Origin of Most Primitive mRNAs and Genetic Codes via Interactions
Genome Informatics 13: 71–81 (2002)
71
Origin of Most Primitive mRNAs and Genetic Codes via Interactions between Primitive tRNA Ribozymes Koji Ohnishi1
Shouken Hokari1
Hiroshi Shutou1
Madoka Ohshima1
Naotaka Furuichi2
Masaki Goda3
[email protected]
[email protected]
[email protected]
[email protected]
1
2 3
Department of Biology, Faculty of Science, Niigata University, Igarashi-2, Niigata 950-2181, Japan Faculty of Agriculture, Niigata University, Igarashi-2, Niigata 950-2181, Japan Faculty of Engineering, Niigata University, Igarashi-2, Niigata 950-2181, Japan
Abstract The origin and early evolution of genetic codon system and early mRNAs were analyzed from a viewpoint of primordial gene theory and the poly-tRNA theory. A hypothetical 25-amino acid (aa)-primordial peptide was deduced from internal aa-sequence homology of adenylate kinases. Theoretical models were made which can reasonably explain how primitive tRNA(s) could have had converted to be earliest mRNAs via interactions between presumptive anticodons and (poly)tRNA ribozyme. Transfer-RNA gene clusters in the trrnD- and rrnB-operons of Bacillus subtilis seemed to be relics of early peptide-synthesizing RNA machine. Detailed analyses revealed that the poly-tRNA regions in these operons are true relics of RNA-machine for making a 16-aa trrnDpeptide and a 21-aa rrnB-peptide, whose aa sequences are in the order of aa specificities of tRNAs in the tRNA gene clusters of the trrnD-operon and rrnB-operon, respectively. The primordial gene-encoded peptide deduced from adenylate kinases were found to be a genuine homologue of the rrnB-peptide. Various protein superfamilies were found to have evolved from either of these two types of primitive peptides. Earlist mRNAs were concluded to have evolved from tRNA Gly (trrnDmRNA) or tRNAHis (rrnB-mRNA) , where trrnD- and rrnB-mRNAs are hypothetical primitive mRNAs complementary to the tandem arrangement of 16 or 21 anticodons of tRNAs in the trrnDoperon and rrnB-operon, respectively. The poly-tRNA model is considered to be an excellent theory, because it can reasonably explain origins of both genetic codes and earliest mRNAs, and because the hypothesis can be statistically evaluated by base-identity levels in proper alignments. The genetic codon system is a typical mature semeiotic system within a cell, and the genesis of the geneic codon system was discussed from an aspect of de Saussure’s semeiology. Arbitrary correspondence between (anti)codon and aa would be most plausibly a result of semeiotic culture system of intracellular tRNA-riboorganismic society.
Keywords: primordial gene, adenylate kinase, poly-tRNA theory, trrnD-peptide, rrnB-peptide, early peptide-synthesis, glyceraldehydes-3-phosphate dehydrogenase, 3-phosphoglycerate transporter protein B, arbitrary correspondence, semeiotic culture
1
Introduction
Before the establishment of the RNA-first theory based on the finding of ribozymes by Cech [2] and Altman [5], Haldane [6] and Eigen [4] predicted that RNAs emerged earlier than DNAs as primitive genes or nucleic acid-replicators. Since an RNA molecule has a structure of “poly-AMP”, an AMP/ATP-dependent energy-obtaining system accompanied by proper metabolism with nucleoside supply must have preceded the emergence of RNA. How first protein-encoding RNAs (most primitive
72
Ohnishi et al.
mRNAs) could have been made throughout evolution has long been a most important unsolved problem in molecular evolutionary biology. A hypothesis, later named “poly-tRNA theory”, proposed by one of us (K.O.) in 1993 [11, 17, 13] seemed to make a possible breakthrough in solving the problem. The poly-tRNA theory revealed that the two tRNA gene clusters (hereafter denoted by trrnD-polytRNA region and rrnD-poly-tRNA region) found in the trrnD-operon and the rrnB-operon, respectively, of a Gram-positive bacteria, Bacillus subtilus, are relics of early peptide-synthesizing RNA apparatus [11, 17, 13, 16, 14, 22, 15, 19, 21, 18, 27]. Detailed comparative analyses have hitherto elucidated that (i) ancient primitive tRNAGly and tRNAHis in the respective early tRNA-clusters would have converted to earliest mRNAs for making 16- aa- and 21-aa-peptides (cald trrnD-peptide and rrnB-peptide) via interaction of the tRNAs and presumptive anticodon regions of the respective poly-tRNAs, and that long DNA regions comprizing several tRNAs and spacers of the trrnD- and rrnB-poly-tRNA regions were found to be most plausibly homologous to the DNA regions encoding Glycyl-tRNA synthetase alpha subunit gene (GlyRS) [17, 16, 14, 18] and the lambd repressor (cI protein) [in preparation]. These findings brought us a new viewpoint to elucidate the evolutionary history of tRNAs and early mRNAs. A proto-tRNA, commonly ancestral to contemporary tRNAs, would have had emerged as a new (possibly intracellular) replicable ribozyme (possibly a kind of “riboorganism”) capable of making dipeptide bonds, and further diverged to different primitive tRNAs with their respective amino acid specificity. Such primitive tRNAs would have interacted and/or co-operated with one another, and constituted an interacting complex or an intracellular symbiotic complex (= riboorganismic society [18, 27]) for making oligopeptides or longer peptides. This paper reports new evidence for the evolution of poly-tRNA-derived mRNAs. Origin of genetic codon system was discussed from the aspects of poly-tRNA theory and de Saussure’s semeiotic culture theory.
2 2.1
Proposing and Proving of the Poly-tRNA Hypothesis Reconsideration of a Primordial Gene as Building Block of Adenylate Kinase Genes
Ohno and Matsunaga [23, 24] first postulated a primordial gene (PG) theory, hypothesizing that relics of some 48-base PGs might have existed as building blocks of modern genes encoding proteins such as serum albumin and immunoglobulin. One of us (K.O.) had a great interest in their theory, and gegan to search possible common amino acid (aa) sequence segments shared by internal repeats of house-keeping proteins [10, 12]. A most beautiful result thus obtained was a common sequence found between duplicated aa sequence regions of adenylate kinase (AK) ([12], p.156). Fig. 1[A] shows common aa’s in the E. coli and pig AK’s shared by the first amino-terminal domain (domain I, corresponding to aa’s 1 − 32 in E. coli) and the second domain (domain II, aa’s 33 − 71 in E. coli). Based on this sharing of 12 aa’s, a primordial gene encoding a reconstructed aa sequence “(N-term) (VXX)GSXXGXQAXXIMEKXXXXLXT (C-term)” (where every of X’s is unknown) can be deduced as an ancient protein sequence encoded by a most primitive primordial gene. The essential portion of this primordial sequence was further confirmed by using other sequence data of adenylate kinases from obtained from GenBank Database. Since adenylate kinase is possessed by every of living organisms, a gene duplication (of the primordial gene) generating the contemporary 1st and 2nd homology domains must have occurred at a time before the latest common ancestor of all living organisms. Among some DNA sequences encoding common aa-sequence segments, there seemed to be some sequences having slight similarities to tRNAs, which suggested an interesting possibility that some PG(s) might have had evolved from some kind(s) of primitive tRNA.
2.2
Towards Building a Tentative RNA Model for Early Peptide-Synthesis [14]
If some PG(s) could be really homologous to tRNAs, how and why such homologous relations could have emerged? If a primitive tRNA (or proto-tRNA) could be an ancestor of a first most primitive mRNA, then it is reasonable to consider that the primitive tRNA must have converted to be a first
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
73
mRNA. In order to consider more exactly, a term “presumptive X” (pres. X) is defined as something to evolve to X in future, and therefore, is an evolutionary ancestor of “primitive X” (prim. X) (X = anticodon, codon, mRNA, etc.).
Figure 1: A primordial gene and its possible origin from a hypothetical rrnB-mRNA. [A] A primordial peptide deduced from an internal aa sequence homology between the 1st and 2nd homology regions of the adenylate kinases (AK’s) from E. coli (adk gene) and pig (myokinase). Amino acids (aa’s) shared by the 1st and 2nd homology regions are boxed. Aa’s identities between the 2nd AK domain and a hypothetical primitive peptide, rrnB-peptide, are also shown. See text and Fig. 4 for rrnB-peptide. [B] Base-sequence homology between a hypothetical rrnB-mRNA and a yeast (S. cervisiae) AK gene (adk1). Aa’s shared by the rrnB-peptide and the yeast AK are boxed. Bases identical to the rrnBmRNA are indicated by “o”. Statistical evaluation of base-identity level (%) by Pnuc (m,n) value (See (Eq.1) in text.) is given. See text and Fig. 4. for definition of rrnB-mRNA. Evolution from a tRNA to a most primi. mRNA means that every three-base-unit (triplet) in a tRNA (which is “pres. mRNA”) would have had converted to a triplet codon complementary (in the sense of Watson-Crick-type base-pairing) to the “pres. anticodon” of some tRNA, via interaction between the triplet (pres. codon) on the pres. mRNA and the pres. anticodon on the tRNA, as shown in Fig. 2. If we make a reasonable assumption that prim. tRNAs before the emergence of mRNAs would have been capable of carrying a (more or less) specific aa (probably by using ATP) at its 3’(CCA)end. The model in Fig. 2 therefore suggests that early peptide synthesis might have had occurred by a co-operative interactions of early tRNAs (i.e., by cooperative behaviour of early tRNA-riboorganisms, in terms of riboorganismic behaviology as discussed by Ohnishi et al. [18, 27]) capable of making peptide-bond by their own ribozymic activity. If different aa-carrying tRNAs interacted with pres. anticodons on a pres. mRNA (which is some kind of prim. tRNA), then their cooperative behaviour would ensure easy making of a peptide bond between the aa’s carried by the neighbouring tRNAs. If
74
Ohnishi et al.
such peptide-bond-making by two-neighbouring tRNAs could occur regularly from 5’-side to 3’-side of the pres. mRNA, then a primitive peptide would be made by co-operative interactions of tRNAs (or by cooperation of tRNA riboorganisms), as shown in Fig. 2. Base-replacement mutations giving rise to base-complementary between pres. codons on the pres. mRNA and the corresponding pres. anticodons would generate a semeiotic correspondence between prim. codons (evolved from pres. codons) and a prim. codon-specific aa carried by the corresponding tRNA. Accordingly, the tentative early peptide-synthesizing model in Fig. 2 seems to give an important basis for considering the origin of the first mRNA possessing earliest codons. Some difficulties encountered are that how such interactions between pres. codons and pres. anticodons could have had occurred so as to satisfy thermodynamical and probabilistic conditions towards accumulating basecomplementarities at reasonable evolutionary rate. These difficulties would be solved in the following sections.
Figure 2: A tentative model for early peptide-synthesis showing a possibility that a primitive tRNA (= presumptive mRNA) could have had converted to an earliest mRNA via interactions with presumptive anticodons of other tRNAs. Watson-Crick-type base-complementarities between presumptive anticodons and presumptive codons would have been selected via interactions between presumptive anticodons (on primitive tRNAs) and presumptive codons on a presumptive mRNA (= a kind of primitive tRNA). See text for details.
2.3
Bacillus subtilis is an Excellent Teacher for Building the Poly-tRNA Model
In the next step, we need to refine the tentative early-peptide-synthesizing RNA model shown in Fig. 2. Structural characteristics of the B. subtilis trrnD- and rrnB-operons were found to kindly teach us a critical solution for this difficult problem. The two operons shown in Fig. 3 mean that the direct transcripts of these operons have a structure of “16S-rRNA-23S-rRNA-5S-rRNA-(tRNA)n ”, where n= 16 and 21 for the trrnD- and the rrnB-operon, respectively. In the tentative model in Fig. 2, aa-carrying prim. tRNAs are composed of tRNA molecules possessing different aa-specificities, and every of them behaves independently, resulting in probabilistic difficulties to reach and interact with their properly corresponding pres. codons on the pres. mRNA in every generation throughout evolution towards the genesis of the first mRNA. On the contrary, each transcript of the two operons has a large RNA region (hereafter called “poly-tRNA(region)”) consisting of tandemly connected tRNAs, which allows pres. codons (of pres. mRNA) to easily interact with their counterpart pres. anticodons (of poly-tRNA). And moreover, such interactions of two consecutive codons with anticodons of the corresponding two consecutive tRNAs (of poly-tRNA) could have tended to repeatedly re-occur (throughout evolution) on the prim. A and P sites on a prim. ribosome consisting exclusively of prim. 16S-, 23S- and
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
75
5S-rRNAs (i.e. without possessing any ribosomal proteins) encoded by the same operon (as found in Fig. 3). This is because early ribosome is considered to have consisted exclusively of rRNA ribozymes on which peptide synthesis (peptidyl-transferation) would have had occurred [9]. These considerations allow us to see that the poly-tRNAs found in the trrnD- and the rrnB-operons are most seemingly relics of early tRNA-complexes capable of cooperatively making most primitive peptides, whose aa sequences are “(N-term.) NSEVMDFTYWHQGCLL (C-term.)” (named trrnD-peptide) and “(Nterm.) VTKLGLRPAMISMDFHGINSE (C-term.)” (named rrnB-peptide), respectively.
Figure 3: Eubacterial t-RNA gene clusters in representative species. Gram-positive bacteria (Bacillus subtilis, Staphylococcus aureus) possess close homologues of B. subtilis trrnD- and/or rrnB-type tRNA gene clusters. In the E. coli rrnB operon (Genbank Locus ECORGNG), a pseudo-tRNA gene cluster closely homologous to the B. subtilis trrnD-type poly-tRNA was found [21]. Note that two types of tri-tRNA segment, “NSE” and “MDF”, can be found in both the trrnD- and the rrnB-operons of B. subtilis, which seems to suggest earlier evolution of tripeptide(NSE or MDF)-making tri-tRNA complexes. If we compare the trrnD- and rrnB-peptides with the first and the second homology domains of AK’s as well as with the above-mentioned PG-encoded 25-aa-peptide, “(N-term)(VXX)GSXXGXQAX XIMEKXXXXLXT (C-term)” (deduced from internal homology of AK’s), remarkable aa sequence similarities were found in the comparison of the rrnB-peptide with these AK domains and therefrom deduced PG-encoded peptide, as is clearly shown in Fig. 1 [A], [B]. Seven out of the 21 aligned aapositions (= 33.3%) are occupied by identical residues in the comparison of the rrnB-peptide with the S. cervisiae AK (encoded by adk1 gene). These aa-similarities strongly suggests that the rrnBtype tRNA gene cluster (or rrnB-poly-tRNA) would be a real relic of an early RNA biomachine for synthesizing the rrnB-peptide. On the other hand, it is also important to note that tRNA-gene clusters homologous to either or both of the B. subtilis trrnD-type and rrnB-type clusters are widely found in most of hitherto analyzed Gram-positive bacteria, as exemplified by a large tRNA cluster in Staphylococcus aureus (as shown in Fig. 3), and by an Acholeplasma laidwai tRNA cluster, “VTKLGLRPAMISMDF” (not shown in Fig. 3). A pseudo-tRNA gene-cluster (consisting of at least 10 pseudo-tRNA genes) homologous to the B. subtilis trrnD-poly-tRNA region has recently been found in the immediate downstream of the 5S rRNA gene in the rrnB-operon (rRNA operon) of a representative Gram-negative bacteria (Proteobacteria), E. coli [21]. E. coli has also a three-tRNA gene-cluster, tRNAGly -tRNACys -tRNALeu , which is evidently a true homologue of the tRNAGly -tRNACys -tRNALeu region of the B. subtilis
76
Ohnishi et al.
trrnD-tRNA gene cluster. This homologue-relationship among bacterial tRNA gene-clusters strongly suggested that the B. sbutilis trrnD- and rrnB-tRNA clusters would well conserve a very ancient, ancestral poly-tRNA structure.
2.4
The Primordial Gene Deduced from Adenylate Kinases is a First Candidate for a Descendant of the rrnB-mRNA
Thus we now reach to a possible, very important hypothesis that the trrnD- and rrnB-peptides could be most plausible earliest peptides made by co-operative interactions (or behaviours) of tRNAriboorganisms. Based on the theoretical considerations mentioned above, a rrnB-type poly-tRNA model for synthesizing a primitive 21-aa-rrnB-peptide can be constituted as found in Fig. 4. In this Figure, the “rrnB-poly-tRNA is a poly-tRNA region of the long RNA-transcript from the rrnB-operon. The “rrnB-mRNA” is defined as an hypothetical 63-base-RNA (a candidate for prim.mRNA) whose sequence is complementary (in Watson-Crick-type) to the 21-anticodons (63 bases) in the rrnB-polytRNA, as shown in Fig. 4. The 21 pres. (or prim.) codons of an early rrnB-mRNA would have had interacted with their respective counterpart pres. (or prim.) anticodons of an early rrnB-poly-tRNA, and thereby helped the catalytic ribozyme function of the poly-tRNA to make a “rrnB-peptide” (on the primitive A and P sites of a primitive 16S-23S-5S-rRNA-ribosome), whose aa sequence is “VTKLGLRPAMISMDFHGINSE”. The rrnB-mRNA would have had most plausibly evolved from some primitive tRNA. If this rrnB-type poly-tRNA model well represents the true evolutionary history of early peptide synthesis, it is strongly expected that the DNA segment (in the adk1 gene) encoding the second homology region (aa’s 47-66) of the S. cerevisiae AK would be homologous to the rrnB-mRNA, an ancestral mRNA for making the rrnB-peptide. The comparison is given in Fig. 1[B], resulting in a finding that a 50.5% (= 29/57) base-match level was observed between these tow base-sequences. Statistical evaluation of base-identity levels of m observrd base-matches in n aligned base-positions (excluding gap-including positions) was made by computing Pnuc (m,n) values defined by X Pnuc (m, n) = Cn,i (1/4)i (3/4)n−i , (Eq.1) P
(i=m,n)
where Cn,i = n!/[i!(n − i)!], and (i=m,n) denotes summation over m to n. Pnuc (m,n) value in (Eq.1) denotes the probability by chance for the occurrance of m- or more-base-matches in an n-base-long alignment of a pair of randomly selected sequences, and gives a statistical evaluation of the homologylevels of the observed base-matche level. For the observed base-match level in Fig. 1[B], Pnuc (29, 57) = 2.4 × 10−5 was obtained, suggesting that the adk1 DNA segment is homologous to the rrnB-mRNA, with a probability of error occurrence of 2.4 × 10−5 . Since another hypothetical mRNA, trrnD-mRNA, can be defined based on the trrnDpoly-tRNA sequence [11, 17, 13, 16, 18, 27], in a similar way as in the case of the rrnB-mRNA, the trrnD-mRNA is also a most plausible candidate for an early mRNA from which various contemporary house-keeping protein genes ( such as those genes encoding 3-phosphoglycerate transporter protein B (pgtB), glycyl-tRNA synthetase alpha subunit, Fi-ATP synthase γ subunit, F0-ATP synthase α subunit, C-type lectin, etc.) could have later evolved, as has been reported in [11, 17, 18, 27].
2.5
Poly-tRNA Models and Its Evolutionary Significance: Proteins Encoded by rrnB-mRNA-Derived Genes and trrnD-mRNA-Derived Genes
These results now bring us with a hopeful strategy for finding early evolution of primitive mRNAs from either of the rrnB- and trrnD-mRNAs. We can make the following predictions; *(Pred. 1) Some aa-sequence segments in contemporary proteins in living organisms might be homologous to either one of the rrnB- and trrnD-peptides. *(Pred. 2) If (Pred.1) is true, then the DNA sequence segments encoding such aa-sequence regions are evolutionary homologues of either of the rrnB-and trrnD-mRNAs. *(Pred. 3) If (Pred. 1) and (Pred. 2.) are true, then the trrnDand the rrnB-mRNAs are considered either *(Pred. 3-1) to have evolved from some kinds of tRNA,
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
77
as predicted from the original considerations of the (poly-)tRNA models in Fig. 2 and Fig. 4, or *(Pred. 3-2) to have evolved from some unknown primitive RNA(s) which would have later interacted with presumptive anticodons of the corresponding poly-tRNA. *(Pred. 4) If (Pred. 3-1) is true, then the contemporary protein-encoding DNA segments homologous to the trrnD- or rrnB-mRNA would be more highly similar to the trrnD- or rrnB-mRNA than to the corresponding homologous tRNA. In the search for protein sequence segments similar to the trrnD- and rrnB-peptides, database search by FASTA, BLAST and psi-BLAST were used by employing GenBank and SWISS-PLOT databases. Querry sequences are extremely short, and therefore, conventional thresholds of E-values based on Gumbel’s statistics of extremes would not represent exact statistical meanings, since BLAST and psi-BLAST searches are based on statistics assuming that the querry is considerably long [1, 8]. However, these algolisms are very useful for finding out most similar very short sequence segments, if we employ adequetly large thresholds of E-value. Thus found most similar short sequences were compared with the querry and also with one another, by employing multialignment ClustalW algolism, and finally obtained alignments (as showen in Fig. 4) were therafter stasistically evaluated by P nuc values. In case of the trrnD-peptide-derived proteins hitherto found and reported, (Predictions 1, 2, 3-1, and 4) have been elegantly proved to be true, as found in references [11, 17, 13, 16, 14, 15, 19, 21, 18, 27]. The trrnD-peptide/trrnD-mRNA shows highest similarities to the Salmonella typhimurium 3phosphoglycerate transporter protein B (PgtB) /pgtB gene, giving 11 aa-identity (66.8%) to the 16-aatrrnD-peptide (60.9% base-identity, Pnuc (28, 46) = 2.8 × 10−7 ), which is lower than than the similarity of pgtB with the tRNAGly -region of trrnD-poly-tRNA (58.1% base-match, Pnuc (25, 43) = 3.9 × 10−6 ), thus satisfying (Pred. 4) [15]. The E. coli glycyl-tRNA synthetase (GlyRS) alpha subunit and its coding DNA region are also genuine homologues of the trrnD-peptide and trrnD-mRNA (63.8%, Pnuc (30, 47) = 2.2 × 10−8 ), respectively [15]. F1 -F0 -ATP-synthase subunits (F1 -gamma, F0 -a, etc.) and C-type lectin (lectin Bra-3) also show significant levels of homology to GlyRS, trrnD-mRNA and trrnD-poly-tRNA (See Fig. 1 and Fig. 2 in the references [18, 27]). A portion of 16S rRNA is also homologous to the tRNA(Gly)-tRNA(Cys)-tRNA(Leu) region [19, 7], which is commonly found in diverge species of eubacteria as found in Fig. 3. As will be published elsewhere, further searches using recent Genbank/SWISSPROT databases elucidated that aa-sequence portions in succinyl-CoA synthetase beta subunit, the JH segment of Immunoglobulin VH domain (encoded by JH exon), and reverse transcriptases are also homologous to trrnD-peptide and its homologues. Figure 4 (Next page): An rrnB-type poly-tRNA model for an early peptide (rrnB-peptide)-synthesizing RNA machine, showing homologies of the rrnB-peptide and/or rrnB-mRNA with the B. subtilis tRNA His , adenylate kinase domains, glyceraldehydes-3-phosphate dehydrogenase, DNA-binding domains (of lambda repressor and the sponge homeoprotein 1), and an arginyl tRNA-synthetase. In the upper portion of this Figure, an rrnB-type poly-tRNA model is given. In the lower portion, the tRNAHis -including portion of the poly-tRNA region of the B. subtilis rrnB-operon , as well as possible candidates of homologues of the rrnB-peptide together with their corresponding coding DNA regions are aligned against the rrnB-peptide and the rrnB-mRNA. The helix regions of HTH (helix-turn-helix) DNA-binding domains (in cI and homeoprotein 1) are indicated by “”. Bases in the spacer regions of the rrnB-poly-tRNA are given in lower-case letters. Amino acid (aa) residues identical to those of the rrnB-peptide are encircled by ellipse. Bases identical to the rrnB-mRNA are boxed by “”. Base identities in the comparison with indicated sequences are marked with “+” or “*” immediately under the sequence. Results of statistical evaluations in representative comparisons are given in the right area, with giving base-identity levels (%) and corresponding Pnuc (m,n) values. Sequence data were obtained from GenBank Database and SWISSPROT. Abbreviations: rrnB-poly-tRNA: poly-tRNA region of the B. subtilis rrnB operon; tRNA-His, the tRNAHis gene in the rrnB-poly-tRNA region., cI, lambda phage cI gene encoding lambda repressor; adk, E. coli (EC) or Saccharomyces cerevisiae (SC) adenylate kinase gene; GAPDH: glyceladehyde-3-phosphate dehydrogenase encoded by S. cerevisiae gpd2 gene; argS: B. subtilis arginyl-tRNA synthetase gene.
78
Ohnishi et al.
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
79
On the other hand, FASTA and BLAST homology search revealed that glyceraldehydes 3-phosphate dehydrogenase (GAPDH) shows a highest aa-sequence similarity to the rrnB-peptide. DNA-binding domains of lambda-repressor and homeoproteins are also homologues of rrnB-peptide [22]. As shown in Fig. 4, a highest level of base-sequece homology with rrnB-mRNA is found in the B. subtilis argS gene segment encoding arginyl tRNA synthetase (aa’s 450-471), giving a 55.9% basse-identity and Pnuc (33, 59) = 4.0 × 10−7 . Since the base-identity level of argS and rrnB-mRNA is 40.0%, and Pnuc (24, 60) = 7.3 × 10−3 , we can conclude that (Pred. 4) is most plausibly proved to be true. Alignment of AK domains (aa- and base-sequences) with trrnD-peptide/trrnD-mRNA also support their genuine homology-relations, confirming that the predicted PG product deduced in Fig. 1[A] has now proved to be a genuine PG product derived from a “more primordial” gene product, i.e., rrnB-peptide. Dot-matrix analysis revealed that the tRNAHis region of the rrnB-poly-tRNA region is most similar to the rrnB-mRNA, showing a possible very low level of homology, a 36.1% base-identity and Pnuc (22,61) = 0.036, which is much weaker than the above-mentioned homology level between argS and the rrnB tRNA region (40.0% base-match and Pnuc (= 7.3 × 10−3 ). Thus (Pred. 4) as well as (Pred. 3-1) would be true in this case. It is also important to note that the sponge homeoprotein 1 (aa’s 50-107) shows evident homology with the aa’s 262-320 of the yeast glyceraldehydes 3-phosphate dehydrogenase (gpd2), giving 29.8% aa- and 46.8% base-matches, Pnuc (80, 171) = 5.9 × 10−10 . Further Database analyses elucidated that the rrnB-peptide shows significant levels of sequence similarity not only to lactate dehydrogenase and anthrocyanidin synthase [20], but also to ftsZ RNA-binding protein, NADHplastoquinone oxidoreductase, NAD+ synthetase, LPS biosynthesis protein, and TGF-beta, which will be published elsewhere.
3
Discussion and Conclusion
All results described in this paper strongly suggest that the trrnD-type and rrnB-type poly-tRNA structures found in the B. subtilis operons are genuine relics of early peptide-synthesizing poly-tRNA biomachines. Since two types of tri-tRNA complexes are shared by these two poly-tRNAs (as shown in Fig. 3 and its legend), tri-tRNA ribozymes for making “NSE” and “MDF” would have had evolved before the emergence of the trrnD-type and rrnB-type poly-tRNA machines. This tells us that aa-specific diversification of proto-tRNA would have had preceded the emergence of early peptide-synthesizing oligo- and poly-tRNA machines. Since anticodons in tri-tRNAs (NSE, MDF) are nearly the same ones in the two B. subtilis poly-tRNAs, base-sequence complementarities between pres. codons and anticodons can be considered to have been evolutionarily established by selection of base replacements occurred mainly in the pres. mRNA, and not in the pres. anticodons. This means that tRNA riboorganisms seem to have “actively selected” suitable base-replacement mutations by using pres. anticodons capable of interacting with pres. codons of pres. mRNA (which was also a member of tRNA-riboorganism). This seemingly active or autopoietic selection of basereplacements is very interesting from the aspect of autopoietic evolution of biotic systems, which is recently discussed in detail by Ohnishi et al. [18, 3, 27]. Genesis of genetic codon-system is a typical process of semeiogenesis, which seems to be a “semeiotic culture” of the intracellular tRNAriboorganismic society [19, 18, 12, 27]. Arbitrariness of the correspondences between codons and amino acids strongly suggests that genetic codon-system is a typical semeitic culture system, because such arbitrary correspondence is known to be the most essential character of semeiotic (culture) systems as de Saussure [3] has first found and established as a basis of general semeiotic phenomenon. Such are also true in bio-semeiotic and zoo-semeiotic systems found in animal behaviology and mating behaviours (such as display signals) [25]. Further discussions from the aspect of evolutionary semeiotics have been made in the references [19, 18, 27]. Such discussions seem to suggest that evolutionary process towards emergence of geneic codon system would be a cultural semeiotic phenomenon of intracellular riboorganismic society. The poly-tRNA theory would therefore be a “semeiotic culture theory” on the origin of genetic codes and earliest mRNAs, which is neither accidental frozen theory
80
Ohnishi et al.
(by Crick) nor a stereochemical theory.
Achnowledgments The authors acknowledges Profs. S. Yokoyama, H. Yanagawa, M. Go, Y. Hushimi, Dan-Sowkawa, Ikegami, T. and S. Odani for useful discussions. This work is partly supported by Grant for the Promotion of Niigata University Projects in 2000.
References [1] Altschul, S.F. and Gish, W., Local alignment statistics, Methods Enzymol., 266:450–480 1996. [2] Cech, T,R., Zaug, A.J., and Grabowski, P.K., In vitro splicing of the ribosomal RNA precursor of Tetrahymena, Cell 27:487–496, 1981. [3] De Saussure, F., Course de linguistique generale. Edition critique Rudolf Engeler, Harrasowitz, 1989. [4] Eigen, M. and Schuster, P., The hypercycle, Natur-wissenschaften, 54:541–565, 341–369, 1977. [5] Guerrier-Takada, C., Gardinner, K., Marsh, T., Pace, N., and Altman, S., The RNA moiety of ribonuclease P is the catalytic subunit of the enzyme, Cell, 35:849–857, 1983. [6] Haldane, J.B.S., Data needed for a blueprint of the first organism, The Origins of Prebiological Systems and of their Molecular Matrices, Academic Press, 11–15, 1965. [7] Hokari, S. and Ohnishi, K., The origin of 16S rRNA, as viewed from poly-tRNA theory, Nucleic Acids Symp. Ser., 39:159–160, 1998. [8] Mount, D.W., Bioinformatics. Sequence and Genome Analysis, Cold Spring Harbor Press, 2001. [9] Noller, H.F., Hoffarth V., and Zimniak L., Unusual resistance of peptidyl transferase to protein extraction procedures, Science, 256:1416–1419, 1992. [10] Ohnishi, K., A most primitive primordial gene as a building block of the genes for the adenylate kinase, F1-ATPase subunits, and core enzyme subunits, Nucleic Acids Symp. Ser., 17:127–130, 1986. [11] Ohnishi, K., Evolution from semi-tRNA to early protein-synthesizing RNA molecule, Endocytobiology V. 5th International Colloquium on Endocytobiology and Symbiosis, Tubingen University Press, 407–414, 1993. [12] Ohnishi, K., Evolutionary meanings of the primary and secondary structure of the “ur-RNA”, Symmetries in Science IV, Plenum Press, 147–176, 1990. [13] Ohnishi, K., Molecular evolution of ion channels in central nervous system: Possible primitive mRNA segments conserved in glutamate and acetylcholin receptor genes, Ann. N.Y. Acad. Sci. 707:524–528, 1993. [14] Ohnishi, K., Non-paradoxical origin of the genetic codes [I], (In Japanese), Seibutsu Kagaku (Biological Science) 47:10–23, 1995. [15] Ohnishi, K., Phosphoglycerate-transporter protein B as a most primitive protein predicted by the poly-tRNA theory, Genome Informatics, 9:393–394, 1998.
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
81
[16] Ohnishi, K., Poly-tRNA structures as early protein-synthesizing apparatus: Evolution from prototRNA to most primitive mRNAs, Viva Origino, 23:237–254, 1995. [17] Ohnishi, K., Poly-tRNA theory on the origin and evolution of mRNA and genetic codes: Evolution from tandem tRNA repeats to premitive mRNA, Genome Informatics 4:325–331, 1993. [18] Ohnishi,K., Hokari, S., Shutou, H., Kanbe, D., Ohshima, M., Sawamura, H., Furuichi, N., Shibuki, K., and Goda,M., Neural-network-like bio-machinogenesis via semeiogenesis : A unified theory on the origin of genetic codes and other semeiotic systems, Viva Origino, 30(2):63–78, 2002. (See URL site in [26]) [19] Ohnishi, K., Hokari, S., and Yanagawa, H., Origin and evolution of early peptide-synthesizing biomachines by means of hierarchical sociogenesis of intracellular primitive tRNA-riboorganisms, Proc. 4th Int. Symp. on Artificial Life and Robotics, 344–349, 1999. [20] Ohnishi, K. and Shutou, H., Autopoietic active selection by cognitive self-systems, Proc. 6th International Symposium on Artificial Life and Robotics, Shuubunsha, 490–495, 2001. [21] Ohnishi, K., Suzuki, T., and Shutou, H., An E. coli pseudo-tRNA gene cluster homologous to the tRNA-gene cluster in the Bacillus subtilis trrnD operon, Genome Informatics, 11:325–326, 2000. (URL site in [27]) [22] Ohnishi, K., Tanaka, H., and Yanagawa, H., The origin of DNA-binding domains, as viewed from poly-tRNA theory, Nucleic Acids Symp. Ser., 39:251–252, 1998. [23] Ohno, S., Original domain for the serum albumin family arose from repeated sequences, Proc. Natl. Acad. Sci. USA, 78:7657–7661, 1981. [24] Ohno, S. and Matsunaga, T., The 48-base-long primordial building block of immunoglobulin light-chain variable regions, Proc. Natl. Acad. Sci. USA, 79:2338–2341, 1982. [25] Sebeok, T.A., Perspectives in Zoosemiotics, Mouton, 1972. [26] http://www.origin-life.gr.jp/3002/3002063/3002063.html (URL site for [18]) [27] http://www.jsbi.org/journal/GIW00/GIW00P045.pdf (URL site for [21])
71
Origin of Most Primitive mRNAs and Genetic Codes via Interactions between Primitive tRNA Ribozymes Koji Ohnishi1
Shouken Hokari1
Hiroshi Shutou1
Madoka Ohshima1
Naotaka Furuichi2
Masaki Goda3
[email protected]
[email protected]
[email protected]
[email protected]
1
2 3
Department of Biology, Faculty of Science, Niigata University, Igarashi-2, Niigata 950-2181, Japan Faculty of Agriculture, Niigata University, Igarashi-2, Niigata 950-2181, Japan Faculty of Engineering, Niigata University, Igarashi-2, Niigata 950-2181, Japan
Abstract The origin and early evolution of genetic codon system and early mRNAs were analyzed from a viewpoint of primordial gene theory and the poly-tRNA theory. A hypothetical 25-amino acid (aa)-primordial peptide was deduced from internal aa-sequence homology of adenylate kinases. Theoretical models were made which can reasonably explain how primitive tRNA(s) could have had converted to be earliest mRNAs via interactions between presumptive anticodons and (poly)tRNA ribozyme. Transfer-RNA gene clusters in the trrnD- and rrnB-operons of Bacillus subtilis seemed to be relics of early peptide-synthesizing RNA machine. Detailed analyses revealed that the poly-tRNA regions in these operons are true relics of RNA-machine for making a 16-aa trrnDpeptide and a 21-aa rrnB-peptide, whose aa sequences are in the order of aa specificities of tRNAs in the tRNA gene clusters of the trrnD-operon and rrnB-operon, respectively. The primordial gene-encoded peptide deduced from adenylate kinases were found to be a genuine homologue of the rrnB-peptide. Various protein superfamilies were found to have evolved from either of these two types of primitive peptides. Earlist mRNAs were concluded to have evolved from tRNA Gly (trrnDmRNA) or tRNAHis (rrnB-mRNA) , where trrnD- and rrnB-mRNAs are hypothetical primitive mRNAs complementary to the tandem arrangement of 16 or 21 anticodons of tRNAs in the trrnDoperon and rrnB-operon, respectively. The poly-tRNA model is considered to be an excellent theory, because it can reasonably explain origins of both genetic codes and earliest mRNAs, and because the hypothesis can be statistically evaluated by base-identity levels in proper alignments. The genetic codon system is a typical mature semeiotic system within a cell, and the genesis of the geneic codon system was discussed from an aspect of de Saussure’s semeiology. Arbitrary correspondence between (anti)codon and aa would be most plausibly a result of semeiotic culture system of intracellular tRNA-riboorganismic society.
Keywords: primordial gene, adenylate kinase, poly-tRNA theory, trrnD-peptide, rrnB-peptide, early peptide-synthesis, glyceraldehydes-3-phosphate dehydrogenase, 3-phosphoglycerate transporter protein B, arbitrary correspondence, semeiotic culture
1
Introduction
Before the establishment of the RNA-first theory based on the finding of ribozymes by Cech [2] and Altman [5], Haldane [6] and Eigen [4] predicted that RNAs emerged earlier than DNAs as primitive genes or nucleic acid-replicators. Since an RNA molecule has a structure of “poly-AMP”, an AMP/ATP-dependent energy-obtaining system accompanied by proper metabolism with nucleoside supply must have preceded the emergence of RNA. How first protein-encoding RNAs (most primitive
72
Ohnishi et al.
mRNAs) could have been made throughout evolution has long been a most important unsolved problem in molecular evolutionary biology. A hypothesis, later named “poly-tRNA theory”, proposed by one of us (K.O.) in 1993 [11, 17, 13] seemed to make a possible breakthrough in solving the problem. The poly-tRNA theory revealed that the two tRNA gene clusters (hereafter denoted by trrnD-polytRNA region and rrnD-poly-tRNA region) found in the trrnD-operon and the rrnB-operon, respectively, of a Gram-positive bacteria, Bacillus subtilus, are relics of early peptide-synthesizing RNA apparatus [11, 17, 13, 16, 14, 22, 15, 19, 21, 18, 27]. Detailed comparative analyses have hitherto elucidated that (i) ancient primitive tRNAGly and tRNAHis in the respective early tRNA-clusters would have converted to earliest mRNAs for making 16- aa- and 21-aa-peptides (cald trrnD-peptide and rrnB-peptide) via interaction of the tRNAs and presumptive anticodon regions of the respective poly-tRNAs, and that long DNA regions comprizing several tRNAs and spacers of the trrnD- and rrnB-poly-tRNA regions were found to be most plausibly homologous to the DNA regions encoding Glycyl-tRNA synthetase alpha subunit gene (GlyRS) [17, 16, 14, 18] and the lambd repressor (cI protein) [in preparation]. These findings brought us a new viewpoint to elucidate the evolutionary history of tRNAs and early mRNAs. A proto-tRNA, commonly ancestral to contemporary tRNAs, would have had emerged as a new (possibly intracellular) replicable ribozyme (possibly a kind of “riboorganism”) capable of making dipeptide bonds, and further diverged to different primitive tRNAs with their respective amino acid specificity. Such primitive tRNAs would have interacted and/or co-operated with one another, and constituted an interacting complex or an intracellular symbiotic complex (= riboorganismic society [18, 27]) for making oligopeptides or longer peptides. This paper reports new evidence for the evolution of poly-tRNA-derived mRNAs. Origin of genetic codon system was discussed from the aspects of poly-tRNA theory and de Saussure’s semeiotic culture theory.
2 2.1
Proposing and Proving of the Poly-tRNA Hypothesis Reconsideration of a Primordial Gene as Building Block of Adenylate Kinase Genes
Ohno and Matsunaga [23, 24] first postulated a primordial gene (PG) theory, hypothesizing that relics of some 48-base PGs might have existed as building blocks of modern genes encoding proteins such as serum albumin and immunoglobulin. One of us (K.O.) had a great interest in their theory, and gegan to search possible common amino acid (aa) sequence segments shared by internal repeats of house-keeping proteins [10, 12]. A most beautiful result thus obtained was a common sequence found between duplicated aa sequence regions of adenylate kinase (AK) ([12], p.156). Fig. 1[A] shows common aa’s in the E. coli and pig AK’s shared by the first amino-terminal domain (domain I, corresponding to aa’s 1 − 32 in E. coli) and the second domain (domain II, aa’s 33 − 71 in E. coli). Based on this sharing of 12 aa’s, a primordial gene encoding a reconstructed aa sequence “(N-term) (VXX)GSXXGXQAXXIMEKXXXXLXT (C-term)” (where every of X’s is unknown) can be deduced as an ancient protein sequence encoded by a most primitive primordial gene. The essential portion of this primordial sequence was further confirmed by using other sequence data of adenylate kinases from obtained from GenBank Database. Since adenylate kinase is possessed by every of living organisms, a gene duplication (of the primordial gene) generating the contemporary 1st and 2nd homology domains must have occurred at a time before the latest common ancestor of all living organisms. Among some DNA sequences encoding common aa-sequence segments, there seemed to be some sequences having slight similarities to tRNAs, which suggested an interesting possibility that some PG(s) might have had evolved from some kind(s) of primitive tRNA.
2.2
Towards Building a Tentative RNA Model for Early Peptide-Synthesis [14]
If some PG(s) could be really homologous to tRNAs, how and why such homologous relations could have emerged? If a primitive tRNA (or proto-tRNA) could be an ancestor of a first most primitive mRNA, then it is reasonable to consider that the primitive tRNA must have converted to be a first
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
73
mRNA. In order to consider more exactly, a term “presumptive X” (pres. X) is defined as something to evolve to X in future, and therefore, is an evolutionary ancestor of “primitive X” (prim. X) (X = anticodon, codon, mRNA, etc.).
Figure 1: A primordial gene and its possible origin from a hypothetical rrnB-mRNA. [A] A primordial peptide deduced from an internal aa sequence homology between the 1st and 2nd homology regions of the adenylate kinases (AK’s) from E. coli (adk gene) and pig (myokinase). Amino acids (aa’s) shared by the 1st and 2nd homology regions are boxed. Aa’s identities between the 2nd AK domain and a hypothetical primitive peptide, rrnB-peptide, are also shown. See text and Fig. 4 for rrnB-peptide. [B] Base-sequence homology between a hypothetical rrnB-mRNA and a yeast (S. cervisiae) AK gene (adk1). Aa’s shared by the rrnB-peptide and the yeast AK are boxed. Bases identical to the rrnBmRNA are indicated by “o”. Statistical evaluation of base-identity level (%) by Pnuc (m,n) value (See (Eq.1) in text.) is given. See text and Fig. 4. for definition of rrnB-mRNA. Evolution from a tRNA to a most primi. mRNA means that every three-base-unit (triplet) in a tRNA (which is “pres. mRNA”) would have had converted to a triplet codon complementary (in the sense of Watson-Crick-type base-pairing) to the “pres. anticodon” of some tRNA, via interaction between the triplet (pres. codon) on the pres. mRNA and the pres. anticodon on the tRNA, as shown in Fig. 2. If we make a reasonable assumption that prim. tRNAs before the emergence of mRNAs would have been capable of carrying a (more or less) specific aa (probably by using ATP) at its 3’(CCA)end. The model in Fig. 2 therefore suggests that early peptide synthesis might have had occurred by a co-operative interactions of early tRNAs (i.e., by cooperative behaviour of early tRNA-riboorganisms, in terms of riboorganismic behaviology as discussed by Ohnishi et al. [18, 27]) capable of making peptide-bond by their own ribozymic activity. If different aa-carrying tRNAs interacted with pres. anticodons on a pres. mRNA (which is some kind of prim. tRNA), then their cooperative behaviour would ensure easy making of a peptide bond between the aa’s carried by the neighbouring tRNAs. If
74
Ohnishi et al.
such peptide-bond-making by two-neighbouring tRNAs could occur regularly from 5’-side to 3’-side of the pres. mRNA, then a primitive peptide would be made by co-operative interactions of tRNAs (or by cooperation of tRNA riboorganisms), as shown in Fig. 2. Base-replacement mutations giving rise to base-complementary between pres. codons on the pres. mRNA and the corresponding pres. anticodons would generate a semeiotic correspondence between prim. codons (evolved from pres. codons) and a prim. codon-specific aa carried by the corresponding tRNA. Accordingly, the tentative early peptide-synthesizing model in Fig. 2 seems to give an important basis for considering the origin of the first mRNA possessing earliest codons. Some difficulties encountered are that how such interactions between pres. codons and pres. anticodons could have had occurred so as to satisfy thermodynamical and probabilistic conditions towards accumulating basecomplementarities at reasonable evolutionary rate. These difficulties would be solved in the following sections.
Figure 2: A tentative model for early peptide-synthesis showing a possibility that a primitive tRNA (= presumptive mRNA) could have had converted to an earliest mRNA via interactions with presumptive anticodons of other tRNAs. Watson-Crick-type base-complementarities between presumptive anticodons and presumptive codons would have been selected via interactions between presumptive anticodons (on primitive tRNAs) and presumptive codons on a presumptive mRNA (= a kind of primitive tRNA). See text for details.
2.3
Bacillus subtilis is an Excellent Teacher for Building the Poly-tRNA Model
In the next step, we need to refine the tentative early-peptide-synthesizing RNA model shown in Fig. 2. Structural characteristics of the B. subtilis trrnD- and rrnB-operons were found to kindly teach us a critical solution for this difficult problem. The two operons shown in Fig. 3 mean that the direct transcripts of these operons have a structure of “16S-rRNA-23S-rRNA-5S-rRNA-(tRNA)n ”, where n= 16 and 21 for the trrnD- and the rrnB-operon, respectively. In the tentative model in Fig. 2, aa-carrying prim. tRNAs are composed of tRNA molecules possessing different aa-specificities, and every of them behaves independently, resulting in probabilistic difficulties to reach and interact with their properly corresponding pres. codons on the pres. mRNA in every generation throughout evolution towards the genesis of the first mRNA. On the contrary, each transcript of the two operons has a large RNA region (hereafter called “poly-tRNA(region)”) consisting of tandemly connected tRNAs, which allows pres. codons (of pres. mRNA) to easily interact with their counterpart pres. anticodons (of poly-tRNA). And moreover, such interactions of two consecutive codons with anticodons of the corresponding two consecutive tRNAs (of poly-tRNA) could have tended to repeatedly re-occur (throughout evolution) on the prim. A and P sites on a prim. ribosome consisting exclusively of prim. 16S-, 23S- and
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
75
5S-rRNAs (i.e. without possessing any ribosomal proteins) encoded by the same operon (as found in Fig. 3). This is because early ribosome is considered to have consisted exclusively of rRNA ribozymes on which peptide synthesis (peptidyl-transferation) would have had occurred [9]. These considerations allow us to see that the poly-tRNAs found in the trrnD- and the rrnB-operons are most seemingly relics of early tRNA-complexes capable of cooperatively making most primitive peptides, whose aa sequences are “(N-term.) NSEVMDFTYWHQGCLL (C-term.)” (named trrnD-peptide) and “(Nterm.) VTKLGLRPAMISMDFHGINSE (C-term.)” (named rrnB-peptide), respectively.
Figure 3: Eubacterial t-RNA gene clusters in representative species. Gram-positive bacteria (Bacillus subtilis, Staphylococcus aureus) possess close homologues of B. subtilis trrnD- and/or rrnB-type tRNA gene clusters. In the E. coli rrnB operon (Genbank Locus ECORGNG), a pseudo-tRNA gene cluster closely homologous to the B. subtilis trrnD-type poly-tRNA was found [21]. Note that two types of tri-tRNA segment, “NSE” and “MDF”, can be found in both the trrnD- and the rrnB-operons of B. subtilis, which seems to suggest earlier evolution of tripeptide(NSE or MDF)-making tri-tRNA complexes. If we compare the trrnD- and rrnB-peptides with the first and the second homology domains of AK’s as well as with the above-mentioned PG-encoded 25-aa-peptide, “(N-term)(VXX)GSXXGXQAX XIMEKXXXXLXT (C-term)” (deduced from internal homology of AK’s), remarkable aa sequence similarities were found in the comparison of the rrnB-peptide with these AK domains and therefrom deduced PG-encoded peptide, as is clearly shown in Fig. 1 [A], [B]. Seven out of the 21 aligned aapositions (= 33.3%) are occupied by identical residues in the comparison of the rrnB-peptide with the S. cervisiae AK (encoded by adk1 gene). These aa-similarities strongly suggests that the rrnBtype tRNA gene cluster (or rrnB-poly-tRNA) would be a real relic of an early RNA biomachine for synthesizing the rrnB-peptide. On the other hand, it is also important to note that tRNA-gene clusters homologous to either or both of the B. subtilis trrnD-type and rrnB-type clusters are widely found in most of hitherto analyzed Gram-positive bacteria, as exemplified by a large tRNA cluster in Staphylococcus aureus (as shown in Fig. 3), and by an Acholeplasma laidwai tRNA cluster, “VTKLGLRPAMISMDF” (not shown in Fig. 3). A pseudo-tRNA gene-cluster (consisting of at least 10 pseudo-tRNA genes) homologous to the B. subtilis trrnD-poly-tRNA region has recently been found in the immediate downstream of the 5S rRNA gene in the rrnB-operon (rRNA operon) of a representative Gram-negative bacteria (Proteobacteria), E. coli [21]. E. coli has also a three-tRNA gene-cluster, tRNAGly -tRNACys -tRNALeu , which is evidently a true homologue of the tRNAGly -tRNACys -tRNALeu region of the B. subtilis
76
Ohnishi et al.
trrnD-tRNA gene cluster. This homologue-relationship among bacterial tRNA gene-clusters strongly suggested that the B. sbutilis trrnD- and rrnB-tRNA clusters would well conserve a very ancient, ancestral poly-tRNA structure.
2.4
The Primordial Gene Deduced from Adenylate Kinases is a First Candidate for a Descendant of the rrnB-mRNA
Thus we now reach to a possible, very important hypothesis that the trrnD- and rrnB-peptides could be most plausible earliest peptides made by co-operative interactions (or behaviours) of tRNAriboorganisms. Based on the theoretical considerations mentioned above, a rrnB-type poly-tRNA model for synthesizing a primitive 21-aa-rrnB-peptide can be constituted as found in Fig. 4. In this Figure, the “rrnB-poly-tRNA is a poly-tRNA region of the long RNA-transcript from the rrnB-operon. The “rrnB-mRNA” is defined as an hypothetical 63-base-RNA (a candidate for prim.mRNA) whose sequence is complementary (in Watson-Crick-type) to the 21-anticodons (63 bases) in the rrnB-polytRNA, as shown in Fig. 4. The 21 pres. (or prim.) codons of an early rrnB-mRNA would have had interacted with their respective counterpart pres. (or prim.) anticodons of an early rrnB-poly-tRNA, and thereby helped the catalytic ribozyme function of the poly-tRNA to make a “rrnB-peptide” (on the primitive A and P sites of a primitive 16S-23S-5S-rRNA-ribosome), whose aa sequence is “VTKLGLRPAMISMDFHGINSE”. The rrnB-mRNA would have had most plausibly evolved from some primitive tRNA. If this rrnB-type poly-tRNA model well represents the true evolutionary history of early peptide synthesis, it is strongly expected that the DNA segment (in the adk1 gene) encoding the second homology region (aa’s 47-66) of the S. cerevisiae AK would be homologous to the rrnB-mRNA, an ancestral mRNA for making the rrnB-peptide. The comparison is given in Fig. 1[B], resulting in a finding that a 50.5% (= 29/57) base-match level was observed between these tow base-sequences. Statistical evaluation of base-identity levels of m observrd base-matches in n aligned base-positions (excluding gap-including positions) was made by computing Pnuc (m,n) values defined by X Pnuc (m, n) = Cn,i (1/4)i (3/4)n−i , (Eq.1) P
(i=m,n)
where Cn,i = n!/[i!(n − i)!], and (i=m,n) denotes summation over m to n. Pnuc (m,n) value in (Eq.1) denotes the probability by chance for the occurrance of m- or more-base-matches in an n-base-long alignment of a pair of randomly selected sequences, and gives a statistical evaluation of the homologylevels of the observed base-matche level. For the observed base-match level in Fig. 1[B], Pnuc (29, 57) = 2.4 × 10−5 was obtained, suggesting that the adk1 DNA segment is homologous to the rrnB-mRNA, with a probability of error occurrence of 2.4 × 10−5 . Since another hypothetical mRNA, trrnD-mRNA, can be defined based on the trrnDpoly-tRNA sequence [11, 17, 13, 16, 18, 27], in a similar way as in the case of the rrnB-mRNA, the trrnD-mRNA is also a most plausible candidate for an early mRNA from which various contemporary house-keeping protein genes ( such as those genes encoding 3-phosphoglycerate transporter protein B (pgtB), glycyl-tRNA synthetase alpha subunit, Fi-ATP synthase γ subunit, F0-ATP synthase α subunit, C-type lectin, etc.) could have later evolved, as has been reported in [11, 17, 18, 27].
2.5
Poly-tRNA Models and Its Evolutionary Significance: Proteins Encoded by rrnB-mRNA-Derived Genes and trrnD-mRNA-Derived Genes
These results now bring us with a hopeful strategy for finding early evolution of primitive mRNAs from either of the rrnB- and trrnD-mRNAs. We can make the following predictions; *(Pred. 1) Some aa-sequence segments in contemporary proteins in living organisms might be homologous to either one of the rrnB- and trrnD-peptides. *(Pred. 2) If (Pred.1) is true, then the DNA sequence segments encoding such aa-sequence regions are evolutionary homologues of either of the rrnB-and trrnD-mRNAs. *(Pred. 3) If (Pred. 1) and (Pred. 2.) are true, then the trrnDand the rrnB-mRNAs are considered either *(Pred. 3-1) to have evolved from some kinds of tRNA,
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
77
as predicted from the original considerations of the (poly-)tRNA models in Fig. 2 and Fig. 4, or *(Pred. 3-2) to have evolved from some unknown primitive RNA(s) which would have later interacted with presumptive anticodons of the corresponding poly-tRNA. *(Pred. 4) If (Pred. 3-1) is true, then the contemporary protein-encoding DNA segments homologous to the trrnD- or rrnB-mRNA would be more highly similar to the trrnD- or rrnB-mRNA than to the corresponding homologous tRNA. In the search for protein sequence segments similar to the trrnD- and rrnB-peptides, database search by FASTA, BLAST and psi-BLAST were used by employing GenBank and SWISS-PLOT databases. Querry sequences are extremely short, and therefore, conventional thresholds of E-values based on Gumbel’s statistics of extremes would not represent exact statistical meanings, since BLAST and psi-BLAST searches are based on statistics assuming that the querry is considerably long [1, 8]. However, these algolisms are very useful for finding out most similar very short sequence segments, if we employ adequetly large thresholds of E-value. Thus found most similar short sequences were compared with the querry and also with one another, by employing multialignment ClustalW algolism, and finally obtained alignments (as showen in Fig. 4) were therafter stasistically evaluated by P nuc values. In case of the trrnD-peptide-derived proteins hitherto found and reported, (Predictions 1, 2, 3-1, and 4) have been elegantly proved to be true, as found in references [11, 17, 13, 16, 14, 15, 19, 21, 18, 27]. The trrnD-peptide/trrnD-mRNA shows highest similarities to the Salmonella typhimurium 3phosphoglycerate transporter protein B (PgtB) /pgtB gene, giving 11 aa-identity (66.8%) to the 16-aatrrnD-peptide (60.9% base-identity, Pnuc (28, 46) = 2.8 × 10−7 ), which is lower than than the similarity of pgtB with the tRNAGly -region of trrnD-poly-tRNA (58.1% base-match, Pnuc (25, 43) = 3.9 × 10−6 ), thus satisfying (Pred. 4) [15]. The E. coli glycyl-tRNA synthetase (GlyRS) alpha subunit and its coding DNA region are also genuine homologues of the trrnD-peptide and trrnD-mRNA (63.8%, Pnuc (30, 47) = 2.2 × 10−8 ), respectively [15]. F1 -F0 -ATP-synthase subunits (F1 -gamma, F0 -a, etc.) and C-type lectin (lectin Bra-3) also show significant levels of homology to GlyRS, trrnD-mRNA and trrnD-poly-tRNA (See Fig. 1 and Fig. 2 in the references [18, 27]). A portion of 16S rRNA is also homologous to the tRNA(Gly)-tRNA(Cys)-tRNA(Leu) region [19, 7], which is commonly found in diverge species of eubacteria as found in Fig. 3. As will be published elsewhere, further searches using recent Genbank/SWISSPROT databases elucidated that aa-sequence portions in succinyl-CoA synthetase beta subunit, the JH segment of Immunoglobulin VH domain (encoded by JH exon), and reverse transcriptases are also homologous to trrnD-peptide and its homologues. Figure 4 (Next page): An rrnB-type poly-tRNA model for an early peptide (rrnB-peptide)-synthesizing RNA machine, showing homologies of the rrnB-peptide and/or rrnB-mRNA with the B. subtilis tRNA His , adenylate kinase domains, glyceraldehydes-3-phosphate dehydrogenase, DNA-binding domains (of lambda repressor and the sponge homeoprotein 1), and an arginyl tRNA-synthetase. In the upper portion of this Figure, an rrnB-type poly-tRNA model is given. In the lower portion, the tRNAHis -including portion of the poly-tRNA region of the B. subtilis rrnB-operon , as well as possible candidates of homologues of the rrnB-peptide together with their corresponding coding DNA regions are aligned against the rrnB-peptide and the rrnB-mRNA. The helix regions of HTH (helix-turn-helix) DNA-binding domains (in cI and homeoprotein 1) are indicated by “”. Bases in the spacer regions of the rrnB-poly-tRNA are given in lower-case letters. Amino acid (aa) residues identical to those of the rrnB-peptide are encircled by ellipse. Bases identical to the rrnB-mRNA are boxed by “”. Base identities in the comparison with indicated sequences are marked with “+” or “*” immediately under the sequence. Results of statistical evaluations in representative comparisons are given in the right area, with giving base-identity levels (%) and corresponding Pnuc (m,n) values. Sequence data were obtained from GenBank Database and SWISSPROT. Abbreviations: rrnB-poly-tRNA: poly-tRNA region of the B. subtilis rrnB operon; tRNA-His, the tRNAHis gene in the rrnB-poly-tRNA region., cI, lambda phage cI gene encoding lambda repressor; adk, E. coli (EC) or Saccharomyces cerevisiae (SC) adenylate kinase gene; GAPDH: glyceladehyde-3-phosphate dehydrogenase encoded by S. cerevisiae gpd2 gene; argS: B. subtilis arginyl-tRNA synthetase gene.
78
Ohnishi et al.
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
79
On the other hand, FASTA and BLAST homology search revealed that glyceraldehydes 3-phosphate dehydrogenase (GAPDH) shows a highest aa-sequence similarity to the rrnB-peptide. DNA-binding domains of lambda-repressor and homeoproteins are also homologues of rrnB-peptide [22]. As shown in Fig. 4, a highest level of base-sequece homology with rrnB-mRNA is found in the B. subtilis argS gene segment encoding arginyl tRNA synthetase (aa’s 450-471), giving a 55.9% basse-identity and Pnuc (33, 59) = 4.0 × 10−7 . Since the base-identity level of argS and rrnB-mRNA is 40.0%, and Pnuc (24, 60) = 7.3 × 10−3 , we can conclude that (Pred. 4) is most plausibly proved to be true. Alignment of AK domains (aa- and base-sequences) with trrnD-peptide/trrnD-mRNA also support their genuine homology-relations, confirming that the predicted PG product deduced in Fig. 1[A] has now proved to be a genuine PG product derived from a “more primordial” gene product, i.e., rrnB-peptide. Dot-matrix analysis revealed that the tRNAHis region of the rrnB-poly-tRNA region is most similar to the rrnB-mRNA, showing a possible very low level of homology, a 36.1% base-identity and Pnuc (22,61) = 0.036, which is much weaker than the above-mentioned homology level between argS and the rrnB tRNA region (40.0% base-match and Pnuc (= 7.3 × 10−3 ). Thus (Pred. 4) as well as (Pred. 3-1) would be true in this case. It is also important to note that the sponge homeoprotein 1 (aa’s 50-107) shows evident homology with the aa’s 262-320 of the yeast glyceraldehydes 3-phosphate dehydrogenase (gpd2), giving 29.8% aa- and 46.8% base-matches, Pnuc (80, 171) = 5.9 × 10−10 . Further Database analyses elucidated that the rrnB-peptide shows significant levels of sequence similarity not only to lactate dehydrogenase and anthrocyanidin synthase [20], but also to ftsZ RNA-binding protein, NADHplastoquinone oxidoreductase, NAD+ synthetase, LPS biosynthesis protein, and TGF-beta, which will be published elsewhere.
3
Discussion and Conclusion
All results described in this paper strongly suggest that the trrnD-type and rrnB-type poly-tRNA structures found in the B. subtilis operons are genuine relics of early peptide-synthesizing poly-tRNA biomachines. Since two types of tri-tRNA complexes are shared by these two poly-tRNAs (as shown in Fig. 3 and its legend), tri-tRNA ribozymes for making “NSE” and “MDF” would have had evolved before the emergence of the trrnD-type and rrnB-type poly-tRNA machines. This tells us that aa-specific diversification of proto-tRNA would have had preceded the emergence of early peptide-synthesizing oligo- and poly-tRNA machines. Since anticodons in tri-tRNAs (NSE, MDF) are nearly the same ones in the two B. subtilis poly-tRNAs, base-sequence complementarities between pres. codons and anticodons can be considered to have been evolutionarily established by selection of base replacements occurred mainly in the pres. mRNA, and not in the pres. anticodons. This means that tRNA riboorganisms seem to have “actively selected” suitable base-replacement mutations by using pres. anticodons capable of interacting with pres. codons of pres. mRNA (which was also a member of tRNA-riboorganism). This seemingly active or autopoietic selection of basereplacements is very interesting from the aspect of autopoietic evolution of biotic systems, which is recently discussed in detail by Ohnishi et al. [18, 3, 27]. Genesis of genetic codon-system is a typical process of semeiogenesis, which seems to be a “semeiotic culture” of the intracellular tRNAriboorganismic society [19, 18, 12, 27]. Arbitrariness of the correspondences between codons and amino acids strongly suggests that genetic codon-system is a typical semeitic culture system, because such arbitrary correspondence is known to be the most essential character of semeiotic (culture) systems as de Saussure [3] has first found and established as a basis of general semeiotic phenomenon. Such are also true in bio-semeiotic and zoo-semeiotic systems found in animal behaviology and mating behaviours (such as display signals) [25]. Further discussions from the aspect of evolutionary semeiotics have been made in the references [19, 18, 27]. Such discussions seem to suggest that evolutionary process towards emergence of geneic codon system would be a cultural semeiotic phenomenon of intracellular riboorganismic society. The poly-tRNA theory would therefore be a “semeiotic culture theory” on the origin of genetic codes and earliest mRNAs, which is neither accidental frozen theory
80
Ohnishi et al.
(by Crick) nor a stereochemical theory.
Achnowledgments The authors acknowledges Profs. S. Yokoyama, H. Yanagawa, M. Go, Y. Hushimi, Dan-Sowkawa, Ikegami, T. and S. Odani for useful discussions. This work is partly supported by Grant for the Promotion of Niigata University Projects in 2000.
References [1] Altschul, S.F. and Gish, W., Local alignment statistics, Methods Enzymol., 266:450–480 1996. [2] Cech, T,R., Zaug, A.J., and Grabowski, P.K., In vitro splicing of the ribosomal RNA precursor of Tetrahymena, Cell 27:487–496, 1981. [3] De Saussure, F., Course de linguistique generale. Edition critique Rudolf Engeler, Harrasowitz, 1989. [4] Eigen, M. and Schuster, P., The hypercycle, Natur-wissenschaften, 54:541–565, 341–369, 1977. [5] Guerrier-Takada, C., Gardinner, K., Marsh, T., Pace, N., and Altman, S., The RNA moiety of ribonuclease P is the catalytic subunit of the enzyme, Cell, 35:849–857, 1983. [6] Haldane, J.B.S., Data needed for a blueprint of the first organism, The Origins of Prebiological Systems and of their Molecular Matrices, Academic Press, 11–15, 1965. [7] Hokari, S. and Ohnishi, K., The origin of 16S rRNA, as viewed from poly-tRNA theory, Nucleic Acids Symp. Ser., 39:159–160, 1998. [8] Mount, D.W., Bioinformatics. Sequence and Genome Analysis, Cold Spring Harbor Press, 2001. [9] Noller, H.F., Hoffarth V., and Zimniak L., Unusual resistance of peptidyl transferase to protein extraction procedures, Science, 256:1416–1419, 1992. [10] Ohnishi, K., A most primitive primordial gene as a building block of the genes for the adenylate kinase, F1-ATPase subunits, and core enzyme subunits, Nucleic Acids Symp. Ser., 17:127–130, 1986. [11] Ohnishi, K., Evolution from semi-tRNA to early protein-synthesizing RNA molecule, Endocytobiology V. 5th International Colloquium on Endocytobiology and Symbiosis, Tubingen University Press, 407–414, 1993. [12] Ohnishi, K., Evolutionary meanings of the primary and secondary structure of the “ur-RNA”, Symmetries in Science IV, Plenum Press, 147–176, 1990. [13] Ohnishi, K., Molecular evolution of ion channels in central nervous system: Possible primitive mRNA segments conserved in glutamate and acetylcholin receptor genes, Ann. N.Y. Acad. Sci. 707:524–528, 1993. [14] Ohnishi, K., Non-paradoxical origin of the genetic codes [I], (In Japanese), Seibutsu Kagaku (Biological Science) 47:10–23, 1995. [15] Ohnishi, K., Phosphoglycerate-transporter protein B as a most primitive protein predicted by the poly-tRNA theory, Genome Informatics, 9:393–394, 1998.
Origin of Most Primitive mRNAs and Genetic Codes via Interactions
81
[16] Ohnishi, K., Poly-tRNA structures as early protein-synthesizing apparatus: Evolution from prototRNA to most primitive mRNAs, Viva Origino, 23:237–254, 1995. [17] Ohnishi, K., Poly-tRNA theory on the origin and evolution of mRNA and genetic codes: Evolution from tandem tRNA repeats to premitive mRNA, Genome Informatics 4:325–331, 1993. [18] Ohnishi,K., Hokari, S., Shutou, H., Kanbe, D., Ohshima, M., Sawamura, H., Furuichi, N., Shibuki, K., and Goda,M., Neural-network-like bio-machinogenesis via semeiogenesis : A unified theory on the origin of genetic codes and other semeiotic systems, Viva Origino, 30(2):63–78, 2002. (See URL site in [26]) [19] Ohnishi, K., Hokari, S., and Yanagawa, H., Origin and evolution of early peptide-synthesizing biomachines by means of hierarchical sociogenesis of intracellular primitive tRNA-riboorganisms, Proc. 4th Int. Symp. on Artificial Life and Robotics, 344–349, 1999. [20] Ohnishi, K. and Shutou, H., Autopoietic active selection by cognitive self-systems, Proc. 6th International Symposium on Artificial Life and Robotics, Shuubunsha, 490–495, 2001. [21] Ohnishi, K., Suzuki, T., and Shutou, H., An E. coli pseudo-tRNA gene cluster homologous to the tRNA-gene cluster in the Bacillus subtilis trrnD operon, Genome Informatics, 11:325–326, 2000. (URL site in [27]) [22] Ohnishi, K., Tanaka, H., and Yanagawa, H., The origin of DNA-binding domains, as viewed from poly-tRNA theory, Nucleic Acids Symp. Ser., 39:251–252, 1998. [23] Ohno, S., Original domain for the serum albumin family arose from repeated sequences, Proc. Natl. Acad. Sci. USA, 78:7657–7661, 1981. [24] Ohno, S. and Matsunaga, T., The 48-base-long primordial building block of immunoglobulin light-chain variable regions, Proc. Natl. Acad. Sci. USA, 79:2338–2341, 1982. [25] Sebeok, T.A., Perspectives in Zoosemiotics, Mouton, 1972. [26] http://www.origin-life.gr.jp/3002/3002063/3002063.html (URL site for [18]) [27] http://www.jsbi.org/journal/GIW00/GIW00P045.pdf (URL site for [21])
