Equally Parsimonious Pathways Through an RNA Sequence Space Are Not Equally Likely

An experimental system for determining the potential ability of sequences resembling 5S ribosomal RNA (rRNA) to perform as functional 5S rRNAs in vivo in the Escherichia coli cellular environment was devised previously. Presumably, the only 5S rRNA sequences that would have been fixed by ancestral populations are ones that were functionally valid, and hence the actual historical paths taken through RNA sequence space during 5S rRNA evolution would have most likely utilized valid sequences. Herein, we examine the potential validity of all sequence intermediates along alternative equally parsimonious trajectories through RNA sequence space which connect two pairs of sequences that had previously been shown to behave as valid 5S rRNAs in E. coli. The first trajectory requires a total of four changes. The 14 sequence intermediates provide 24 apparently equally parsimonious paths by which the transition could occur. The second trajectory involves three changes, six intermediate sequences, and six potentially equally parsimonious paths. In total, only eight of the 20 sequence intermediates were found to be clearly invalid. As a consequence of the position of these invalid intermediates in the sequence space, seven of the 30 possible paths consisted of exclusively valid sequences. In several cases, the apparent validity/invalidity of the intermediate sequences could not be anticipated on the basis of current knowledge of the 5S rRNA structure. This suggests that the interdependencies in RNA sequence space may be more complex than currently appreciated. If ancestral sequences predicted by parsimony are to be regarded as actual historical sequences, then the present results would suggest that they should also satisfy a validity requirement and that, in at least limited cases, this conjecture can be tested experimentally. Received: 27 August 1996 / Accepted: 14 April 1997     

Common 5S rRNA Variants Are Likely To Be Accepted in Many Sequence Contexts

Over evolutionary time RNA sequences which are successfully fixed in a population are selected from among those that satisfy the structural and chemical requirements imposed by the function of the RNA. These sequences together comprise the structure space of the RNA. In principle, a comprehensive understanding of RNA structure and function would make it possible to enumerate which specific RNA sequences belong to a particular structure space and which do not. We are using bacterial 5S rRNA as a model system to attempt to identify principles that can be used to predict which sequences do or do not belong to the 5S rRNA structure space. One promising idea is the very intuitive notion that frequently seen sequence changes in an aligned data set of naturally occurring 5S rRNAs would be widely accepted in many other 5S rRNA sequence contexts. To test this hypothesis, we first developed well-defined operational definitions for a Vibrio region of the 5S rRNA structure space and what is meant by a highly variable position. Fourteen sequence variants (10 point changes and 4 base-pair changes) were identified in this way, which, by the hypothesis, would be expected to incorporate successfully in any of the known sequences in the Vibrio region. All 14 of these changes were constructed and separately introduced into the Vibrio proteolyticus 5S rRNA sequence where they are not normally found. Each variant was evaluated for its ability to function as a valid 5S rRNA in an E. coli cellular context. It was found that 93% (13/14) of the variants tested are likely valid 5S rRNAs in this context. In addition, seven variants were constructed that, although present in the Vibrio region, did not meet the stringent criteria for a highly variable position. In this case, 86% (6/7) are likely valid. As a control we also examined seven variants that are seldom or never seen in the Vibrio region of 5S rRNA sequence space. In this case only two of seven were found to be potentially valid. The results demonstrate that changes that occur multiple times in a local region of RNA sequence space in fact usually will be accepted in any sequence context in that same local region.     

Defining 5S rRNA structure space: point mutation data can be used to predict the phenotype of multichange variants

A portion of the 5S ribosomal RNA (rRNA) structure space in the vicinity of the Vibrio proteolyticus 5S rRNA sequence is explored in detail with the intention of establishing principles that will allow a priori prediction of which sequences would be valid members of a particular RNA structure space. Four hundred and one sequence variants differing from the V. proteolyticus 5S rRNA wild-type sequence in 1-7 positions were characterized using an in vivo assay system. Most significantly, it was found that in general, the phenotypic effects of single changes were independent of the phenotypic effect of a second change. As a result, it was possible to use the new data in conjunction with results from prior studies of the same RNA to develop "truth tables" to predict which multiple change variants would be functional and which would be nonfunctional. The actual phenotype of 93.8% of the multichange variants studied was consistent with the predictions made using truth tables thereby providing for perhaps the first time an upper limit estimate of how frequent unexpected interactions are. It was also observed that single changes at positions involved in secondary structure were no more likely to be invalid than changes in other regions. In particular, internal changes in long standard stems were in fact almost always tolerated. Changes at positions that were hypervariable in the context of an alignment of related sequences were, as expected, usually found to be valid. However, the potential validity of changes that were idiosyncratic to a single lineage of related sequences when placed in the V. proteolyticus 5S rRNA context was unpredictable.     

The nucleotide sequence of 5S rRNA from bovine liver.

We have determined the nucleotide sequence of ribosomal 5S RNA from bovine liver. The comparison of this sequence with those from other eukaryotic sources shows that a common secondary structure model for all eukaryotic 5S rRNAs may exist. Analysis of the evolutionary conserved nucleotides in metazoan 5S rRNAs suggests that the tertiary interactions, proposed earlier for plant 5S rRNA, are also possible.     

Escherichia coli 16S rRNA 3''-end formation requires a distal transfer RNA sequence at a proper distance.

The 16S rRNA species in bacterial precursor rRNAs is followed by two evolutionarily conserved features: (i) a double-stranded stem formed by complementary sequences adjacent to the 5' and 3' ends of the 16S rRNA; and (ii) a 3'-transfer RNA sequence. To assess the possible role of these features, plasmid constructs with precursor-specific features deleted were tested for their capacity to form mature rRNA. Stem-forming sequences were dispensable for both 5' and 3' terminus formation; whereas an intact spacer tRNA positioned greater than 24 nucleotides downstream of the 16S RNA sequence was required for correct 3'-end maturation. These results suggest that spacer tRNA at an appropriate location helps form a conformation obligate for pre-rRNA processing, perhaps by binding to a nascent binding site in preribosomes. Thus, spacer tRNAs may be an obligate participant in ribosome formation.     

Identifying constraints on the higher-order structure of RNA: continued development and application of comparative sequence analysis methods.

Comparative sequence analysis addresses the problem of RNA folding and RNA structural diversity, and is responsible for determining the folding of many RNA molecules, including 5S, 16S, and 23S rRNAs, tRNA, RNAse P RNA, and Group I and II introns. Initially this method was utilized to fold these sequences into their secondary structures. More recently, this method has revealed numerous tertiary correlations, elucidating novel RNA structural motifs, several of which have been experimentally tested and verified, substantiating the general application of this approach. As successful as the comparative methods have been in elucidating higher-order structure, it is clear that additional structure constraints remain to be found. Deciphering such constraints requires more sensitive and rigorous protocols, in addition to RNA sequence datasets that contain additional phylogenetic diversity and an overall increase in the number of sequences. Various RNA databases, including the tRNA and rRNA sequence datasets, continue to grow in number as well as diversity. Described herein is the development of more rigorous comparative analysis protocols. Our initial development and applications on different RNA datasets have been very encouraging. Such analyses on tRNA, 16S and 23S rRNA are substantiating previously proposed associations and are now beginning to reveal additional constraints on these molecules. A subset of these involve several positions that correlate simultaneously with one another, implying units larger than a basepair can be under a phylogenetic constraint.     

Prediction of common folding structures of homologous RNAs.

We have developed an algorithm and a computer program for simultaneously folding homologous RNA sequences. Given an alignment of M homologous sequences of length N, the program performs phylogenetic comparative analysis and predicts a common secondary structure conserved in the sequences. When the structure is not uniquely determined, it infers multiple structures which appear most plausible. This method is superior to energy minimization methods in the sense that it is not sensitive to point mutation of a sequence. It is also superior to usual phylogenetic comparative methods in that it does not require manual scrutiny for covariation or secondary structures. The most plausible 1-5 structures are produced in O(MN2 + N3) time and O(N2) space, which are the same requirements as those of widely used dynamic programs based on energy minimization for folding a single sequence. This is the first algorithm probably practical both in terms of time and space for finding secondary structures of homologous RNA sequences. The algorithm has been implemented in C on a Sun SparcStation, and has been verified by testing on tRNAs, 5S rRNAs, 16S rRNAs, TAR RNAs of human immunodeficiency virus type 1 (HIV-1), and RRE RNAs of HIV-1. We have also applied the program to cis-acting packaging sequences of HIV-1, for which no generally accepted structures yet exist, and propose potentially stable structures. Simulation of the program with random sequences with the same base composition and the same degree of similarity as the above sequences shows that structures common to homologous sequences are very unlikely to occur by chance in random sequences.     

The excision of intervening sequences from Salmonella 23S ribosomal RNA

Novel, approximately 90 bp intervening sequences (IVs) were discovered within the 23S rRNA genes of S. typhimurium and S. arizonae. These non-rRNA sequences are transcribed and then excised during rRNA maturation. The rRNA fragments that result from the excision of the extra sequences are not religated. This results in fragmented 23S rRNAs. The excision of one IVS was shown to be catalyzed in vivo and in vitro by ribonuclease III. These IVSs are highly volatile evolutionarily, sometimes occurring in only some of the multiple rRNA operons of a particular cell. The sporadic nature of the occurrence of fragmented rRNAs among closely related organisms argues that such fragmentation is a derived state, not a primitive one. Possible sources of these IVSs, their parallels with internal transcribed spacers and introns in eukaryotes, and their possible roles in the evolutionary process are discussed.     

tRNA genes are found between 16S and 23S rRNA genes in Bacillus subtilis.

There are at least nine, and probably ten, ribosomal RNA gene sets in the genome of Bacillus subtilis. Each gene set contains sequences complementary to 16S, 23S and 5S rRNAs. We have determined the nucleotide sequences of two DNA fragments which each contain 165 base pairs of the 16S rRNA gene, 191 base pairs of the 23S rRNA gene, and the spacer region between them. The smaller space region is 164 base pairs in length and the larger one includes an additional 180 base pairs. The extra nucleotides could be transcribed in tRNAIIe and tRNA Ala sequences. Evidence is also presented for the existence of a second spacer region which also contains tRNAIIe and tRNA Ala sequences. No other tRNAs appear to be encoded in the spacer regions between the 16S and 23S rRNA genes. Whereas the nucleotide sequences corresponding to the 16S rRNA, 23S rRNA and the spacer tRNAs are very similar to those of E. coli, the sequences between these structural genes are very different.     

Fine structure of ribosomal RNA. III. Location of evolutionarily conserved regions within ribosomal DNA

In order to define functional regions within ribosomal RNA, we have identified areas of the molecule which have been conserved during evolution. Our previous studies showed that there is evolutionary conservation between the rRNAs of different eukaryotes and that the sequences conserved between distantly related species are a subset of those conserved between closely related species. In the present work, we have employed DNA-DNA and DNA-RNA hybridization techniques to localize these conserved regions to mapped restriction fragments 50 to 300 base-pairs in length within cloned Xenopus laevis ribosomal DNA. Our experiments have detected evolutionary conservation only within the coding regions, suggesting that if there is any conservation within the spacers, these sequences must be very short. Regions of conservation can be classified either by evolutionary distance or by the extent of conservation between two species. Three regions, including one near the 3' end of 18 S and two near the 3' end of 28 S rRNA are conserved over great evolutionary distance, that is between Escherichia coli and X. laevis. In addition, several fragments in the central portions of the 188 and 28 S rRNAs are exceptional in the extent of their conservation between yeast and Xenopus. We have been able to correlate the regions we have defined as conserved with certain structural or functional roles, such as initiation of translation, possible interaction with transfer RNA, rRNA methylation, and the site where intervening sequences interrupt some eukaryotic rRNAs. As a result, these studies serve to define relatively short (less than 300 base-pairs) segments within the almost 11,000 base X. laevis rDNA repeat unit which are worthy of further investigation.     

16S rRNA sequences of uncultivated hot spring cyanobacterial mat inhabitants retrieved as randomly primed cDNA.

Cloning and analysis of cDNAs synthesized from rRNAs is one approach to assess the species composition of natural microbial communities. In some earlier attempts to synthesize cDNA from 16S rRNA (16S rcDNA) from the Octopus Spring cyanobacterial mat, a dominance of short 16S rcDNAs was observed, which appear to have originated only from certain organisms. Priming of cDNA synthesis from small ribosomal subunit RNA with random deoxyhexanucleotides can retrieve longer sequences, more suitable for phylogenetic analysis. Here we report the retrieval of 16S rRNA sequences from three formerly uncultured community members. One sequence type, which was retrieved three times from a total of five sequences analyzed, can be placed in the cyanobacterial phylum. A second sequence type is related to 16S rRNAs from green nonsulfur bacteria. The third sequence type may represent a novel phylogenetic type.     

Walking through protein sequence space

Following the original idea of Maynard Smith on evolution of the protein sequence space, a novel tool is developed that allows the "space walk", from one sequence to its likely evolutionary relative and further on. At a given threshold of identity between consecutive steps, the walks of many steps are possible. The sequences at the ends of the walks may substantially differ from one another. In a sequence space of randomized (shuffled) sequences the walks are very short. The approach opens new perspectives for protein evolutionary studies and sequence annotation.     

16S rRNA sequences of uncultivated hot spring cyanobacterial mat inhabitants retrieved as randomly primed cDNA.

Cloning and analysis of cDNAs synthesized from rRNAs is one approach to assess the species composition of natural microbial communities. In some earlier attempts to synthesize cDNA from 16S rRNA (16S rcDNA) from the Octopus Spring cyanobacterial mat, a dominance of short 16S rcDNAs was observed, which appear to have originated only from certain organisms. Priming of cDNA synthesis from small ribosomal subunit RNA with random deoxyhexanucleotides can retrieve longer sequences, more suitable for phylogenetic analysis. Here we report the retrieval of 16S rRNA sequences from three formerly uncultured community members. One sequence type, which was retrieved three times from a total of five sequences analyzed, can be placed in the cyanobacterial phylum. A second sequence type is related to 16S rRNAs from green nonsulfur bacteria. The third sequence type may represent a novel phylogenetic type.     

Ribosomal RNA homologies and the evolution of the filamentous blue-green bacteria

Summary Ribosomal RNA (rRNA) sequence homology (as determined by comparisons of T1 oligonucleotide catalogs of³²P-labeled 16S rRNAs) has been used to assess phylogenetic relationships within the filamentous and unicellular blue-green bacteria, and to identify regions of evolutionary conservatism within blue-green bacterial 16S rRNAs.Nostoc andFischerella, representatives of two morphologically distinct and highly differentiated orders, are shown to be as closely related (on the basis of RNA sequence homology) as typical members of the non-blue-green bacterial genusBacillus. They are further shown to be (on the same basis) indistinguishable from typical unicellular members of a subgroup of the unicellular blue-green bacterial order Chroococcales. These results have general implications for studies of the origin of differentiated prokaryotes and of evolutionary change in prokaryotic macromolecules. In particular, they provide indirect evidence that the divergences of contemporary major prokaryotic groups are truly ancient ones.     

The nucleotide sequences of the 5 S rRNAs of seven molds and a yeast and their use in studying ascomycete phylogeny.

The sequences of the 5 S rRNAs isolated from 8 ascomycete species belonging to the genera Aspergillus, Penicillium, Acremonium and Candida are reported. Two of the examined strains each yielded a mixture of 3 slightly different 5 S RNAs, which were individually sequenced after fractionation. A previously published sequence for Aspergillus nidulans 5 S RNA was found to contain errors. Reconstruction of an evolutionary tree based on 5 S RNA sequences showed that the 16 presently examined ascomycetes form three clusters. The same threefold partition can be observed in the secondary structure pattern, each cluster showing a slightly different variant of the general 5-helix model for 5 S rRNA (De Wachter, Chen and Vandenberghe (1982) Biochimie 64, 311-329), and different sets of secondary structure equilibrium forms in helices C and E of the aforementioned model.     

Sequence Heterogeneity between the Two Genes Encoding 16s Rrna from the Halophilic Archaebacterium Haloarcula Marismortui

The halophilic archaebacterium, Haloarcula marismortui, contains two nonadjacent ribosomal RNA operons, designated rrnA and rrnB, in its genome. The 16S rRNA genes within these operons are 1472 nucleotides in length and differ by nucleotide substitutions at 74 positions. The substitutions are not uniformly distributed but rather are localized within three domains of 16S rRNA; more than two-thirds of the differences occur within the domain bounded by nucleotides 508 and 823. This domain is known to be important for P site binding of aminoacylated tRNA and for 30-50S subunit association. Using S1 nuclease protection, it has been shown that the 16S rRNAs transcribed from both operons are equally represented in the functional 70S ribosome population. Comparison of these two H. marismortui sequences to the 16S gene sequences from related halophilic genera suggests that (i) in diverging genera, mutational differences in 16S gene sequences are not clustered but rather are more generally distributed throughout the length of the 16S sequence, and (ii) the rrnB sequence, particularly within the 508-823 domain, is more different from the out group sequences than is the rrnA sequence. Several possible explanations for the evolutionary origin and maintenance of this sequence heterogeneity within 16S rRNA of H. marismortui are discussed.     

Wheat embryo mitochondrial 18S ribosomal RNA: evidence for its prokaryotic nature.

We present a catalog of sequences of oligonucleotides produced by T1 ribonuclease digestion of 32P-labeled small-ribosomal-subunit RNA ("18S rRNA) isolated from purified wheat embryo mitochondria. This catalog is compared to catalogs published for prokaryotic and chloroplast 16S rRNAs and to preliminary results for wheat cytosol 18S rRNA. These comparisons indicate that: (1) wheat mitochondrial 18S rRNA is clearly prokaryotic in nature, showing significantly more sequence homology with 16S rRNAs than can be expected to arise by chance (p less than 0.000001); (2) shared oligonucleotide sequences include an especially high proportion of those identified as conserved in the evolution of prokaryotic rRNAs; and (3) wheat embryo mitochondrial and cytosol 18S rRNAs retain no more, and perhaps less, than the minimum sequence homology detectable by this sensitive method. These results argue in favor of an endosymbiotic origin for mitochondria.     

The contribution of DNA slippage to eukaryotic nuclear 18S rRNA evolution

Six of 204 eukaryotic nuclear small-subunit ribosomal RNA sequences analyzed show a highly significant degree of clustering of short sequence motifs that indicates the fixation of products of replication slippage within them in their recent evolutionary history. A further 72 sequences show weaker indications of sequence repetition. Repetitive sequences in SSU rRNAs are preferentially located in variable regions and in particular in V4 and V7. The conserved region immediately 5 to V7 (C7) is also consistently repetitive. Whereas variable regions vary in length and appear to have evolved by the fixation of slippage products, C7 shows no indication of length variation. Repetition within C7 is therefore either not a consequence of slippage or reflects very ancient slippage events. The phylogenetic distribution of sequence simplicity in small-subunit rRNAs is patchy, being largely confined to the Mammalia, Apicomplexa, Tetrahymenidae, and Trypanosomatidae. The regions of the molecule associated with sequence simplicity vary with taxonomic grouping as do the sequence motifs undergoing slippage. Comparison of rates of insertion and substitution in a lineage within the genus Plasmodium confirms that both rates are higher in variable regions than in conserved regions. The insertion rate in variable regions is substantially lower than the substitution rate, suggesting that selection acts more strongly on slippage products than on point mutations in these regions. Patterns of coevolution between variable regions may reflect the consequences of selection acting on the incorporation of slippage-derived sequences across the gene.     

Composite structure of the chloroplast 23 S ribosomal RNA genes of Chlamydomonas reinhardii: Evolutionary and functional implications

The chloroplast ribosomal unit of Chlamydomonas reinhardii displays two features which are not shared by other chloroplast ribosomal units. These include the presence of an intron in the 23 S ribosomal RNA gene and of two small genes coding for 3 S and 7 S rRNA in the spacer between the 16 S and 23 S rRNA genes (Rochaix & Malnoë, 1978). Sequencing of the 7 S and 3 S rRNAs as well as their genes and neighbouring regions has shown that: (1) the 7 S and 3 S rRNA genes are 282 and 47 base-pairs long, respectively, and are separated by a 23 base-pair A + T-rich spacer. (2) A sequence microheterogeneity exists within the 3 S RNA genes. (3) The sequences of the 7 S and 3 S rRNAs are homologous to the 5′ termini of prokaryotic and other chloroplast 23 S rRNAs, indicating that the C. reinhardii counterparts of 23 S rRNA have a composite structure. (4) The sequences of the 7 S and 3 S rRNAs are related to that of cytoplasmic 5.8 S rRNA, suggesting that these RNAs may perform similar functions in the ribosome. (5) Partial nucleotide sequence complementarity is observed between the 5′ ends of the 7 S and 3 S RNAs on one hand and the 23 S rRNA sequences which flank the ribosomal intron on the other. These data are compatible with the idea that these small rRNAs may play a role in the processing of the 23 S rRNA precursor.