Abstract
We have sequenced genomic fragments which encode the N-terminus of the TIMELESS (TIM) clock protein in Drosophila simulans and D.yakuba . We observe that in these two species, the initiating methionine appears to lie downstream of the one proposed to encode the translational start in D.melanogaster , thereby truncating the N-terminus by 23 amino acids. We then sequenced the corresponding 5′ fragment in a number of D.melanogaster individuals from different strains. We observed a polymorphism which strongly suggests that the originally proposed start site cannot be utilised in some individuals, and that these flies will initiate translation of TIM at the downstream ATG. Given the current interest in TIM regulation in D.melanogaster , it is important to correctly define the N-terminus in this species.
Introduction
Circadian rhythmicity is a widespread and ancient adaptation of living organisms to the rotation of the earth. Mutations affecting circadian rhythms have been identified in several organisms ranging from prokaryotes ( 1 ) to plants ( 2 ) and mammals ( 3 ). In Drosophila melanogaster two clock genes which encode an integral component of the pacemaker have been cloned and characterized, period ( per ) ( 4–7 ) and timeless ( tim ) ( 8 , 9 ). It has been demonstrated that PER and TIM physically interact ( 9–11 ) and both proteins are involved in a mutual negative feedback regulation of their own expression ( 12 ). Moreover, tim+ activity is necessary for the nuclear localization of a PER reporter ( 13 ).
The D.melanogaster tim cDNA sequence was obtained by sequencing several clones from independently derived head specific cDNA libraries as well as genomic DNA ( 8 ). The data reveal that the sequence contains two open reading frames (ORFs) coding for two conceptual proteins of 1122 or 1389 amino acids ( 8 ). Consequently, there may be two forms of the TIM protein. The two ORFs share the residues 1–1104 but have different C-termini depending on the removal or not of a particular intron.
As part of our continuing comparative analyses of clock genes in diptera ( 14–16 ), we designed a series of oligoprimers in order to amplify fragments of the tim gene from other species. Primers which focused on the region encoding the N-terminus, amplified a fragment in the D.melanogaster sibling species, D.simulans , and a more distantly related member of the melanogaster subgroup of species, D.yakuba . We observe in both species a different translational start from the one identified by Myers et al. ( 8 ) in D.melanogaster . Furthermore, the same alternative translational start is also observed in some D.melanogaster individuals. Given the current interest in PER and TIM ( 17 , 18 ) and the ongoing biochemical analysis of these two proteins in D.melanogaster ( 10 , 11 ), particularly with regard to TIM degradation, the N-terminus of TIM may play an important role in the regulation of the two proteins.
Materials and Methods
Drosophila strains
Single D.melanogaster individuals were obtained from the following laboratory strains: Canton-S, per+ , per S and per 01 mutant lines. In addition, isofemale lines were collected from a natural population from Cognac in France (see ref. 19 for details). Individual male descendants were crossed to attached -X females to generate self perpetuating iso -X chromosome lines. Two of these lines, Co-TG14 and Co-TG20 , were used. Consequently the second chromosomes in these lines will carry either the natural Cognac-derived or the attached -X strain-derived tim allele.
A D.yakuba strain was obtained from J. David (University of Paris) and the D.simulans strain was obtained from the Bowling Green Drosophila Species Stock Center, USA.
PCR amplification and DNA sequencing
For the amplification of the tim fragment from D.melanogaster the 5′ primer used was 5′-CACAATCACATCTGGAATAA-3′ corresponding to nucleotide positions 264–283 in the sequence from Myers et al. ( 8 ). The 3′ primer corresponded to positions 384–403 (5′-GCATTCGGGTTGACCACATA-3′). For the amplification from D.simulans and D.yakuba the following degenerate primers were used: 5′ primer was 5′- CAYAAYCAYATHTGGAAYAA-3′ (264–283); 3′ primer was 5′-GCRTTSGGRTTNACNACR-3′ (384–403).
Single fly DNA was prepared using the method of Gloor and Engels ( 20 ). DNA amplification by PCR was carried out according to Jeffreys et al. ( 21 ) in a MJ Research thermocycler for 30 cycles: 95°C for 1 min, 65°C for 1 min, 72°C for 1 min for D.melanogaster and 95°C for 1 min, 48°C for 1 min 10 s, 72°C for 1 min 20 s, for D.simulans and D.yakuba . BioTaq polymerase from Bioline was used.
Nucleotide sequences of a tim fragment (139–140 bp) from six different laboratory lines of D.melanogaster and from one strain of D.simulans and D.yakuba (see Material and Methods). The fragments were obtained by PCR with oligos amplifying from positions 264 to 403 in the cDNA sequence of Myers et al. ( 8 ) which is shown above for comparison from nucleotide 1 to 403. The ATG motif in bold is the downstream translation initiation codon for D.simulans, D.yakuba and for the D.melanogaster per01 and per+ sequences. The ATG underlined represents the upstream initiating motif previously proposed by Myers et al. ( 8 ). Two further upstream ATG sequences which are represented in bold italics generate premature translational stops. The tga motif in lower case represents the stop codon which is generated by the G deletion in position 294. The boxed area corresponds to the position of the Kozak sequence (see text).
The PCR amplified fragments to be sequenced were cloned into the pMOSBlueT-vector (Amersham). The recombinant DNAs were used to transform MOSBlue competent cells (Amersham). A minimum of three clones was sequenced from two independent PCR amplifications for each fly's DNA. The DNA sequencing was carried out either using the ThermoSequenase cycle sequencing kit US78500 (Amersham) or by an automatic system using the Perkin-Elmer Turbo-Catalyst robotic station and ABI 377 sequencing apparatus. The M13R and −40 oligos were used to sequence the cloned fragment.
Results
Figure 1 illustrates the genomic sequences obtained from the 139 or 140 bp tim fragment. The 5′ stretch (from nucleotide 1 to 403) of the tim cDNA sequence of D.melanogaster obtained by Myers et al. ( 8 ) is also reported for comparison (GenBank accession number U37018). Note that there are a number of ATG motifs. The first two (represented in bold italics in Fig. 1 , positions 208–210 and 227–229) give rise to premature translational stop codons, and the third (positions 243–245, underlined in Fig. 1 ), has been suggested by Myers et al. ( 8 ) to represent the initiating motif.
The D.melanogaster genomic tim sequences show both nucleotide and length polymorphisms. In particular, the tim sequences in D.melanogaster lines per+ and per 01 show a substitution (A→C) in position 297 and a single base deletion (G) in position 294, when compared with the other D.melanogaster genomic and cDNA sequences. The D.simulans sequence differs from the D.melanogaster cDNA adopted as a reference, with nucleotide substitutions in position 297 (A→C) and in position 359 (C→A). The same single base deletion (G) in position 294 observed in the D.melanogaster per + and per 01 lines, is also found in D.simulans . Finally, the D.yakuba sequence shows substitutions in positions 288 (A→G), 297 (A→C), 332 (T→C), 342 (T→C), 359 (C→T), 365 (T→C), 374 (G→A), 377 (C→G) and, again, the single base deletion (G) in position 294. Among the intra- and interspecific nucleotide polymorphisms observed in the sequences reported in Figure 1 , only one produces an amino acid substitution (position 342), giving the Y→H substitution present in the putative N-terminus region of the yakuba TIM protein ( Fig. 2 ).
Putative N-terminus of the TIM protein for three Drosophila species ( D.melanogaster, D.simulans and D.yakuba ). The D.melanogaster TIM sequence indicated as a reference (ref.) is the one proposed by Myers et al. ( 8 ) from cDNA analysis. The other protein sequences are derived from the data in Figure 1 which suggest a different initiating methionine (in bold) for the TIM protein in the D.melanogaster strains per + and per 01 , and D.simulans and D.yakuba . The amino acid substitution observed in D.yakuba (Y to H) is also highlighted in bold.
Consequently, the G deletion in position 294 and the A→C substitution in position 297 are shared by D.simulans and D.yakuba , as well as the tim sequences in the D.melanogaster per + and per 01 strains. Moreover, the first and second ATG motifs in the cDNA (in bold italics in Fig. 1 ) generate precocious stop codons. The sequence of the D.melanogaster TIM putative protein published by Myers et al. ( 8 ) represents the translation of an ORF which starts from the third ATG motif (nucleotide positions 243–245, underlined in Fig. 1 ) present in the cDNA sequence. If we consider the reading frame generated by this ATG, it is dramatically affected by the G deletion in position 294, causing a frame-shift and a consequent premature translational stop codon (TGA, nucleotides 301–303, lower case in Fig. 1 ). Consequently, the next downstream in-frame ATG is found in positions 312–314 (bold in Fig. 1 ).
Discussion
The intraspecific and interspecific pattern of sequences described above, strongly suggest that the first methionine of the TIM protein is coded by the ATG motif (positions 312–314, represented in bold in Fig. 1 ). This is located downstream from position 294 where the polymorphism for the deletion of the corresponding base (G) has been observed. This scenario would maintain the same reading frame initiated by the upstream ATG motif (position 243–245) suggested by Myers et al. ( 8 ), but would truncate the N-terminus by 23 amino acids ( Fig. 2 ).
The sequences surrounding the downstream ATG (positions 306–315, boxed in Fig. 1 ) are in agreement with the most stringent requirements dictated by Kozak's consensus translation initiation sequence (GCC R CC ATGG ) ( 22 , 23 ). In fact, in this case, a purine (G) is present three bases before the ATG motif, and a G immediately following it. These two bases are the most important in the consensus sequence as they can influence the efficiency of translation 10-fold, while the other positions have much smaller effects ( 22 ). The upstream ATG motif favoured by Myers et al. ( 8 ) does not have the same characteristics, since a purine (A) is present three bases before it, but there is an A instead of a G immediately downstream. We cannot exclude that TIM protein synthesis occurs with more than one N-terminus, through a context-dependent leaky scanning mechanism ( 23 ) in those D.melanogaster individuals which do not show the G deletion at position 294. However, for those individuals of D.melanogaster and D.simulans and D.yakuba carrying the deletion, the putative TIM protein(s) should be 23 amino acids ‘shorter’ than previously predicted, as shown in Figure 2 .
The regulation of PER and TIM is currently of great interest ( 10 , 11 , 17 , 18 ) and the stability and degradation of these two proteins has major implications for the negative feedback model of circadian timing ( 24 ). A precise definition of the N-terminus of TIM may therefore be of considerable importance.
Acknowledgements
We acknowledge a NERC grant to CPK, a CEC grant from the Biotechnology programme (ERBBIO4CT960096), and a MURST-British Council grant to CPK and RC for Anglo-Italian cooperation, and a MURST studentship to FS.
Comments