, 269). The 5S rRNA moleculesof both prokaryotes (31, 32, 76) and eukaryotes(6, 30, 89, 223, 315) lack modified nucleosides.The structures of the minor nucleosides of E.

coli rRNA, excepting those substituted only onthe ribose, and their distributions between the16 and 23S molecules are shown in Fig. 1. Nodetailed cataloging of the modified nucleosidespresent in the rRNA components of prokaryotesother than E. coli has been carried out, althoughmethylation of the high-molecular-weightrRNA components has been reported for avariety of diverse organisms (129, 273). Thefraction of modified nucleosides does not appearto be constant, however. For example, therRNA of Mycoplasma hominis contains onlyone-half as many methyl-substituted nucleo-sides as that of E. coli (129).The methylated nucleosides of the prokar-

yotic rRNA molecules are not randomly distrib-uted, but rather are associated with particularnucleotide sequences (86, 273); the methylationevents are highly specific and, therefore, pre-sumably have some function. Moreover, Soginet al. (273) have found that most of the methyl-ated nucleosides and the nucleotide sequencesin their immediate vicinity are identifiable inthe 16S rRNA components of bacterial specieswhich otherwise have relatively little similarityin primary structure. The occurrence of thesefeatures within the 16S rRNA molecules fromphylogenetically diverse bacterial species im-plies that not only the methylated nucleosidesbut also the nucleotide sequences surroundingthem are somehow important in the construc-tion or the function of ribosomes; they otherwisewould not have been conserved evolutionarily.Although some of the methylated nucleosides

appear to be important in rRNA function, thenature of the involvement is not known. Somemethylated components may be indispensable;others may merely refine properties alreadyinherent in unmodified nucleotide sequences(273). Still other modifications appear superfi-cially to be inconsequential. For example, themutation of E. coli to resistance to the antibi-otic kasugamycin is attributable (117, 118) tothe loss of a methylase specific for 16$ rRNA.16S rRNA from kasugamycin-resistant cellsconsequently lack the methylated, T1 ribonu-clease (RNase)-generated oligonucleotide me2-Ame2ACCUG, and yet the cells appear to grownormally; this methylated sequence plays little,

NH H,, CH3

~~N N

HC3 NiN>

ribose ribose

2-methyladenosine [23] N6-methyadonosine [23]

(72.327) (723.277)

H>, CH 3N

O -NHOH2Co,

OH OCH3

N4,02-dimthycytidine [16]

05,13)

0

KNN

H.-N N N

CH3 ri60s0

NH2

N. CH3

0 N .

ribose

H3C,,N,,CH3N

ribose

N6,N6-dimethyladenosine [16](72.35135277)

0

H3C-N N0[>

H2N N N)ribose

5-methylcytidino [16,23] l-methylguanosine [23]

(72,85) 35.277)

0

HN N

H 3C, "be

H 3C ribose

N2-methylguanosine [16,23] N2,N2-dimethylguanosine [?](72.35) 1277)

N

ibose

0 ,_ CH3

3 I~iribose

7-methylguanosine [16,23]05)

0

Hribose

5-methyluridine [23] 5-tB-D-Ribofuranosyl)uracil [16,23](72.5) (Psudouridine)

(723)

FIG. 1. Base-methylated nucleosides in the rRNAof Escherichia coli. Numbers in brackets after thenames of the compounds are the rRNA components inwhich the modified nucleosides have been positivelyidentified. References are listed parenthetically belowthe names of the compounds.

if any, role in ribosome formation or function.Nevertheless, these methyl substituents obvi-ously influence the properties of the kasugamy-cin-resistant ribosomes; they lose the capacityto interact with the drug.Another indication that the methylation of

the rRNA components may subtly influence thefunction of ribosomes comes from studies ofinduced drug resistance. Certain strains ofStaphylococcus aureus are physiologically in-duced by erythromycin to resistance to this andsome other drugs which interact with the 50Sribosome subunit. Lai and Weisblum (149) haveobserved that the induction to drug resistance isaccompanied by the appearance of N6-dimeth-yladenosine in the S. aureus 23S rRNA; thismodified nucleoside is not present in 23S rRNAof uninduced cultures. It has not yet beenproven that the presence of the N6-dimeth-yladenosine indeed confers resistance to theantibiotic, but certainly the rRNA is involved.23S rRNA purified from induced drug-resistantcells, but not that from sensitive cells, confers

VOL. 37, 1973 567


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

drug resistance upon reconstituted 50S ribo-somal subunits (150). Still another example ofan apparently unimportant modified nucleosideis 1-methylguanosine. Bjork and Isaksson (12)have isolated a mutant of E. coli which lacksthis component in its rRNA, and yet the mutantgrows normally, at least in the laboratory.

It should be obvious that our understandingof the role of the modified nucleosides in rRNAis limited at best. One hopes that as investiga-tors continue to chip away at the intricacies ofthe ribosome, this understanding will broaden.It is probable that studies of modified nucleo-sides in tRNA will also provide information onthe mechanisms by which methylation mayaffect polynucleotide structure and function.

Primary and Secondary Structures of therRNA

Many questions regarding the structures ofthe rRNA molecules, their interactions with theribosomal proteins, and their function duringprotein synthesis will only be understood whenthe complete nucleotide sequences of the mole-cules are available. This is truly a formidabletask because of the large sizes of 16 and 23SrRNA's and the tedious, albeit elegant, natureof the experimental techniques involved. Thistechnology has been reviewed by Sanger andBrownlee (252) and by Gilham (97). Neverthe-less, several laboratories have undertaken thedetermination of the primary structures of therRNA molecules, and a limited amount ofinformation has become available.At this time, the only completely known

nucleotide sequences of prokaryotic rRNA arethose of the 5S rRNA molecules of E. coli (31,32) and of Pseudomonas fluorescens (76). Thestructures of the 5S rRNA components fromthese two ostensibly phylogenically distant bac-teria differ substantially, but certain commonfeatures are present. Most notably, the middletwo-thirds of the molecules show a high degreeof similarity in nucleotide sequence, and ap-proximately 75% of the identical bases in thetwo molecules can be aligned (76). Also, the 5'-and 3'-terminal eight or nine nucleotides of both5S rRNA species are structurally capable offorming antiparallel, hydrogen-bonded"stalks", which confer amphora configurationson the molecules (31, 76). This same feature isinherent in the primary structures of 5S rRNAfrom eukaryotes (30, 89, 223, 315). The 5S rRNAmolecules of both the prokaryotes and eukary-otes also contain the sequence -G-A-A-C-,which at least in principle could interact byhydrogen bonding with the -G-T-I-C- sequenceof the common arm of the known tRNA mole-

cules (32). It is premature to assert that thesecommon features are somehow involved in 5SrRNA function, but as the structures of 5SrRNA from other prokaryotes become available,comparative examinations should prove usefulin defining the functional regions of the mole-cule.The 5S rRNA populations of E. coli and P.

fluorescens are largely homogeneous with regardto their primary structures, but minor heteroge-neity is evident. For example, E. coli strainMRE600 produces two forms of 5S rRNA (32),differing only at position number 13 (nucleotidepositions are conventionally numbered from the5'-terminus of the polynucleotide); one formcontains a G at that position and the secondcontains a U. Another E. coli strain, CA265(32), has two 5S rRNA forms which differ atposition 12. Similarly, the 5S rRNA populationof P. fluorescens is composed of two classes,differing at two positions (76). It is unlikely thatthese minor structural variations within the 5SrRNA populations of a given organism have anyfunctional significance. All of the rRNA mole-cules are defined by multiple genes (see below),which presumably are evolving more or lessindependently. Point mutations in any one ofthe multiple cistrons therefore would lead to amixed population of rRNA structures.

Unlike the example presented by studies oftransfer RNA, the availability of the primarystructures of some 5S rRNA molecules has notled to precise information regarding their sec-ondary structures. After the determination ofthe exact nucleotide sequences of 5S rRNAmolecules from E. coli and humans and thepresentation of physical data indicating thatabout 70 to 80% of the nucleotides in 5S rRNAare hydrogen bonded (17, 36, 53, 156), severalpossible models for the secondary and tertiarystructures of the 5S molecules were conceived(35, 156, 168, 237). When the nucleotide se-quence of 5S rRNA from P. fluorescens thenbecame available, attempts to fit it to theproposed models were more or less unsuccessful(76). Jordan (131) and Mirzabekov and Griffin(187) have evaluated experimentally these vari-ous models by subjecting E. coli 5S rRNA tovery gentle digestions with ribonuclease andexamining the products of hydrolysis. SinceRNase cleaves only at nucleotides which are notinvolved in hydrogen bonding, the regions of themolecule which are relatively resistant to diges-tion should be those which are involved in somesort of stable secondary structure. The resultswere not completely compatible with any of theproposed models. Therefore, the secondarystructure of 5S rRNA remains elusive, and in

568 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

fact Richards has concluded (242) from studiesof molecular models that there may be nounique secondary structure for 5S rRNA insolution. Also, it must be appreciated that thesecondary and tertiary structures of any of therRNA molecules in solution are probably signif-icantly different from their conformations whenpackaged within the ribosome in juxtapositionwith potentially interacting proteins. The nu-merous physical studies of the rRNA moleculesin solution and in the ribosome have beenreviewed recently by Spirin and Gavrilova(277). The 16 and 23S rRNA molecules haveabout the same degree of secondary structure insolution and in the ribosome, but the associa-tion with the ribosomal proteins imparts to theduplex regions of the RNA considerably morestability to thermal denaturation than observedwith RNA in solution (96, 182). Furthermore,23S rRNA in solution occupies about twice theeffective volume that it assumes in the ribosome(182); the tertiary structures of the molecule insolution and in the particle are certainly quitedifferent. Evidence also has been presented(291) that the binding of ribosomal proteins to16S rRNA may markedly influence its second-ary structure.The nucleotide sequences of the 16 and 23S

rRNA molecules are very difficult to evaluatebecause of their large sizes. However, throughthe efforts of several investigators (23, 80, 84,253, 254), the sequence of 16S rRNA from E. colishould be available soon. The nucleotide se-quence of about 70% of the molecule, distrib-uted over several large fragments, is more or lesscompletely known (80, 84, 85). The determinedsequence is compatible with physical measure-ments (52, 111) indicating that about 60 to 70%of bases in 16 and 23S rRNA's are involved inhydrogen bonding. Fellner and his colleagueshave suggested that the consequent secondarystructure is likely predominantly "local" (85).That is, adjacent sequences of 10 to 30 nucleo-tides tend to be approximately complementary,so that the molecule in solution, and presum-ably in the ribosome, might be envisioned as abewildering cluster of hairpin loops.The partial sequence of the 16S rRNA of E.

coli, as that of the prokaryotic 5S rRNA mole-cules, contains several positions which are vari-able (85, 192, 329). These, too, probably are aconsequence of the multiple genes specifyingrRNA. It is noteworthy that most if not all of thevariable positions are clustered. This implieseither that the variable regions are functionallyunimportant and therefore structural drift ispermissible, or that these areas of the 16Smolecule require very precise primary struc-

tures and that any successful variation must beaccompanied by compensatory changes in theimmediate vicinity. Fellner and his associates(85) also have pointed out that the primarystructure of 16S rRNA, as that of 5S (32) andpossibly 23S rRNA's (86), appears to contain asignificant amount of repetitive sequences. Themeaning of these sequences is not known, butprobably the repetitive sequences have equiva-lent functions, for example as common portionsof binding sites for the ribosomal proteins.Two approaches toward defining the regions

within the 16S rRNA sequence which interactspecifically with particular ribosomal proteinsare being made by a number of investigators.One of these experimental approaches involvesthe binding of purified ribosomal proteins topurified 16S rRNA and then treating the com-plex with ribonuclease. By virtue of its associa-tion with the protein, the specific, interactingregion of the rRNA is protected from digestion.The aggregate of the rRNA fragment and theribosomal protein then may be isolated, and theinvolved RNA sequence may be determined(260, 261). In a second type of approach, 16SrRNA fragments of known sequence are testedfor their capacity to bind particular ribosomalproteins (333). This type of analysis, coupledwith continued evaluation of the 16S rRNAprimary structure, should provide within thenext few years a reasonably complete pictureof the smaller ribosomal subunit of E. coli.

Question of Functional Heterogeneity in therRNA Molecules

The finding that certain of the proteins asso-ciated with the 30S ribosomal subunit of bacte-ria are present in less than unimolar concentra-tions led to some consideration that individualribosomes might be differentiated in their func-tion (209). If extensive heterogeneity in rRNAprimary structure exists, then this also wouldsuggest that there is functional differentiationof ribosomes. However, there is no evidence atthis time for such heterogeneity within therRNA population of a given prokaryote. It wasnoted above that minor variability exists in therRNA primary structures, but that this proba-bly reflects the independent evolutionary driftof the multiple rRNA genes. There is onefrequently cited report (3) that nutritional con-ditions may influence rRNA structure, but thisfinding is almost certainly incorrect for techni-cal reasons.As an assay for rRNA heterogeneity, Aronson

and Holowczyk (3) examined the relative quan-tities of particular oligonucleotides released bypancreatic RNase digestion of 31P-labeled 16S

569VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

rRNA purified from ribosomal subunits of bothE. coli and Pseudomonas aeruginosa. No signifi-cant differences were found in the oligonucleo-tide contents of 16S rRNA from cells grown in"poor" versus "rich" media when the isotopiclabeling times were lengthy. When, however,cultures were shifted from a low growth rate to ahigher one by addition of a better carbon sourceand the cultures were labeled with [32P]ortho-phosphate for only 5 to 6 min, then the quan-tities of certain labeled oligonucleotides re-covered from 16S rRNA varied significantlyfrom those labeled for lengthy time periodsduring steady-state growth. It was cautiouslyconcluded that perturbation of balanced growthresults in a temporary change in the pattern ofrRNA synthesis.

In retrospect, the observed differences inoligonucleotide recoveries almost certainly werenot a consequence of differences in the structureof the 16S molecules. Instead, the brief labelingperiod, coupled with the isolation of labeledRNA from 30S ribosomes, would have resultedin the purification of 16S rRNA which was notuniformly labeled along its molecular length.The time required for synthesis of a 16S rRNAmolecule is about 30 s, and an additional 4 to 5min are required for the completed rRNA chainto appear in mature 30S ribosomes. Therefore,in a pulse labeling experiment, the synthesis ofmost radioactive RNA molecules isolated frommature ribosomal particles would have beeninitiated before addition of the isotope. Conse-quently, a steep gradient would exist in thespecific radioactivity (32P/mole of nucleotide) ofindividual nucleotides along the molecularlength of the 16S rRNA, with the gradientincreasing toward the 3'-terminus (the terminalpoint of transcription) of the molecule. Thismeans that Aronson and Holowczyk were effec-tively comparing radioactive oligonucleotidesfrom the (pulse-labeled) 3'-proximal regions ofthe 16S molecule to those of the uniformlylabeled (steady-state) molecule. They of coursediffered. This conclusion is borne out by theobservation of these investigators that withprogressively increasing labeling times, the oli-gonucleotide content of pulse-labeled 16S rRNAprogressively approached that of the uniformlylabeled molecule.

Phylogenetic Differences in the rRNA ofProkaryotes

The rRNA molecules from taxonomically di-verse prokaryotes are superficially very similarin their properties. Equivalent rRNA compo-nents are similar in base composition and size(see above), and, more important, ribosomes

reconstituted from RNA and proteins fromdifferent prokaryotes are more or less functional(211). These facts imply that the evolution ofthe rRNA molecules has been rather conserva-tive. Nevertheless, it is known that extensivedepartures in the primary structures of therRNA components of diverse prokaryotes haveoccurred.

Determinations of the primary structures ofthe rRNA components from diverse organismswill be invaluable in achieving an understand-ing of rRNA function. Those portions of therRNA molecules which are required for func-tion, whether they are primary or secondarystructural features, should be ubiquitous,whereas less important regions will appearhighly diverse in structure. It is desirable forthis sort of comparison to have detailed, pri-mary structures available, but for the timebeing this is experimentally feasible only with5S rRNA. In principle, considerable useful in-formation could be obtained from phylogeneti-cally diverse 16S and 23S rRNA molecules bycomparison of large (and therefore probablyunique) oligonucleotides released from the mol-ecules by nuclease digestions and by compari-son of the regions of the rRNA molecules whichinteract with specific ribosomal proteins. Theseapproaches are currently being undertaken byWoese and his colleagues (273, 274).Although it is not yet possible to compare the

detailed structures of the high-molecular-weight rRNA molecules from diverse prokary-otes, the extent of their similarities and dissimi-larities can be evaluated approximately byDNA-RNA hybridization. The methodologyand interpretation of nucleic acid hybridizationexperiments have been reviewed recently byGillespie (98) and Kennell (136). Two types ofDNA-RNA hybridization experiments, whichprovide comparable results (296), have beenused most frequently in evaluating the related-ness of rRNA from various prokaryotes. Thefirst of these is hybridization saturation, inwhich constant amounts of DNA are annealedwith increasing quantities of radioactive rRNA.When no more radioactivity enters a "true"(RNase-resistant) hybrid, then the complemen-tary sequences in the DNA are saturated andthe fraction of DNA complementary to therRNA may be calculated. When DNA andrRNA from different organisms are annealed,then the amount of RNA hybridizing with theheterologous DNA provides some measure ofthe relatedness of the primary structures of therRNA molecules from the two organisms. Thesecond approach, that of hybridization compe-tition, is to anneal DNA of a given organism

570 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

with radioactive, homologous rRNA in the pres-ence of a large molar excess of unlabeled,heterologous rRNA. If the rRNA's from the twoorganisms have any sequence similarities, thenthe unlabeled rRNA competes with the labeledrRNA for the complementary DNA sequences.The difference in the amounts of labeled, ho-mologous RNA hybridizing in the absence andin the presence of the competing, heterologousrRNA defines the extent of similarity betweenthe two rRNA populations.

Several groups of investigators have usedthese methods to delineate the extent of diver-sity of prokaryotic rRNA primary structures(67, 75, 219, 296). Perhaps surprisingly, in viewof their superficial similarities, the primarystructures of the prokaryotic rRNA moleculesmay vary considerably. For example, E. coliand B. subtilis 16 and 23S rRNA's have about20 to 30% of homology (75, 219, 296); Strep-tomyces griseus and B. subtilis have only about2% common sequences in their rRNA's. Bacte-ria from the same genera possess extensiverRNA homology, but often the results indicate,once again, the arbitrary nature of bacterialclassification. For example, Marmur and hiscolleagues (74) and Doi and Igarashi (67) haveconsidered rRNA homologies within the genusBacillus, which is, to be sure, a catch-all genus.The observed rRNA homologies ranged from 50to 100%, whereas total DNA (as represented bythe RNA products of in vitro transcription)from the various species displayed far less simi-larity in primary structure (75); rRNA structureindeed is more conservative, evolutionarily,than the bulk of the information residing in theDNA. Quantitatively the rRNA of B. polymyxa(75) is about as closely related to that of B.subtilis as is the RNA of Alcaligenes faecalis orof Staphylococcus epidermidis (296). Therefore,by the criterion of DNA-RNA hybridization, thegenus Bacillus should be broken into severalother genera. On the other hand, certain orga-nisms bearing little superficial resemblancemay have considerable homology in their rRNAcomponents. E. coli rRNA saturates Vibriometschnicovii DNA to the same extent as thehomologous RNA, although the two organismshave little structural similarity in their messen-ger RNA (mRNA) (pulse-labeled RNA) popula-tions (227). Pace and Campbell (219) havecompared by hybridization competition therRNA homologies of numerous organisms to twolittle-related members of the eubacteriales, E.coli and Bacillus stearothermophilus, and theyfound that rRNA populations more closely re-lated to the rRNA of B. stearothermophiluswere less related to that of E. coli, and vice

versa. This reciprocal relationship implies thatthere exists a spectrum of relatedness amongthe rRNA molecules of the prokaryotes exam-ined and that these organisms may have acommon ancestor. Therefore, the phylogeneticapproach to evaluating structure-function rela-tionships in prokaryotic rRNA is a valid one.The outcome of heterologous DNA-rRNA hy-

bridization competition or saturation experi-ments is generally expressed as "percent relat-edness," and it measures the degree to whichheterologous rRNA is capable of competitivelypreventing a certain fraction of homologousrRNA from annealing to complementary DNAsequences, or alternatively, the extent to whichheterologous RNA is capable of interacting withthe complementary DNA sequences in a suffi-ciently precise fashion to confer RNase resist-ance on portions of the RNA molecule. It is notyet clear whether exact base pairing of a hetero-polymer DNA-RNA hybrid is necessary forRNase resistance of the hybrid (98, 136), so the"homologous" regions of the rRNA moleculesfrom different organisms in fact may not beidentical. Also, it is effectively impossible todecide whether or not the "homologous" regionsof the rRNA of different organisms have thesame linear organization in the molecules. Thedifficulties of interpreting DNA-RNA hybridi-zation results in terms of primary structuralhomologies between organisms is exemplifiedby two sorts of experiments. The first of theseis the observation by Pederson and Kjeldgaard(227) that rRNA from E. coli will form RNase-resistant hybrids with the entire rDNA of Vibriometschnicovii. However, the rate of formationof the heterologous DNA-RNA hybrid is onlyabout one-fourth that of the homologous pair.This means that more DNA-RNA collisions arerequired to produce a stable heterologoushybrid than with either homologous pair. Thehomologous regions are scattered throughoutthe rDNA (227), or, alternatively, the heterolo-gous DNA-RNA hybrids in fact are not com-posed of perfect complements, even thoughthey are resistant to ribonuclease digestion.Therefore, each collision between E. coli rRNAand V. metschnicovii DNA would have a lowerprobability of forming a hybrid than would thehomologous pair.A second example of experimental evidence

for the ambiguity of heterologous DNA-RNAhybridization experiments is the apparent ho-mology between 16 and 23S rRNA's in someorganisms. It is fortunate that the first DNA-RNA hybridization tests for the uniqueness ofthe 16 and 23S genes were performed withBacillus megaterium (327). With the rRNA of

VOL. 37, 1973 571


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

this organism (327) and other members of thegenus Bacillus (75; B. Pace, Ph.D. thesis, Univ.of Ill., Urbana, 1968), the 16 and 23S rRNAmolecules form hybrids with only their respec-tive genes, and neither molecule competes withthe other in hybridization tests. On the otherhand, 16 and 23S rRNA from E. coli (5, 171; B.Pace, Ph.D. thesis, Univ. of Ill., Urbana, 1968)or Alicaligenes faecalis (B. Pace, Ph.D. thesis,Univ. of Ill., Urbana, 1968) do display consider-able homology in hybridization saturation orcompetition experiments. This result is notattributable to cross-contamination of therRNA preparations, but careful hybridizationcompetition experiments do show that the affin-ities of 16S rRNA for the 23S genes and 23SrRNA for the 16S genes are only a fraction of thehomologous affinities (171, 231; B. Pace, Ph.D.thesis, Univ. of Ill., Urbana, 1968) and thatheterologous hybrids are substantially moreunstable than the homologous ones. One is leftwith the feeling that primary structural homolo-gies determined by DNA-RNA hybridization donot necessarily mean that the "homologous"sequences in fact are identical, but rather thatthey are only similar at best.McCarthy and his colleagues (9, 189) realized

the difficulties of evaluating the structural re-latedness of rRNA molecules from differentorganisms by DNA-RNA hybridization, andthey have proposed a different approach, onewhich involves annealing rRNA from a givenorganism to DNA from other organisms andthen testing the stability of any resulting hy-brids to dissociation by heat. The temperaturesrequired to release RNA from hybrids is somefunction of the number and type of interacting,complementary base pairs; DNA-RNA pairswhich conform more perfectly to one anotherare considerably more stable to thermal denatu-ration than if the complementary fit is a poorone. Therefore, a hierarchy of relatedness of anyrRNA to that of other organisms may be estab-lished on the basis of the temperatures at whichthe rRNA is released from hybrids formed withthe DNA of those organisms. Unfortunately,this method does not lend itself well to quanti-tative evaluation because the thermal transi-tions are not sharp. This implies that thesequences distributed along a heterologousrRNA-DNA hybrid have a wide spectrum ofrelatedness.To summarize, it appears that although the

rRNA molecules of prokaryotes have manysuperficial resemblances, they in fact may differconsiderably in their detailed, primary struc-tures. Nevertheless, the rRNA molecules fromdiverse organisms do possess more or less exten-

sive similarities in their nucleotide sequences,as revealed by DNA-RNA hybridization. Atfirst glance it might appear that rRNA struc-tural similarity would provide a useful index forclassification purposes, but the available meansfor determining sequence homologies are notvery precise. However, DNA-RNA hybridiza-tion competition is sufficiently straightforwardthat it can be used as a taxonomic tool, particu-larly at the family level. The method, in fact,has been used to bolster a reclassificationscheme for the genus Desulfovibrio (218).

Prokaryotic Origin for the EukaryoticOrganelles?

In passing, I would like to consider briefly thegeneral features of the ribosomes and rRNAfrom mitochondria and chloroplasts, for thereason that both of these organelles often areconsidered to be derived evolutionarily fromprokaryotic "parasites" (45, 175, 196, 241). Thissupposition is based in large part upon ostensi-ble similarities between the translation appa-ratus of the organelles and that of prokaryotes.The case for the prokaryotic origin of chloro-plasts is a reasonably sound one. The ribosomesof chloroplasts are clearly prokaryotic in charac-ter; the 70S chloroplast ribosomes (285) dissoci-ate into 50 and 30S subunits which contain,respectively, bacterial-like 23S (153, 161) and5S rRNA's (285) and 16S rRNA (153, 161).Similar to some of the photosynthetic prokary-otes, the 23S chloroplast rRNA is subject topost-transcriptional scission (153). Probablymost important is the fact that the ribosomalsubunits of chloroplasts from both lower (154)and higher (104) plants are capable of formingactive hybrid ribosomes with the ribosomalsubunits of E. coli. Furthermore, the rRNAcomponents of Euglena chloroplasts display arelatively high degree of structural similarity tothose of certain blue-green algae, as assayed inDNA-RNA hybridization tests (234).The ribosomes of mitochondria, on the other

hand, are quite unlike those of the prokaryotes;their physical properties have been reviewedrecently by Borst and Grivell (22). The mitori-bosomes of the ascomycetes are the most similarto the ribosomes of prokaryotes, but they arestructurally distinct. The mitoribosomes, theirsubunits, and their high-molecular-weightrRNA components are 10 to 20% larger inmolecular size than those of E. coli (22), theylack an identifiable 5S rRNA molecule (159),and, unlike the ribosomal subunits of chloro-plasts, those of the ascomycetes cannot formfunctional hybrid ribosomes with the subunitsof E. coli ribosomes (104). The mitoribosomes of

572 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

the vertebrates are even more unlike the prokar-yotic ribosomes. The unit ribosomes are 50 to60S in size, depending upon the animal source.They dissociate into subunits of about 40 and30S which contain, respectively, rRNA mole-cules only about 16 to 18S and 12 to 14S inparticle size (22). In brief, the properties of themitochondrial ribosomes and their rRNA com-ponents do not support the notion that mito-chondria are derived from a prokaryotic symbi-ote. This thesis has been critically reviewedrecently by Raff and Mahler (238). An alterna-tive explanation is, of course, that the mito-chondrial ribosomes have evolved in the host toa sufficient extent that they no longer arerecognizable as prokaryotic.

GENES SPECIFYING rRNA

Gene Dosage for rRNAWhen purified ribosomal RNA molecules

from several bacterial species initially werehybridized with their homologous DNA by Yan-kofsky and Spiegelman (325, 326), it was notedthat considerably more rRNA specifically an-nealed than would be anticipated if only onegene were specifying each rRNA component. Inall prokaryotes examined, the fraction of thegenome which is complementary to rRNA isremarkably constant, comprising about 0.3 to0.4% of the total DNA (135, 213, 220, 251), orabout 0.6 to 0.8% of the potentially availablegenetic information. In E. coli, in which thegenome size is known with some precision, theamount of DNA hybridizing with rRNA corre-sponds to about six copies of each of the rRNAgenes. For Mycoplasma, whose genome is onlyabout 20 to 25% the size of that of E. coli, theamount of rDNA present is equivalent to onlyone gene for each of the rRNA components(251). Furthermore, DNA-RNA hybridizationsaturation experiments performed with the in-dividual rRNA components, 16 and 23S (5, 213,271; B. Pace, Ph.D. thesis, Univ. of Ill., Urbana,1968) and 5S (220) rRNA's, have shown that thenumber of genes specifying each is the same.These observations satisfactorily account forthe equimolar accumulation of the three rRNAmolecules in the cell without requiring thepreferential production of any of them.

Mapping the rRNA Genes in E. coliDetermination of the locations of the rRNA

genes within the bacterial chromosome hasproved to be technically difficult. There are noknown bacterial mutants which are ascribableto the rRNA genes, and any search for suchmutants is a rather speculative undertaking

because of the multiplicity of the rRNA genes;most deleterious mutations would be recessiveand therefore undetectable. Even in the absenceof appropriate mutants, however, considerableexperimental athletics have provided a rela-tively good map position for at least part of therDNA of E. coli.

Initial attempts to define the location of therDNA within the genome of E. coli were madeby Chargaff and his colleagues (248, 249), whosupposed that as the rDNA is replicated moregenes specifying rRNA would be available, andthat an increase in rRNA output would immedi-ately result. Consequently, synchronous cul-tures of E. coli were pulse-labeled with [32p]-orthophosphate at different intervals during acell division cycle, and the base compositions ofthe radioactive RNA synthesized during thepulse periods were evaluated. Two waves ofsynthesis ofRNA with base compositions biasedtoward rRNA were observed. After examiningtwo Hfr strains, which were thought to havedifferent initiation points for DNA replication,map positions were suggested (249), assumingunidirectional DNA synthesis at a uniform rate(but see below). Re-evaluation (11) of theorigins of DNA replication in the Hfr strainsemployed in the experiments places one of theseclusters in the 70- to 80-min region of the Taylorand Trotter map of the E. coli chromosome(299) and the second cluster at 35 to 50 min. Inretrospect, these experiments of Chargaff andhis collaborators yielded approximately the cor-rect answer, at least with regard to the majorrDNA cluster in the 70- to 80-min region, but itis not clear why. Doubling the amount of rDNAin cells apparently does not double the con-tinued output ofrRNA (56, 62, 64), although therate of rRNA production may be enhancedtemporarily as the rDNA is duplicated (56).

Cutler and Evans (56) also have used syn-chronously dividing populations of E. coli toevaluate the map position(s) of the rDNA.Unique portions of E. coli DNA were pulse-labeled with 5-bromouracil at intervals duringsynchronous division and then separated, byCsCl buoyant density centrifugation, from theless dense portions of the DNA which were notbeing replicated during the pulse interval. Tenunique parts of the E. coli DNA, comprising theentire genome (55), were isolated, and theirabilities to form hybrids with purified rRNAwere tested. Two regions were found to do so; itwas proposed that these are located in the 60- to75- and 10- to 25-min regions of the linkagemap.The use of synchronously replicating bacte-

rial populations to define the map position of

VOL. 37, 1973 573


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

the genes specifying rRNA is, at best, animprecise experimental probe. More impor-tantly, the basic assumptions of this approach,that the genome DNA replicates at a constantrate and in a unidirectional fashion, are proba-bly incorrect (224). An alternative avenue to-ward the determination of the map positions ofthe rRNA genes became available with theconstruction, by several investigators, of a largecollection of F-merogenotes. The basic ideahere, which first was exploited by Yu et al.(330), is that any strain containing an F-merogenote is partially diploid for a givenregion, and if the episome contains the rDNA,then the rRNA gene dosage, as measured byDNA-RNA hybridization saturation experi-ments, is increased over the haploid value by aquantity corresponding to the number of rRNAgenes present in the episome. Yu et al. (330)examined the rRNA gene dosage for strains ofE. coli carrying a variety of F-merogenotes,which collectively spanned the entire chromo-some. Contrary to the results of the indirectexperiments suggesting two locales for therDNA, it was concluded that only one region ofthe chromosome was complementary to rRNA.By comparison of the amounts of rRNA hybrid-izing to the DNA of strains carrying variousepisomes, the genes defining rRNA could belocalized in the 74- to 77-min region of the E.coli map. However, Yu et al. examined only twoF-merogenotes which span the 10- to 70-minregion of the linkage map. Both of these epi-somes were about 30 min in length, correspond-ing to approximately 30% of the total genome. Ifthese episomes did not contain rDNA, then therRNA/DNA ratios observed in hybridizationsaturation experiments should have been lowerin the diploids than in the corresponding hap-loid strains; the episome ostensibly contributesno rDNA, but it does contribute to the totalmass of DNA in the cell. And yet, one of thediploids (15 to 40 min) contained normal pro-portions of rDNA, and the second (40 to 74min) had only slightly reduced levels of rDNA.It therefore might appear that both of theseepisomes in fact contain rDNA. Unfortunately,the communication presented does not containthe detailed hybridization saturation data, andso it is difficult to judge the seriousness of thediscrepancy between expected and observedresults.

It also was proposed that the rDNA mighteven be confined to the rbs-ilv span (74 to 74.7min), but this almost certainly is not the case.Purdom et al. (235) have presented reasonablyconvincing evidence that individual clusters ofrRNA genes, each probably comprised of one 16,

23, and 5S gene (148, 149), are separated bystretches of DNA at least 10 x 106 daltons. Thismeans that the total length of the rDNA mapregion, in order to contain five to six genes foreach of the rRNA molecules, would span at leasttwo map minutes.Birnbaum and Kaplan (11) also used F-

merogenotes in DNA-RNA hybridization exper-iments, and they, too, have determined thatpart of the rRNA genes are scattered throughoutthe 74- to 77-min region, excluding the rbs-ilvarea. This conclusion was drawn by hybridizingrRNA with the DNA of several E. coli F-me-rogenotes purified from Proteus mirabilis carry-ing the episomes. The G+C content of the P.mirabilis DNA is significantly lower than thatof E. coli F-merogenote DNA, and so theepisomes may be more or less purified from thegenome DNA. These investigators noted, how-ever, that only about half of the total rRNAcistrons could be accounted for within the 74- to77-min region, an observation which also proba-bly is compatible with the data of Yu et al., somore rRNA gene clusters remained to be addedto the map. Kaplan and his colleagues (312),through further experiments with E. coli F-merogenotes isolated from P. mirabilis, haveproposed that two doses of each of the rRNAgenes are located somewhere in the 54- to59-min span of the chromosome. The findingsupports the earlier suggestion of Jarry andRosset (128) that some of the genes specifying5S rRNA might lie in the 40- to 66-min region ofthe linkage map, in addition to the 74- to77-min area, and equivalent numbers of 16 and23S genes should accompany the 5S rDNA (seebelow). Also, Gorelic (102) has provided evi-dence for increased rDNA levels in an E. colistrain which is diploid for the 60- to 66-minregion, but this result could not be confirmed byUnger et al. (312).

It is not completely certain whether copies ofthe rDNA exist outside the 74- to 77- and 54- to59-min regions of the linkage map; they con-ceivably all are accounted for. However, epi-somes containing DNA outside these spansshould be examined more closely for any rDNAcontent. This would be done best by partiallypurifying the episomal DNA, perhaps as de-scribed by Birnbaum and Kaplan (11), beforehybridizing with rRNA. If the episome is iso-lated, its presence in the hybridization assays isguaranteed. Bacterial strains which are diploidfor rDNA conceivably are unstable with regardto at least those genes and possibly the entireepisome (330); the multiple rRNA genes in thechromosome probably are stabilized againstrecombinational deletion by the presence of

574 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

essential genes within the non-rDNA stretcheswhich separate the individual rDNA clusters.Further, the determination of the rDNA contentof an episome by hybridization saturation ex-

periments with the entire diploid genome isprobably not a very reliable method. It is our

experience that rRNA gene dosage values, de-termined through hybridization saturation ex-

periments with different preparations of rRNAand DNA from even the same strains of bacte-ria, may vary by 10 to 20% or even more.

Isolation of the episomal DNA before the hy-bridization tests minimizes quantitative uncer-

tainties implicit in the technique.

Mapping the rRNA Genes in B. subtilisThe applicability of DNA transformation in

B. subtilis has made the location of its rDNA,relative to other markers, rather more simplethan with E. coli, but the answer is less precisethan information gained from F-merogenotes ofE. coli. Oishi and Sueoka (213) and Marmurand his collaborators (74, 271) have used den-sity labeling of synchronously replicating B.subtilis DNA to locate at least most of therRNA genes. These experiments were per-

formed by inducing synchronous DNA replica-tion through germination of spores (213, 271) in5-bromouracil or D20-containing medium, or

by diluting stationary cultures pregrown inD20-containing medium into water-containingmedium (74). DNA of density indicative ofnewly synthesized material could then be puri-fied pycnographically from the remainder of thegenome, and its content of rDNA could bedetermined by DNA-RNA hybridization. Thetime of appearance of the rRNA genes in thenewly synthesized DNA then was correlated, bygenetic transformation, to the replication timesof various nutritional and antibiotic resistancemarkers. About 60 to 80% (271) of the rDNA was

located within the 25% of the genome replicat-ing initially; the remaining 20 to 40% of therRNA genes appeared to be replicated with theterminal 25% of the genome. The positioning ofthe major rDNA cluster possibly could be even

more narrowly located within the DNA regionlying 15 to 20% of the replication distance fromthe origin (74), but it certainly is distinct fromthe str locus (267). Therefore, in B. subtilis as inE. coli, there probably are at least two regions ofthe chromosome containing the rDNA clusters.

All of these mapping experiments with the B.subtilis genome are compatible with the inter-mingling or close association of the individualgenes specifying 16 and 23S (74, 213) as well as5S (271) rRNA's. Colli and Oishi (46) and Colliet al. (48) have refined these observations by

shearing B. subtilis DNA to small size (about3 x 105 daltons), isolating the fragments capa-ble of annealing with either 16S or 23S rRNA's,and then testing their ability also to hybridizewith the other high-molecular-weight rRNAcomponent or 5S rRNA. It was determined that16S-specific DNA fragments are also capable ofhybridizing with 23S, but not 5S, rRNA. On theother hand, 23S-specific DNA could be an-nealed with both 16 and 5S rRNA's. Thisdemonstrates that the three rRNA genes areclosely adjacent to one another and linked inthe order 16S-23S-5S. Presumably these indi-vidual clusters in B. subtilis, as in the gram-negative bacteria examined (235), are separatedby lengthy stretches of non-ribosomal DNA,which may contain various other genes (99) notnecessarily related to the ribosome (11, 299).

Isolation of the rDNAThe ready availability of quantities of the

rRNA molecules makes possible the isolation ofthe DNA regions complementary to them, and,indeed, the rDNA has been substantially puri-fied from several organisms (47, 140, 275, 311).Isolation procedures generally begin by shearingthe cellular DNA to fragments of sizes not toomuch larger than the combined molecularweights of the rRNA molecules. The fragmentsof DNA are then annealed with rRNA, and thehybrids are isolated from the remainder of theDNA by various means, including CsCl (275) orCs2SOI (containing Hg2e ions; 47) buoyantdensity centrifugation or chromatography oncolumns of hydroxylapatite (140) or deoxycho-late-saturated benzoylated diethylaminoethylcellulose (311). All of these procedures more orless cleanly resolve the various products of thehybridization reactions, which include single-strand DNA and RNA, duplex DNA, and thecoveted DNA-RNA hybrid. Also, the hybridiza-tion step may be prefaced by enrichment for therDNA by chromatographic means (47) or bypreferentially denaturing the majority of thegenome DNA, if its G+C content is sufficientlyless than that of the rDNA (295). This latterenrichment scheme is based on the fact that thetemperature at which duplex DNA denaturesincreases as a function of increasing G+Ccontent (176). Therefore, if the genome DNA isof lower average G+C content than the rDNA,then the bulk DNA may be heated to a temper-ature sufficiently high to denature a large por-tion of the non-ribosomal DNA. The remainingduplex DNA, now considerably enriched forrDNA, may easily be separated from the single-strand DNA.As their rationale for seeking to purify DNA

575VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

complements to the rRNA molecules, someinvestigators have offered the possibility ofusing these as templates for in vitro studies ofrRNA synthesis. It is unlikely that the productsof the rDNA purification schemes devised thusfar will be useful for such experiments. Only theDNA complements of the rRNA can be isolatedin quantity, and these are of variable length andcompleteness as a consequence of the necessaryshearing procedures involved in the purificationschedules. A more fruitful avenue toward invitro studies of rDNA transcription appears tobe the construction of the appropriate rDNA-containing F-merogenotes, even though sub-stantial quantities of non-ribosomal DNA mustbe tolerated.

TRANSCRIPTION OF THE rRNA GENES

rRNA Transcriptional UnitIn eukaryotic cells, coordinate synthesis of

the 28, 18, and probably 28S-A rRNA moleculesis guaranteed by the fact that the genes specify-ing each of these rRNA components are all partof the same "transcriptional unit". A transcrip-tional unit is a segment of DNA bounded bytranscription initiation and termination sites,which is read without interruption by any RNApolymerase molecule effectively beginning RNAsynthesis at the initiation site. A transcriptionalunit may be comprised of one or many genes,and if it contains a demonstrable operator locus,then it appropriately is termed an operon. In themammalian cell, the initial product of therRNA transcriptional unit is about 4 x 106daltons and 45S in molecular size, and it iscleaved in a series of endonucleolytic steps toyield the individual, mature rRNA molecules(4, 33). This compound precursor of the rRNAmolecules is readily demonstrable in pulse-labeling experiments (33).Such an elaborate scheme for production of

the rRNA molecules in prokaryotes was notsuspected until rather recently, since the pre-cursors of the mature prokaryotic rRNA compo-nents, as identified by pulse-labeling experi-ments, are only slightly larger than the maturemolecules (see below). This was generally con-sidered as weakly implying that the genesdefining the rRNA molecules were independenttranscriptional units. The first solid indicationthat this probably was not the case was pro-vided by the "transcriptional mapping" experi-ments of Woese and his collaborators (13, 14),which used the drug actinomycin D (AMD) tomeasure the sizes of transcriptional units gener-ating rRNA and tRNA in Bacillus subtilis.AMD binds at random to the DNA, interrupt-

ing the progress of RNA polymerase moleculesduring transcription, and so the sensitivity ofthe synthesis of any given RNA molecule toAMD should be a function of the size of theDNA segment preceding and including the genedefining that RNA molecule in its transcrip-tional unit. If cells are exposed to AMD concen-trations sufficient to inhibit RNA synthesis onlypartially, then the larger transcriptional unitsshould be preferentially inactivated by virtue ofthe larger target size which they present to thedrug. When the quantities of the individualrRNA molecules accumulating in the presenceof low concentrations of AMD were tested (14),it was found that the synthesis of 5S rRNA wasinhibited concomitantly with 23S rRNA andthat both of these were more sensitive than 16SrRNA to the effects of the drug. In view of thevery small size of the gene for 5S rRNA, if it infact were an independent transcriptional unit,then the synthesis of 5S rRNA should have beenrefractory to AMD concentrations sufficient toretard 90% of the 23S rRNA synthesis. Bleymanet al. consequently proposed (13) that the 5Sgenes in B. subtilis are promoter distal intranscriptional units containing at least the 23Sgenes as well.Armed with this information and the observa-

tions of Colli and Oishi (46) that the genes for 16and 23S rRNA's in B. subtilis are physicallylinked, other investigators used rifampin toprovide evidence that in E. coli the rRNAmolecules are derived from transcriptional unitsconsisting each of one 16S, one 23S, and one 5Sgene, which are read by the RNA polymerase inthat order (71, 72, 225). The drug rifampin (316)inhibits the initiation of RNA synthesis by theRNA polymerase, but it does not affect comple-tion of nascent chains. When rifampin and aradioactive precursor of RNA are simultane-ously added to growing bacterial populations,the amount of radioactivity appearing in anygiven RNA molecule is a consequence of the sizeof that molecule and the number of RNApolymerase molecules which transcribe its geneafter addition of the drug. In analyzing thetranscriptional organization of the rRNA genes,the necessary considerations are that equalnumbers of the genes defining 16, 23, and 5SrRNA's serve as templates in producing equalnumbers of the rRNA molecules. Therefore, thefrequencies at which the three genes are tran-scribed are the same, and the numbers of RNApolymerase molecules associated with each ofthe rRNA genes at any instant are proportionalto the sizes of the genes. To an acceptableapproximation, the genes specifying the rRNAmolecules are of the same length as the mature

576 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

rRNA species. Therefore, the amounts of iso-topic label appearing in any of the rRNAcomponents after residual RNA synthesis in thepresence of rifampin are predictable for anyconceivable transcriptional organization. Theoutcome of such experiments with E. coli isthat, whereas the 16S gene apparently is readonly by those RNA polymerase molecules resid-ing upon it at the time of addition of the drug,the amount of label appearing in 23S rRNA,relative to the amount in 16S RNA, could onlybe accounted for if the 23S gene is preceded inits transcriptional unit by a segment of DNAcorresponding in size to the 16S gene; not onlydo polymerase molecules associated with the23S rDNA deposit label in 23S rRNA, but theyalso associate with polymerase molecules readinto the 23S rDNA from the DNA segmentpreceding it (72, 225). Furthermore, considera-bly more label accumulates in 5S rRNA thanwould be expected if it comprised an independ-ent transcriptional unit; the amount of labeldeposited in 5S rRNA in the presence of rifam-pin is compatible only with the 5S gene beingpreceded in its transcriptional unit by a DNAsegment about the size of the combined 16 and23S genes (71, 225). Equivalent experimentswith B. subtilis (N. R. Pace and M. L. Pato,unpublished observations) have the same out-come, in support of the findings of Bleyman etal. (13). In this organism, too, the rRNA mole-cules are derived from transcriptional unitscontaining genes for 16, 23, and 5S rRNA's,which are read by the RNA polymerase in thatorder.The experiments with rifampin are cor-

roborated by other results. When rRNA synthe-sis is induced by restoration of an essentialamino acid to a starved amino acid auxotroph ofE. coli, the synthesis of 16S rRNA, as identifiedby DNA-RNA hybridization tests, commencesbefore that of 23S rRNA (144). Also, the timeduration of 16 and 23S rRNA synthesis in E. coliafter addition of rifampin is compatible withtheir genes being arranged in compound tran-scriptional units (24, 127, 186). Therefore, theorganization of the prokaryotic rRNA genes inpolycistronic transcriptional units is quite anal-ogous to the situation in eukaryotes, except thatthe prokaryotic 5S gene is included in thetranscriptional unit, whereas in eukaryotes agene defining the 28S-A ("7S") rRNA moleculeis included. The genes responsible for eukary-otic 5S rRNA production are not even associ-ated with the nucleolus, which is the site ofsynthesis of the 18, 28, and 28S-A rRNA mole-cules (4, 33). If this complex transcriptionalorganization for the rRNA genes has any func-

tional basis other than guaranteeing coordinatesynthesis of all of the rRNA molecules, then it isconceivable that the eukaryotic 28S-A rRNA,and not the 5S rRNA, is the evolutionaryhomologue of the bacterial 5S rRNA (72).

Interaction of the RNA Polymerase with therDNA

About 40 to 50% of the RNA being synthe-sized at any instant by rapidly, exponentiallygrowing cultures of E. coli is identifiable asrRNA in DNA-RNA hybridization competitionexperiments (135, 231). This implies that thesix or so transcriptional units specifying therRNA components in E. coli are read by theRNA polymerase with a frequency which is, onthe average, 50 to 100 times greater than thefrequency of transcription of genes specifyingthe several hundred (135) mRNA species re-quired for cell growth. Since one knows thenumber and size of the rRNA transscriptionalunits, the rate at which transcription occurs(about 50 nucleotides/s; 26, 173), and the num-ber of ribosomes (104 to 2 x 104) which must besynthesized during a division cycle of 30 to 40min, then the approximate number of RNApolymerase molecules associated with therDNA at any time may be calculated. Thisexercise yields a value of about 100 to 150 RNApolymerase molecules per rRNA transcriptionalunit. Since RNA polymerase molecules areabout 7.5 nm in diameter (185) and the rRNAtranscriptional unit is about 1.7 ,um in length(5,000 base pairs x 0.34 nm per base pair),about 70% of the absolute maximal number ofpolymerase molecules which can be accom-modated within the transcriptional unit in factare loaded on during rapid growth. The distancebetween adjacent polymerase molecules there-fore mus.t be only about 10 to 12 base pairs,whereas the polymerase itself occupies in theorder of 20 to 25 base pairs.The in vivo rate at which the RNA polymer-

ase synthesizes RNA in E. coli is generallyaccepted to be about 50 nucleotides/s (26, 173)at 37 C, so the times required for the construc-tion of the 16 and 23S rRNA chains would be,respectively, about 30 and 60 s. In contrast tothese expected times for the synthesis of therRNA components, Mangiarotti et al. (171) andAdesnik and Levinthal (1) have presented dataregarding the rates of appearance of exoge-nously supplied isotopic label in 16 and 23SrRNA's which, in their simplest interpretation,indicate that the times required for constructionof the two molecules in E. coli might be thesame. This conclusion is almost certainly incor-rect, however, and the reasons underlying the

577VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

observations are not yet completely clear. AsMangiarotti et al. (171) have discussed, thereare two possible models whereby the timesrequired for the synthesis of 16 and 23S rRNA'scould be the same. The first of these is that theprocession rate of the RNA polymerase on the16S genes is only half that of polymerasemolecules associated with the 23S genes. Sinceequimolar quantities of 16 and 23S rRNA's areproduced, equal numbers of polymerase mole-cules must then be associated with the twoclasses of genes, and the packing density ofpolymerase molecules on the 16S gene would betwice that of the 23S gene. Alternatively, equaltranscription times for 16 and 23S rRNA's couldbe accounted for if 23S rRNA were assembledfrom two, 16S-sized molecules which originatefrom different DNA segments. This sort ofscheme for the synthesis of 23S rRNA has alsobeen considered by Midgley (184). Therefore,two polymerase molecules would be active inproduction of the 23S rRNA halves for eachpolymerase reading the 16S gene. Both of thesemodels have definite predictions regarding theamounts of radioactivity deposited in 16 and23S rRNA's during residual synthesis in thepresence of rifampin, and neither model isindicated by such experiments (72, 225), whichwere described above. Furthermore, Bremerand Berry (24) have found that the persistenceof 23S rRNA synthesis after rifampin additionto cultures of E. coli is three times that of 16SrRNA, an observation compatible only with the23S rRNA synthesis time being twice that of the16S molecules and with the arrangement of the16 and 23S genes in a tandem transcriptionalunit. These results also suggested that onlyabout 20 s is required for the synthesis of a 16SrRNA molecule in E. coli. This is about thesame length of time (18 s) proposed by Zimmer-mann and Levinthal (332) for the synthesis of16S rRNA in B. subtilis. Both estimates yield achain elongation rate for the RNA polymerase ofabout 75 to 85 nucleotides/s, which is somewhatmore rapid than the generally accepted velocityof 50 nucleotides/s. This latter value was ob-tained by measuring the rate of synthesis of thetotal cellular RNA, however, and it thereforeessentially averages the chain elongation ratesof mRNA and rRNA. Therefore, the rate ofnucleotide addition to growing mRNA chainsindeed may be substantially slower than 50/s,whereas the rate of chain elongation of rRNAcould be in the order of 75 to 85 nucleotides/sduring rapid growth. Diffusion of ribosomesassociated with the nascent mRNA conceivablyslows the RNA polymerase in its procession,whereas the nascent rRNA chains, which are

not associated with ribosomes, would minimallyfetter the polymerase.

Initiation of rDNA TranscriptionThe RNA polymerase holoenzyme is com

posed of a "core" enzyme, which is constructedof four subunits and is responsible for RNAsynthesis, and a dissociable protein factor, a,which is required for meaningful initiation oftranscription. The chemistry of the RNA po-lymerase and the process of RNA synthesis hasbeen reviewed recently by several authors (8, 40,95, 162). One facet of RNA synthesis whichremains to be defined is that which determinesthe distribution of the RNA polymerase amongdifferent genes. This question is particularlypertinent to the synthesis of rRNA, since therRNA transcriptional units can be read withfrequencies 50 to 100 times greater than those ofmost genes.There are two obvious mechanisms which

might govern the very high relative rate atwhich the rRNA genes are read by the RNApolymerase. The first of these supposes that theaffinity of the RNA polymerase for the promoterof the rRNA transcriptional unit is 50 to 100times greater than its affinity for promoters ofDNA segments which define mRNA, and there-fore the frequency at which the rDNA is tran-scribed is correspondingly higher than the rateof production of any given mRNA molecule.The second possible scheme requires the exist-ence of a protein factor which would bind eitherto the RNA polymerase, activating it so that therDNA promoter may be recognized, or to therDNA promoter, to form a complex for whichthe RNA polymerase would have an exception-ally high affinity.At least in principle, a decision between these

two alternative models could be made by con-fronting purified E. coli DNA with purifiedRNA polymerase and examining the quantity ofrRNA which is produced relative to total RNA.When this sort of experiment was performed, itwas observed initially (231, 310) that little ifany of the in vitro product could be preventedfrom hybridizing with DNA by the presence ofcompeting, authentic rRNA. This indicatedthat little rRNA was produced and, therefore,that some sort of auxiliary factor probably wasrequired in order to enhance the affinity of theRNA polymerase for the rDNA. Travers and hiscolleagues (310) then reported the isolation of aprotein, termed #r, which could stimulate puri-fied RNA polymerase to preferentially producerRNA in vitro. This observation did not provereproducible (112), however, and #,r is nowconsidered to be a nonspecific stimulator of in

578 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

vitro RNA synthesis (162) whose exact functionin normal cells is unknown.More recently, several investigators have de-

termined that, contrary to the earlier reports,the purified RNA polymerases from E. coli(112), 230) and B. subtilis (125) are, in fact,capable of preferentially transcribing therDNA. With reasonably intact DNA, about 10%of the in vitro RNA transcripts are homologouswith authentic rRNA in DNA-RNA hybridiza-tion competition tests. It is not clear why rRNAsynthesis was not observed in previous experi-ments, but there were probably at least twocontributing factors (112). The first of these isthat the purified DNA originally used as thetemplate for in vitro RNA synthesis might havebeen partially single-stranded and sufficientlynicked by deoxyribonuclease so that only a verysmall fraction of the total RNA chains producedwas "physiologically meaningful". Disruptionof the DNA duplex causes aberrant transcrip-tion with regard to initiation and probablytermination. Consequently, any rRNA pro-duced in vitro would have been diluted by theaberrant transcripts beyond the sensitivity ofthe hybridization assays. The second factor inthe inability to detect in vitro rRNA synthesismight have been the nature of the DNA-RNAhybridization tests used in searching for therRNA. Experimenters who failed to detect anyrRNA among the in vitro transcripts (231, 310)used unlabeled rRNA in hybridization competi-tion with the RNA which was isotopically la-beled in vitro. Even the presence of 10% rRNAsequences among the in vitro RNA might nothave been detected since this value lies withinthe expected error of the method. Investigatorswho subsequently detected the synthesis ofrRNA in vitro used much more sensitive tests.One of these measured the ability of the in vitroRNA to compete with isotopically labeled, au-thentic rRNA in DNA-RNA hybridization com-petition tests (112, 125). In this protocol, onlythe quantity of rRNA among the in vitro prod-ucts was measured in terms of the amount of invitro product required to prevent a knownamount of labeled rRNA from hybridizing.Another method used purified rDNA in search-ing for rRNA among the in vitro products (230).This, too, minimized the quantitative uncer-tainties of discriminating between the amountsof rRNA produced in vitro and the non-riboso-mal RNA sequences, which are the majority.We are left with the conclusion that highly

purified RNA polymerase is capable of preferen-tially transcribing the rDNA without additionalcellular components. It should be realized thatthe observed 10% rRNA among the total in vitro

transcripts probably represents much morethan 10% of the physiologically significant RNAproducts present, since it is likely that many, ifnot most, RNA polymerase molecules utilizeincorrect initiation sites and ignore proper ter-mination sites in vitro with the DNA templatesused. Consequently, a minority of the in vitroRNA products may be equivalent to thoseproduced by the cell, and it is this minoritywhich the quantity of rRNA synthesized mustbe related to. There is no way of judging whatfraction of the total in vitro product is physio-logically relevant, so it can be asserted only thatthe high frequency of rDNA transcription isgoverned largely by the affinity of the RNApolymerase (holoenzyme) for the promoter ofthe rRNA transcriptional unit. The existence ofmodulating devices similar to Ir, which couldenhance the affinity of the polymerase for therDNA promoter, is not excluded, however.

Regulation of rRNA Synthesis DuringBalanced Growth

The availability of nutrients determines thegrowth potential of any population of orga-nisms. As a result of their evolution in anenvironment which presents, alternately, feastand famine, the prokaryotes have devisedmeans for modulating macromolecular synthe-sis according to the suitability of the mediumfor proliferation. If the growth rate of a cellpopulation is limited by the availability ofnutrients rather than the limitations imposedby the potential of the biosynthetic machinery,then it behooves the cells to restrict theirproduction of the components of this machin-ery. The stable RNA molecules of the prokar-yotic cell are components of the protein synthe-sizing apparatus, and as such their rates ofsynthesis are governed by the growth rate of thecell population.

Early studies of variations in the RNA con-tent of bacterial cells as a function of growthrate have been reviewed by Maaloe and Kjeld-gaard (167), and more recent efforts have beenconsidered in detail by Koch (138). The funda-mental observation regarding variation in cellu-lar RNA content, derived from the work ofSchaechter et al. (77, 256), is that, to a goodapproximation, the RNA content and the num-ber of ribosomes per cell are proportional to therate of cellular protein synthesis and the rate ofgrowth of the propulation. For example, inSalmonella typhimurium (77) or E. coli (138),the amount of RNA accumulating per cell massincreases two- to threefold over a fourfold in-crease in growth rate. Total cellular RNA ofcourse includes mRNA, rRNA, and tRNA, so

579VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

the question we must here address is: are therates of synthesis of these RNA species coordi-nately or independently controlled by the rateof growth?

Coordinate control of RNA synthesis could beachieved by cellular manipulation of the availa-bility of nucleoside triphosphates or of activeRNA polymerase molecules. Probably intracel-lular levels of nucleoside triphosphates do notdetermine the RNA synthetic capacity of thecell (139, 197-9), but the role of active versusinactive RNA polymerase molecules is difficultto assess. Some evidence suggests (225) thatslowly growing cells possess a pool of polymer-ase molecules which are not engaged in RNAsynthesis but which become active upon shift toa more rapid growth rate. It is not knownwhether the pool of unengaged polymerase isinherently inactive or if it is a consequence ofrestricted availability of promoter sites in theDNA template. Independent regulation of thesynthesis of rRNA, mRNA, and tRNA of coursewould require regulatory devices capable ofinteracting directly with the transcriptionalunits involved, and their effects would be super-imposed upon any restraints which affect allRNA synthesis during balanced growth.

Early, rather indirect measurements of varia-tion in the rates of synthesis of mRNA, rRNA,and tRNA at varying rates of balanced growthprovided suggestive evidence for the independ-ent control of these classes of RNA (137, 167,246), but this conclusion was not unanimous(90). More recent measurements, largely byrather convincing DNA-RNA hybridizationcompetition techniques, indicate that the ratesof synthesis of at least mRNA and rRNA maydiffer during rapid and slow cell growth, butthat the differences in the synthetic rates of thetwo classes of RNA are insufficient to accountfor differences in their observed quantities dur-ing growth.Norris and Koch (212) have adjusted by

glucose limitation the balanced growth rates ofE. coli cultures in chemostats and have mea-sured the instantaneous rates of synthesis ofmRNA, rRNA, and tRNA at varying growthrates by DNA-RNA hybridization competitionof pulse-labeled RNA with unlabeled rRNA,tRNA, or a mixture of the two. The conclusionwas reached that the newly synthesizedmRNA:rRNA ratio in a population doublingeach 10 h is about twice that of cultures withgeneration times of 55 min. The composition ofthe medium was the same at both growth rates,and so presumably the mRNA complementrequired for growth was the same; the differencein mRNA: rRNA ratios, therefore, suggests that

independent controls of mRNA and rRNA syn-thesis exist. This same general conclusion wasdrawn by Lazzarini and Winslow (152), whoemployed similar analytical techniques but dif-ferent carbon sources (glucose versus lactate) toachieve a four- to fivefold difference in thegeneration times of cultures. The observed,pulse-labeled, total RNA: rRNA ratio at thelower growth rate was about twice that at thehigh rate, again indicating a preferential restric-tion of rRNA output at reduced growth rates.The twofold difference in the rates of synthe-

sis of mRNA and rRNA at widely differentgrowth rates apparently is not the only factorinfluencing rRNA accumulation in the cell,however. Julien et al. (132) and Norris andKoch (212) have pointed out that a significantfraction of the rRNA synthesized at low growthrates does not eventually appear among thestable RNA components of the cell; the fractionapparently is degraded soon after its synthesis.This observation is strengthened by the findingsof Pederson (personal communication), who hascompared by DNA-RNA hybridization theamounts of rRNA synthesized residually in thepresence of rifampin to the amounts whicheventually are stabilized. During rapid culturegrowth, all of the rRNA produced is stable,whereas at very much lower growth rates, onpoor carbon sources, only about half of therRNA is stabilized.The suggestions that part of the rRNA pro-

duced during slow growth is unstable conceiva-bly could cast doubt on the measurements ofthe relative rates of synthesis of mRNA andrRNA at varying growth rates. The elevatedmRNA: rRNA ratios at low compared with highgrowth rates could be a consequence of rRNAdegradation if the half-life of rRNA destined tobe destroyed is significantly shorter than that ofmRNA. However, Pato and von Meyenburg(225) found no obvious differences in the half-life of unstable RNA over a wide range of growthrates. The lower growth rates examined fellwithin the range where appreciable rRNA deg-radation should occur, and so if rRNA break-down is rapid then some perturbation of thedecay kinetics might be expected. Instead, thedata suggest that any unstable rRNA decays atabout the same rate as mRNA, and so theconclusion seems correct that the rate of synthe-sis of rRNA, relative to that of mRNA, isrestricted by some controlling factor other thanthose which limit total RNA synthesis at re-duced growth rates. The accumulation of ribo-somes in the cell, therefore, appears to begoverned at several levels.One control mechanism affects overall cellu-

580 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

lar RNA synthesis and might involve the availa-bility of active RNA polymerase molecules. Asecond control seems to restrict rRNA synthesisrelative to that of mRNA and, at its simplest,would preferentially restrict utilization of therRNA transcriptional units by the RNA polym-erase. It has been proposed that unchargedtRNA (147, 282) or disengaged ribosomes (191,245) might mediate in suppression of rDNAtranscription, but no substantiation for either ofthese proposals has been offered. Still a thirdmeans for regulating the accumulation of ribo-somes appears to function subsequent to tran-scription: rRNA is degraded. Conservation ofrRNA conceivably could be determined by theavailability of ribosomal proteins. At relativelyrapid growth rates, ribosomal protein and rRNAsynthesis may be coordinate (109), and ribo-somal protein synthesis is reduced relative tototal protein synthesis as the rate of balancedgrowth slows (262), but the relative rates ofrRNA and ribosomal protein synthesis have notbeen examined during slow but balancedgrowth. If the availability of the ribosomalproteins were to become limiting at reducedgrowth rates, then a fraction of the newlysynthesized rRNA would not be incorporatedinto rRNA-protein aggregates, and therefore itwould be susceptible to enzymatic degradation.

It is not clear at this time whether thesynthetic rates of rRNA and tRNA are inde-pendently controlled. Judging the relative, in-stantaneous rates of rRNA and tRNA synthesisby the available DNA-RNA hybridization com-petition experiments (212) is difficult. Mea-sured variations in the tRNA content of pulse-labeled RNA at different cell growth ratesprobably fall within the inaccuracies of themethodology, since tRNA comprises such aminor component of the RNA being synthesizedat any instant. The cellular tRNA: rRNA ratioduring very slow, balanced growth may be 1.5 to2.0 times that during rapid cell growth (63, 137,246), but these measurements reflect only theamounts of tRNA and rRNA which accumulate.Therefore, the disparity in the tRNA: rRNAratios may simply be a consequence of thepreferential degradation of rRNA.

Regulation of rRNA Synthesis DuringPerturbation of Balanced Growth

The evidence for noncoordinate control of thesynthesis of rRNA and mRNA is rather clearwhen studies are performed with cultures whosebalanced growth rate is abruptly reduced by achange in available nutrients or whose growth ishalted by removal of an essential amino acid.The regulatory mechanisms invoked by these

unbalancing events differ at least in part.The influence of growth rate transitions upon

bacterial RNA synthesis has been the subject ofnumerous investigations. The basic facts (167,201) are that, if a culture is removed from amedium permitting rapid proliferation and isplaced in one supporting only poor growth("step-down"), then rRNA synthesis is tem-porarily reduced until the ribosome:cell massratio characteristic of the slower growth rate isachieved. For example, in one study (152), afour-to fivefold reduction in the growth rate ofE. coli (achieved by changing the carbon andenergy sources from glucose to lactate) immedi-ately resulted in a 75% decrease in the instan-taneous rate of rRNA synthesis but only a 25%reduction in mRNA production, both as mea-sured in DNA-RNA hybridization competitionexperiments. This is strong and convincingevidence for regulatory mechanisms which arecapable of independently influencing rRNA andmRNA production. Conversely, during "shift-up", from relatively poor to rich media, thestable RNA (rRNA plus tRNA):unstable RNA(mRNA) ratio rapidly becomes characteristic ofthe faster growth rate (225), and there is evi-dence for the transient, preferential synthesis ofrRNA (205, 206).A second type of growth perturbation leading

to noncoordinate reduction of rRNA and mRNAsynthesis results from the removal of an essen-tial amino acid from auxotrophic cultures. Thishas immediate and profound consequences formacromolecular synthesis; the phenomenologyhas been reviewed recently by Edlin and Broda(78) and Ryan and Borek (250). Most importantin the present context is that net RNA accumu-lation abruptly slows upon removal of therequired amino acid. Restriction of rRNA syn-thesis is more profound than the decrease in therate of mRNA production; Lazzarini and Wins-low (152), for example, find by DNA-RNAhybridization competition experiments that therates of synthesis of rRNA and mRNA arereduced upon amino acid deprivation to 10 and40%, respectively, of the rates observed ingrowing cells. The disproportionate inhibitionsof rRNA and mRNA synthesis imply, again,that cells possess the capacity to regulate inde-pendently the transcription of genes specifyingthe two classes of RNA. The synthesis of rRNA,at least, is retarded under these conditions byprevention of the initiation of transcription ofthe rDNA by the RNA polymerase (279). Thisspecific reduction in rRNA production is magni-fied by other consequences of amino acid depri-vation which might affect RNA synthesis.These include, for example, a decrease in the

VOL. 37, 1973 581


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

purine nucleoside triphosphate pools (79, 94)and a reduction in the nucleotide step time ofthe RNA polymerase (320). Moreover, Doninihas pointed out (68) that the reductions in therates of synthesis of rRNA and mRNA duringstarvation for an essential amino acid are prob-ably insufficient to account for the reduction inthe rate of RNA accumulation. This suggeststhat much, if not all, of the rRNA'which isproduced during starvation is rapidly degraded.Presumably the degradation is a consequence ofthe lack of synthesis of the ribosomal proteins;therefore, the rRNA-molecules are not packagedinto nuclease-resistant particles.The ability of cells to respond to amino acid

deprivation is governed at least in part by thegene rel (250). Wild-type strains are designatedas "stringent" with regard to the control ofRNAsynthesis, whereas reh strains, in which RNAsynthesis continues unabated after removal of arequired amino acid, are termed "relaxed". Todate, mutants at the rel locus are known for E.coli (20, 282), Salmonella typhimurium (178),and Bacillus subtilis (288). One striking meta-bolic difference between rel+ and reh strains isthat immediately after amino acid deprivationthe rel+, but not the rel, strain accumulates asignificant quantity of an unusual nucleotide(38), which has been identified as guanosine,5'-diphosphate, 2'-(or 3')-diphosphate (ppGpp;39). The function of the rel gene product isprobably limited to conditions of growth whichbring about starvation for an amino acid, butthe enhanced production of ppGpp is observedconcomitantly with any perturbation of growthwhich results in the restriction of rRNA synthe-sis. Both rel+ and relh strains of E. coli restricttheir output of rRNA during step-down imbal-ance of growth (200) and during starvation for acarbon source (110, 151) or an inorganic nitro-gen source (126). These same conditions pro-mote the accumulation of ppGpp in both therel+ and the relh genotypes (110, 126, 151).At concentrations observed within rel+ cells

during amino acid deprivation, ppGpp inhibitsin vitro RNA synthesis by purified RNA polym-erase with regard to both chain elongation andinitiation. The inhibition of chain initiationapparently is directed largely toward RNAchains with guanosine triphosphate (GTP) attheir 5'-terminal (37). Cashel has suggested(37), on the basis of this latter observation, thatthe presence of ppGpp in starved, stringent cellscould markedly alter the specificity of the RNApolymerase. If the transcript of the rDNA con-tains 5'-terminal GTP and if a portion of themRNA population does not, then the preferen-tial restriction of rRNA production during

amino acid starvation might be accounted for.However, this suggestion is not substantiatedby the findings of Ikemura and Dahlberg (J. E.Dahlberg, personal communication), who haveobserved that rel+ strains of E. coli, starved foran essential amino acid, preferentially accumu-late at least one RNA molecule with GTP at its5,-terminus.

Regulation of rRNA Synthesis DuringMorphogenesis

Certain prokaryotes have the capacity to altersignificantly their morphology and physiologyin response to environmental effects and toenter dormant states. Examples of this simpleform of morphogenesis include sporulation bythe genus Bacillus and microcyst formation inMyxococcus species. Both of these genera re-strict the net formation of ribosomes during thedifferentiation events, but they probablyachieve this through different mechanisms.The regulation of rRNA synthesis in sporulat-

ing B. subtilis has been reviewed briefly byLosick (162). During sporulation the synthesisof certain classes of mRNA required for vegeta-tive growth is halted (65, 324), and the produc-tion of rRNA is dramatically reduced (124).Losick and his collaborators have providedevidence that the restriction of rRNA synthesisis not a consequence of some regulatory mecha-nism directed specifically toward the rDNA,but rather the template specificity of the RNApolymerase itself is modified. RNA polymerasepurified from vegetative B. subtilis can utilizeas template both phage 4e DNA andpoly(dA-dT), whereas the enzyme from sporu-lating cells is capable of actively transcribingonly the synthetic DNA (164). The mechanismsresponsible for altering the template specificityof the RNA polymerase from sporulating cells,are not yet understood (158, 163, 169), but thealteration is also reflected in the inability of thepolymerase to synthesize rRNA in vitro (125)and, presumably, in sporulating cells. This veryneat scheme for regulating the expression of therDNA in sporulating B. subtilis remains to beconfirmed, however. Szulmajster and his col-leagues have reported (18) that the basic prem-ise of the scheme, that rRNA production haltsin sporulating cells, may be incorrect. They findthat although net RNA synthesis indeed isdrastically reduced in populations of sporulat-ing cells, the proportion which is attributable torRNA is only about 30% lower than in culturesgrowing exponentially. Unfortunately they didnot offer compelling evidence that all membersof the cell populations studied in fact wereundergoing sporulation. Since net RNA synthe-

582 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

sis is reduced in the sporulating cultures, if onlya few percent of the cells escaped the commit-ment to sporulation and continued to growexponentially, then these could account for theobserved quantities of residual rRNA synthesis.

In contrast to B. subtilis, populations ofMyxococcus xanthus which are induced to formmicrocysts continue to synthesize RNA, includ-ing ribosomal RNA, at almost the rate observedin vegetative cells (214, 239). This synthesisresults in little net accumulation of RNA,however (7, 239). Therefore, either the ribo-somes which existed before induction are de-graded during cyst formation, or the newlysynthesized rRNA is degraded soon after itssynthesis. Probably both of these events occur,but degradation of rRNA synthesized duringmicrocyst formation apparently is more impor-tant in balancing RNA synthesis with theabsence of net accumulation. The immediateprecursor of 16S rRNA, which is somewhatlarger than the mature molecule (see below), isidentifiable by sedimentation analysis in cellsinduced to form microcysts. However, mature16S rRNA is not evident in induced cells afteristopic labeling periods sufficiently lengthy todemonstrate its presence in vegetative cells (7).This implies that most (but possibly not all; 7)of the rRNA synthesized during cyst formationdoes not enter ribosomes, but rather it is de-graded, probably at the precursor stage.

POST-TRANSCRIPTIONALPROCESSING OF rRNA

Hypothetical Tandem Transcript of therDNA

In mammalian cells, the transcriptional unitsspecifying the rRNA molecules are composed ofone gene each for 18, 28, and 28S-A rRNA's. Theimmediate product of transcription is a 45Smolecule containing the sequences of thesethree rRNA components plus considerable ex-cess material which is discarded during thepost-transcriptional maturation events (4, 33).As described above, the rDNA of bacteria also isarranged in compound transcriptional unitsconsisting of one gene each for 16, 23, and 5SrRNA's; any RNA polymerase molecule whichreads the 16S DNA continues on to produce one23S and one 5S rRNA sequence. However, noprecursor of bacterial rRNA analogous to themammalian 45S RNA has been isolated fromnormally growing cells. Instead, the identifiableprecursors of 16 and 23S rRNA's are onlyslightly larger than the mature molecules. Fol-lowing the nomenclature introduced by Hechtand Woese (116), I henceforth shall refer to the

mature rRNA components as m16, m23, and m5rRNA's, and to their precursors as p16, p23, andp5 rRNA's. These will be discussed at somelength below.There are two conceivable explanations for

the absence of a compound rRNA precursor inbacteria. The first possibility is that, as theRNA polymerase molecules pass from, for ex-ample, the 16S gene into the 23S gene, theyencounter some sort of signal to release thenewly completed p16 molecule before com-mencing the synthesis of p23 rRNA. Alterna-tively, the cell might possess a specific and veryactive endonuclease which clips the completedp16 molecule from the nascent RNA chain verysoon after the polymerase reads into the 23Sgene. The available evidence strongly supportsthe latter possibility.

Pettijohn and his collaborators (231) haveisolated the intact genetic apparatus from E.coli; the "nucleoids" retain the RNA polymer-ase molecules and nascent RNA strands whichare associated with the DNA at the time ofdisruption of the cells. When the four nucleosidetriphosphates then are added to preparations ofthe nucleoids, the resident RNA polymerasemolecules complete the reading of the transcrip-tional unit with which they are associated, andthey release identifiable classes of RNA mole-cules (233). The released RNA polymerase mol-ecules are incapable of reinitiating synthesis,however, since the initiation factor a is notpresent in the nucleoids. The two major RNAclasses completed in vitro include one whichapparently is slightly larger than p23 rRNA insize and the second, which has been termed p30,has the electrophoretic mobility in polyacryl-amide gels expected of a molecule with thecombined molecular lengths of 16 and 23SrRNA's. The in vitro reaction yields virtually noRNA that is the size of p16 rRNA. DNA-RNA hybridization competition experimentshave shown that the p23-sized material indeedcontains m23 rRNA sequences, but the p30RNA includes both m16 and m23 rRNA; p30apparently is a transcript of the entire rDNA(232; D. E. Pettijohn and C. R. Kossman,personal communication). If cells are radi-oisotopically pulse-labeled immediately beforepreparation of nucleoids, then both the p30 andp23 products completed in vitro contain thepulse label, but the pulse-labeled sequencesassociated with the p30 product are predomi-nantly 16S rRNA (232). Since the 16S se-quences are associated only with p30, which iscompleted in vitro, and these 16S sequences areinitiated in vivo, then the 16S sequences mustbe in the 5' portion of the p30 molecule. Little

VOL. 37, 1973 583


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

p30-associated 23S RNA is labeled in vivo, somost of the p30 RNA must be the product ofpolymerase molecules associated with the 16Sgene at the time of nucleoid isolation, which invitro continue on to read through the 23S geneand, presumably, the 5S gene. The associationof 5S sequences with p30 has not yet beendemonstrated.The production of a tandem transcript of the

rDNA in vitro suggests strongly that the mecha-nism for the release of p16 from the RNApolymerase as it moves into the 23S gene is notencoded into the DNA in some fashion, butrather the release is effected by a specificendonuclease which is soluble and is removedfrom the nucleoids during purification. Thus,all of the 16S sequences completed in vitro areassociated with p30; no cleavage enzyme ispresent to dissociate the p16 from the nascentRNA chain. The absence of the requisite, spe-cific cleavage enzyme also may explain theobservaticn (232) that the p23-sized materialgenerated by the nucleoids is somewhat largerin size than the p23 rRNA appearing in vivo.The in vitro p23 (and p30 as well) probablycontains the 5S rRNA in addition to the p23sequence. The linked, p23-p5 RNA species hasonly a very brief lifetime in growing cells beforeit is cleaved to the longer-lived p23 and p5molecules, and probably only part of the com-pleted p5 population passes through the linkedstate. Isotopic label added to growing culturesof E. coli appears in p5 rRNA with higher orderkinetics than the labeling of total RNA (1, 221),implying the existence of a rate-limiting stepbetween the transcription of the rDNA and therelease of p5 rRNA (27). The rate-limiting steppresumably is the cleavage of the linked, p23-p5rRNA. However, the difference in the rates oflabeling of total RNA and p5 rRNA probably isnot sufficiently great that all of the p5 se-quences pass through the p23-p5 intermediate(1, 221), so it seems likely that many of the p23rRNA molecules are clipped from the growingRNA chain soon after the polymerase migratesinto the 5S gene.The existence of endonucleases which sepa-

rate p16 and p23 rRNA from the growing rDNAtranscript also is implicit in the results ofChantrenne (44) and Grunberger et al. (105),who observed that Bacillus cereus culturestreated with 8-azaguanine accumulate a stableRNA component which is about the size of thetandem transcript of the rDNA. It has not beenproven by the appropriate DNA-RNA hybridi-zation experiments that the observed RNAmolecule in fact contains both the 16 and 23Ssequences, but in light of the data of Pettijohn

and his colleagues it probably does. Presumablythe 8-azaguanine is incorporated into the RNA,and its presence perturbs the polynucleotidestructure such that the substrate sequence canno longer be recognized by the specific cleavageenzyme(s). The presence of the analogue wouldnot be expected to interfere with p16 and p23rRNA release if the RNA polymerase itself weresomehow involved in the release of the com-pleted chains.

Therefore, there are probably at least threeendonuclease cuts made in the RNA productduring the course of transcription of the rDNAor very soon thereafter. One of these is theremoval of the hypothetical 5'-terminus of therDNA transcript; the event is evidenced only bythe observation discussed below that the 5'-ter-minus of p16 rRNA does not have the propertiesexpected of an immediate product of the tran-scription reaction. A second scission event re-leases the p16 sequence from the growing 23Smolecule, and a third frees the p23 moleculefrom the nascent or just-completed 5S chain.The enzymes involved in severing the growingrDNA transcript are highly sophisticated intheir action; they must scan polynucleotides,pick a sequence, and cleave it precisely. It isnot yet known whether only one enzyme cata-lyzes these three scission events or if eachrequires a specific enzyme. In either case, thenuejeases effecting scission during transcriptionare probably distinct from those participatingin post-transcriptional cleavage. The latter ap-pear to require ribonucleoprotein aggregates astheir substrates (see below), whereas the formerare capable of functioning normally with moreor less naked polynucleotide chains; inhibitorsof protein synthesis do not retard the accumula-tion of p16, p23, and p5 rRNA's, but they doprohibit the subsequent cleavage events whichgenerate the mature rRNA molecules.

Immediate Precursors of the rRNAMolecules

Soon after the advent of studies of ribosomebiosynthesis, it was determined that the pulse-labeled RNA components of the ribonucleo-protein precursors of ribosomes differ somewhatfrom mature 16 and 23S rRNA's in their sedi-mentation velocities (141, 180) and in theirbehavior during chromatography on columns ofmethylated albumin-kieselguhr (157). In partic-ular, the "immature" 16 and 23S rRNA mole-cules were observed to sediment slightly morerapidly than their mature counterparts, imply-ing either that pl6 and p23 are larger inmolecular weight than m16 and m23, or that theprecursors possess a more compact secondary

584 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

structure. When the precursor and matureforms of the rRNA molecules of B. subtilis (116)and E. coli (1, 221) were examined by polyacryl-amide gel electrophoresis, it was apparent thatthe maturation of both rRNA components in-volved a reduction in size. Both p23 and p16migrate more slowly in the gels than do themature molecules; they therefore have either ahigher molecular weight or a larger molecularcross section than do the mature rRNA compo-nents. However, if the precursors were equal tothe mature forms in molecular weight but ofgreater cross section, then they would sedimentmore slowly than the mature molecules. Theconverse is the observation, and so p16 and p23in principle must be of greater molecular length.The electrophoretic mobilities of the p16 andp23 products of both B. subtilis (116) and E. coli(1,221) suggest that they are, respectively,about 10 and 5% longer than the mature mole-cules in chain length. A minor portion (10%o orso) of the p16 of E. coli is of slightly more rapidelectrophoretic mobility than the majority (1,59). This difference may not reflect size differ-ences, but rather conformational alternatives ofthe p16 molecules (29, 59). Pulse-labeled mole-cules similar in size to p16 and p23, andtherefore presumably occupying precursor rela-tionships to m16 and m23 rRNA's, have beenidentified in several other diverse prokaryotes,including, for example, Salmonella typhimu-rium (306), Myxococcus xanthus (7), the blue-green alga Anacystis nidulans (69, 293), andeven the parasitic trachoma agent (107). Thewidespread occurrence of transient p23- andp16-like molecules suggests that similar precur-sors are involved in rRNA formation in allprokaryotes. A search for exceptions would be ofsome interest.The precursor forms of the rRNA molecules

are identifiable not only in pulse-labeling exper-iments, but they also accumulate under anyconditions which prohibit protein synthesiswhile permitting continued RNA production.Antibiotic inhibitors of protein synthesis suchas chloramphenicol (122, 148) or puromycin (58,122, 259) prevent rRNA maturation, as doesremoval of an essential amino acid from auxo-trophic strains carrying the rel- lesion, so thatrRNA synthesis is not halted (57, 194, 202).Other conditions which retard the formation offunctional proteins and lead to the accumula-tion of p16 and p23 in E. coli include fluoroura-cil addition to cultures (142), toluene treatmentof cells (229), probably potassium depletion(82), and treatment of cells with cobalt chloride(15). The precursor rRNA molecules also accu-mulate under restrictive growth conditions in

conditionally lethal mutants of E. coli which aredefective in the assembly of ribosomes (sadmutants; 106, 195). The rRNA precursors pro-duced by cells treated with chloramphenicol orstarved for an essential amino acid have re-ceived considerable attention because they arerecoverable as ribonucleoprotein aggregateswhich at one time were thought to be intermedi-ates in ribosome assembly (148, 194). Osawa(215) and Nomura (208) have reviewed thesestudies in some detail. The only relationshipthat these particles have with the precursors ofribosomes is their rRNA content, however. Theassociated proteins are extremely heterogeneousin their electrophoretic properties, and theybear little resemblance to the true ribosomalproteins (263, 328). Presumably the particlesare random aggregates of basic proteins and thepolyanionic rRNA molecules. The importantpoint is that inhibition of protein productionprevents rRNA maturation. This implies eitherthat the enzymes involved in the post-transcrip-tional cleavage of p16 and p23 recognize as theirsubstrates not the free RNA, but rather aspecific rRNA-ribosomal protein aggregate, orthat the cleavage enzymes turn over rapidly.The inability of the cold-sensitive sad mutantsto carry out the maturation of rRNA (195)suggests that the former possibility is correct.These mutants are incapable of assemblingribosomal subunits at restrictive temperatures,but protein synthesis is not immediately af-fected.The comparisons of the sedimentation and

electrophoretic properties of the precursor andthe mature rRNA molecules is strong evidencethat they differ in their molecular weights, butthe data do not prove such differences. In fact,it has been reported that heat denaturation ofp16 and m16 renders them indistinguishableby methylated albumin-Kieselguhr chromatog-raphy (292) and that denaturation of p23 andm23 converts them to forms which can no longerbe separated by polyacrylamide gel electropho-resis (59). These observations would seem toimply that the precursor and mature rRNAmolecules are merely conformational isomers ofone another, but this conclusion is certainlyincorrect, and the basis for the experimentaldata is not clear. Recent detailed comparisonsof oligonucleotides released by T1 RNase diges-tion of the precursor and mature rRNA mole-cules of E. coli (29, 115, 165, 272), Bacillusmegaterium, and B. subtilis (M. L. Sogin,Ph.D. thesis, Univ. of Ill., Urbana, 1972) haveproven beyond doubt that rRNA maturationinvolves size transitions. T1 RNase releasesfrom any given RNA molecule a specific array of

585VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

oligonucleotides, which is characteristic of theprimary structure of that RNA molecule; thequantitative yield of the various oligonucleo-tides is diagnostic of the size of the molecule.The p16 and p23 products of both genera provedto contain nucleotide sequences not present inthe mature molecules, and the observed sizes ofthe precursor and mature molecules confirmthe results suggested by gel electrophoresis. Thep16 rRNA molecule is about 10% longer thanm16 in all organisms examined, whereas p23 isabout 5% longer than the mature molecule in E.coli and about 7 to 8% longer in the Bacillusspecies. Therefore, during maturation about 150to 200 nucleotides are cleaved from each of theprecursor rRNA chains.The p16 and m16 rRNA molecules of E. coli

have received the most detailed attention withregard to their structures (29, 115, 165, 272).The precursor contains sequences at both the 3'-and 5'-termini which are not present in themature molecule, so the conversion of p16 tom16 rRNA involves at least two cleavageevents. However, the distribution of the excessmaterials at either end of the p16 molecule isnot yet known. The sequences associated specif-ically with the p16 rRNA of E. coli, at least, arenot particularly distinguishing except that theirpyrimidine content is rather higher than that ofthe mature rRNA molecule as a whole (M. L.Sogin, Ph.D. thesis, Univ. of Ill., Urbana, 1972).It is also noteworthy that the 5'-terminal nu-cleotide of p16 rRNA is 5'-monophosphorylateduridylic acid (29, 115, 165, 272); it is not a 5'-triphosphorylated purine, as might be expectedif it were the initial nucleotide deposited bythe RNA polymerase during transcription(25, 170). This suggests that a segment maybe removed from the 5'-terminal portion ofthe RNA transcript of the 16S gene before orvery soon after the release of p16 from thepolymerase-DNA complex.

Rather less detail is available regarding theconversion of p23 to m23 rRNA. It is only knownwith certainty that the p23 molecules of theorganisms examined are larger than their ma-ture counterparts and that the 5'-termini of theprecursor and mature components differ; se-quences must be removed from at least the5'-termini during maturation (M. L. Sogin,Ph.D. thesis, Univ. of Ill., Urbana, 1972). Thedifficulty in precisely defining the p23-specificsequences is that the large sizes of the precursorand mature molecules severely complicate theiranalysis with the available technology. Struc-tural changes occurring during the maturationof p16 rRNA are more easily monitored only byvirtue of the smaller sizes of the moleculesinvolved, but even 16S rRNA is sufficientlylarge that it is not an ideal system for exploring

the mechanics and the rationale of the rRNAmaturation process.The kinetics of the maturation of p16 and p23

have been monitored in B. subtilis (116) and E.coli (1, 221) by polyacrylamide gel electrophore-sis. Both p23 and p16 are abruptly reduced tothe sizes of the mature rRNA components;transition products intermediate in size be-tween the precursor and mature forms whichare sufficiently long-lived to be detected do notappear to exist within the limits of resolution ofpolyacrylamide gel electrophoresis. In E. coligrowing at 37 C, the average p16 moleculeremains in the precursor form for about 4 to 5minutes before it is converted to m16 rRNA,whereas the lifetime of p23 is only about 2 to 3min (1, 221). It is not yet possible to decideconclusively whether the enzymes effecting ma-turational cleavage draw upon the pools of theprecursors at random, or if each precursormolecule must retain the precursor-specific nu-cleotide sequences for a definite length of time,while certain operations are performed upon therRNA. Probably the latter possibility is correct;the precursor rRNA components at least mustinteract with numerous ribosomal proteins be-fore they become suitable substrates for thecleavage enzymes.

Relatively little effort has been expended onthe identification of the enzyme or enzymesinvolved in the cleavage events accompanyingrRNA maturation. RNase II, a processive 3'-exonuclease (268, 276), has been implicated inthe cleavage of p16 to m16 rRNA in E. coli, butrecent evidence does not support its involve-ment. Corte et al. (50) and Yuki (331) firstnoted that E. coli N464, which produces atemperature-sensitive (ts) RNase II, is incapa-ble of producing mature ribosomes or maturerRNA at restrictive temperatures. Moreover,purified p16 rRNA was converted in vitro byostensibly highly purified RNase II to materialsmigrating in polyacrylamide gels coincidentlywith m16 rRNA, with the release of a fragmentat least 100 nucleotides in length (50).The simplest interpretation of these results

would be that RNase II mediates the conversionof p16 to m16, although the action of theenzyme would seem to be more complex thanexpected of a simple exonuclease; it also mustbe able to serve as a highly specific endonucle-ase. However, two observations invalidate thisconclusion. First, Weatherford and his col-leagues (314) have examined the RNase II of 40independent, ts+ transductants of a tempera-ture-sensitive strain of E. coli, which initiallyproduced a thermolabile RNase II. All isolatesretained the temperature-sensitive RNase II butwere capable of growing normally at elevatedtemperatures and, therefore, of carrying out

58-006 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

normal rRNA metabolism. 16S rRNA purifiedfrom the ts+ transductants grown at elevatedtemperatures was definitely m16 in size (314; D.Apirion, personal communication), and so theobserved accumulation of p16 rRNA in E. coliN464 at restrictive temperatures certainly wasnot a consequence of the defective RNase II.Furthermore, Weatherford et al. (314) recoveredessentially normal levels of RNase II from thets+ transductants grown at elevated tempera-tures; the RNase its mutation apparently is notmanifested in vivo.

Since RNase II apparently is not responsiblefor the cleavage of p16 rRNA, then the mutationof rRNA maturation to temperature sensitivityfortuitously accompanied the mutation toRNase II thermolability, and the rRNA cleav-age enzyme more or less co-purified with RNaseII. These unlikely accidents seem to have oc-curred. Schlessinger and his co-workers havenoted (313) that fractions from diethylamino-ethyl-cellulose chromatography of "purified"RNase II do not have constant proportions ofendonuclease and exonuclease activities, sug-gesting that the two reactions are effected bydifferent proteins. The endonuclease activitywhich ostensibly is capable of converting p16 tom16 rRNA (50) has not yet been characterizedin detail. If this endonuclease in fact is responsi-ble for 16S rRNA maturation, then it is curiousthat it is capable in vitro of utilizing purified p16rRNA as a substrate, whereas in the cell p16-ribosomal protein aggregates are required.The involvement of post-transcriptional

cleavage steps in the maturation of rRNA is notlimited to 16 and 23S rRNA's. The mature 5SrRNA components of all prokaryotes so farexamined also are fragments of larger precur-sors. Since 5S rRNA is small in size (120nucleotides in E. coli), the details of its matura-tion are more or less amenable to experimentaldefinition. Monier and his colleagues (87, 88,188) have focused considerable attention on thep5 rRNA of E. coli, which is identifiable inpulse-labeled cells and which accumulates inthe absence of protein synthesis. As isolated,this precursor is a mixed population of RNAmolecules which are only one, two, or maxi-mally three nucleotides larger than the mature5S rRNA. All of the additional nucleotides areassociated with the 5'-terminus (87, 88), andkinetics experiments (87, 93) have demon-strated that during maturation the additionalnucleotides are removed in a stepwise fashionuntil the mature 5S length is achieved. 5SrRNA precursors similar in size to those of E.coli have been detected in Salmonellatyphimurium (240) and the blue-green algaAnacystis nidulans (W. F. Doolittle, personalcommunication), so this scheme for 5S matura-

tion may be rather widespread among theprokaryotes. Unfortunately, the scheme is sosimple that it probably cannot provide a usefulmodel system for understanding the tailoringevents involved in the production of the mature,high-molecular-weight rRNA components.The maturation of 5S rRNA in the genus

Bacillus (222) is very much more complex thanthat described for E. coli, and it resembles to aconsiderable degree the maturation of p16rRNA. In the absence of protein synthesis or inpulse-labeling experiments, two precursors of5S rRNA, present in about equimolar quan-tities, are evident in B. subtilis (222). One ofthese, p5A, is about 180 nucleotides in length,and the second, p5B, is composed of approxi-mately 150 nucleotides. As with the p16 rRNAof E. coli, the non-5S sequences of p5A and p5Bare associated with the 5'- (216) as well as the3'- (M. L. Sogin and N. R. Pace, unpublishedobservations) termini of the 5S sequence. Bothp5A and p5B are rapidly and apparently inde-pendently converted to the mature 5S rRNA, sop5B does not appear to be an intermediate onthe p5A maturation pathway. There seem to betwo possible explanations for the existence ofthese independent 5S precursors: either they arethe products of independent 5S genes, or theyare the result of two, alternate substrate sitesfor the cleavage enzyme that releases p23 rRNAfrom the rDNA transcript as the RNA polymer-ase passes from the 23S into the 5S gene duringrRNA synthesis. Whether p5A and p5B haveindependent functions during their integrationinto the immature ribosomes remains to bediscovered.

Function of the Precursor-SpecificSequences

Each precursor rRNA species thus far studiedhas been found to be largely homogeneous withregard to the T1 oligonucleotides attributable toits precursor-specific regions. This homogeneityimplies that the sequences cleaved from theprecursors during maturation are functional;since multiple genes are involved as templatesfor the rRNA molecules, the precursor-specificsequences would be expected to diverge in theirdetailed structures unless held more or lessconstant by functional constraints. Two sorts ofroles might be envisaged. The first of these isthat the precursor-specific sequences might fa-cilitate the coalescing of the ribosomal proteinswith the rRNA to form the ribosomal subunits.For example, the termini of the precursors mightinteract in some fashion with the mature rRNAsequences, permitting them to assume requiredconformations which otherwise would be ther-modynamically unfavorable. After removal of

VOL. 37, 1973 587


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

the excess sequences, the conformation of themature rRNA chain would be stabilized byinteraction with the ribosomal proteins. Theincorporation of the rRNA molecules into theribosomal subunits certainly does not requirethe presence of the precursor-specific termini,however. The purified, mature rRNA moleculesand ribosomal proteins spontaneously unite toform 50 and 30S ribosomal subunits under theappropriate in vitro conditions (146, 209).Nevertheless, the assembly rate apparently ismore efficient in the cell. Traub and Nomura(309) have investigated the temperature de-pendence of the in vitro assembly of the 30Ssubunit from mature 16S rRNA and the 30Ssubunit proteins from E. coli, and they find thatabove 30 C or so, assembly occurs at about thesame rate as in vivo, but below that tempera-ture the in vitro assembly is extremely slow.This suggests that cells possess some mech-anisms for facilitating ribosome constructionwhich are not inherent in the structures of themature rRNA molecules and the ribosomalproteins. The excess sequences associated withthe precursor rRNA molecules conceivably par-

ticipate in this process.The second possibility for the role of the

precursor-specific sequences is that they merelyare part of the substrate sites recognized by theenzymes which cleave the tandem transcript ofthe rDNA into the component rRNA molecules.With either of these two conjectures, removal ofthe excess sequences would be required if thetermini of the mature rRNA molecules were insome manner required for ribosome function.

Regardless of the function of the precursor-

specific sequences, at least the p16 rRNA whichaccumulates in met-, rel- E. coli starved formethionine is incapable of fulfilling the role ofm16 rRNA. Wireman and Sypherd (personalcommunication) have reconstituted 30S parti-cles from the requisite proteins and relaxed p16rRNA. The particles formed have certain of theattributes of normal 30S subunits, including theability to bind polyuridylic acid, but they lackthe capacity to bind protein S3, and conse-

quently they cannot bind tRNAphe (146) andassociate with 50S subunits to form the func-tional 70S complex. However, whether the ina-bility of p16 to bind all of the ribosomal proteinsis due to the presence of the precursor-specificsequences or to the lack of methylation in therelaxed RNA is not yet known.

Post-Transcriptional Modification ofNucleosides

The post-transcriptional metabolism of therRNA components includes not only the cleav-

age events, but also the methylation of selectnucleosides and 5-ribosyluracil formation.Borek and Srinivasan (21, 278) and Hall (108)have discussed these events at length. The basicfact is that nucleosides are not methylatedbefore incorporation into RNA; methyl groupsare transferred from S-adenosylmethionine to apolynucleotide acceptor. Most studies of themethylation of RNA have been directed towardtRNA, but some information is available re-garding the rRNA of E. coli. The deposition ofmethyl groups in the 16S rRNA of E. coliprobably occurs in two stages. The first stage isduring the formation and lifetime of p16 rRNA,and the second either immediately precedes orimmediately follows the size transition from p16to m16 rRNA. The evidence for the first stage ofmethylation is that the p16 rRNA which accu-mulates in cells treated with chloramphenicol(73) or in rel- strains starved for an essentialamino acid (other than methionine, the methyldonor; 290) contains only about 30 to 50% of thenormal complement of methyl groups. Themethylated nucleosides present in the p16rRNA are not representative of the total presentin m16 rRNA, however; only specific sequencesare methylated (115, 272, 273). These methyla-tion events are effected in the absence of proteinsynthesis, and so the responsible methylatingenzymes must be capable of recognizing the freepolynucleotide chain, even if the rRNA is some-what cluttered by association with the non-ribosomal, basic proteins which are constituentsof the "relaxed" or "chloramphenicol" parti-cles. It is not yet certain whether methylation isinitiated during completion of the nascentrRNA chains or if the completed p16 or p23molecules are required as substrates, but proba-bly methylation begins before the completion ofthe transcription process; at least one of thespecific methylating enzymes has no require-ment for intact rRNA (see below). The secondstage of methylation, during which the remain-der of the methylated nucleosides characteristicof mature 16S rRNA are formed, occurs onlyduring active protein synthesis; restoration of arequired amino acid to starved rel- cells appar-ently permits completion of methylation of atleast some of the submethylated rRNA whichaccumulates during starvation (290). So, theenzymes effecting these final methyl substitu-tions presumably require as their substrates notthe free polynucleotide chains, but rather theribonucleoprotein precursors of the ribosomalsubunits.Some attention has been devoted to the

isolation and characterization of the specificmethylases involved in rRNA metabolism, butmost studies have pursued the tRNA methyl-

588 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

ases (108). Gordon and Boman (101) first re-ported that submethylated rRNA, from relhcells starved for methionine (the ultimatemethyl donor), is capable of accepting methylgroups in vitro, but no attempts were made toisolate the specific methylases involved. Re-cently, two rRNA methylases have been par-tially characterized in vitro. One of these is anexample of the enzymes which are capable ofmethylating free polynucleotides, and the sec-ond functions only with rRNA-ribosomal pro-tein aggregates.

Sipe et al. (270) have extensively purifiedfrom E. coli strain B an adenine (N6-) methyl-transferase, which is responsible for the forma-tion of N-methyladenine in at least m23 rRNA(73, 86, 281). The enzyme utilizes S-adenosyl-methionine as a donor, and it deposits a limitednumber of methyl groups in purified, methyl-deficient (isolated from rel-, met-, methionine-starved cells) rRNA from E. coli B, but not inmature rRNA from the same strain of E. coli.This methylase, therefore, appears to be quitespecific in its action with homologous rRNA,although it has not been demonstrated rigor-ously that the adenine residues methylated invitro are associated with the same sequences asin mature rRNA. However, the specificity of thein vitro reaction apparently is reduced if themethyl acceptor is rRNA from MicrococcusIysodeikticus or B. subtilis; the mature rRNAmolecules from these organisms are effectiveacceptors, and so either spurious methylation isoccurring or nonmethylated sequences are pres-ent, which in the rRNA of E. coli would bemethylated.A second type of methylase, that responsible

for dimethyladenosine formation in E. coli, hasbeen partially purified by Helser et al. (118). Asdiscussed above, the lack of this enzyme incertain mutants confers resistance to the antibi-otic kasugamycin; the me2Ame2ACCUG se-

quence in 16S rRNA, which somehow partici-pates in binding the drug, is not produced in theresistant strains (117). Protein preparationscontaining this methylase use S-adenosylmeth-ionine as a methyl donor to specifically di-methylate the appropriate adenine residues in16S rRNA from kasugamycin-resistant strainsof E. coli, as demonstrated by oligonucleotidefingerprint analysis, but the 16S rRNA fromnormal cells, which already possesses the di-methylated sequence, is not an acceptor. Thismethylase is distinct from that described bySipe et al. (270) not only in its reaction prod-ucts, but also in its substrate requirements. Thedimethylase functions only with rRNA-riboso-mal protein aggregates; purified rRNA is not an

effective substrate. This finding is in concert

with observations that the methyl substitutionsin the me2Ame2ACCUG sequence are not pres-ent in p16 rRNA prepared by any of variousprocedures (115, 165, 273); they have beenobserved only in 16S rRNA from mature 30Sparticles. Therefore, this dimethylase mustfunction either immediately before or immedi-ately after the cleavage of p16. In their studies,Helser et al. used as substrates for the di-methylase the "core" ribosomal particles ob-tained by buoying mature ribosomal subunitsin CsCl gradients (307). This procedure stripsmany of the ribosomal proteins from the sub-units. These core particles may be useful asassay substrates for many of the modifying en-zymes, if suitable mutants lacking the modifi-cation of interest may be isolated or if enzymesand ribosomal core substrates from differentorganisms do not possess the same specifici-ties as the homologous pairs (cf. 270).The majority of methyl substitutions in the

rRNA of prokaryotes is on the purine or pyrimi-dine bases. Only a few 2'-0-positions on theribose moiety of the polynucleotides are methyl-ated. Nichols and Lane (204) have determinedthat the enzymes involved in these methyla-tions also are amenable to in vitro study. Crudeextracts of E. coli transfer methyl groups fromS-adenosylmethionine to the precursor rRNAcomponents from methionine-starved rel- cellsto form 021-methylguanosine and Q2'-methyl-cytidine. This system cannot generate N4, 02Idimethylcytidine or 02"-methyluridine, how-ever Possibly the enzymes responsible for theselatter modifications cannot function with thefree polynucleotides, but rather require therRNA-ribosomal protein aggregates.Even less information is available regarding

the formation of 5-ribosyluracil (pseudouridine)in rRNA. The chemistry (41) and biochemistry(100) of this modified nucleoside have beenreviewed. Dubin and GUnalp (73) have deter-mined that chloramphenicol inhibits the forma-tion of pseudouridine in the rRNA of E. coli, soit must be formed post-transcriptionally, andpresumably the responsible enzymes requireribonucleoprotein precursors of ribosomes assubstrates. There have been no reports of invitro studies of pseudouridine formation inrRNA, but Johnson and Soll (130) have definedan in vitro system for its formation in unmodi-fied tRNA.

Interaction of the rRNA with the RibosomalProteins

During their maturation the rRNA moleculesmust aggregate with the ribosomal proteins toform the functional ribosomes. This aggregation

VOL. 37, 1973 589


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

process is nonrandom; specific proteins areadded to the rRNA-protein complex at specificstages in the formation of the mature ribosomalsubunits. The sequence of addition of the ribo-somal proteins to the maturing 30S subunit ofE. coli has been more or less completely deline-ated by a number of investigators through thestepwise in vitro reconstitution of the subunitfrom its isolated components. These resultshave been reviewed recently (146, 209). Only sixof the 20 to 21 ribosomal proteins from 30Ssubunits are capable of binding to 16S rRNA(258, 259, 284); the remainder bind only torRNA-ribosomal protein aggregates. Similarly,only a fraction of the proteins of the 50S subunitwill bind in a specific fashion to 23 or 5S rRNA's(146, 284). Four, and possibly five, of theRNA-binding proteins from the 30S subunitinteract with the 5'-terminal half of the 16Smolecule, and they have no requirement forintact rRNA as a binding substrate (333).Therefore, in the growing cell the ribosomalproteins could begin to aggregate with therRNA chains even before their synthesis iscomplete. Protein aggregation with the nascentrRNA probably has no influence upon thetranscription process or upon the specific scis-sion of the growing tandem transcript of therRNA transcriptional unit; these events appearto occur normally in the absence of proteinsynthesis.

Early investigators of the formation of ribo-somes suggested that the rRNA moleculesmight serve as mRNA in defining some or all ofthe ribosomal proteins (215, 243). Although themature rRNA components are incapable ofserving in vitro as mRNA for the synthesis ofproteins (207, 303, 319), there were some indica-tions that the precursor rRNA molecules, whichaccumulate during treatment of cells withchloramphenicol (216) or during starvation foran essential amino acid (193), indeed might beable to do so. However, Manor and Haselkorn(174) and Sypherd (289) subsequently foundthat the template activity attributed to therRNA precursors probably was due to con-taminating, bona fide mRNA. When suffi-ciently highly purified, the immature rRNAspecies were ineffective as templates for poly-peptide synthesis. More recently, Miller and hiscolleagues (185) have examined, by electronmicroscopy, complexes of rDNA and nascentrRNA, which are identifiable in lysates of E.coli on the basis of the size of the rRNAtranscriptional units and the relatively highpacking density of the nascent RNA strands. Incontrast to the nascent mRNA molecules, noneof the growing RNA chains associated with theputative rDNA has ribosomes attached to it.

This observation provides visual assurance thatrRNA is never translated.

CONCLUDING COMMENTSIt should by now be apparent that a consider-

able body of descriptive information regardingthe rRNA molecules of prokaryotes is available.However, this information is superficial, andfew fundamental questions have been answered.For example, at this time we do not know thefunction (if any) of the rRNA molecules duringprotein synthesis beyond their role as skeletalstructures for the ribosome. The evolutionarilyconservative character of certain features ofrRNA implies that the rRNA molecules enjoyrather more importance than that of a mereframework. It is my feeling that the roles of therRNA molecules probably as well include dy-namic participation in the exquisitely complexoperation of the ribosome. A few other facets ofthe rRNA molecules whose functions remain tobe discovered are the modified nucleosides andthe "excess" nucleotide sequences associatedspecifically with the precursors of the rRNAmolecules. Conceivably these latter features aretrivial in function, but I suspect not. In animalcells, about one-half of the rRNA transcript isdiscarded during maturation of the rRNA-con-taining segments. The fact that these precursor-specific segments are so extensive implies theirutility to the animal cell and, by extension, theutility of their prokaryotic counterparts.As more information becomes available re-

garding the interactions of the ribosome andother elements of the translation apparatus, wecan look forward to the challenge of intellec-tually reducing this apparatus to its essentialcomponents and speculating on the mechanicsof its origin and function in the pre- or protobi-otic milieu. The central importance of theprotein-synthesizing machinery must have dic-tated its very early appearance in rudimentaryform. Crick has suggested (54) that the requi-site, individual RNA components of the primi-tive translation machinery arose independentlyin the prebiotic sea. This is not unreasonable,except for the choice of polymers. Divalentcation-catalyzed hydrolysis of polyribonucleo-tides (via a 2',3-cyclic phosphate intermedi-ate) probably precluded the random abioticconstruction and survival of a sufficiently com-plex variety of RNA chains. Therefore, theoriginal elements of the pre- or protobiotictranslation apparatus likely were DNA, notRNA.The information now available represents

only a first step, but the prospect of achievingan understanding of the detailed mechanics ofthe translation apparatus appears bright.

590 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

SUMMARY

General Features of Prokaryotic Ribosomes

The ribosomes of phylogenetically diverseprokaryotes are superficially similar in theirproperties. The active ribosomes at physiologi-cal Mg2+ concentrations sediment as particleswhich are about 70S in size. At low Mg2+concentrations or when not engaged in proteinsynthesis, each 70S particle dissociates into one30S and one 50S particle, which are composedof, respectively, about 60 and 75% RNA. Theremaining mass in the ribosomal subunits iscontributed by numerous ribosomal proteins.The collections of ribosomal proteins from di-verse prokaryotes are similar in the numbersand sizes of their components, but the corre-sponding proteins from different organisms donot necessarily display close antigenic relation-ships. Therefore, the prokaryotes have divergedconsiderably in the detailed structures of thecomponents of their ribosomes. Nevertheless,functional homology has been retained; hybridribosomal subunits which are active in proteinsynthesis may be reconstituted from the compo-nent proteins and rRNA from little-relatedorganisms.

Structures of the Prokaryotic rRNAMolecules

The ribosomes of prokaryotes generally con-

tain three metabolically stable rRNA compo-

nents. Two of these, 23 and 5S rRNA's, are

derived from 50S ribosomal subunits. The third,16S rRNA, originates from the 30S subunit. Themolecular weights of 23, 16, and 5S rRNA's are,

respectively, about 1.1 million, 0.55 million,and 0.04 million; these weights correspond tochain lengths of about 3,300, 1,650, and 120nucleotides, respectively. A few photosyntheticprokaryotes, for example, Rhodopseudomonasspheroides or Anacystis nidulans, possess 23SrRNA only transiently. The molecule is specifi-cally cleaved into two fragments, apparentlywithout affecting the capacity of the ribosomesto carry out protein synthesis. The utility of thecleavage of 23S rRNA to the organisms remainsobscure.The nucleotide compositions of the total

rRNA from diverse prokaryotes are remarkablyconstant, even though the base compositions oftheir genomes vary widely. The similar basecompositions do not necessarily reflect similarprimary structure, however. The rRNA compo-

nents of phylogenetically dissimilar organismsfrequently possess little homology in nucleotidesequence, as defined in DNA-RNA hybridiza-

tion experiments. Therefore, the evolutionarilyconservative features of the rRNA moleculeswhich are reflected in their base compositionsare probably their secondary or tertiary struc-tures.The 16 and 23S (but not 5S) rRNA compo-

nents of all prokaryotes thus far examined eachcontain minor quantities (about 1%) of modifiednucleosides, including methylated nucleosidesand 5-ribosyluridine. About 90% of the methyl-ated nucleosides are base substituted; the re-mainder are 2'-O-methylated. The function ofthe modified nucleosides remains to be deter-mined, but their utility is implicit in the factthat many of the methylated nucleosides andthe nucleotide sequences in their immediatevicinities are identifiable in the rRNA compo-nents of bacterial species which otherwise havelittle similarity in their primary structures. Thefunctions of certain of the methyl substitutionsare sufficiently subtle that they are superficiallydispensable; mutants have been isolated whichlack some individual modifications but whichare capable of apparently normal growth.

Evaluation of the detailed primary, second-ary, and tertiary structures of the rRNA mole-cules is only beginning. The nucleotide se-quences of the 5S rRNA molecules of E. coli andPseudomonas fluorescens are known, and about70% of the sequence of the 16S rRNA of E. coli isknown to an approximation. The determinedstructures of the rRNA molecules are compati-ble with physical measurements indicating that60 to 70% of the bases are involved in hydrogenbinding. In the case of the 16S rRNA of E. coli,the consequent secondary structure apparentlyis predominantly local; adjacent, short nucleo-tide sequences tend to be complementary, sothat the molecule in solution and presumably inthe ribosome might be viewed as a complexseries of hairpin loops. The individual rRNAclasses of a given strain of bacteria are largelyhomogeneous in their structures, but minordifferences, in only a few nucleotides, are evi-dent. These presumably occur because eachorganism possesses multiple copies of the rRNAgenes, and their individual structures are evolv-ing more or less independently. There is noevidence for functional heterogeneity within therRNA population of individual prokaryotic or-ganisms.

Genes Specifying rRNA

DNA-RNA hybridization saturation experi-ments have revealed that most prokaryotescontain multiple genes specifying the rRNAcomponents. Generally, about 0.3 to 0.4% of the

591VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

total genome DNA is complementary to rRNA.For example, in E. coli this corresponds toabout six copies of each of the genes specifying16, 23, and 5S rRNA's; Mycoplasma speciespossess only one copy of each of the rRNAgenes.The lack of appropriate mutants in the rRNA

genes has retarded locating them within thelinkage map of E. coli, but at least most of therDNA loci now have been mapped by DNA-rRNA hybridization experiments using epi-somes which span defined regions of the chro-mosome. One cluster of rDNA, comprisingabout three genes for each of the rRNA compo-nents, is located at about 74 to 77 min on thelinkage map of E. coli. A second cluster, con-taining probably two genes for each of the rRNAspecies, is somewhere in the 54- to 59-min span.It is not yet certain whether all of the rDNA ofE. coli is accounted for.The rDNA in the Bacillus subtilis genome

also is distributed between at least two regions,but the location of these has not been deter-mined accurately. About 70% of the rDNA islocated in the first 25% of the genome toreplicate during spore germination; the remain-der of the rDNA appears to be replicated withthe terminal 25% of the genome. In both E. coliand B. subtilis, the individual rRNA genes arearranged in clusters containing one gene eachfor 16, 23, and 5S rRNA's, linked in that order.These clusters are separated by lengthy (at least6 to 10 times the length of the 16-23-5S unit)stretches of non-ribosomal DNA.

Transcription of the rRNA GenesA major fraction of the RNA being synthe-

sized at any instant by E. coli is identifiable asrRNA. The rDNA is read by the RNA polymer-ase with a frequency 50 to 100 times that atwhich the average mRNA gene is read. Therelatively high rate at which the rDNA is readapparently is determined by the affinity of theRNA polymerase for the promoter of the rDNA;no auxiliary protein factors are required for thepreferential in vitro synthesis of rRNA by puri-fied RNA polymerase holoenzyme in vitro.To a good approximation, the RNA content

and the number of ribosomes per cell areproportional to the rate of cellular proteinsynthesis and the rate of growth of the popula-tion. The accumulation of rRNA in growing E.coli is governed at several levels. The firstcontrol mechanism affects overall cellular RNAsynthesis and might involve the availability ofactive RNA polymerase molecules. A secondcontrol seems to noncoordinately modulaterRNA synthesis relative to that of mRNA; the

rRNA: mRNA ratio among pulse-labeled RNAmay vary two- to threefold, depending upon thegrowth rate of cultures. The mechanism bywhich the frequency of transcription of therDNA is specifically modulated is not known.Still a third means for regulating the accumula-tion of rRNA functions post-transcriptionally:rRNA is degraded. This process conceivablyis controlled by the availability of ribosomalproteins.

Certain prokaryotes reduce their rates ofaccumulation of ribosomes during "differentia-tion" events. During endospore formation, Ba-cillus subtilis halts rRNA synthesis, apparentlybecause the RNA polymerase is proteolyticallymodified so that it no longer is capable ofrecognizing the rDNA promoter. Myxococcusxanthus undergoing microcyst formation alsorestricts the net accumulation of ribosomes, butin this case there apparently is little or noretardation of the relative rate of rRNA synthe-sis. Instead, the newly synthesized rRNA, aswell as probably some pre-existing rRNA, isdegraded.

Post-Transcriptional Processing of rRNAThe genes specifying the rRNA molecules

comprise transcriptional units; any RNA po-lymerase molecule which begins to read the 16Sgene continues on to read the 23S gene and thenthe 5S gene before terminating transcription.This transcriptional organization guaranteesproduction of equimolar quantities of the rRNAcomponents. However, a tandem transcript ofthe rDNA does not normally appear in growingcells. Probably a specific endonuclease cleaves acompleted rRNA chain from the nascent RNAsoon after the RNA polymerase moves into theadjacent, "downstream" gene in the transcrip-tional unit.The immediate products of transcription of

the rDNA are not structurally identical to themature rRNA components. In general, the pre-cursor (p16) of mature 16S rRNA (m16) is about10% larger than the mature molecule; p23 is 5 to8% greater in molecular length than m23. Dur-ing the maturation of p16 rRNA in E. coli,excess sequences are removed from both the 3'-and the 5'-termini. Sequences are removed fromat least the 5'-terminus of p23 during its conver-sion to m23. The mature 5S rRNA moleculesalso are not the immediate products of thetranscription. The p5 rRNA molecule of E. coliis maximally only three nucleotides larger thanm5, but the maturation of 5S rRNA in Bacillussubtilis bears resemblance to events affectingthe high-molecular-weight rRNA components;extensive, precursor-specific sequences are re-

592 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

moved from both termini. The function (if any)of the sequences associated specifically with therRNA precursors is not known. The character ofthe enzyme(s) which affects the maturationalcleavages also is unknown. The substrates forthis cleavage enzyme(s) probably are not thefree, precursor rRNA molecules, but ratheraggregates of the precursors and specific ribo-somal proteins; inhibitors of protein synthesisconcomitantly prohibit rRNA maturation.

Post-transcriptional modifications of therRNA molecules also include methylation ofselect nucleosides and the formation of 5-ribosyluracil. In E. coli, certain of the methyla-tion events may occur in the absence of cellularprotein synthesis, and therefore the enzymescatalyzing them are capable of utilizing assubstrates polynucleotide chains which are notcomplexed with ribosomal proteins. Othermethylation reactions, as well as the formationof 5-ribosyluracil, do not occur in the absence ofprotein synthesis; rRNA-ribosomal protein ag-gregates are required substrates for the respon-sible enzymes. Examples of methyl transfer-ases, which employ as substrates either freerRNA or rRNA complexed with some ribosomalproteins, have been partially purified and char-acterized. All use S-adenosylmethionine as amethyl donor and specific nucleosides withinthe polynucleotide chains as acceptors.

ACKNOWLEDGMENTSDuring the assembly of this review I relied heavily

upon the discussions and criticisms of my colleagues,Mel Averner, Martin Pato, Bernadette Pace, CarlWoese, and Ford Doolittle. Considerable gratitudealso is extended to Howard Rickenberg, GeorgeMunro, Frank Harold, Nelda Sullivan, Sam Kaplan,and Jack Sadler for their critical review of a prelimi-nary draft of the manuscript. I am indebted toKatherine Lamb for invaluable assistance with thebibliography and to Gail Garrett and Helga Forsterfor their typing expertise. This review is dedicated toCarl R. Woese for his 41st birthday.

ADDENDUM IN PROOFSubsequent to the assembly of the above informa-

tion, several pertinent papers have been publishedor submitted to press. I would like to call attention totwo noteworthy items. First, Travers (A. Travers.1973. Control of ribosomal RNA synthesis in vitro.Nature (London) 244:15-18) has again implicated thefactor Or in control of rRNA production. He findsthat there exist two states for the rRNA promoter,"open" or "closed," depending upon the conforma-tion of the DNA template. If in vitro RNA synthesisis carried out at low salt concentrations (0.01 M) orat elevated temperatures (38 C), then the rRNA pro-moters are "open" and the RNA polymerase holo-enzyme is capable of preferentially transcribing

the rRNA genes. At higher salt concentrations (0.1M) and at a lower temperature (34 C), the rRNApromoters are "closed," and rRNA synthesis bypurified RNA polymerase is substantially enhancedby the presence of 'r. Moreover, Travers haspresented evidence that the 4r-stimulated rRNAsynthesis is inhibited by physiological concentrationsof ppGpp.The second finding that I would like to point out

is that of Schlessinger and his colleagues (N. Nikolaev,L. Silengo, and D. Schlessinger, 1973. Synthesisof a large precursor to ribosomal RNA in a mutantof E. coli. Proc. Nat. Acad. Sci. U.S.A., in press),who have presented evidence that RNase III isresponsible for cleaving the tandem transcript ofthe rDNA into the immediate precursors of themature rRNA molecules.

LITERATURE CITED1. Adesnik, M., and C. Levinthal. 1969. Synthesis

and maturation of ribosomal RNA in Esche-richia coli. J. Mol. Biol. 46:281-303.

2. Altenburg, L. C., and G. F. Saunders. 1971.Properties of hybrid ribosomes formed fromthe subunits of mesophilic and thermophilicbacteria. J. Mol. Biol. 55:487-502.

3. Aronson, A. I., and A. Holowczyk. 1965. Compo-sition of bacterial ribosomal RNA: heterogene-ity within a given organism. Biochim. Bio-phys. Acta 95:217-231.

4. Attardi, G., and F. Amaldi. 1970. Structure andsynthesis of ribosomal RNA. Annu. Rev. Bio-chem. 39:183-226.

5. Attardi, G., P. C. Huang, and S. Kabat. 1965.Recognition of ribosomal RNA sites in DNA.1. Analysis of the Escherichia coli system.Proc. Nat. Acad. Sci. U.S.A. 53:1490-1498.

6. Averner, M. J., and N. R. Pace. 1972. Thenucleotide sequence of marsupial 5S ribo-somal ribonucleic acid. J. Biol. Chem.247:4491-4493.

7. Bacon, K., and E. Rosenberg. 1967. Ribonucleicacid synthesis during morphogenesis in Myx-ococcus xanthus. J. Bacteriol. 94:1883-1889.

8. Bautz, E. K. F. 1972. Regulation of RNA synthe-sis. Progr. Nucl. Acid Res. Mol. Biol.12:129-160.

9. Bendich, A. J., and B. J. McCarthy. 1970.Ribosomal RNA homologies among distantlyrelated organisms. Proc. Nat. Acad. Sci.U.S.A. 65:349-356.

10. Bickle, T. A., and R. R. Traut. 1971. Differencesin size and number of 80S and 70S ribosomalproteins by dodecyl sulfate gel electophoresis.J. Biol. Chem. 246:6828-6834.

11. Birnbaum, L. S., and S. Kaplan. 1971. Localiza-tion of a portion of the ribosomal RNA genesin Escherichia coli. Proc. Nat. Acad. Sci.U.S.A. 68:925-929.

12. Bjbrk, G. R., and L. A. Isaksson. 1970. Isolationof mutants of Escherichia coli lacking 5-methyluracil in transfer ribonucleic acid or1-methylguanine in ribosomal RNA. J. Mol.Biol. 51:83-100.

593VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

13. Bleyman, M., M. Kondo, N. Hecht, and C. R.Woese. 1969. Transcriptional mapping: func-tional organization of the ribosomal and trans-fer ribonucleic acid cistrons in the Bacillussubtilis genome. J. Bacteriol. 99:535-543.

14. Bleyman, M., and C. R. Woese. 1969. "Tran-scriptional mapping." I. Introduction to themethod and the use of actinomycin D as atranscriptional mapping agent. Proc. Nat.Acad. Sci. U.S.A. 63:532-539.

15. Blundell, M. R., and D. G. Wild. 1969. Inhibi-tion of bacterial growth by metal salts. Theaccumulation of ribonucleic acid during inhi-bition of Escherichia coli by cobalt chloride.Biochem. J. 115:213-223.

16. Boedtker, H. 1971. Conformation independentmolecular weight determinations of RNA bygel electrophoresis. Biochim. Biophys. Acta240:448-453.

17. Boedtker, H., and D. G. Kelling. 1967. Theordered structure of 5S RNA. Biochem. Bio-phys. Res. Commun. 29:758-766.

18. Bonamy, C., L. Hirschbein, and J. Szulmajster.1973. Synthesis of ribosomal ribonucleic acidduring sporulation of Bacillus subtilis. J. Bac-teriol. 113:1296-1306.

19. Borda, L., M. H. Green, and M. D. Kamen.1969. Sedimentation properties of ribonucleicacid from Rhodopseudomonas spheroides. J.Gen. Microbiol. 56:345-351.

20. Borek, E., A. Ryan, and J. Rockenbach. 1955.Nucleic acid metabolism in relation to thelysogenic phenomenon. J. Bacteriol.69:460-467.

21. Borek, E., and P. R. Srinivasan. 1966. Themethylation of nucleic acids. Annu. Rev. Bio-chem. 35:275-298.

22. Borst, P., and L. A. Grivell. 1971. Mitochondrialribosomes. FEBS Lett. 13:73-78.

23. Bowman, C. M., J. E. Dahlberg, T. Ikemura, J.Konisky, and M. Nomura. 1971. Specific inac-tivation of 16S ribosomal RNA induced bycolicin E3 in vivo. Proc. Nat. Acad. Sci.U.S.A. 68:964-968.

24. Bremer, H., and L. Berry. 1971. Co-transcriptionof 16S and 23S ribosomal RNA in Escherichiacoli. Nature N. Biol. 234:81-83.

25. Bremer, H., M. W. Konrad, K. Gaines, and G. S.Stent. 1965. Direction of chain growth inenzymatic RNA synthesis. J. Mol. Biol.13:540-553.

26. Bremer, H., and D. Yuan. 1968. RNA chaingrowth-rate in Escherichia coli. J. Mol. Biol.38:163-180.

27. Britten, R. J., and B. J. McCarthy. 1962. Thesynthesis of ribosomes in Escherichia coli. II.Analysis of the kinetics of tracer incorporationin growing cells. Biophys. J. 2:49-55.

28. Brown, G. M., and G. Attardi. 1965. Methyla-tion of nucleic acids in Hela cells. Biochem.Biophys. Res. Commun. 20:298-302.

29. Brownlee, G. G., and E. Cartwright. 1971. Se-quence studies on precursor 16S ribosomalRNA of Escherichia coli. Nature N. Biol.232:50-52.

30. Brownlee, G. G., E. Cartwright, T. McShane,and R. Williamson. 1972. The nucleotide se-quence of somatic 5S RNA from Xenopuslaevis. FEBS Lett. 25:8-12.

31. Brownless, G. G., F. Sanger, and B. G. Barrell.1967. Nucleotide sequence of 5S-ribosomalRNA from Escherichia coli. Nature (London)215:735-736.

32. Brownlee, G. G., F. Sanger, and B. G. Barrell.1968. The sequence of 5S ribosomal ribonu-cleic acid. J. Mol. Biol. 34:379-412.

33. Burdon, R. H. 1971. Ribonucleic acid matura-tion in Animal cells. Progr. Nucl. Acid Res.Mol. Biol. 11:33-79.

34. Cahn, F., E. M. Schachter, and A. Rich. 1970.Polypeptide synthesis with ribonuclease-digested ribosomes. Biochim. Biophys. Acta209:512-520.

35. Cantor, C. R. 1967. Possible conformations of5S ribosomal RNA. Nature (London) 216:513-514.

36. Cantor, C. R. 1968. The extent of base pairing in5S ribosomal RNA. Proc. Nat. Acad. Sci.U.S.A. 59:478-483.

37. Cashel, M. 1970. Inhibition of RNA polymeraseby ppGpp, a nucleotide accumulated duringthe stringent response to amino acid starva-tion in Escherichia coli. Cold Spring HarborSymp. Quant. Biol. 35:407-413.

38. Cashel, M., and J. Gallant. 1969. Two com-pounds implicated in the function of the RCgene of Escherichia coli. Nature (London)221:838-841.

39. Cashel, M., and B. Kalbacher. 1970. The controlof ribonucleic acid synthesis in Escherichiacoli. V. Characterization of a nucleotide asso-ciated with the stringent response. J. Biol.Chem. 245:2309-2318.

40. Chamberlin, M. 1970. Transcription 1970: asummary. Cold Spring Harbor Symp. Quant.Biol. 35:851-873.

41. Chambers, R. W. 1966. The chemistry of pseu-douridine. Progr. Nucl. Acid Res. Mol. Biol.5:349-398.

42. Champney, W. S., and P. S. Sypherd. 1970.Absence of thiolated nucleotides in ribosomalRNA of Escherichia coli. Biochem. Biophys.Res. Commun. 41:1328-1334.

43. Chang, F. N., C. J. Sih, and B. Weisblum.1966. Lincomycin, an inhibitor of aminoacylsRNA binding to ribosomes. Proc. Nat. Acad.Sci. U.S.A. 55:431-438.

44. Chantrenne, H. 1964. Changes in the patterns ofnucleic acid and protein synthesis caused by apurine analogue. J. Cell. Comp. Physiol.65(Suppl. 1):149-164.

45. Cohen, S. S. 1970. Are/were mitochondria andchloroplasts microorganisms? Amer. Sci.58:281-289.

46. Colli, W., and M. Oishi. 1969. Ribosomal RNAgenes in bacteria: evidence for the nature ofthe physical linkage between 16S and 23SRNA genes in Bacillus subtilis. Proc. Nat.Acad. Sci. U.S.A. 64:642-649.

594 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

47. Colli, W., and M. Oishi. 1970. A procedure forgene purification: the purification of the ribo-somal RNA genes of Bacillus subtilis or DNA-RNA hybrids. J. Mol. Biol. 51:657-669.

48. Colli, W., I. Smith, and M.. Oishi. 1971. Physicallinkage between 5S, 16S and 23S ribosomalRNA genes in Bacillus subtilis. J. Mol. Biol.56:117-127.

49. Colwell, R. R., and M. Mandel. 1964. Basecomposition of deoxyribonucleic acid of ma-

rine and nonmarine vibrios deduced frombuoyant-density measurements in cesiumchloride. J. Bacteriol. 88:1816-1817.

50. Corte, G., D. Schlessinger, D. Longo, and P.Venkov. 1971. Transformation of 17S to 16Sribosomal RNA using ribonuclease II of Esche-richia coli. J. Mol. Biol. 60:325-338.

51. Cotter, R. I., and W. B. Gratzer. 1970. Thiolatedbases in ribosomal RNA of Escherichia coli.Biochem. Biophys. Res. Commun. 39:766-773.

52. Cox, R. A. 1970. A spectrophotometric study ofthe secondary structure of ribonucleic acidisolated from the smaller and larger ribosomalsubparticles of rabbit reticulocytes. Biochem.J. 117:101-118.

53. Cramer, F., and V. A. Erdmann. 1968. Amountof adenine and uracil base pairs in Escherichiacoli 23S, 16S and 5S ribosomal RNA. Nature(London) 218:92-93.

54. Crick, F. H. C. 1968. The origin of the geneticcode. J. Mol. Biol. 38:367-379.

55. Cutler, R. G., and J. E. Evans. 1967. Isolation ofselected segments from the genome of HfrEscherichia coli. J. Mol. Biol. 26:81-90.

56. Cutler, R. G., and J. E. Evans. 1967. Relativetranscription activity of different segments ofthe genome throughout the cell division cycleof Escherichia coli. The mapping of ribosomaland transfer RNA and the determination ofthe direction of replication. J. Mol. Biol.26:91-105.

57. Dagley, S., G. Turnock, and D. G. Wild. 1963.The accumulation of ribonucleic acid by a

mutant of Escherichia coli. Biochem. J.88:555-566.

58. Dagley, S., A. E. White, D. G. Wild, and J.Sykes. 1962. Synthesis of protein and ribo-somes by bacteria. Nature (London)194:25-27.

59. Dahlberg, A. E., and A. C. Peacock. 1971.Studies of 16 and 23S ribosomal RNA ofEscherichia coli using composite gel electro-phoresis. J. Mol. Biol. 55:61-74.

60. Davies, J., and M. Nomura. 1972. The geneticsof bacterial ribosomes. Annu. Rev. Genetics6:203-204.

61. Delihas, N. 1970. Effect of ribonuclease onEscherichia coli ribosomes. Biochem. Bio-phys. Res. Commun. 39:905-910.

62. Dennis, P. P. 1971. Regulation of stable RNAsynthesis in Escherichia coli. Nature N. Biol.232:43-47.

63. Dennis, P. P. 1972. Regulation of ribosomal andtransfer ribonucleic acid synthesis in Esche-

richia coli B/r. J. Biol. Chem. 247:2842-2845.64. Dennis, P. P. 1972. Stable ribonucleic acid

synthesis during the cell division cycle inslowly growing Escherichia coli B/r. J. Biol.Chem. 247:204-208.

65. Doi, R. H., and R. T. Igarashi. 1964. Genetictranscription during morphogenesis. Proc.Nat. Acad. Sci. U.S.A. 52:755-762.

66. Doi, R. H., and R. T Igarashi. 1964. Ribonucleicacids of Bacillus subtilis spores and sporulat-ing cells. J. Bacteriol. 87:323-328.

67. Doi, R. H., and R. T. Igarashi. 1965. Conserva-tion of ribosomal and messenger ribonucleicacid cistrons in Bacillus species. J. Bacteriol.90:384-390.

68. Donini, P. 1972. Turnover of ribosomal RNAduring the stringent response in Escherichiacoli. J. Mol. Biol. 72:553-569.

69. Doolittle, W. F. 1972. Ribosomal ribonucleicacid synthesis and maturation in the blue-green alga Anacystis nidulans. J. Bacteriol.111:316-324.

70. Doolittle, W. F. 1973. Postmaturational cleavageof 23S ribosomal ribonucleic acid and itsmetabolic control in the blue-green alga Ana-cystis nidulans. J. Bacteriol. 113:1256-1263.

71. Doolittle, W. F., and N. R. Pace. 1970. Synthesisof 5S ribosomal RNA in Escherichia coli afterrifampicin treatment. Nature (London)228:125-129.

72. Doolittle, W. F., and N. R. Pace. 1971. Tran-scriptional organization of the ribosomal RNAcistrons in Escherichia coli. Proc. Nat. Acad.Sci. U.S.A. 68:1786-1790.

73. Dubin, D. T., and A. Giinalp. 1967. Minornucleotide composition of ribosomal precur-sor, and ribosomal ribonucleic acid in Esche-richia coli. Biochim. Biophys. Acta134:106-123.

74. Dubnau, D., I. Smith, and J. Marmur. 1965.Gene conservation in Bacillus species. II. Thelocation of genes concerned with the synthesisof ribosomal components and soluble RNA.Proc. Nat. Acad. Sci. U.S.A. 54:724-730.

75. Dubnau, D., I. Smith, P. Morell, and J. Mar-mur. 1965. Gene conservation in Bacillus spe-cies. I. Conserved genetic and nucleic acidbase sequence homologies. Proc. Nat. Acad.Sci. U.S.A. 54:491-498.

76. DuBuy, B., and S. M. Wissman. 1971. Nucleo-tide sequences of Pseudomonas fluorescens 5Sribonucleic acid. J. Biol. Chem. 246:747-761.

77. Ecker, R. E., and M. Schaechter. 1963. Bacterialgrowth under conditions of limited nutrition.Ann. N.Y. Acad. Sci. 102:549-563.

78. Edlin, G., and P. Broda. 1968. Physiology andgenetics of the "ribonucleic acid control" locusin Escherichia coli. Bacteriol. Rev.32:206-226.

79. Edlin, G., and J. Neuhard. 1967. Regulation ofnucleoside triphosphate pools in Escherichiacoli. J. Mol. Biol. 24:225-230.

80. Ehresmann, C., P. Stiegler, P. Fellner, and J.-P.Ebel. The determination of the primary struc-ture of the 16S ribosomal RNA of Escherichia

595VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

coli. II. Nucleotide sequences of products frompartial enzymatic hydrolysis. Biochemie54:901-967.

81. Eliceiri, G. L. 1971. 35S incorporation into bacte-rial RNA. Biochem. Biophys. Res. Commun.42:784-788.

82. Ennis, H. L., and M. Lubin. 1965. Pre-ribosomalparticles formed in potassium-depleted cells.Biochim. Biophys. Acta 95:605-623.

83. Fahnestock, S., V. Erdmann, and M. Nomura.1973. Reconstitution of 50S ribosomal sub-units from protein-free ribonucleic acid. Bio-chemistry 12:220-224.

84. Fellner, P., C. Ehresmann, and J.-P. Ebel. 1972.The determination of the primary structure ofthe 16S ribosomal RNA of Escherichia coli. I.Nucleotide sequence analysis of TI and pan-creatic ribonuclease digestion products. Bi-ochemie 54:853-900.

85. Fellner, P., C. Ehresmann, P. Stiegler, and J.-P.Ebel. 1972. Partial nucleotide sequence of 16Sribosomal RNA from Escherichia coli. NatureN. Biol. 239:1-5.

86. Fellner, P., and F. Sanger. 1968. Sequenceanalysis of specific areas of the 16S and 23Sribosomal RNAs. Nature (London) 219:236-238.

87. Feunteun, J., B. R. Jordan, and R. Monier. 1972.Study of the maturation of 5S RNA precursorsin Escherichia coli. J. Mol. Biol. 70:465-474.

88. Forget, B. G., and B. Jordan. 1970. 5S RNAsynthesized by Escherichia coli in the pres-ence of chloramphenicol: different 5'-terminalsequences. Science 167:382-384.

89. Forget, B. G., and S. M. Weissman. 1967.Nucleotide sequence of KB cell 5S RNA.Science 158:1695-1699.

90. Friesen, J. D. 1966. Control of messenger RNAsynthesis and decay in Escherichia coli. J.Mol. Biol. 20:559-573.

91. Fujisawa, Y., and A. Sibatani. 1954. Is there anyquantitative relationship between the synthe-sis and the breakdown of nucleic acids inliving cells? Experientia 10:172-180.

92. Furano, A. V., and M. I. harris. 1971. Theactivity of ribosomes whose RNA has beendegraded by incubation in the presence orabsence of oxidized glutathione. Biochim. Bio-phys. Acta 247:291-303.

93. Galibert, F., P. Tiollais, F. Sanfourche, and M.Boiron. 1971. Coordination de la transcriptiondes RNA 5S et 23S et coordination de lamaturation du RNA 5S et de la sous-uniteribosomique 50S chez Escherichia coli. Eur. J.Biochem. 20:381-391.

94. Gallant, J., and B. Harada. 1969. The control ofribonucleic acid synthesis in Escherichia coli.III. The functional relationship between pu-rine ribonucleoside triphosphate pool sizesand the rate of ribonucleic acid accumulation.J. Biol. Chem. 244:3125-3132.

95. Geiduschek, E. P., and R. Haselkorn. 1969.Messenger RNA. Annu. Rev. Biochem.38:647-676.

96. Gesteland, R. F. 1966. Unfolding of Escherichia

coli ribosomes by removal of magnesium. J.Mol. Biol. 18:356-371.

97. Gilham, P. T. 1970. RNA sequence analysis.Annu. Rev. Biochem. 39:227-250.

98. Gillespie, D. 1968. The formation and detectionofDNA-RNA hybrids, p. 641-668. In L. Gross-man and K. Moldave (ed.), Methods in enzy-mology, vol. 12, part B. Academic Press Inc.,New York.

99. Goldthwaite, C., D. Dubnau, and I. Smith. 1970.Genetic mapping of antibiotic resistancemarkers in Bacillus subtilis. Proc. Nat. Acad.Sci. U.S.A. 65:96-103.

100. Goldwasser, E., and R. L. Hendrikson. 1966. Thebiochemistry of pseudouridine. Progr. Nucl.Acid Res. Mol. Biol. 5:399-416.

101. Gordon, J., and H. G. Boman. 1964. Studies onmicrobial RNA. II. Transfer of methyl groupsfrom methionine to the RNA of a ribonucleo-protein particle. J. Mol. Biol. 9:638-653.

102. Gorelic, L. 1970. Chromosomal location of ribo-somal RNA cistrons in Escherichia coli. Mol.Gen. Genet. 106:323-327.

103. Grienenberger, J.-M., and D. Simon. 1972.Etude de la degradation des acides ribonucle-iques ribosomaux de quelques bacteries dugenre Agrobacterium. C. R. Acad. Sci. (Paris)274:1399-1402.

104. Grivell, L. A., and H. L. Walg. 1972. Subunithomology between Escherichia coli, mitochon-drial and chloroplast ribosomes. Biochem.Biophys. Res. Commun. 49:1452-1458.

105. Griunberger, D., R. N. Maslova, and F. L. Sorm.1964. Effect of 8-azaguanine on the synthesisof pulse-labeled ribonucleic acid in Bacilluscereus. Collect. Czech. Chem. Commun.29:152-160.

106. Guthrie, C., H. Nashimoto, and M. Nomura.1969. Studies on the assembly of ribosomes invivo. Cold Spring Harbor Symp. Quant. Biol.34:69-75.

107. Gutter, B., and Y. Becker. 1972. Synthesis andmaturation of ribosomal RNA during the de-velopmental cycle of trachoma agent, a pro-karyotic obligate parasite of eukaryocytes. J.Mol. Biol. 66:239-253.

108. Hall, R. H. 1971. The modified nucleosides innucleic acids. Columbia University Press,New York.

109. Harney, C. E., and D. Nakada. 1970. Coordinatesynthesis of ribosomal RNA and messenger forribosomal proteins. Biochim. Biophys. Acta213:529-531.

110. Harshman, R. B., and H. Yamazaki. 1971.Formation of ppGpp in a relaxed and stringentstrain of Escherichia coli during diauxie lag.Biochemistry 10:3980-3982.

111. Hartman, K. A., and G. J. Thomas, Jr. 1970.Secondary structure of ribosomal RNA.Science 170:740-741.

112. Haseltine, W. A. 1972. In vitro transcription ofEscherichia coli ribosomal RNA genes. Nature(London) 235:329-333.

113. Hayashi, Y., S. Osawa, and K. Miura. 1966. Themethyl groups in ribosomal RNA from Esche-

596 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

richia coli. Biochim. Biophys. Acta 129:519-531.

114. Hayashi, M., and S. Spiegelman. 1961. Theselective synthesis of informational RNA inbacteria. Proc. Nat. Acad. Sci. U.S.A.47:1564-1580.

115. Hayes, F., D. Hayes, P. Fellner, and C. Ehres-mann. 1971. Additional nucleotide sequencesin precursor 16S ribosomal RNA from Esche-richia coli. Nature N. Biol. 232:54-55.

116. Hecht, N. B., and C. R. Woese. 1968. Separationof bacterial ribosomal ribonucleic acid from itsmacromolecular precursors by polyacrylamidegel electrophoresis. J. Bacteriol. 95:986-990.

117. Helser, T. L., J. E. Davies, and J. E. Dahlberg.,1971. Change in methylation of 16S ribosomalRNA associated with mutation to kasugamy-cin resistance in Escherichia coli. Nature N.Biol. 233:12-14.

118. Helser, T. L., J. E. Davies, and J. E. Dahlberg.1971. Mechanism of kasugamycin resistancein Escherichia coli. Nature'N. Biol. 235:6-9.

119. Higo, S., M. Higo, and S. Tanifuji. 1971. Spe-cific dissociation of Pea 25-s rRNA by hot-phenol treatment. Biochim. Biophys. Acta246:499-506.

120. Hill, L. R. 1966. An index to deoxyribonucleicacid base compositions of bacterial species. J.Gen. Microbiol. 44:419-437.

121. Hill, W. E., G. P. Rossetti, and K. E. Van Holde.1969. Physical studies of ribosomes from Esch-erichia coli. J. Mol. Biol. 44:263-277.

122. Hosokawa, K., and M. Nomura. 1965. Incom-plete ribosomes produced in chloramphenicol-and puromycin-inhibited Escherichia coli. J.Mol. Biol. 12:225-241.

123. Howland, G. P., and J. Ramus. 1971. Analysis ofblue-green and red algal ribo3omal-RNAs bygel electrophoresis. Arch. Mikrobiol.76:292-298.

124. Hussey, C., R. Losick, and A. L. Sonnenshein.1971. Ribosomal RNA synthesis is turned offduring, sporulation of Bacillus subtilis. J. Mol.Biol. 57:59-70.

125. Hussey, C., J. Pero, R. Shorenstein, and R.Losick. 1972. In vitro synthesis of ribosomalRNA by Bacillus subtilis RNA polymerase.Proc. Nat. Acad. Sci. U.S.A. 69:407-411.

126. Irr, J. D. 1972. Control of nucleotide metabolismand ribosomal ribonucleic acid synthesis dur-ing nitrogen starvation of Escherichia coli. J.Bacteriol. 110:554-561.

127. Jacobson, L. A. 1971. Evidence for the polycis-tronic transcription of ribosomal RNA inEscherichia coli. Biochem. Biophys. Res.Commun. 43:372-377.

128. Jarry, B., and R. Rosset. 1970. An attempt tolocalize 5S RNA cistrons in Escherichia coli.Biochem. Biophys. Res. Commun. 41:789-796

129. Johnson, J. D., and J. Horowitz. 1971. Charac-terization of ribosomes and RNAs from Myco-plasma hominis. Biochim. Biophys. Acta247:262-279.

130. Johnson, L., and D. Soll. 1970. In vitro biosyn-thesis of pseudouridine at the polynucleotide

level by an enzyme extract from Escherichiacoli. Proc. Nat. Acad. Sci. U.S.A. 67:943-950.

131. Jordan, B. R. 1971. Studies on 5S RNA confor-mation by partial ribonuclease hydrolysis. J.Mol. Biol. 55:423-439.

132. Julien, J., R. Rosset, and R. Monier. 1968.Metabolisme des acides ribonucleiques chezEscherichia coli carenee en phosphate. Bull.Soc. Chem. Biol. 49:131-145.

133. Kaempfer, R., and M. Meselson. 1968. Perma-nent association of 5S RNA molecules with50S ribosomal subunits in growing bacteria. J.Mol. Biol. 34:703-707.

134. Kaempfer, R. O., R. M. Meselson, and H. J.Raskas. 1968. Cyclic dissociation into stablesubunits and re-formation of ribosomes duringbacterial growth. J. Mol. Biol. 31:277-289.

135. Kennell, D. 1968. Titration of the gene sites onDNA by DNA-RNA hybridization. II. TheEscherichia coli chromosome. J. Mol. Biol.34:85-103.

136. Kennell, D. E. 1971. Principles and practices ofnucleic acid hybridization. Progr. Nucl. AcidRes. Mol. Biol. 11:259-301.

137. Kjeldgaard, N. O., and C. G. Kurland. 1963. Thedistribution of soluble and ribosomal RNA asa function of growth rate. J. Mol. Biol.6:341-348.

138. Koch, A. L. 1970. Overall controls on the biosyn-thesis of ribosomes in growing bacteria. J.Theor. Biol. 28:203-231.

139. Koch, A. L., and C. S. Deppe. 1971. In vivo assayof protein synthesizing capacity of Escherichiacoli from slowly growing chemostat cultures.J. Mol. Biol. 55:549-562.

140. Kohne, D. E. 1968. Isolation and characteriza-tion of bacterial ribosomal RNA cistrons.Biophys. J. 8:1104-1118.

141. Kono, M., and S. Osawa. 1964. Intermediarysteps of ribosome formation in Escherichiacoli. Biochim. Biophys. Acta 87:326-334.

142. Kono, M., E. Otaka, and S. Osawa. 1964.Changes in sedimentation properties of ribo-somal ribonucleic acids during the course ofribosome formation in Escherichia coli. Bio-chim. Biophys. Acta 91:612-618.

143. Koser, R. B., and J. R. Collier. 1971. Themolecular weight and thermolability ofIlyanassa ribosomal RNA. Biochim. Biophys.Acta 254:272-277.

144. Kossman, C. R., T D. Stamato, and D. E.Pettijohn. 1971. Tandem synthesis of the 16Sand 23S ribosomal RNA sequences of Esche-richia coli. Nature N. Biol. 234:102-104.

145. Kurland, C. G. 1960. Molecular characterizationof ribonucleic acid from Escherichia coli ribo-somes. I. Isolation and molecular weights. J.Mol. Biol. 2:83-91.

146. Kurland, C. G. 1972. Structure and function ofthe bacterial ribosome. Annu. Rev. Biochem.41:377-408.

147. Kurland, C. G., and 0. Maal0e. 1962. Regulationof ribosomal and transfer RNA synthesis. J.Mol. Biol. 4:193-210.

148. Kurland, C. G., M. Nomura, and J. D. Watson.

597VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

1962. The physical properties of the chloro-mycetin particles. J. Mol. Biol. 4:388-394.

149. Lai, C. J., and B. Weisblum. 1971. Alteredmethylation of ribosomal RNA in an eryth-romycin-resistant strain of Staphylococcusaureus. Proc. Nat. Acad. Sci. U.S.A.68:856-860.

150. Lai, C.-J., B. Weisblum, S. R. Fahnestock, andM. Nomura. 1973. Alteration of 23S ribosomalRNA and erythromycin-induced resistance tolincomycin and spiromycin in Staphylococcusaureus. J. Mol. Biol. 74:67-72.

151. Lazzarini, R. A., M. Cashel, and J. Gallant.1971. On the regulation of guanosine tetra-phosphate levels in stringent and relaxedstrains of Escherichia coli. J. Biol. Chem.246:4381-4385.

152. Lazzarini, R. A., and R. M. Winslow. 1970. Theregulation of RNA synthesis during growthrate transitions and amino acid deprivation inEscherichia coli. Cold Spring Harbor Symp.Quant. Biol. 35:383-390.

153. Leaver, C. J., and J. Ingle. 1971. The molecularintegrity of chloroplast ribosomal ribonucleicacid. Biochem. J. 123:235-243.

154. Lee, S. G., and W. R. Evans. 1971. Hybridribosome formation from Escherichia coli andchloroplast ribosome subunits. Science173:241-242.

155. Lessie, T. G. 1965. The atypical ribosomal RNAcomplement of Rhodopseudomonas spher-oides. J. Gen. Microbiol. 39:311-320.

156. Lewis, J. B., and P. Doty. 1970. Derivation of thesecondary structure of 5S RNA from its bind-ing of complementary oligonucleotides. Na-ture (London) 225:510-512.

157. Lindigkeit, R., and W. Handschack. 1965. Someproperties of ribonucleic acid obtained fromribosomal precursors of Escherichia coli. Bio-chim. Biophys. Acta 103:241-251.

158. Linn, T. G., A. L. Greenleaf, R. G. Shorenstein,and R. Losick. 1973. Loss of the sigma activityof RNA polymerase of Bacillus subtilis duringsporulation. Proc. Nat. Acad. Sci. U.S.A.70:1865-1869.

159. Lizardi, P. M., and D. J. Luck. 1971. Absence ofa 5S RNA component in the mitochondrialribosomes of Neurospora crassa. Nature N.Biol. 229:140-142.

160. Loening, U. E. 1968. Molecular weights of ribo-somal RNA in relation to evolution. J. Mol.Biol. 38:355-365.

161. Loening, U. E., and J. Ingle. 1967. Diversity ofRNA components in green plant tissues. Na-ture (London) 215:363-367.

162. Losick, R. 1972. In vitro transcription. Annu.Rev. Biochem. 41:409-446.

163. Losick, R., R. G. Shorenstein, and A. L. Sonen-shein. 1970. Structural alteration of RNApolymerase during sporulation. Nature (Lon-don) 227:910-913.

164. Losick, R., and A. L. Sonenshein. 1969. Changein the template specificity of RNA polymeraseduring sporulation of Bacillus subtilis. Nature(London) 224:35-37.

165. Lowry, C. V., and J. E. Dahlberg. 1971. Struc-tural differences between the 16S ribosomalRNA of Escherichia coli and its precursor.Nature N. Bidl. 232:52-54.

166. Lucas-Lenard, J., and F. Lipmann. 1971. Pro-tein biosynthesis. Annu. Rev. Biochem.40:409-448.

167. Maaloe, O., and N. 0. Kjeldgaard. 1966. Controlof macromolecular synthesis. W. A. Benjamin,Inc. New York.

168. Madison, J. T. 1968. Primary structure of RNA.Annu. Rev. Biochem. 37:131-148.

169. Maia, J. C. C., P. Kerjan, and J. Szulmajster.1971. DNA-dependent RNA polymerase fromvegetative cells and from spores of Bacillussubtilis. FEBS Lett. 13:269-274.

170. Maitra, U., and J. Hurwitz. 1965. The role ofDNA in RNA synthesis. IX. Nucleoside tri-phosphate termini in RNA polymerase prod-ucts. Proc. Nat. Acad. Sci. U.S.A. 54:815-822.

171. Mangiarotti, G., D. Apirion, D. Schlessinger,and L. Silengo. 1968. Biosynthetic precursorsof 30S and 50S ribosomal particles in Esche-richia coli. Biochemistry 7:456-472.

172. Mangiarotti, G., and D. Schlessinger. 1967.Polyribosome metabolism in Escherichia coli.II. Formation and lifetime of messenger RNAmolecules, ribosomal subunit couples andpolyribosomes. J. Mol. Biol. 29:395-418.

173. Manor, H., D. Goodman, and G. Stent. 1969.RNA chain growth rates in Escherichia coli. J.Mol. Biol. 39:1-29.

174. Manor, H., and R. Haselkorn. 1967. Properties ofribonucleic acid components in ribonucleo-protein particle preparations obtained from anEscherichia coli RC relaxed strain. J. Mol.Biol. 24:269-288.

175. Margulis, L. 1970. Origin of eukaryotic cells.Yale University Press, New Haven.

176. Marmur, J., and P. Doty. 1959. Heterogeneity indeoxyribonucleic acids. I. Dependence oncomposition of the configurational stability ofdeoxyribonucleic acids. Nature (London)183:1427-1429.

177. Marrs, B., and S. Kaplan. 1970. 23S precursorribosomal RNA of Rhodopseudomonasspheroides. J. Mol. Biol. 49:297-317.

178. Martin, R. 1968. Polarity in relaxed strains ofSalmonella typhimurium. J. Mol. Biol.31:127-134.

179. Maruta, H., T. Tsuchiya, and D. Mizuno. 1971.In vitro reassembly of functionally active 50Sribosomal particles from ribosomal proteinsand RNA's of Escherichia coli. J. Mol. Biol.61: 123-134.

180. McCarthy, B. J., R. J. Britten, and R. B.Roberts. 1962. The synthesis of ribosomes inEscherichia coli. III. Synthesis of ribosomalRNA. Biophys. J. 2:57-82.

181. Meselson, M., M. Nomura, S. Brenner, C. Dav-ern, and D. Schlessinger. 1964. Conservationof ribosomes during bacterial growth. J. Mol.Biol. 9:696-711.

182. Miall, S. H., and I. 0. Walker. 1969. Structuralstudies on ribosomes. II. Denaturation and

598 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

sedimentation of ribosomal subunits unfoldedin EDTA. Biochim. Biophys. Acta 174:551-560.

183. Midgley, J. E. M. 1962. The nucleotide basecomposition of ribonucleic acid from severalmicrobial species. Biochim. Biophys. Acta61:513-525.

184. Midgley, J. E. M. 1965. Studies on the structureof 23S ribosomal ribonucleic acid from Esche-richia coli. Biochim. Biophys. Acta 108:348-354.

185. Miller, 0. L., Jr., B. R. Beatty, B. A. Hamkalo,and C. A. Thomas, Jr. 1970. Electron micro-scopic visualization of transcription. ColdSpring Harbor Symp. Quant. Biol.35:505-512.

186. Miller, 0. L., Jr., and B. A. Hamkalo. 1972.Visualization of RNA synthesis on chromo-somes. Int. Rev. Cytol. 33:1-25.

187. Mirzabekov, A. D., and B. E. Griffin. 1972. 5SRNA conformation. Studies of its partial T,ribonuclease digestion by gel electrophoresisand two-dimensional thin-layer chromatogra-phy. J. Mol. Biol. 72:633-643.

188. Monier, R., J. Feunteun, B. Forget, B. Jordan,M. Reynier, and F. Verricchio. 1969. 5S RNAand the assembly of bacterial ribosomes. ColdSpring Harbor Symp. Quant. Biol.34:139-148.

189. Moore, R. L., and B. J. McCarthy. 1967. Com-parative study of ribosomal ribonucleic acidcistrons in enterobacteria and myxobacteria.J. Bacteriol. 94:1066-1074.

190. Morell, P., I. Smith, D. Dubnau, and J. Marmur.1967. Isolation and characterization of lowmolecular weight ribonucleic acid species fromBacillus subtilis. Biochemistry 6:258-265.

191. Morris, D. W., and J. A. DeMoss. 1966. Poly-some transitions and the regulation of ribonu-cleic acid synthesis in Escherichia coli. Proc.Nat. Acad. Sci. U.S.A. 56:262-268.

192. Muto, A. 1970. Nucleotide distribution of Esche-richia coli 16S ribosomal ribonucleic acid.Biochemistry 9:3683-3694.

193. Nakada, D. 1965. Formation of ribosomes by a"relaxed" mutant of Escherichia coli. J. Mol.Biol. 12:695-725.

194. Nakada, D., I. A. C. Anderson, and B. Magasa-nik. 1964. Fate of the ribosomal RNA pro-duced by a "relaxed" mutant of Escherichiacoli. J. Mol. Biol. 9:472-488.

195. Nashimoto, H., and M. Nomura. 1970. Structureand function of bacterial ribosomes. XI. De-pendence of 50S ribosomal assembly on simul-taneous assembly of 30S subunits. Proc. Nat.Acad. Sci. U.S.A. 67:1440-1447.

196. Nass, S. 1969. The significance of the structuraland functional similarity of bacteria and mito-chondria. Int. Rev. Cytol. 25:55-129.

197. Nazar, R. N., L. A. Tyfield, and J. T.-F. Wong.1972. Regulation of ribonucleic acid accumu-lation in vivo by nucleoside triphosphates. J.Biol. Chem. 247:798-804.

198. Nazar, R. N., and J. T.-F. Wong. 1972. Nucleo-tide changes and the regulation of ribonucleic

acid accumulation during growth rate shifts inEscherichia coli. J. Biol. Chem. 247:790-797.

199. Neidhardt, F. C. 1963. Effects of environment onthe composition of bacterial cells. Annu. Rev.Microbiol. 17:61-86.

200. Neidhardt, F. C. 1963. Properties of a bacterialmutant lacking amino acid control of RNAsynthesis. Biochim. Biophys. Acta 68:365-379.

201. Neidhardt, F. C. 1964. The regulation of RNAsynthesis in bacteria. Progr. Nucl. Acid Res.Mol. Biol. 3:145-181.

202. Neidhardt, F. C., and Eidlic. 1963. Characteri-zation of the RNA formed under conditions ofrelaxed amino acid control in Escherichia coli.Biochim. Biophys. Acta 68:380-388.

203. Nichols, J. L., and B. G. Lane. 1966. N4-meth-yl-2'-O-methyl cytidine and other methyl-substituted nucleoside constituents of Esche-richia coli ribosomal and soluble RNA. Bio-chim. Biophys. Acta 119:649-651.

204. Nichols, J. L., and B. G. Lane. 1968. The in vitroO2'-methylation of RNA in the ribonucleo-protein precursor-particles from a relaxed mu-tant of Escherichia coli. Can. J. Biochem.46:109-115.

205. Nierlich, D. P. 1972. Regulation of ribonucleicacid synthesis in growing bacterial cells. I.Control over the total rate of RNA synthesis.J. Mol. Biol. 72:751-764.

206. Nierlich, D. P. 1972. Regulation of ribonucleicacid synthesis in growing bacterial cells. II.Control over the composition of the newlymade RNA. J. Mol. Biol. 72:765-777.

207. Nirenberg, M. W., and J. H. Matthaei. 1961.The dependence of cell-free protein synthesisin Escherichia coli upon naturally-occurringor synthetic polyribonucleotides. Proc. Nat.Acad. Sci. U.S.A. 47:1588-1602.

208. Nomura, M. 1970. Bacterial ribosome. Bacteriol.Rev. 34:228-277.

209. Nomura, M. 1972. Assembly of bacterial ribo-somes. Fed. Proc. 31:18-20.

210. Nomura, M., and V. A. Erdmann. 1970. Recon-stitution of 50S ribosomal subunits from dis-sociated molecular components. Nature (Lon-don) 228:744-748.

211. Nomura, M., P. Traub, and H. Bechmann. 1968.Hybrid 30S ribosomal particles reconstitutedfrom components of different bacterial origins.Nature (London) 219:793-799.

212. Norris, T. E., and A. L. Koch. 1972. Effect ofgrowth rate on the relative rates of synthesis ofmessenger, ribosomal and transfer RNA inEscherichia coli. J. Mol. Biol. 64:633-649.

213. Oishi, M., and N. Sueoka. 1965. Location ofgenetic, loci of ribosomal RNA on Bacillussubtilis chromosome. Proc. Nat. Acad. Sci.U.S.A. 54:483-491.

214. Okano, P., K. Bacon, and E. Rosenberg. 1970.Ribonucleic acid synthesis during microcystformation in Myxococcus xanthus: characteri-zation by deoxyribonucleic acid-ribonucleicacid hybridization. J. Bacteriol. 104:275-282.

215. Osawa, S. 1968. Ribosome formation and struc-ture. Annu. Rev. Biochem. 37:109-130.

599VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

216. Otaka, E., S. Osawa, and A. Sibatani. 1964.Stimulation of "4C-leucine incorporation intoprotein in vitro' by ribosomal RNA of Esche-richia coli. Biochem. Biophys. Res. Commun.15:568-574.

217. Pace, B., and L. L. Campbell. 1967. Correlationof maximal growth temperature and ribosomeheat stability. Proc. Nat. Acad. Sci. U.S.A.57:1110-1116.

218. Pace, B., and L. L. Campbell. 1971. Homology ofribosomal ribonucleic acid of Desulfovibriospecies with Desulfovibrio vulgaris. J. Bacte-riol. 106:717-719.

219. Pace, B., and L. L. Campbell. 1971. Homology ofribosomal ribonucleic acid of diverse bacterialspecies with Escherichia coli and Bacillusstearothermophilus. J. Bacteriol. 107:543-547.

220. Pace, B., and N. R. Pace. 1971. Gene dosage for5S ribosomal ribonucleic acid in Escherichiacoli and Bacillus megaterium. J. Bacteriol.105:142-149.

221. Pace, B., R. L. Peterson, and N. R. Pace. 1970.Formation of all stable RNA species in Esche-richia coli by posttranscriptional modifica-tion. Proc. Nat. Acad. Sci. U.S.A. 65:1097-1104.

222. Pace, N. R., M. L. Pato, J. McKibbin, and C. W.Radcliffe. 1973. Precursors of 5S ribosomalRNA in Bacillus subtilis. J. Mol. Biol.75:619-631.

223. Pace, N. R., T. Walker, B. Pace, and R. L.Erikson. 1974. The nucleotide sequence ofchicken 5S ribosomal ribonucleic acid. Sub-mitted to J. Biol. Chem.

224. Pato, M. L. 1972. Regulation of chromosomereplication and the bacterial cell cycle. Annu.Rev. Microbiol. 26:347-368.

225. Pato, M. L., and K. von Meyenburg. 1970.Residual RNA synthesis in Escherichia coliafter inhibition of initiation of transcriptionby rifampicin. Cold Spring Harbor Symp.Quant. Biol. 35:497-504.

226. Payne, P. J., and T. A. Dyer. 1971. Characteriza-tion of cytoplasmic and chloroplast 5S ribo-somal ribonucleic acid from broad-beanleaves. Biochem. J. 124:83-89.

227. Pederson, S., and N. 0. Kjeldgaard. 1972. Ahybridization assay specific for ribosomalRNA from Escherichia coli. Mol. Gen. Genet.118:85-91.

228. Pbne, J. J., E. Knight, Jr., and J. E. Darnell, Jr.1968. Characterization of a new low molecularweight RNA in HeLa cell ribosomes. J. Mol.Biol. 33:609-623.

229. Peterson, R. L., C. W. Radcliffe, and N. R. Pace.1971. Ribonucleic acid synthesis in bacteriatreated with toluene. J. Bacteriol. 107:585-588.

230. Pettijohn, D. E. 1972. Ordered and preferentialinitiation of ribosomal RNA synthesis in vitro.Nature N. Biol. 235:204-206.

231. Pettijohn, D. E., K. Clarkson, C. R. Kossman,and 0. G. Stonington. 1970. Synthesis ofribosomal RNA on a protein-DNA complex

isolated from bacteria: a comparison of ribo-somal RNA synthesis in vitro and in vivo. J.Mol. Biol. 52:281-300.

232. Pettijohn, D. E., and C. R. Kossman. 1973. Thetandem precursor of ribosomal RNA in Esche-richia coli: organization, initial cleavage andchain termination. Submitted to J. Mol. Biol.

233. Pettijohn, D. E., 0. G. Stonington, and C. R.Kossman. 1970. Chain termination of ribo-somal RNA synthesis in vitro. Nature (Lqn-don) 228:235-239.

234. Pigott, G. H., and N. G. Carr. 1972. Homologybetween nucleic acids of blue-green algae andchloroplasts of Euglena gracilis. Science175: 1259-1261.

235. Purdom, I., J. 0. Bishop, and M. L. Birnstiel.1970. Scattered arrangement of the bacterialribosomal cistrons. Nature (London)227:239-242.

236. Quash, G. 1963. The distribution of proteinantigens among Escherichia coli ribosomes.Biochim. Biophys. Acta 68:330-332.

237. Raacke, I. D. 1968. 'Cloverleaf conformation for5S RNAs. Biochem. Biophys. Res. Commun.31:528-533.

238. Raff, R. A., and H. R. Mahler. 1972. Thenon-symbiotic origin of mitochondria. Science177:575-582.

239. Ramsey, W. S., and M. Dworkin. 1970. Stablemessenger ribonucleic acid and germination ofMyxococcus xanthus microcysts. J. Bacteriol.101:531-540.

240. Raue, H. A., and M. Gruber. 1971. Control ofstable-RNA synthesis and maturation of 5SRNA in Salmonella typhimurium RC'tr andRCrSI. Biochim. Biophys. Acta 232:314-323.

241. Raven, P. H. 1970. A multiple origin for plastidsand mitochondria. Science 169:641-646.

242. Richards, E. G. 1969. 5S RNA. An analysis ofpossible base-pairing schemes. Eur. J. Bio-chem. 10:36-42.

243. Roberts, R. B. 1965. The synthesis of ribosomalprotein. J. Theor. Biol. 8:49-53.

244. Robinson, A., and J. Sykes. 1971. A study of theatypical ribosomal RNA components ofRhodopseudomonas spheroides. Biochim. Bio-phys. Acta 238:99-115.

245. Ron, E. Z., R. E. Kohler, and B. D. Davis. 1966.Increased stability of polysomes in an Esche-richia coli mutant with relaxed control of RNAsynthesis. Proc. Nat. Acad. Sci. U.S.A.56:471-475.

246. Rosset, R., J. Julien, and R. Monier. 1966.Ribonucleic acid composition of bacteria as afunction of growth rate. J. Mol. Biol.18:308-320.

247. Rosset, R., and R. Monier. 1963. A propos de lapresence d'acide ribonucleique de faible poidsmoleculaire dans les ribosomes d'Escherichiacoli. Biochim. Biophys. Acta 68:653-656.

248. Rudner, R., B. Prokop-Schneider, and E. Char-gaff. 1964. Rhythmic alternations in the rateof synthesis and the composition of rapidlylabelled ribonucleic acid during the synchro-

600 PACE


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

nous growth of bacteria. Nature (London)203:479-483.

249. Rudner, R., E. Rejman, and E. Chargaff. 1965.Genetic implications of periodic pulsations ofthe rate of synthesis and the composition ofrapidly labeled bacterial RNA. Proc. Nat.Acad. Sci. U.S.A. 54:904-911.

250. Ryan, A. M., and E. Borek. 1971. The relaxedcontrol phenomenon. Progr. Nucl. Acid Res.Mol. Biol. 11:193-228.

251. Ryan, J. L., and H. J. Morowitz. 1969. Partialpurification of native rRNA and tRNA cis-trons from Mycoplasma sp. (KID). Proc. Nat.Acad. Sci. U.S.A. 63:12828-1289.

252. Sanger, F., and G. G. Brownlee. 1967. A two-dimensional fractionation method for radioac-tive nucleotides, p. 361-381. In L. Grossmanand K. Moldave (ed.), Methods in enzymol-ogy, vol. 12, part A. Academic Press Inc., NewYork.

253. Santer, M., and U. V. Santer. 1972. Characteri-zation of the 5' and 3' ends of the 16Sribonucleic acid from Tl-ribonuclease-treated30S ribosomes. Biochemistry 11:786-791.

254. Santer, U. V., and M. Santer. 1972. The se-quence of the 3-OH and of the 16S RNA ofEscherichia coli. FEBS Lett. 21:311-314.

255. Sarov, I., and Y. Becker. 1968. RNA in theelementary bodies of trachoma agent. Nature(London) 217:849-852.

256. Schaechter, M., 0. Maaloe, and N. 0. Kjeld-gaard. 1958. Dependency on medium and tem-perature of cell size and chemical compositionduring balanced growth of Salmonellatyphimurium. J. Gen. Microbiol. 19:592-606.

257. Schaup, H. W., J. B. Best, and A. B. Goodman.1969. Evidence for heterogeneity of 23S RNAin Escherichia coli. Nature (London)221:864-865.

258. Schaup, H. W., M. Green, and C. G. Kurland.1970. Molecular interactions of ribosomalcomponents. I. Identification of RNA bindingsites for individual 30S ribosomal proteins.Mol. Gen. Genet. 109:193-205.

259. Schaup, H. W., M. Green, and C. G. Kurland.1971. Molecular interactions of ribosomalcomponents. II. Site-specific complex forma-tion between 30S proteins and ribosomalRNA. Mol. Gen. Genet. 112:1-8.

260. Schaup, H. W., and C. G. Kurland. 1972.Molecular interactions of ribosomal compo-nents. III. Reconstitution of a ribosomal pro-tein nucleic acid binding site. Mol. Gen.Genet. 114:350-357.

261. Schaup, H. W., M. Sogin, C. Woese, and C. G.Kurland. 1971. Characterization of an RNA"binding site" for a specific ribosomal proteinof Escherichia coli. Mol. Gen. Genet. 114:1-8.

262. Schleif, R. 1967. Control of production of ribo-somal protein. J. Mol. Biol. 27:41-55.

263. Schleif, R. F. 1968. Origin of chloramphenicolparticle protein. J. Mol. Biol. 37:119-129.

264. Schlessinger, D., and D. Apirion. 1969. Esche-richia coli ribosomes: recent developments.

Annu. Rev. Microbiol. 23:387-426.265. Sells, B. H. 1964. RNA synthesis and ribosome

production in puromycin-treated cells. Bio-chim. Biophys. Acta 80:230-241.

266. Sendecki, W., K. Mikulik, and A. T. Matheson.1971. Some properties of ribosomes fromMyxobacter 495. Can. J. Biochem. 49:1333-1339.

267. Sgaramella, V., and M. Polsinelli. 1967. On therelationship between the streptomycin andribosomal RNA loci. Biochim. Biophys. Acta149:496-505.

268. Singer, M. F., and G. Tolbert. 1965. Purificationand properties of a potassium-activated phos-phodiesterase (RNAase II) from Escherichiacoli. Biochemistry 4:1319-1330.

269. Singh, H., and B. G. Lane. 1964. The alkali-sta-ble dinucleotide sequences in "'S and 2Sribonucleates from wheat germ. Can. J. Bio-chem. 42:1011-1021.

270. Sipe, J. E., W. M. Anderson, Jr., C. N. Remy,and S. H. Love. 1972. Characterization ofS-adenosylmethionine: ribosomal ribonucleicacid-adenine (N'-) methyltransferase of Esch-erichia coli strain B. J. Bacteriol. 110:81-91.

271. Smith, I., D. Dubnau, P. Morell, and J. Mar-mur. 1968. Chromosomal location of DNAbase sequences complementary to transferRNA and to 5S, 16S, and 23S ribosomal RNAin Bacillus subtilis. J. Mol. Biol. 33:123-140.

272. Sogin, M., B. Pace, N. R. Pace, and C. R. Woese.1971. Primary structural relationship of p16 tom16 ribosomal RNA. Nature N. Biol.232:48-49.

273. Sogin, M. L., K. J. Pechman, L. Zablen, B. J.Lewis, and C. R. Woese. 1972. Observations onthe post-transcriptionally modified nucleo-tides in the 16S ribosomal ribonucleic acid. J.Bacteriol. 112:13-16.

274. Sogin, S. J., M. L. Sogin, and C. R. Woese. 1972.Phylogenetic measurement in prokaryotes byprimary structural characterization. J. Mol.Evol. 1: 173-184.

275. Spadari, S., V. Sgaramella, G. Mazza, and A.Falaschi. 1971. Enzymic purification of a hy-brid between ribosomal ribonucleic acid anddeoxyribonucleic acid. Eur. J. Biochem.19:294-300.

276. Spahr, P. F., and D. Schlessinger. 1963. Break-down of messenger ribonucleic acid by apotassium-activated phosphodiesterase fromEscherichia coli. J. Biol. Chem. 238:PC2251-PC2253.

277. Spirin, A. S., and L. P. Gavrilova. 1969. Theribosome. Springer-Verlag, Inc., New York.

278. Srinivasan, P. R., and E. Borek. 1966. En-zymatic alteration of macromolecular struc-ture. Proc. Nucl. Acid. Res. Mol. Biol.5:157-189.

279. Stamato, T. D., and D. E. Pettijohn. 1971.Regulation of ribosomal RNA synthesis instringent bacteria. Nature N. Biol.234:99-102.

280. Stanley, W. M., Jr., and R. M. Bock. 1965.

601VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


BACTERIOL. REV.

Isolation and physical properties of the ribo-somal ribonucleic acid of Escherichia coli.Biochemistry 4:1302-1311.

281. Starr, J. L., and R. Fefferman. 1964. The occur-rence of methylated bases in ribosomal ribo-nucleic acid of Escherichia coli K12 W-6. J.Biol. Chem. 239:3457-3461.

282. Stent, G. S., and S. Brenner. 1961. A geneticlocus for the regulation of ribonucleic acidsynthesis. Proc. Nat. Acad. Sci. U.S.A.47:2005-2014.

283. Stevens, A. R., and P. F. Pachler. 1972. Discon-tinuity of 26S rRNA in Acanthamoebacastellani. J. Mol. Biol. 62:225-237.

284. Stbffler, G., L. Daya, K. H. Rak, and R. A.Garrett. 1971. Ribosomal proteins. XXVI. Thenumber of specific protein binding sites on 16Sand 23S RNA of Escherichia coli. J. Mol. Biol.62:411-414.

285. Stutz, E., and H. Noll. 1967. Characterization ofcytoplasmic and chloroplast polysomes inplants: evidence for three classes of ribosomalRNA in nature. Proc. Nat. Acad. Sci. U.S.A.57:774-781.

286. Subramanian, A. R., E. Z. Ron, and B. D. Davis.1968. A factor required for ribosome dissocia-tion in Escherichia coli. Proc. Nat. Acad. Sci.U.S.A. 61:761-767.

287. Sun, T.-T., T. A. Bickle, and R. R. Traut. 1972.Similarity in the size and number of ribosomalproteins from different prokaryotes. J. Bacte-riol. 111:474-480.

288. Swanton, M., and G. Edlin. 1972. Isolation andcharacterization of an RNA relaxed mutant ofBacillus subtilis. Biochem. Biophys. Res.Commun. 46:583-588.

289. Sypherd, P. S. 1967. Message activity of RNAderived from immature ribosomes. J. Mol.Biol. 24:329-332.

290. Sypherd, P. S. 1968. Ribosome development andthe methylation of ribosomal ribonucleic acid.J. Bacteriol. 95:1844-1850.

291. Sypherd, P. S. 1971. Ribosomal proteins and theconformation of ribosomal ribonucleic acid. J.Mol. Biol. 56:311-318.

292. Sypherd, P. S., and B. S. Fansler. 1967. Struc-tural transitions in ribonucleic acid duringribosome development. J. Bacteriol.93:920-929.

293. Szalay, A., D. Munsche, R. Wollgiehn, and B.Parthier. 1972. Ribosomal ribonucleic acidand ribosomal precursor ribonucleic acid inAnacystis nidulans. Biochem. J. 129:135-140.

294. Szilagyi, J. F. 1968. 16S and 23S components inthe ribosomal ribonucleic acid of Rhodop-seudomonas spheroides. Biochem. J.109:191-195.

295. Takahashi, H. 1969. An attempt to concentrateribosomal RNA cistrons in Bacillus subtilis byMillipore filtration. Biochim. Biophys. Acta190:214-216.

296. Takahashi, H., H. Saito, and Y. Ikeda. 1967.Species specificity of the ribosomal RNA cis-trons in bacteria. Biochim. Biophys. Acta134:124-133.

297. Takai, M., and N. Kondo. 1962. Studies on theribosomal ribonucleic acid from Bacilluscereus. Biochim. Biophys. Acta 55:875-879.

298. Takeda, M., and F. Lipmann. 1966. Comparisonof amino acid polymerization in B. subtilisand Escherichia coli cell-free systems. Hybrid-ization of their ribosomes. Proc. Nat. Acad.Sci. U.S.A. 56:1875-1882.

299. Taylor, A. L., and C. D. Trotter. 1972. Linkagemap of Escherichia coli strain K-12. Bacteriol.Rev. 36:504-524.

300. Taylor, M. M., J. E. Glasgow, and R. Storck.1967. Sedimentation coefficients of RNA from70S and 80S ribosomes. Proc. Nat. Acad. Sci.U.S.A. 57:164-169.

301. Taylor, M. M., and R. Storck. 1964. Uniquenessof bacterial ribosomes. Proc. Nat. Acad. Sci.U.S.A. 52:958-965.

302. Thompson, H. A., 0. G. Baca, and D. Paretsky.1971. Presence of ribosomal ribonucleic acid inthe rickettsia Coxiella burnetti. Biochem. J.125:365-366.

303. Tissieres, A., and J. W. Hopkins. 1961. Factorsaffecting amino acid incorporation into pro-teins by Escherichia coli ribosomes. Proc. Nat.Acad. Sci. U.S.A. 47:2015-2023.

304. Tissieres, A., D. Schlessinger, and F. Gros. 1960.Amino acid incorporation into proteins byEscherichia coli ribosomes. Proc. Nat. Acad.Sci. U.S.A. 46:1450-1463.

305. Tissieres, A., J. D. Watson, D. Schlessinger, andB. R. Hollingworth. 1959. Ribonucleoproteinparticles from Escherichia coli. J. Mol. Biol.1:221-233.

306. Tomlins, R. I., and Z. J. Ordal. 1971. Precursorribosomal ribonucleic acid and ribosome accu-mulation in vivo during the recovery of Sal-monella typhimurium from thermal injury. J.Bacteriol. 107:134-142.

307. Traub, P., and M. Nomura. 1968. Structure andfunction of Escherichia coli ribosomes. I. Par-tial fractionation of the functionally activeribosomal proteins and reconstitution of artifi-cial subribosomal particles. J. Mol. Biol.34:575-593.

308. Traub, P., and M. Nomura. 1968. Structure andfunction of Escherichia coli ribosomes. V.Reconstitution of functionally active 30S ribo-somal particles from RNA and proteins. Proc.Nat. Acad. Sci. U.S.A. 59:777-784.

309. Traub, P., and M. Nomura. 1969. Structure andfunction of Escherichia coli ribosomes. VI.Mechanisms of assembly of 30S ribosomesstudied in vitro. J. Mol. Biol. 40:391-413.

310. Travers, A. A., R. I. Kamen, and R. F. Schleif.1970. Factor necessary for ribosomal RNAsynthesis. Nature (London) 228:748-751.

311. Udvardy, A., and P. Venetianer. 1971. Isolationof the ribosomal RNA genes of Salmonellatyphimurium. Eur. J. Biochem. 20:513-517.

312. Unger, M., L. S. Birnbaum, S. Kaplan, and A.Pfister. 1972. Location of the ribosomal RNAcistrons of Escherichia coli: a second site. Mol.Gen. Genet. 119:377-380.

313. Venkov, P., L. Waltschewa, and D. Schlessinger.

PACE602


http://mm

br.asm.org/

Dow

nloaded from


rRNA OF PROKARYOTES

1972. Polyribosome metabolism in Esche-richia coli. Fed. Eur. Biochem. Soc. Symp.23:379-394.

314. Weatherford, S. C., L. Rosen, L. Gorelic, and D.Apirion. 1972. Escherichia coli strains withthermolabile ribonuclease II activity. J. Biol.Chem. 247:5404-5408.

315. Wegnez, M., R. Monier, and H. Denis. 1972.Sequence heterogeneity of 5S RNA inXenopus laevis. FEBS Lett. 25:13-20.

316. Wehrli, W., and M. Staehelin. 1971. Actions ofthe rifamycins. Bacteriol. Rev. 35:290-309.

317. Weinberg, R. A., and S. Penman. 1968. Smallmolecular weight monodisperse nuclear RNA.J. Mol. Biol. 38:289-304.

318. Wilcox, G., K. J. Clemetson, D. V. Santi, and E.Englesberg. 1971. Purification of the araCprotein. Proc. Nat. Acad. Sci. U.S.A.68:2145-2148.

319. Willson, C., and F. Gros. 1964. Protein synthesiswith an Escherichia coli system in vitro.Biochim. Biophys. Acta 80:478-496.

320. Winslow, R. M., and R. A. Lazzarini. 1969.Amino acid regulation of the rates of synthesisand chain elongation of ribonucleic acid inEscherichia coli. J. Biol. Chem. 244:3387-3392.

321. Wittmann, H. G., G. Stoffler, E. Kaltschmidt,V. Rudloff, H. G. Janda, M. Dzionara, D.Donner, K. Nierhaus, M. Cech, I. Hinden-nach, and B. Wittmann. 1970. Protein-chemi-cal and serological studies on ribosomes ofbacteria, yeasts and plants. Fed. Eur. Bio-chem. Soc. Symp. 21:33-46.

322. Woese, C. R. 1961. Composition of various ribo-nucleic acid fractions from microorganisms ofdifferent deoxyribonucleic acid composition.Nature (London) 189:920-921.

323. Worcel, A., D. S. Goldman, and I. B. Sacks.1968. Properties and fine structure of theribosomes from Mycobacterium tuberculosis.Proc. Nat. Acad. Sci. U.S.A. 61:122-129.

324. Yamagishi, H., and I. Takahashi. 1968. Genetic

transcription in asporogenous mutants of Ba-cillus subtilis. Biochim. Biophys. Acta155:150-158.

325. Yankofsky, S. A., and S. Spiegelman. 1962. Theidentification of the ribosomal RNA cistron bysequence complementarity. I. Specificity ofcomplex formation. Proc. Nat. Acad. Sci.U.S.A. 48:1069-1078.

326. Yankofsky, S. A., and S. Spiegelman. 1962. Theidentification of the ribosomal RNA cistron bysequence complementarity. II. Saturation ofand competitive interaction at the RNA cis-tron. Proc. Nat. Acad. Sci. U.S.A. 48:1466-1472.

327. Yankofsky, S. A., and S. Spiegelman. 1963.Distinct cistrons for the two ribosomal RNAcomponents. Proc. Nat. Acad. Sci. U.S.A.49:538-544.

328. Yoshida, K., and S. Osawa. 1968. Origin of theprotein component of chloramphenicol parti-cles in Escherichia coli. J. Mol. Biol.33:559-569.

329. Young, R. J. 1968. The fractionation of quater-nary ammonium complexes of nucleic acids.Evidence for heterogeneity of ribosomal ribo-nucleic acid. Biochemistry 7:2263-2272.

330. Yu, M. T., C. W. Vermeulen, and K. C. Atwood.1970. Location of the genes for 16S and 23Sribosomal RNA in the genetic map of Esche-richia coli. Proc. Nat. Acad. Sci. U.S.A.67:26-31.

331. Yuki, A. 1971. Tentative identification of a"maturation enzyme" for precursor 16S ribo-somal RNA in Escherichia coli. J. Mol. Biol.62:321-329.

332. Zimmermann, R. A., and C. Levinthal. 1967.Messenger RNA and RNA transcription time.J. Mol. Biol. 30:349-370.

333. Zimmermann, R. A., A. Muto, P. Fellner, C.Ehresmann, and C. Branlant. 1972. Locationof ribosomal protein binding sites on 16Sribosomal RNA. Proc. Nat. Acad. Sci. U.S.A.69:1282-1286.

603VOL. 37, 1973


http://mm

br.asm.org/

Dow

nloaded from


Documents

Structure Synthesis the Ribosomal Ribonucleic Prokaryotes · amongthe prokaryotes. Similarities among the ribosomes isolated from prokaryotes of different taxonomic orders andtheirstructural