No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
The primary structure of tRNAPhe from Bacillus stearothermophilus
Before the evolutionary relations of tRNAs can be examined, those specific sites in tRNA molecular structure that are of possible importance in a particular function need examination, in order to distinguish functional and evolutionary influences. Among the activities in which a given nucleotide may be especially significant are included the maintenance of a secondary or tertiary structure, recognition by the ligase, and attachment to the ribosome, as well as those codon—anticodon interactions that have already received attention (Chapter 7; Section 7.2).
If not among all the universal components of organisms, then among the nucleic acids at least, the transfer RNAs are unique in having both a relatively constant molecular size and configuration. These consistencies are even more noteworthy because the class of substances (amino acids) they transport is totally lacking in both traits. Thus contrary to what might be expected from physicochemical considerations, no molecular relationships are in evidence between the carrier and transported compounds, a point that merits serious attention as the discussion proceeds. This chapter principally compares the characteristics of the base sequences of over 95 different species of tRNA. These results are then analyzed from an evolutionary standpoint in Chapter 8 in order to seek evidence relevant to the origin of life.
Transfer ribonucleic acid (tRNA) is the term used to describe a group of low molecular weight RNA molecules that play a vital part in protein synthesis Transfer RNAs were assigned the role of adaptor molecules by Hoagland et al. When, in the early 1960s, techniques for determining the nucleotide sequence of RNA molecules were developed, the tRNAs—the smallest species of RNA present in the cell—attracted further attention as relatively simple, naturally occurring nucleic acids that, it was hoped, would help in the elucidation of the general principles governing nucleic acid–protein interactions. The study of the biological role of tRNA in the past decade has not only given a fuller understanding of the mechanism of protein synthesis but has also shown that the role of tRNA in the cell is not confined to that of an adaptor molecule. However, attempts to relate the structure of tRNA even to its principal known function as an adaptor have met with mixed success. One of the major difficulties in relating tRNA structure to its function was that, until recently, the three-dimensional structure of tRNA was not known in detail. Physical and chemical studies of tRNAs in solution were limited to providing structural data that was either at a low resolution, giving only an indication of general shape, or resolved to the nucleotide level but restricted to selected regions of the molecule.
Bacteria are classified into at least 3 groups depending on their cardinal temperatures. Bacteria with growth spans from 40°C to 80°C are classified as thermophiles. It is the physiology of this group which is the subject of this review which has nearly 400 references. Topics covered in some detail include a description of the properties of some thermophiles, the properties of macromolecules and structures in thermophiles, the metabolism and enzymes of thermophiles, and thermoadaptation, and the genetic transfer of thermophilicity.
Publisher Summary
Transfer ribonucleic acid (tRNA) is the term used to describe a group of low molecular weight RNA molecules that play a vital part in protein synthesis Transfer RNAs were assigned the role of adaptor molecules by Hoagland et al. When, in the early 1960s, techniques for determining the nucleotide sequence of RNA molecules were developed, the tRNAs—the smallest species of RNA present in the cell—attracted further attention as relatively simple, naturally occurring nucleic acids that, it was hoped, would help in the elucidation of the general principles governing nucleic acid–protein interactions. The study of the biological role of tRNA in the past decade has not only given a fuller understanding of the mechanism of protein synthesis but has also shown that the role of tRNA in the cell is not confined to that of an adaptor molecule. However, attempts to relate the structure of tRNA even to its principal known function as an adaptor have met with mixed success. One of the major difficulties in relating tRNA structure to its function was that, until recently, the three-dimensional structure of tRNA was not known in detail. Physical and chemical studies of tRNAs in solution were limited to providing structural data that was either at a low resolution, giving only an indication of general shape, or resolved to the nucleotide level but restricted to selected regions of the molecule.
The nucleotide sequence of tRNATYr from B. stearothermophilus has been determined: pG-G-A-G-G-G-G-S4U-A-G-C-G-A-A-G-U-Gm-G-C-U-A-A-m1A-C-G-C-G-G-C-G-G-A-C-U-Q-U-A-ms2i6A-A-Ψ-C-C-G-C-U-C-C-C-U-U-U-G-G-C-G-G-T-Ψ-C-G-A-A-U-C-C-G-U-C-C-C-C-C-U-C-C-A-C-C-AOH. A combination of classical fingerprinting methods, partial nuclease P1 digestion and two-dimensional homochromatography and a rapid “read off” sequencing gel technique were used to establish the
complete nucleotide sequence.
This chapter reviews that organelles contain a complete apparatus for the synthesis of proteins. It is clear that, at least in most cases, organelle tRNAs are transcripts of the organelle genome. The limited number of tRNAs present in organelles in contrast to prokaryotes and the eukaryotic cytoplasm-poses an interesting problem. The chapter discusses the two hypotheses: (1) That organelles evolved via endosymbiosis, and (2) that organelles evolved through invagination and compartmentalization of function. It suggests that either certain codons are not utilized by organelles or that, because of “wobble” at the third position of the anticodon, or a few isoacceptors-all 61 “sense” codons may be translated. As codon recognition by aminoacyl-tRNAs is determined, solely by the tRNA component, these elements of the protein synthetic machinery are in large part, responsible for maintaining the fidelity of the genetic code. The aim of this chapter is to review the development and progress of organelle tRNA research and the aminoacyl-tRNA synthetases.
The nucleotide sequence of tRNA(Phe) from Bacillussubtilis W 23 has been determined using (32)P labeled tRNA. This is the second B. subtilis tRNA so far reported. The nucleotide sequence was found to be pG-G-C-U-C-G-G-U-A-G-C-U-C-A-G-U-D-G-G-D-A-G-A-G-C-A-A-C-G-G-A-C-U-Gm-A-A- ms(2)i(6)A-A-psi-C-C-G-U-G-U-m(7)G-U-C-G-G-C-G-G-T-psi- C-G-A-U-U-C-C-G-U-C-C-C-G-A-G-C-C-A-C-C-A(OH).Images
Highly purified tRNAPhe from rabbit liver, calf liver and bovine liver were completely digested with pancreatic ribonuclease and ribonuclease T1. The oligonucleotides were separated and identified. The tRNAPhe from rabbit liver and calf liver were partially cleaved with ribonuclease T1 or by action of lead acetate. We describe the analyses of the large fragments and the derivation of the primary structure of these mammalian tRNAsPhe.
The primary sequence of phenylalanine tRNA (tRNAphe) from the chloroplasts of Euglena gracilis has been determined: pG-C-U-G-G-G-A-U-A-G-C-U-C-A-G-D-U-G-Gm-U(U)-A-G-A-G-C-G-G-A-G-G-A-C-U-G-A-A-A-A-PSI-C-C-U-U-G-U-m7G-Py-C-A-C-C-A-G-T-psi-C-A-A-A-U-C-U-G-G-U-U-C-C-U-A-G-C-A-C-C-A. This represents the first nucleotide sequence determined for an organelle tRNA. As do all other tRNAPhes thus far sequenced, chloroplastic tRNAPhe contains 76 nucleotides. Both in the nature of its modified nucleotides and in its sequence (although the sequence of all known tRNAPhes is quite similar), chloroplastic tRNAPhe more closely resembles procaryotic tRNAPhes than it resembles those from the cytoplasm of eucaryotes. There are eight positions in the tRNAPhe molecule where nucleotides are invariant in procaryotes but differ from invariant nucleotides in eucaryotes; at five of these positions, chloroplastic tRNAPhe is similar to procaryotes. The possible evolutionary significance of this intermediate type of structure is discussed.
Among amino acid codons that require a third-position pyrimidine, there is a significant bias favoring the use of cytidine
over uracil in MS2 phage RNA. This could arise from selection against wobble pairing in the interaction of transfer RNA and
messenger RNA. Among amino acid codons with fourfold degeneracy, there is a bias favoring pyrimidines over purines.
A tRNA(Val) (GAC) gene is located in opposite orientation 552 nucleotides (nt) down-stream of the cytochrome oxidase subunit III (coxIII) gene in sunflower mitochondria. The comparison with the homologous chloroplast DNA revealed that the tRNA(Val) gene is part of a 417 nucleotides DNA insertion of chloroplast origin in the mitochondrial genome. No tRNA(Val) is encoded in monocot mitochondrial DNA (mtDNA), whereas two tRNA(Val) species are coded for by potato mtDNA. The mitochondrial genomes of different plant species thus seem to encode unique sets of tRNAs and must thus be competent in importing the missing differing sets of tRNAs.
A large number of post-transcriptional base modifications in transfer RNAs have been described (Sprinzl, M., Horn, C., Brown, M., Ioudovitch, A., and Steinberg, S. (1998) Nucleic Acids Res. 26, 148-153). These modifications enhance and expand tRNA function to increase cell viability. The intermediates and genes essential for base modifications in many instances remain unclear. An example is wyebutosine (yW), a fluorescent tricyclic modification of an invariant guanosine situated on the 3'-side of the tRNA(Phe) anticodon. Although biosynthesis of yW involves several reaction steps, only a single pathway-specific enzyme has been identified (Kalhor, H. R., Penjwini, M., and Clarke, S. (2005) Biochem. Biophys. Res. Commun. 334, 433-440). We used comparative genomics analysis to identify a cluster of orthologous groups (COG0731) of wyosine family biosynthetic proteins. Gene knock-out and complementation studies in Saccharomyces cerevisiae established a role for YPL207w, a COG0731 ortholog that encodes an 810-amino acid polypeptide. Further analysis showed the accumulation of N(1)-methylguanosine (m(1)G(37)) in tRNA from cells bearing a YPL207w deletion. A similar lack of wyosine base and build-up of m(1)G(37) is seen in certain mammalian tumor cell lines. We proposed that the 810-amino acid COG0731 polypeptide participates in converting tRNA(Phe)-m(1)G(37) to tRNA(Phe)-yW.
The nucleotide sequence of Myaoplasma sp. (Kid) phenylalanine tRNA was determined to be pG-G-U-C-G-U-G-U-A-G-U-C-A-G-U-C-G-D-A-G-A-G-C-A-G-C-A-G-A-C-U-G-A-A-m1G-C-G-U-φ-C-G-G-C-G-G-U-C-A-A-U-C-C-GU-C-C-A-C-G-A-C-C-A-C-C-AOH. It is characterized by the absence of ribothymidine and
the presence of only few modified nucleotides.
Purified yeast phenylalanyl-tRNA synthetase can aminoacylate (yeast) tRNAPhe, (wheat) tRNAPhe, and (Escherichia coli) tRNA1Val (1, 2). We now report that this synthetase can also aminoacylate (E. coli) tRNAPhe and (E. coli) tRNA1Ala. Highly purified (E. coli) tRNAPhe is heterologously aminoacylated to approximately 90% of the extent achieved with the homologous enzyme (crude E. coli phenylalanyl-tRNA synthetase). Pure (E. coli) tRNA1Ala (the major species) is heterologously aminoacylated to 70% of the extent achieved with the homologous synthetase (crude E. coli alanyl-tRNA synthetase).
(E. coli) tRNAPhe is the fourth purified transfer RNA of known sequence to be shown to be an acceptable substrate for purified yeast phenylalanyl-tRNA synthetase. A comparison of these sequences shows that only one region is extremely similar in all four tRNAs. This region is located adjacent to the dihydrouridine loop, and consists of the nucleotides [Formula: see text] We conclude that this is the synthetase recognition site for yeast phenylalanyl-tRNA synthetase.
This conclusion is further supported by partial fragment analysis of (E. coli) tRNA1Ala.
Highly purified tRNATrp from brewer's yeast prepared by countercurrent distribution followed by column chromatography on BD cellulose has been completely digested with pancreatic ribonuclease and T1 ribonuclease. The separation and the identification of the products are described. Analyses indicate that this tRNA is composed of 75 nucleotide residues including 17 minor nucleotides. This tRNA contains the G-T-Ψ-C sequence common to all tRNAs of known structure concerned in protein biosynthesis. The 5′terminal sequence is pG-A-A-G-Cp and the 3′terminal end is U-U-U-C-A-C-C-A. Only one endonucleotidic sequence, Cm-C-A, is able to form hydrogen bonds with the tryptophan codon U-G-G. The primary structure of this tRNA has been determined after partial digestion with pancreatic ribonuclease as stated in the following paper.
Evolution by gene duplication and subsequent divergence is indicated by similarities common to 43 different transfer RNAs. Pairwise comparisons of these tRNAs reveal additional similarity, greatest for certain pairs of tRNAs for the same amino acid in the same organism, and also occurring in certain pairs of tRNAs for different amino acids in the same organism. Although tRNAs functionally interact with several other molecules, there have been surprisingly few restrictions on the divergence of their primary structures. This divergence has proceeded so far that clear phylogenetic separations are absent in most cases: it it impossible to construct a coherent phylogeny for most of the 43. Selection and stochastic processes have both been active in the evolution of tRNA. Selection has favored moderate change more than expected and has reduced radical change below that expected from stochastic processes alone. Two obvious effects of selection are nine invariant loci, another five that are always purines and five others that are always pyrimidines, in the tRNAs involved in protein synthesis. In addition to these constraints in the primary nucleotide sequence, the method of “identical site equivalents”, introduced here, demonstrates that further constraints exist equivalent to about 12 additional invariant loci. These “invisible” restraints reflect disperse chemical forces maintaining the tertiary structure and reducing evolutionary divergence to an extent quantitatively comparable to that of the nine observable invariant loci. The average divergence (49·4%) for pairs of tRNAs for different amino acids involved in protein synthesis represents an equilibrium between natural selection and stochastic processes. These tRNAs have had time to diverge nearly to the 75% maximum expected from stochastic process alone; this is shown by comparing the two glycine tRNAs involved in peptidoglycan synthesis with tRNAs for different amino acids participating in polypeptide synthesis. The rates of nucleotide replacements in genes coding for the tRNAs and the cytochromes c are about the same: 2 × 10 −10 replacements per nucleotide site per year.
A method for the preparative stepwise degradation of polyribonucleotides has resulted from the development of a new reaction sequence employing the periodate oxidation-β elimination-dephosphorylati on procedure. The method consists of a series of degradation cycles each of which involves (i) the periodate oxidation of the 3′-terminal of the polynucleotide, (ii) the removal of the excess periodate by reaction with rhamnose, (iii) the cleavage of the oxidized terminal nucleoside from the chain and the enzymatic dephosphorylation of the polynucleotide product in one step, and (iv) the separation of the polynucleotide from the reactants and other products. The reaction sequence has a number of advantages over previously used procedures in which the periodate oxidation and the β-elimination steps were usually combined in the one reaction mixture and the dephosphorylation was effected in a separate reaction. The conditions for quantitative reaction in each step of the degradation cycle have been perfected by using mononucleotides, dinucleoside phosphates, and adenosine tetra-, penta-, hexa-, and heptanucleotides as substrates, and the degradation procedure has been tested by carrying out nine cycles on a dodecanucleotide, T-ψ-C-A-A-U-U-C-C-C-C-G, obtained from a ribonuclease T1 digest of tRNAAsp. Polynucleotide sequence information may also be obtained during the stepwise degradation by isolating and analyzing the nucleoside fragment produced in each cycle. Thus, from the analyses of nine degradative cycles, the 3′-terminal nonadecanucleotide from yeast initiator tRNA was shown to be m1A-A-A-C-C-G-A*-G*-C-G-G-C-G-C-U-A-C-C-A, a result that eliminates the sequence ambiguity in the stem region of the published sequence of this tRNA. The reaction sequence used in the degradative procedure also allows for the removal of more than one nucleotide from a polynucleotide chain in a single reaction mixture, and this technique has been demonstrated by the conversion of pA-A5-A to A-A2-A in 80% yield.
Thirty-six transfer RNAs were aligned on the basis of the “clover-leaf” model and compared for homology. All these tRNAs have identical residues at eleven loci. Additional homology may occur in pairs of tRNAs for different amino acids in the same organism. This type of homology indicates that tRNAs for 2 different amino acids may evolve from a common ancestor by gene duplication. Such an evolutionary separation must include changes in the anticodon and in the recognition site for the activating enzyme.
It has been found that purified phenylalanyl- and valyl-tRNA synthetases from yeast catalyse the mischarging of numerous heterologous Escherichia coli and even of homologous yeast tRNAs when special aminoacylation conditions are used. For instance E. coli tRNAAla, tRNAVal, tRNALys and yeast tRNAAla, tRNAVal can be incorrectly aminoacylated by yeast phenylalanyl-tRNA synthetase, whereas E. coli tRNAAla, tRNAMet, tRNAIle, tRNAThr and yeast tRNAAla, tRNAPhe can be mischarged by yeast valyl-tRNA synthetase. The production of errors is highly dependant on the experimental conditions used, as their number and their level is increased when the Mg2+/ATP ratio or the enzyme concentration in the medium is increased or when the reactions are performed in the presence of dimethylsulfoxide. In the case of phenylalanyl-tRNA synthetase, conditions can be found where practically all E. coli tRNAs are wrongly aminoacylated, although at different levels. On the other hand, in the case of homologous systems some errors can only be detected when purified tRNAs are used, thus suggesting, when total tRNA is used, a competition between the cognate and non-cognate tRNAs which minimises the mischarging.
The comparison of the sequences of the cognate and non cognate tRNAs which are aminoacylated either by the yeast phenylalanyl- or valyl-tRNA synthetase led us to ascribe some importance to several regions inside of these tRNAs, for instance (a) the dihydrouracil arm, (b) the terminal part of the amino-acid acceptor stem and (c) the extra-loop. These regions should be essentially necessary for the establisment of the correct tri-dimensional conformation necessary for the recognition by the aminoacyl-tRNA synthetases.
Calf and beef liver tRNAPhe were obtained by fractionation of calf and beef tRNA on BD-cellulose columns and selected regions were further purified by hydroxylapatite chromatography. The tRNAPhe was then incubated at pH 2.9 to liberate the strongly fluorescent base of this tRNA, which was purified by silica gel thin-layer chromatography. The structure of the new base obtained was determined by comparing its absorption, fluorescent, and high-resolution mass spectra to that of the similar base of tRNAyeastPhe (originally termed "Y" base). Results of this analysis indicate that the new base of bovine liver tRNAPhe ("peroxy base") differs from the one from yeast only by the presence of a unique hydroperoxide group on the β carbon of the side chain. To exclude the possibility that the peroxy base is an artifact formed during the isolation procedure, the following studies were performed: (1) incubation of synthetic d,l base of yeast and yeast tRNAPhe with a postribosomal supernatant fraction from liver; (2) addition of yeast tRNA to the liver homogenate and subsequent isolation of the tRNAPhe; (3) isolation of calf liver tRNAPhe in the presence of sodium azide. The fluorescent bases isolated from these experiments remained unchanged. Therefore, the hydroperoxide moiety in the side chain of the peroxy base is not merely an artifact of isolation.
It is shown that yeast tRNAPhe, chemically coupled by its oxidized 3′CpCpA end behaves exactly as free tRNAPhe in its ability to form a specific complex with tRNA2Glu having a complementary anticodon. The results support models of tRNA in which the 3′CpCpAOH end and the anticodon are not closely associated in the tertiary structure, and provide a convenient tool of general use to characterize others pairs of tRNA having complementary anticodons, as well as for highly selective purification of certain tRNA species.
Yeast phenylalanyl-tRNA synthetase (PRS) can aminoacylate highly purified phenylalanine tRNAs of E. coli yeast and wheat and E. coli tRNA1Val, tRNA2AVal, tRNA2BVal, tRNAMMet, tRNAIle, tRNA1Ala, tRNA2Ala, and tRNALys. The sequences of the first eight tRNAs have been compared, and based, in part, upon their unique sequence similarities, it was proposed that two regions of the tRNA molecule are primarily involved in the PRS recognition site. The double-stranded region adjacent to the dihydrouridine loop and adenosine as the fourth residue from the 3′ end were proposed to be directly involved in the recognition of a tRNA by this particular synthetase (B. Roe and B. Dudock, Biochem. Biophys. Res. Commun. 49, 399 (1972)). In this paper we report: (a) the purification of PRS essentially to homogeneity; (b) the pH optima for the PRS aminoacylation reactions; and (c) the kinetics of aminoacylation with PRS for each of the 11 tRNAs mentioned above at pH 6.0 (cacodylate buffer) and pH 8.2 (Tris buffer). This study shows that these tRNAs differ only slightly (tenfold) in their Km's for the PRS aminoacylation reaction but differ considerably (200-fold) in their Vmax values. At pH 8.2 (Tris buffer), these 11 tRNAs fall into three distinct classes characterized by order of magnitude differences in their Vmax for the PRS aminoacylation reaction. These three classes are a fast class (Vmax = 0.4-0.5), an intermediate class (Vmax = 0.01-0.09) and a slow class (Vmax = 0.003-0.008). Each of these classes can be correlated primarily with two structural features of the tRNA molecule, specifically the size of the dihydrouridine loop (8 or 9 nucleotides) and the presence of an N2-methylguanine or an unmodified guanine at position 10 from the 5′ end. It is further shown that E. coli tRNA2Ala, which can be aminoacylated by PRS, can act as a competitive inhibitor of the homologous aminoacylation of yeast tRNAPhe by PRS. Furthermore, under identical conditions, E. coli tRNA3Gly, which cannot be aminoacylated by PRS, cannot act as a competitive inhibitor of the homologous aminoacylation of yeast tRNAPhe. E. coli tRNA3Gly has the "correct" nucleotides in the double-stranded region adjacent to the dihydrouridine loop (diHU stem) but lacks adenosine at the fourth position from the 3′ end (it contains uridine at this position). This lack of competitive inhibition suggests that in an intact tRNA specific nucleotides in both the diHU stem region and adenosine as the fourth residue from the 3′ end are required for binding to PRS. The model of the PRS recognition site is discussed as well as its applicability to other synthetases.
tRNAphe (anticodon GAA) and tRNAglu (anticodon presumably UUC) have been found to form a complex with an association constant of about 5×105M−1 at 0°C. This binding is much stronger than the binding of trinucleotide UUC to tRNAphe but has a weaker temperature dependence. This suggests that the anticodon regions of tRNA have similar and complementary structures, such as Watson and Crick helices.
The complete nucleotide sequence of wheat germ phenylalanine transfer RNA (tRNA(Phe)) is presented. This RNA, which is an acceptable substrate for yeast phenylalanine tRNA synthetase, has a structure very similar to that of yeast tRNA(Phe). Only 16 of the 76 nucleotides are different, and all but two of the nucleotide changes are located in regions that are doublestranded in the cloverleaf model. The two changes in single-stranded regions involve minor modifications of the same nucleotide. The dihydrouridine loop and its supporting stem are completely free of nucleotide changes.
A method is described for the two-dimensional fractionation of ribonuclease digests of 32P-labelled RNA. High-voltage ionophoresis is used in both dimensions. The first is on cellulose acetate at pH 3·5, the second on DEAE-paper at an acid pH. The method is capable of resolving the di- and tri- and most of the tetra-nucleotides in digests prepared by the action of ribonuclease T1 or pancreatic ribonuclease. It has been applied to the 16 s and 23 s components of ribosomal RNA which show significant quantitative differences, and to sRNA from Escherichia coli and from yeast. A method is described for the determination of the sequence of a nucleotide by partial digestion with spleen phosphodiesterase.
- I L Grosjean
- C Takada
- J Petre
Grosjean, IL, Takada, C. and Petre, J. (1973) Biochem.
Biophys. Res. Commun. 53, 882.
- J Eisinger
Eisinger, J. (1971) Biochem. Biophys. Res. Commun.
43, 854.
- D Kern
- R Giege
- J P Ebel
Kern, D., Giege, R. and Ebel, J. P. (1972) Eur. J.
Biochem. 31, 148.
Takata-Gucrrier,
C., Grosjean, I-I., Dirheimer, G. and
Keith, G., in preparation.
- T H Jukes
- R Holmquist
Jukes, T. H. and Holmquist,
R. (1972) Biochem.
Biophys. Rcs. Commun 49, 212.
- D Kern
- R Giege
- J P Ebel
- C Takata-Gucrrier
- I-I Grosjean
- G Dirheimer
- G Keith
Kern, D., Giege, R. and Ebel, J. P. (1972) Eur. J.
Biochem. 31, 148.
Takata-Gucrrier,
C., Grosjean, I-I., Dirheimer, G. and
Keith, G., in preparation.
- B G Barrell
- Sanger
Barrell, B. G. and Sanger, 1;. (1969) FEBS Lett. 3, 275.
Kimball, M. E., Szeto, K. S. and Soil, D. (1974)
Nucleic Acids Research 1, 172 1.
- U L Rajbhandary
- S H Chang
- A Stuart
- R D Faulkner
- R M Hoskinson
- Ii G Khorana
RajBhandary,
U. L., Chang, S. H., Stuart, A.,
Faulkner,
R. D., Hoskinson,
R. M. and Khorana, II. G.
(1967) Proc. Natl. Acad. Sci. U.S. 57, 751.
- T H Jukes
- R Holmquist
Jukes, T. H. and Holmquist,
R. (1972) Biochem.