- Email: [email protected]
Towards a reconciliation of the introns early or late views: triosephosphate isomerase genes from insects
Biochimica et Biophysica Acta 1353 Ž1997. 131–136
Towards a reconciliation of the introns early or late views: triosephosphate isomerase genes from insects Michael G. Tyshenko, Virginia K. Walker
)
Department of Biology, Queen’s UniÕersity, Kingston, Ont. K7L 3N6, Canada Received 7 January 1997; revised 13 March 1997; accepted 7 April 1997
Abstract The gene encoding the glycolytic enzyme, triosephosphate isomerase ŽTPI; EC 5.3.1.1., is a favourite model for molecular evolutionists who either subscribe to the theory that introns co-evolved with the ancestral gene, the introns early view, or alternatively, that introns are more recent immigrants. The discovery of an intron in the TPI gene of Culex mosquitoes at a site which was predicted by proponents of the intron early school supported that theory. More recently, the discovery of additional intron sites in several eukaryotes was presented as evidence supporting the introns late school. We have found the ‘Culex intron’ in two closely related mosquitoes, but not in two more evolutionary primitive Dipterans, suggesting that, if it is an ‘ancient intron’, loss may be more frequent than that supposed by the intron late school. In addition, we have found that three introns punctuating the TPI gene from the Lepidopteran, Heliothis, appear to be ancestrally related and may be the result of transposable element insertion, 50–90 million years ago. It is argued that both opposing schools in the intron debate be reconciled — some introns may have been early and certainly others have arrived subsequent to the appearance of the TPI gene. q 1997 Elsevier Science B.V. Keywords: Triosephosphate isomerase; Intron; Transposable element; Molecular evolution
1. Introduction The triosephosphate isomerase ŽTPI. gene evolved before the prokaryotic–eukaryotic divergence a few billion years ago. According the introns late argument, the introns which punctuate the coding sequence have been inserted relatively recently w1–3x. In contrast, the introns early view w4,5x holds that the introns are as old as the gene itself and predicts that the ancestral TPI gene was interrupted by introns at sites corresponding to polypeptide modules w6x or microgenes w7x. During evolution many of these in)
Corresponding author. Fax: Žq1. 613-545-6617; E-mail: [email protected]
trons have been lost but organisms have been found that, together, represent every ancient intron site in the TPI gene, and most recently in the mosquito, Culex tarsalis w8x. The introns late school argues that because introns can be found at additional, non-predicted sites, and that the loss of introns, to accommodate the introns early view, would have to be more frequent than instances of intron gain, parsimony arguments favour their theory. We have further investigated our previous observations Žcited in w9x. and those of others w10x that a closely related mosquito has the ‘Culex TPI intron’. In addition, we have sequenced the gene in two more primitive Dipterans and a member of a sister order, the Lepidoptera. The TPI gene from these insects
0167-4781r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 7 - 4 7 8 1 Ž 9 7 . 0 0 0 6 5 - 1
132
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
does not have this intron, underscoring the flexibility of the insects’ genome and, incidentally, the sometimes inflexibility of the ‘early’ or ‘late’ proponents.
2. Materials and methods 2.1. DNA preparation Genomic DNA Ž0.5 m g. from decapitated adult Culex pipiens and Aedes togoi were used as templates for amplification of TPI gene ŽTpi. fragments in separate experiments by the polymerase chain reaction Ž PCR.. Oligonucleotides for the reactions using mosquito DNA were sense primers: 5X TGGAAGATGAAYGGNGAYAAUGC 3X and antisense primers, 5X GTCTGGCGTTGACAATCT 3X. Whole adult insects were used to prepare DNA from the tabanid, Chrysops Õittatus and the moth, Heliothis Õirescens, using the sense primer: 5X GTNGGNGGNACYTGGAA 3X and the antisense primers: Heliothis 5X CCNGTNCCYATNGCCCA 3X and Chrysops 5X GGCCANACNGGYTCUTA 3X. Additional 3X Heliothis sequence was obtained using a nondegenerate sense sequence 5X TGAAGGTCATTGCCTGTATC 3X and a degenerate antisense oligonucleotide 5X CCACCGACGARRAANCC 3X. Forty cycles of amplification were performed, 1 min each at 948, 588 and 728C, followed by a 728C 8-min extension using Taq polymerase and standard reaction conditions provided by the supplier ŽPromega.. Tpi fragments were subcloned by blunt end cloning with Klenow fragment and ligation into the Sma 1 site of pBS SK q ŽStratagene.. The A. merus TPI gene was isolated by high stringency screening of an A. merus EMBL 3A genomic library using C. tarsalis Tpi as probe. A positive phage was plaque purified and a single hybridizing 9.7 Xho 1 fragment was subcloned into the Xho 1 site of pBS SK q . Sequence was obtained using the degenerate mosquito PCR primers. All sequencing reactions were done at least three times on a 373A automatic DNA sequencer ŽApplied Biosystems. using dye terminators. 2.2. Analysis The mosquito Tpi intron and exon sequences were aligned using PCGene NALIGN with an open gap
cost of 80 and a unit gap cost of 10. The three Heliothis intron sequences were aligned using PCGene CLUSTAL analysis and a consensus sequence obtained. Characteristic intronic 5X and 3X dinucleotides ŽGT and AG, respectively. were omitted from the analysis and as a control, intronic sequences from other organisms were also analysed in a similar manner. Divergence rate is expressed as the number of substitutionsrsitermillion years and was calculated using the Jukes and Cantor model w11x which corrects for overlapping substitutions which would otherwise result in an underestimate of evolutionary distance. The calculated divergence rate was used to estimate the divergence times of the Heliothis introns by solving for time using the Jukes and Cantor model. In the analysis of the mosquito coding regions by pairwise alignment, no additional corrections were made for synonymous and non-synonymous changes and thus these calculations are approximations. Calculations of the rate of Tpi evolution in different orders did take in account those changes and have been previously described w12x.
3. Results and discussion Genomic DNAs from Culex pipiens, Aedes togoi and Chrysops Õittatus were used as templates for the amplification of TPI gene fragments by PCR. Amplified DNA of an appropriate size Ž Fig. 1. was subcloned and sequenced w8x to confirm identity and to
Fig. 1. PCR amplification of insect Tpi sequences from genomic DNAs: lane 1, Heliothis Õirescens; 2, Chrysops Õittatus; 3, Culex pipiens; 4, Aedes togoi; 5, genomic fragment containing Anopheles merus Tpi. Arrows indicate the Tpi DNA subcloned into pBS SKq.
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
reveal any intron positions between the two primers ŽFig. 2.. Anopheles merus Tpi was isolated by hybridizing a genomic library with C. tarsalis Tpi DNA. Promising DNAs were subcloned and sequenced as above. The coding region of Culex tarsalis is interrupted by a single intron of 298 bp between the codons for Gln 62 and Asn 63 w8x. Culex pipiens contains a 319-bp intron at the same site and Aedes togoi, a member of the same Culicidae subfamily, has a 154-bp intron at this position ŽFig. 2.. As expected, there is a faster rate of sequence divergence in introns than exons Ž 6–7 = for mosquito Tpis; Table 1. , but stretches of identity remain. Estimating that the two
133
Culex species diverged approximately 50 million years ŽMy. ago w13x, and that the two mosquito genera diverged 100 My ago, there has been a divergence rate of 0.005 substitutionsrsiterMy for the intron in each species. Thus the intronic sequence appears to be ancestral and at least 100 My old, but it is its position which has been used to support an ‘ancient intron’ origin. Another mosquito within the family Culicidae, Anopheles merus Ža member of the gambaie complex and a malarial vector. , does not have an intron at this position in Tpi Ž Fig. 2. . The absence of this intron has also been reported in an undesignated Anopheles species w10x. Thus according to the introns early view, the more evolutionary
Fig. 2. Amino acid sequences and intron positions of deduced TPIs aligned with TPIs from Homo sapiens, Drosophila melanogaster and Culex tarsalis using CLUSTAL program of PCGene v. 6.0. Intron sites are indicated by shading over an amino acid where the intron site interrupts the codon or over two amino acids where it falls between the codons. Conserved amino acids are indicated with asterisks while similar amino acids are indicated with dots underneath the sequence. Gaps inserted to optimize alignments are indicated by dashes. The sequences are available from Genbank under accession number a U23080 H. Õirescens, a U82708 A. togoi, a U82709 C. pipiens, a U82706 C. Õittatus, a U82707 A. merus.
134
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
Table 1 Divergence of mosquito TPI gene coding and non-coding regions by pair-wise alignment of the mosquito TPI intron and exon regions showing percent identity and divergence rate Ž D, substitutionsrsitermillion years., calculated using the Jukes–Cantor model w11x TPI alignments C. pipiens vs. A. togoi C. tarsalis vs. A. togoi C. pipiens vs. C. tarsalis
Exon
Intron
Rate divergence intronrexon
% identity
D
% identity
D
85.2 86.6 93.4
0.0008 0.0007 0.0007
44.8 42.4 64.2
0.0050 0.0055 0.0049
‘primitive culicid’ w14,15x, Anopheles, has lost this intron which would also have been lost at least 10 independent times during evolution w9,10x. The most parsimonious explanation would be that this intron was gained in the Culicidae subfamily; an interpretation also suggested by others. However, we do not consider the absence of this intron as definitive evidence of an introns late origin w10x for there seems to have been a particularly strong drive to streamline the Anopheles genome; these mosquitoes have only 20% of the nuclear DNA content of the Culex and Aedes subfamily w15x. The absence of introns within Anopheles led us to isolate Tpi fragments from a member of the primitive tabanid fly family which presumably arose from the ancestral Dipterans 150–200 My ago in the Lower Jurassic w16x. Chrysops Õittatus, a deer-fly, also lacks introns at this site within the recovered genomic TPI sequence ŽFig. 2. as does Calliphora Õicinia, a blow fly w10x, and the more recently evolved fruit-fly, Drosophila melanogaster w17x, with a genome size comparable to that of Anopheles. A compact genome, only twice as large as these Dipterans w18x, has also evolved in the pufferfish, Fugu rubripes. In contrast, however, this species generally seems to show a reduction in the size of introns, rather than intron loss. Genes from Fugu have been reported to be 4–6 = smaller than their counterparts from humans, although there are exceptions w19–21x. Mammalian genes which are imprinted, and thus maternally or paternally expressed and important for embryonic growth, are also small and in this case it appears to have been accomplished in a similar manner as the compact insect genomes; they have few and typically small introns w22x. It is possible that both in imprinted genes w22x and in the Dipterans, selection may have favoured loss of introns and reduced intronic sequences so as to effectively increase the rate of
6.2 7.8 7.0
translation product per gene. This seeming ease with which genome size can be reduced by intron loss seriously undermines the introns late parsimony argument: that intron loss, as evoked by the introns early proponents, is far too frequent and therefore introns must have been gained recently. Dipterans form a clade with Lepidoptera, which appeared during the radiation of flowering plants during the Cretaceous, about 100 My ago w16x. Heliothis Õirescens, tobacco budworm, and a member of this family, also does not have an intron at the Culex site ŽFig. 2.. However, this Tpi has at least four other introns, three of which are 78, 79 and 80 bp, and found at novel sites Žtwo between codons, phase 0, and one phase 1; w23x.. It is unlikely that all these sites were generated by ‘ancient intron’ drift or sliding w9,24,25x and thus we believe that these introns are recent introductions, perhaps due to the insertion of a repetitive or transposable element. Indeed, the three intervening sequences can be readily aligned and share considerable sequence similarity ŽFig. 3..
Fig. 3. Consensus sequence for three intervening sequences in the Heliothis TPI gene. The Heliothis intron sequences were aligned using PCGene CLUSTAL and a consensus sequence obtained. Conserved nucleotides are indicated with asterisks and a lack of consensus by a dash. Two gaps are inserted to align the 79-bp intron Žintron 1; corresponding to C. tarsalis amino acid position 36r37. with the 78-bp intron Ž2; position 74. and the 80-bp intron Ž3; position 110r111.. The hypothesis that the sequences are not related was rejected Ž P - 0.001, x 2 s 22, n s 2, with an intron wAqTx of 75%.. Characteristic intronic 5X and 3X dinucleotides ŽGT and AG, respectively. were omitted from the analysis.
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
Assuming that these sequences are homologous and u sin g a rate o f d iv erg en ce o f 0 .0 0 5 substitutionsrsiterMy ŽTable 1. they would have been introduced 50–90 My ago, the time of rapid diversification of angiosperms and specialized lepidopterans. Thus, depending upon the species, Tpi has been used as a model gene for ancestral, and possibly ancient introns, as well as late additions. TPI performs a basic step in glycolysis but in insects it has an additional role to maintain the a-glycerophosphate cycle important for flight w26x. A comparison of amino acid sequences from the TPI of insect species Ž Fig. 2; 10. reveals a surprising divergence. The common ancestor for birds and mammals and the ancestral Dipteran arose during the Triassic, yet Tpi from Drosophila and C. tarsalis have 73% identity compared to the 88% identity of Tpi from chicken and chimpanzee w12x. Nucleotide sequence analysis indicated that the nonsynonomous substitution rate was faster in the dipteran TPI genes compared to the vertebrate sequences, suggesting that the rate of TPI evolution was faster in the insect order. Again, this underscores the divergence of the insect genes. We stated w8x that if the C. tarsalis intron, which was originally predicted by W. Gilbert w5x, was indeed ancient it should be observed in other mosquitoes or their ancestors. We and others w10x have seen it only in only in mosquitoes which shared a common ancestor 100 My ago. However, we have also argued that the insect genome is very dynamic. Although this possibly ancient intron may have been retained in some mosquitoes, the drive to streamline the genome could have eliminated even this single intron from extant members of the more primitive subfamily. The ancient origin of an intron is all but impossible to prove due to the fast rate of intron divergence, intron slippage and intron loss but certainly the introns early view cannot be refuted on a parsimony argument which assumes that the cost of loss and gain are approximately equal. These data and the recent analysis of imprinted genes w22x suggest that there may be little cost to intron loss. Conversely, investigations on transposition events have been shown to be detrimental w27x and may only be retained, depending on their site of insertion Žw28x; H. Campbell and C. Claudianos, pers. commun... Despite this apparent cost, however, the Heliothis TPI gene shows evidence of
135
recent DNA insertion and is a striking example supporting the intron late view, and, incidentally, the prediction made by Cavalier-Smith w29x and Roger and Doolittle w30x that intron spread occurred by the retroposition of a spliceosomal element into different sites. Introns, then, can be lost or gained and the early or late assignment of a particular sequence cannot be undertaken until the relative costs are determined. This may be even more challenging for insect TPIs which appear to have diverged faster than expected. Until this time, we urge that proponents of both sides of this argument be reconciled; some Tpi introns could be ancient, and thus may have been important for exon shuffling, and others are clearly more recent additions.
Acknowledgements We thank A.E.R. Downe and D. Hickey for mosquitoes, R. Chagaturu for Heliothis larvae, and T. Inalsingh for the DNA sequencing. We are grateful for the suggestions of C. Claudianos, D. Forsdyke, W. Gilbert, D. Hickey, J. Logsdon, F. Sperling and C. Tittiger, as well as the encouragement and dedication of an anonymous reviewer. Supported by NSERC ŽCanada..
References w1x T. Cavelier-Smith, Nature ŽLondon. 315 Ž1985. 283–284. w2x J.D. Palmer, J.M. Logsdon Jr., Curr. Opin. Genet. Dev. 1 Ž1971. 470–477. w3x A. Stoltzfus, D.F. Spencer, M. Zuker, J.M. Logsdon Jr., W.F. Doolittle, Science 265 Ž1994. 202–207. w4x W.F. Doolittle, Am. Nat. 130 Ž1987. 915–928. w5x W. Gilbert, M. Marchionni, G. McKnight, Cell 46 Ž1986. 151–154. w6x M. Go, M. Nosaka, Cold Spring Harbor Symp. of Quant. Biol. 52 Ž1987. 915–924. w7x H.M. Seidel, D.I. Pompliano, J.R. Knowles, Science 257 Ž1992. 1449–1450. w8x C. Tittiger, S. Whyard, V.K. Walker, Nature ŽLondon. 361 Ž1993. 470–472. w9x J.M. Logsdon Jr., M.G. Tyshenko, C. Dixon, J. D.-Jafari, V.K. Walker, J.D. Palmer, Proc. Natl. Acad. Sci. USA 92 Ž1995. 8507–8511. w10x J. Kwiatowski, M. Krawczyk, M. Kornacki, K. Bailey, F.J. Ayala, Proc. Natl. Acad. Sci. USA 92 Ž1995. 8503–8506.
136
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
w11x T.H. Jukes, C.R. Cantor, in: H.N. Munro ŽEd.., Mammalian Protein Metabolism III, Academic Press, New York, pp. 21–123. w12x S. Whyard, C. Tittiger, V.K. Walker, Insect Biochem. Mol. Biol. 24 Ž1994. 819–827. w13x G.O. Poinar, Jr., Life in Amber. Stanford University Press, Stanford, CA, 1992. w14x H.H. Ross, Mosquito News 11 Ž1952. 128–132. w15x P.N. Rao, K.S. Rai, Hereditas 113 Ž1991. 139–144. w16x W. Hennig, Insect Phylogeny ŽEd. A.C. Pont., Wiley, Chichester, 1981. w17x R. Shaw-Lee, J. Lissemore, D. Sullivan, Mol. Gen. Genet. 230 Ž1991. 225–229. w18x S. Brenner, G. Elgar, R. Sanford, A. Macrae, B. Venkatesh, S. Aparicio, Nature ŽLondon. 366 Ž1993. 265–267. w19x P.J. Mason, D.J. Stevens, L. Luzzatto, S. Brenner, S. Aparicio, Genomics 26 Ž1995. 587–591. w20x S. Baxendale, S. Abdulla, G. Elgar, D. Buck, M. Berks, G.
w21x w22x w23x w24x w25x w26x w27x w28x w29x w30x
Micklem, R. Durbin, G. Bates, S. Brenner, S. Beck, Nat. Genet. 10 Ž1995. 67–76. A.D. Macrae, S. Brenner, Genomics 25 Ž1995. 436–446. L.D. Hurst, G. McVean, T. Moore, Nat. Genet. 12 Ž1996. 234–237. M.Y. Long, C. Rosenberg, W. Gilbert, Proc. Natl. Acad. Sci. USA 92 Ž1995. 12495–12499. P. Martinez, W. Martin, R. Cerff, J. Mol. Biol. 208 Ž1989. 551–565. L.D. Hurst, Nature ŽLondon. 371 Ž1994. 381–382. B. Sactor, in: M. Rockstein ŽEd.., The Physiology of Insects, Vol. 2, 1965, pp. 483–580. W.F. Eanes, C. Wesley, J. Hey, D. Houle, Genet. Res. 52 Ž1988. 17–26. J.S. Mattick, Curr. Opin. Genet. Dev. 4 Ž1994. 823–831. T. Cavalier-Smith, Trends Genet. 7 Ž1991. 145–148. A.J. Roger, W.F. Doolittle, Nature ŽLondon. 368 Ž1993. 289–290.
Towards a reconciliation of the introns early or late views: triosephosphate isomerase genes from insects Michael G. Tyshenko, Virginia K. Walker
)
Department of Biology, Queen’s UniÕersity, Kingston, Ont. K7L 3N6, Canada Received 7 January 1997; revised 13 March 1997; accepted 7 April 1997
Abstract The gene encoding the glycolytic enzyme, triosephosphate isomerase ŽTPI; EC 5.3.1.1., is a favourite model for molecular evolutionists who either subscribe to the theory that introns co-evolved with the ancestral gene, the introns early view, or alternatively, that introns are more recent immigrants. The discovery of an intron in the TPI gene of Culex mosquitoes at a site which was predicted by proponents of the intron early school supported that theory. More recently, the discovery of additional intron sites in several eukaryotes was presented as evidence supporting the introns late school. We have found the ‘Culex intron’ in two closely related mosquitoes, but not in two more evolutionary primitive Dipterans, suggesting that, if it is an ‘ancient intron’, loss may be more frequent than that supposed by the intron late school. In addition, we have found that three introns punctuating the TPI gene from the Lepidopteran, Heliothis, appear to be ancestrally related and may be the result of transposable element insertion, 50–90 million years ago. It is argued that both opposing schools in the intron debate be reconciled — some introns may have been early and certainly others have arrived subsequent to the appearance of the TPI gene. q 1997 Elsevier Science B.V. Keywords: Triosephosphate isomerase; Intron; Transposable element; Molecular evolution
1. Introduction The triosephosphate isomerase ŽTPI. gene evolved before the prokaryotic–eukaryotic divergence a few billion years ago. According the introns late argument, the introns which punctuate the coding sequence have been inserted relatively recently w1–3x. In contrast, the introns early view w4,5x holds that the introns are as old as the gene itself and predicts that the ancestral TPI gene was interrupted by introns at sites corresponding to polypeptide modules w6x or microgenes w7x. During evolution many of these in)
Corresponding author. Fax: Žq1. 613-545-6617; E-mail: [email protected]
trons have been lost but organisms have been found that, together, represent every ancient intron site in the TPI gene, and most recently in the mosquito, Culex tarsalis w8x. The introns late school argues that because introns can be found at additional, non-predicted sites, and that the loss of introns, to accommodate the introns early view, would have to be more frequent than instances of intron gain, parsimony arguments favour their theory. We have further investigated our previous observations Žcited in w9x. and those of others w10x that a closely related mosquito has the ‘Culex TPI intron’. In addition, we have sequenced the gene in two more primitive Dipterans and a member of a sister order, the Lepidoptera. The TPI gene from these insects
0167-4781r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 7 - 4 7 8 1 Ž 9 7 . 0 0 0 6 5 - 1
132
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
does not have this intron, underscoring the flexibility of the insects’ genome and, incidentally, the sometimes inflexibility of the ‘early’ or ‘late’ proponents.
2. Materials and methods 2.1. DNA preparation Genomic DNA Ž0.5 m g. from decapitated adult Culex pipiens and Aedes togoi were used as templates for amplification of TPI gene ŽTpi. fragments in separate experiments by the polymerase chain reaction Ž PCR.. Oligonucleotides for the reactions using mosquito DNA were sense primers: 5X TGGAAGATGAAYGGNGAYAAUGC 3X and antisense primers, 5X GTCTGGCGTTGACAATCT 3X. Whole adult insects were used to prepare DNA from the tabanid, Chrysops Õittatus and the moth, Heliothis Õirescens, using the sense primer: 5X GTNGGNGGNACYTGGAA 3X and the antisense primers: Heliothis 5X CCNGTNCCYATNGCCCA 3X and Chrysops 5X GGCCANACNGGYTCUTA 3X. Additional 3X Heliothis sequence was obtained using a nondegenerate sense sequence 5X TGAAGGTCATTGCCTGTATC 3X and a degenerate antisense oligonucleotide 5X CCACCGACGARRAANCC 3X. Forty cycles of amplification were performed, 1 min each at 948, 588 and 728C, followed by a 728C 8-min extension using Taq polymerase and standard reaction conditions provided by the supplier ŽPromega.. Tpi fragments were subcloned by blunt end cloning with Klenow fragment and ligation into the Sma 1 site of pBS SK q ŽStratagene.. The A. merus TPI gene was isolated by high stringency screening of an A. merus EMBL 3A genomic library using C. tarsalis Tpi as probe. A positive phage was plaque purified and a single hybridizing 9.7 Xho 1 fragment was subcloned into the Xho 1 site of pBS SK q . Sequence was obtained using the degenerate mosquito PCR primers. All sequencing reactions were done at least three times on a 373A automatic DNA sequencer ŽApplied Biosystems. using dye terminators. 2.2. Analysis The mosquito Tpi intron and exon sequences were aligned using PCGene NALIGN with an open gap
cost of 80 and a unit gap cost of 10. The three Heliothis intron sequences were aligned using PCGene CLUSTAL analysis and a consensus sequence obtained. Characteristic intronic 5X and 3X dinucleotides ŽGT and AG, respectively. were omitted from the analysis and as a control, intronic sequences from other organisms were also analysed in a similar manner. Divergence rate is expressed as the number of substitutionsrsitermillion years and was calculated using the Jukes and Cantor model w11x which corrects for overlapping substitutions which would otherwise result in an underestimate of evolutionary distance. The calculated divergence rate was used to estimate the divergence times of the Heliothis introns by solving for time using the Jukes and Cantor model. In the analysis of the mosquito coding regions by pairwise alignment, no additional corrections were made for synonymous and non-synonymous changes and thus these calculations are approximations. Calculations of the rate of Tpi evolution in different orders did take in account those changes and have been previously described w12x.
3. Results and discussion Genomic DNAs from Culex pipiens, Aedes togoi and Chrysops Õittatus were used as templates for the amplification of TPI gene fragments by PCR. Amplified DNA of an appropriate size Ž Fig. 1. was subcloned and sequenced w8x to confirm identity and to
Fig. 1. PCR amplification of insect Tpi sequences from genomic DNAs: lane 1, Heliothis Õirescens; 2, Chrysops Õittatus; 3, Culex pipiens; 4, Aedes togoi; 5, genomic fragment containing Anopheles merus Tpi. Arrows indicate the Tpi DNA subcloned into pBS SKq.
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
reveal any intron positions between the two primers ŽFig. 2.. Anopheles merus Tpi was isolated by hybridizing a genomic library with C. tarsalis Tpi DNA. Promising DNAs were subcloned and sequenced as above. The coding region of Culex tarsalis is interrupted by a single intron of 298 bp between the codons for Gln 62 and Asn 63 w8x. Culex pipiens contains a 319-bp intron at the same site and Aedes togoi, a member of the same Culicidae subfamily, has a 154-bp intron at this position ŽFig. 2.. As expected, there is a faster rate of sequence divergence in introns than exons Ž 6–7 = for mosquito Tpis; Table 1. , but stretches of identity remain. Estimating that the two
133
Culex species diverged approximately 50 million years ŽMy. ago w13x, and that the two mosquito genera diverged 100 My ago, there has been a divergence rate of 0.005 substitutionsrsiterMy for the intron in each species. Thus the intronic sequence appears to be ancestral and at least 100 My old, but it is its position which has been used to support an ‘ancient intron’ origin. Another mosquito within the family Culicidae, Anopheles merus Ža member of the gambaie complex and a malarial vector. , does not have an intron at this position in Tpi Ž Fig. 2. . The absence of this intron has also been reported in an undesignated Anopheles species w10x. Thus according to the introns early view, the more evolutionary
Fig. 2. Amino acid sequences and intron positions of deduced TPIs aligned with TPIs from Homo sapiens, Drosophila melanogaster and Culex tarsalis using CLUSTAL program of PCGene v. 6.0. Intron sites are indicated by shading over an amino acid where the intron site interrupts the codon or over two amino acids where it falls between the codons. Conserved amino acids are indicated with asterisks while similar amino acids are indicated with dots underneath the sequence. Gaps inserted to optimize alignments are indicated by dashes. The sequences are available from Genbank under accession number a U23080 H. Õirescens, a U82708 A. togoi, a U82709 C. pipiens, a U82706 C. Õittatus, a U82707 A. merus.
134
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
Table 1 Divergence of mosquito TPI gene coding and non-coding regions by pair-wise alignment of the mosquito TPI intron and exon regions showing percent identity and divergence rate Ž D, substitutionsrsitermillion years., calculated using the Jukes–Cantor model w11x TPI alignments C. pipiens vs. A. togoi C. tarsalis vs. A. togoi C. pipiens vs. C. tarsalis
Exon
Intron
Rate divergence intronrexon
% identity
D
% identity
D
85.2 86.6 93.4
0.0008 0.0007 0.0007
44.8 42.4 64.2
0.0050 0.0055 0.0049
‘primitive culicid’ w14,15x, Anopheles, has lost this intron which would also have been lost at least 10 independent times during evolution w9,10x. The most parsimonious explanation would be that this intron was gained in the Culicidae subfamily; an interpretation also suggested by others. However, we do not consider the absence of this intron as definitive evidence of an introns late origin w10x for there seems to have been a particularly strong drive to streamline the Anopheles genome; these mosquitoes have only 20% of the nuclear DNA content of the Culex and Aedes subfamily w15x. The absence of introns within Anopheles led us to isolate Tpi fragments from a member of the primitive tabanid fly family which presumably arose from the ancestral Dipterans 150–200 My ago in the Lower Jurassic w16x. Chrysops Õittatus, a deer-fly, also lacks introns at this site within the recovered genomic TPI sequence ŽFig. 2. as does Calliphora Õicinia, a blow fly w10x, and the more recently evolved fruit-fly, Drosophila melanogaster w17x, with a genome size comparable to that of Anopheles. A compact genome, only twice as large as these Dipterans w18x, has also evolved in the pufferfish, Fugu rubripes. In contrast, however, this species generally seems to show a reduction in the size of introns, rather than intron loss. Genes from Fugu have been reported to be 4–6 = smaller than their counterparts from humans, although there are exceptions w19–21x. Mammalian genes which are imprinted, and thus maternally or paternally expressed and important for embryonic growth, are also small and in this case it appears to have been accomplished in a similar manner as the compact insect genomes; they have few and typically small introns w22x. It is possible that both in imprinted genes w22x and in the Dipterans, selection may have favoured loss of introns and reduced intronic sequences so as to effectively increase the rate of
6.2 7.8 7.0
translation product per gene. This seeming ease with which genome size can be reduced by intron loss seriously undermines the introns late parsimony argument: that intron loss, as evoked by the introns early proponents, is far too frequent and therefore introns must have been gained recently. Dipterans form a clade with Lepidoptera, which appeared during the radiation of flowering plants during the Cretaceous, about 100 My ago w16x. Heliothis Õirescens, tobacco budworm, and a member of this family, also does not have an intron at the Culex site ŽFig. 2.. However, this Tpi has at least four other introns, three of which are 78, 79 and 80 bp, and found at novel sites Žtwo between codons, phase 0, and one phase 1; w23x.. It is unlikely that all these sites were generated by ‘ancient intron’ drift or sliding w9,24,25x and thus we believe that these introns are recent introductions, perhaps due to the insertion of a repetitive or transposable element. Indeed, the three intervening sequences can be readily aligned and share considerable sequence similarity ŽFig. 3..
Fig. 3. Consensus sequence for three intervening sequences in the Heliothis TPI gene. The Heliothis intron sequences were aligned using PCGene CLUSTAL and a consensus sequence obtained. Conserved nucleotides are indicated with asterisks and a lack of consensus by a dash. Two gaps are inserted to align the 79-bp intron Žintron 1; corresponding to C. tarsalis amino acid position 36r37. with the 78-bp intron Ž2; position 74. and the 80-bp intron Ž3; position 110r111.. The hypothesis that the sequences are not related was rejected Ž P - 0.001, x 2 s 22, n s 2, with an intron wAqTx of 75%.. Characteristic intronic 5X and 3X dinucleotides ŽGT and AG, respectively. were omitted from the analysis.
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
Assuming that these sequences are homologous and u sin g a rate o f d iv erg en ce o f 0 .0 0 5 substitutionsrsiterMy ŽTable 1. they would have been introduced 50–90 My ago, the time of rapid diversification of angiosperms and specialized lepidopterans. Thus, depending upon the species, Tpi has been used as a model gene for ancestral, and possibly ancient introns, as well as late additions. TPI performs a basic step in glycolysis but in insects it has an additional role to maintain the a-glycerophosphate cycle important for flight w26x. A comparison of amino acid sequences from the TPI of insect species Ž Fig. 2; 10. reveals a surprising divergence. The common ancestor for birds and mammals and the ancestral Dipteran arose during the Triassic, yet Tpi from Drosophila and C. tarsalis have 73% identity compared to the 88% identity of Tpi from chicken and chimpanzee w12x. Nucleotide sequence analysis indicated that the nonsynonomous substitution rate was faster in the dipteran TPI genes compared to the vertebrate sequences, suggesting that the rate of TPI evolution was faster in the insect order. Again, this underscores the divergence of the insect genes. We stated w8x that if the C. tarsalis intron, which was originally predicted by W. Gilbert w5x, was indeed ancient it should be observed in other mosquitoes or their ancestors. We and others w10x have seen it only in only in mosquitoes which shared a common ancestor 100 My ago. However, we have also argued that the insect genome is very dynamic. Although this possibly ancient intron may have been retained in some mosquitoes, the drive to streamline the genome could have eliminated even this single intron from extant members of the more primitive subfamily. The ancient origin of an intron is all but impossible to prove due to the fast rate of intron divergence, intron slippage and intron loss but certainly the introns early view cannot be refuted on a parsimony argument which assumes that the cost of loss and gain are approximately equal. These data and the recent analysis of imprinted genes w22x suggest that there may be little cost to intron loss. Conversely, investigations on transposition events have been shown to be detrimental w27x and may only be retained, depending on their site of insertion Žw28x; H. Campbell and C. Claudianos, pers. commun... Despite this apparent cost, however, the Heliothis TPI gene shows evidence of
135
recent DNA insertion and is a striking example supporting the intron late view, and, incidentally, the prediction made by Cavalier-Smith w29x and Roger and Doolittle w30x that intron spread occurred by the retroposition of a spliceosomal element into different sites. Introns, then, can be lost or gained and the early or late assignment of a particular sequence cannot be undertaken until the relative costs are determined. This may be even more challenging for insect TPIs which appear to have diverged faster than expected. Until this time, we urge that proponents of both sides of this argument be reconciled; some Tpi introns could be ancient, and thus may have been important for exon shuffling, and others are clearly more recent additions.
Acknowledgements We thank A.E.R. Downe and D. Hickey for mosquitoes, R. Chagaturu for Heliothis larvae, and T. Inalsingh for the DNA sequencing. We are grateful for the suggestions of C. Claudianos, D. Forsdyke, W. Gilbert, D. Hickey, J. Logsdon, F. Sperling and C. Tittiger, as well as the encouragement and dedication of an anonymous reviewer. Supported by NSERC ŽCanada..
References w1x T. Cavelier-Smith, Nature ŽLondon. 315 Ž1985. 283–284. w2x J.D. Palmer, J.M. Logsdon Jr., Curr. Opin. Genet. Dev. 1 Ž1971. 470–477. w3x A. Stoltzfus, D.F. Spencer, M. Zuker, J.M. Logsdon Jr., W.F. Doolittle, Science 265 Ž1994. 202–207. w4x W.F. Doolittle, Am. Nat. 130 Ž1987. 915–928. w5x W. Gilbert, M. Marchionni, G. McKnight, Cell 46 Ž1986. 151–154. w6x M. Go, M. Nosaka, Cold Spring Harbor Symp. of Quant. Biol. 52 Ž1987. 915–924. w7x H.M. Seidel, D.I. Pompliano, J.R. Knowles, Science 257 Ž1992. 1449–1450. w8x C. Tittiger, S. Whyard, V.K. Walker, Nature ŽLondon. 361 Ž1993. 470–472. w9x J.M. Logsdon Jr., M.G. Tyshenko, C. Dixon, J. D.-Jafari, V.K. Walker, J.D. Palmer, Proc. Natl. Acad. Sci. USA 92 Ž1995. 8507–8511. w10x J. Kwiatowski, M. Krawczyk, M. Kornacki, K. Bailey, F.J. Ayala, Proc. Natl. Acad. Sci. USA 92 Ž1995. 8503–8506.
136
M.G. Tyshenko, V.K. Walkerr Biochimica et Biophysica Acta 1353 (1997) 131–136
w11x T.H. Jukes, C.R. Cantor, in: H.N. Munro ŽEd.., Mammalian Protein Metabolism III, Academic Press, New York, pp. 21–123. w12x S. Whyard, C. Tittiger, V.K. Walker, Insect Biochem. Mol. Biol. 24 Ž1994. 819–827. w13x G.O. Poinar, Jr., Life in Amber. Stanford University Press, Stanford, CA, 1992. w14x H.H. Ross, Mosquito News 11 Ž1952. 128–132. w15x P.N. Rao, K.S. Rai, Hereditas 113 Ž1991. 139–144. w16x W. Hennig, Insect Phylogeny ŽEd. A.C. Pont., Wiley, Chichester, 1981. w17x R. Shaw-Lee, J. Lissemore, D. Sullivan, Mol. Gen. Genet. 230 Ž1991. 225–229. w18x S. Brenner, G. Elgar, R. Sanford, A. Macrae, B. Venkatesh, S. Aparicio, Nature ŽLondon. 366 Ž1993. 265–267. w19x P.J. Mason, D.J. Stevens, L. Luzzatto, S. Brenner, S. Aparicio, Genomics 26 Ž1995. 587–591. w20x S. Baxendale, S. Abdulla, G. Elgar, D. Buck, M. Berks, G.
w21x w22x w23x w24x w25x w26x w27x w28x w29x w30x
Micklem, R. Durbin, G. Bates, S. Brenner, S. Beck, Nat. Genet. 10 Ž1995. 67–76. A.D. Macrae, S. Brenner, Genomics 25 Ž1995. 436–446. L.D. Hurst, G. McVean, T. Moore, Nat. Genet. 12 Ž1996. 234–237. M.Y. Long, C. Rosenberg, W. Gilbert, Proc. Natl. Acad. Sci. USA 92 Ž1995. 12495–12499. P. Martinez, W. Martin, R. Cerff, J. Mol. Biol. 208 Ž1989. 551–565. L.D. Hurst, Nature ŽLondon. 371 Ž1994. 381–382. B. Sactor, in: M. Rockstein ŽEd.., The Physiology of Insects, Vol. 2, 1965, pp. 483–580. W.F. Eanes, C. Wesley, J. Hey, D. Houle, Genet. Res. 52 Ž1988. 17–26. J.S. Mattick, Curr. Opin. Genet. Dev. 4 Ž1994. 823–831. T. Cavalier-Smith, Trends Genet. 7 Ž1991. 145–148. A.J. Roger, W.F. Doolittle, Nature ŽLondon. 368 Ž1993. 289–290.