Purine and pyrimidine composition in 5S rRNA and its mutational significance

Purine and pyrimidine composition in 5S rRNA and its mutational significance

Sandra Maria Rodrigues Subacius¹ and Wilton de Oliveira Bussab²

¹ Departamento de Biotecnologia, Faculdade Estadual de Engenharia Química de Lorena, Rodovia Itajubá Lorena, km 74.5, Caixa Postal: 116, 12600-000 Lorena, SP, Brasil. Send correspondence to S.M.R.S. ² Instituto de Matemática and Estatística, Universidade de São Paulo, São Paulo, SP, Brasil.

ABSTRACT

Variations observed in 5S rRNA base compositions are almost entirely due to fixation of point mutations. As a consequence, 5S rRNA size has remained relatively constant during evolution and, therefore, dependencies among the four bases can be predicted. In order to characterize the nature and to determine the degree of such dependencies, correlation analysis followed by principal component factorial analysis was conducted on a large sample of 5S rRNA sequences. The results show that the purine and pyrimidine contents tend to remain constant, so that A + G = Kpur and C + U = Kpyr. The composition of the four bases expressed now by Kpur/Kpyr relationships is also constant (Ks). These relationships imply that the behavior of the mutations in the variable sites of the molecule follows rules imposed by the chemical nature of the bases involved. Consequently, transition mutations would be more favored than substitutions in transversion sites and also than insertion-deletion (rare in 5S rRNAs), since transitions would not significantly alter the values of the Ks-index.

INTRODUCTION

5S ribosomal RNA (5S rRNA) has been structurally and functionally well conserved throughout evolution. Nevertheless, systematic variations are observed in base composition. These variations are almost entirely due to fixation of point mutations (transitions or transversions). As a consequence, 5S rRNA has remained relatively constant in size during evolution and, therefore, dependencies among the four bases can be predicted. Recent research has focused on the pattern of nucleotide substitution at specific sites of the molecule as evidenced through alignment. It has been found that transitions are the most frequent fixed mutations followed by transversions, whereas insertions-deletions are quite uncommon (Sankoff et al., 1976; see also Vogel and Kopun, 1977). With very few exceptions insertions-deletions in general involve few nucleotides, altering only discretely the length of the primary chain among the universal residues, mainly in eubacteria (Delihas et al., 1984; Erdmann and Wolters, 1986; Specht et al., 1990, 1991). Studies of systematic variations in the composition of the four bases, considering the molecule "as a whole", are scarce (Vogel and Kopun, 1977; Gomes et al., 1985; Digweed et al., 1986; Guo-rong et al., 1988; Guimarães and Erdmann, 1992; Subacius, 1994). This article presents the results obtained from the correlation analysis and multiple factor analysis among the variables expressed as observed frequencies (Fobs) of adenine (A), cytosine (C), guanine (G) and uracil (U) in order to understand better the nature and degree of their interdependency.

MATERIAL AND METHODS

The sample under study consists of 494 sequences of bacteria, achaebacteria, protist, fungi, animal and plant 5S rRNAs obtained from the Berlin RNA DataBank (BRDb; updated in December, 1987). The base composition identified by the A, C, G and U variables was submitted to statistical analysis after transformation of the variables using arcsin Öx (where x = relative frequency). A simple correlation suggests a relationship between the transformed variables A'-G' (r = -0.7511) and C'-U' (r = -0.7566). This was tested by multiple factor analysis using SPSSPC⁺ software (SPSS Inc., Chicago, USA). It was found that two factors are sufficient to explain more than 80% of the total variability. The highest factor loads with opposite signs were observed for the A'-G' and C'-U' variables (Table I). The linear transformation coefficient extracted from the multiple factor analysis shows that factors I and II are dependent on the chemical nature of the bases. This means that the compositions of the A and G purines (pur) and of the C and U pyrimidines (pyr) do not vary at random in the molecule, in our case, the sense chain of the 5S-gene. Rather, an increase (or decrease) in the amount of purine (or pyrimidine) implies a similar decrease (or increase) of the other purine (or pyrimidine), so that A + G = Kpur and C + U = Kpyr. That being so, a new relation between the bases is obtained:

Table I -
Correlation matrix among A, C, G and U variables and factor loads. The highest loads are observed between pyr (C and U) for factor 1 and pur (A and G) for factor 2.

A’		C’	G’	U’	factor 1	factor 2
A’	0	-	-	-	-0,2219	0,9161
C’	-0,5357	0	-	-	0,8907	-0,2656
G’	-0,7511	0,4053	0	-	0,3105	-0,8777
U’	0,3604	-0,7566	-0,5959	0	-0,9081	0,2591

In most species analyzed, the Ks constant tends to remain close to 1.0, indicating that the pyrimidine and the purine contents are balanced (or in equilibrium). Ks assumes an average sample value of 1.048 with a difference of 0.048 in favor of the purines, owing to a slight excess of G in the 5S rRNA data set.

RESULTS AND DISCUSSION

Variations in 5S rRNA composition are exclusively due to genetic mutation, since 5S is not extensively modified after transcription (Erdmann 1976; Korn 1982). These mutations can be of two types: point mutations (transitions or transversions) or insertion-deletion events. A point mutation can, a priori, affect any position of the 5S gene. If these mutations are of the transition type, they do not alter the Ks-values in accordance with the A + G/C + U relation. In contrast, transversion events can displace the Ks-values depending upon the direction of the mutations: pur Þ pyr or vice versa. The change in magnitude of the Ks-values, due to transversions, depends on the total number of sites involved in transversion mutation events and the degree to which there is an excess (or deficiency) of transversion from pur to pyr versus pyr to pur.

Only some of the mutations that occur in the gene will be fixed on the molecule. Selection will prevent fixation of substitutions that impair 5S rRNA function. Thus, Ks reflects in fact the fraction of the mutations that were successful in the variable sites of the molecule. These mutations, in most species, maintain an equilibrium between the pur and pyr contents, so that Ks-values oscillate near the average sample value of 1.048 ± 0.05 (1 standard deviation). This behavior might be explained by the preferential fixation of transition mutations, especially in unpaired regions of the 5S rRNA, and also of the compensatory transversion mutation involving substitution of nucleotide pairs in opposite directions. In spite of the slow mutation rates of the 5S rRNAs, in the order of 1.0-1.8 x 10^-10 mutations/nucleotide/year (Hori et al., 1977; DeWachter et al., 1985), double substitutions would be mainly fixed in paired-regions of the molecule owing to the structural and evolutionary dependencies between the complementary bases. As a consequence the Ks-values of the 5S rRNA tend to equilibrium.

In 90% of the species analyzed, the Ks-values fluctuated around the average sample value by ± 0.10, which corresponds to two standard deviation units. A variation of 2s in the Ks is equivalent to a maximum number of 4 or 5 uncompensated transversions fixed at the variable sites of the molecule, apparently transparent to the structural restrictions. Considering the small molecule size and its high evolutionary conservatism, this number of uncompensated transversion is high. Regardless of this fact, more extreme Ks-values were observed (Table II) especially among the prokaryotes. Ks-values more than two deviation units, for instance, characterized the 5S rRNA of the methanogen archaebacteria belonging to the methanobacter and methomicrobium groups and only two species of Gram-negative eubacteria (EUB G-), Isosphaera pallida (planctomyces) and Prochloron didemni (cyanobacteria). Such behavior reveals the fixation of strongly unbalanced substitutions in the 5S of these species in transversion sites having the pur Þ pyr direction (Table II). On the other hand, the variations of the Ks-values followed a distinct pattern among the Gram-positive eubacteria (EUB G+), with a prevalent fixation of transversions that enriched the molecule with the purinic bases, A and G (see Table II). Significantly positive Ks-values were observed in some mycoplasm species (EUB G+,¯ G + C), in only one clostridiobacterium species (EUB G+, intermediary G + C) and in all actinomycete species (EUB G+, ­ G + C) from the soil. The values obtained for the 5S rRNA of both the mycoplasms and the actinomycetes suggest a preferential incorporation of A and G, respectively (see Table II) .

Table II -
Observed frequencies of A, C, G and U and Ks-values (A + G/C + U) in 5S rRNA of outlying groups. Note that the Ks-values, above and below the sample average (1.048 ± 0.10), are due to base composition strongly shifted from the most probable content neutrality (frequencies equal to 30 for each base). These Ks-values reflect an unbalance of the mutation rate of the purine in relation to pyrimidine, or vice-versa, in the 5S rRNA of these species. Abbreviations are described in the text except for MB = Mycobacteria, CLO = Clostridiobacteria, FIR = Firmibacteria, DINO = Dinophyta, PLA = Planctomyces and CYA = Cyanobacteria.

Group	Species	A	C	G	U	Ks-values
ARC-MBAC	Methanobacterium formicicum	22	31	39	33	0,967
	Methanobacterium thermoformicicum	17	29	38	40	0,872
	Methanobacterium thermoautotrophicum A1	18	30	41	38	0,912
	Methanobacterium thermoautotrophicum B	18	26	38	41	0,899
	Methanobacterium thermoautotrophicum A2	17	28	41	43	0,883
	Methanobrevibacter ruminantium	20	24	35	51	0,847
MMIC	Methanosarcina vacuolata	27	36	34	30	0,958
	Methanosarcina barkeri	27	40	33	30	0,920
	Methanosarcina acetivorans	25	36	32	30	0,922
EUB/PLAN	Isosphaera pallida	19	39	32	21	0,916
MB	Mycobacterium nonchromogenicum	33	36	33	13	1,208
	Mycobacterium flavescens	29	38	39	12	1,219
	Mycobacterium bovis BCG	27	39	37	13	1,150
	Mycobacterium bovis ATCC35737	29	40	34	13	1,131
ACT	Rhodococcus erythropollis	28	36	40	17	1,160
	Streptomyces griseus	26	31	42	21	1,159
	Pimelobacter simplex	27	36	43	14	1,224
	Micrococcus lysodeikticus A	24	36	43	16	1,160
	Micrococcus lysodeikticus B	25	33	44	18	1,184
	Curtobacterium citreum	27	31	45	19	1,224
	Corynebacterium aquaticum	26	32	42	20	1,161
	Cellulomonas biazotea	26	36	41	17	1,149
	Aureobacterium testaceum	26	33	42	19	1,163
	Arthrobacter luteus	28	35	42	15	1,221
	Arthrobacter globiformis	26	36	43	17	1,166
CLO	Enterococcus faecallis	28	26	38	25	1,151
MC	Acholeplasma modicum	31	25	30	21	1,169
	Mycoplasma gallisepticum	35	23	27	22	1,192
	Mycoplasma hyopneumoniae	40	19	22	27	1,162
FIR	Clostridium butyricum 1	29	23	37	27	1,165
	Clostridium butyricum 2	29	24	37	26	1,164
CYA	Prochlorom didemni	24	29	32	37	0,913
EUK/DINO	Crypthecodinium cohnii	22	34	42	24	1,187

A higher frequency of outlying Ks-values in the EUB G+ situated within the limits of the variation interval of the G + C genomic contents is expected, since the Ks index reflects bias in the behavior of the mutations.

ACKNOWLEDGMENTS

We are grateful to Dr. George E. Fox (University of Houston, Texas, USA) for his comments on this manuscript, to Dr. Romeu C. Guimarães (UFMG - Belo Horizonte, BR) and Dr. Thomas R. Fairchild (IG-USP) for their suggestions, to Dr. Volker A. Erdmann (Freire Universitat fur Biochemie-Dahlem, Germany) for the diskettes with the BRDb and to FAPESP and CNPq for their financial support. Publication supported by FAPESP.

RESUMO

A cadeia primária da molécula de RNAr 5S exibe variações na sua seqüência nucleotídica. Tais variações são devidas quase que exclusivamente a fixação de mutações pontuais (transversão e transição). Conseqüentemente, o tamanho da molécula tem permanecido constante e dependências entre as bases são previsíveis. Visando caracterizar a natureza e determinar o grau dessas dependências, procedemos a uma análise de correlação seguida por uma análise fatorial por componente principal da amostragem disponível no Berlin RNA Data Bank. Os resultados demonstram que o conteúdo das purinas e das pirimidinas tende a permanecer constante, de tal forma que A + G = Kpur e C + U = Kpir. Nesses termos, a composição das quatro bases dada agora pela razão Kpur/Kpyr é também igual a uma constante (Ks). O índice Ks implica que o comportamento das mutações nos sítios variáveis da molécula segue regras impostas pela natureza química das bases envolvidas, explicando assim porque as mutações do tipo transição são mais favorecidas na molécula em relação às transversões e às inserções-deleções (raras no 5S), visto que as primeiras não alteram o Ks.

(Received August 18, 1996)

• Delihas, N., Andersen, J. and Singhal, R.P. (1984). Structure, function and evolution of 5S ribossomal RNAs. Prog. Nucleic Acid Res. Mol. Biol. 31: 161-191.
• DeWachter, R., Huysmans, E. and Vandenberghe, A. (1985). 5S ribosomal RNA as a tool for studying evolution. In: Evolution of Prokaryotes (Schleifer, K.H. and Stackebrandt, E., eds). Academic Press, New York, pp. 115.
• Digweed, M., Pieler, T. and Erdmann, V.A. (1986). A comparative analysis of structural dynamics in 5S rRNA. In: Structure and Dynamic of RNA (van Knippenberg, P.H. and Hilbers, C.W., eds). Plenum Publishing Corporation, New York, pp. 205.
• Erdmann, V.A. (1976). Structure and function of 5S and 5,8S RNA. Prog. Nucleic Acid Res. Mol. Biol. 18: 45-90.
• Erdmann, V.A. and Wolters, J. (1986). Collection of published 5S, 5,8S and 4,5S ribosomal RNA sequences. Nucleic Acids Res. 14: r1-r57.
• Gomes, E., Guimarăes, R.C., Padovani, C.A. and Bloch, D.P. (1985). The universal dinucleotides asymmetry rule does not hold for the rRNA 5S. Arq. Biol. Tecnol. 28: 3 (Abstract).
• Guimarăes, R.C. and Erdmann, V.A. (1992). Evolution of adenine clustering in 5S rRNA. Endocyt. Cell. Res. 9: 13-45.
• Guo-rong Qi, Going-jie Cao, Peng Jiang, Xiao li Fing and Xiang-rong Gu (1988). Studies on the sites expressing evolutionary changes in the structure of eukaryotic 5S ribosomal RNA. J. Mol. Evol. 27: 336-340.
• Hori, H., Higo, K. and Osawa, S. (1977). The rates of evolution in some ribosomal components. J. Mol. Evol. 9: 191-201.
• Korn, L.J. (1982). Transcription of Xenopus 5S ribosomal RNA genes. Nature 295: 101-105.
• Sankoff, D., Cedergren, R.J. and Lapalme, G. (1976). Frequency of insertion-deletion, transversion and transition in the evolution of 5S ribosomal RNA. J. Mol. Evol. 7: 133-149.
• Specht, T., Wolters, J. and Erdmann, V.A. (1990). Compilation of 5S rRNA and 5S RNA gene sequence. Nucleic Acids Res. 18 (Suppl.): 2215-2230.
• Specht, T., Wolters, J. and Erdmann, V.A. (1991). Compilation of 5S rRNA and 5S gene sequence. Nucleic Acid Res. 19 (Suppl.): 2189-2191.
• Subacius, S.M.R. (1994). The evolution of 5S rRNA. Doctoral thesis, Universidade de Săo Paulo, SP.
• Vogel, F. and Kopun, M. (1977). Higher frequencies of transitions among point mutation. J. Mol. Evol. 9: 159-180.

Publication Dates

• Publication in this collection
  06 Jan 1999
• Date of issue
  June 1998

History

• Received
  18 Aug 1996