Features and exceptions in the tRNA cloverleaf model from the ten genomes. The sequence presented under the cloverleaf (2D) model is the consensus sequence of the 274 tRNA genes from S.cerevisiae. The conventional IUB/IUPAC degenerate DNA (or RNA with T changed for U) alphabet (32) is used in this and following figures: R (purine), A or G; Y (pyrimidine), C or T; S (strong), G or C; W (weak), A or T; M (amino), A or C; K (keto), G or T, B (not A), C, G or T; D (not C), A, G or T; H (not G), A, C or T; V (not T), A, C or G; N (any), A, C, G or T; small dots also indicate n (any nucleotide). The key to base pairing symbols is: ‘+’,Watson–Crick base pairing only; ‘*’,Watson–Crick pairing or GT/TG pairing; ‘#’, Watson–Crick pairing or mismatch; ‘-’, Watson–Crick pairing or GT/TG pairing or mismatch. The arrows indicate the four variable positions of the D-loop which are (or not) occupied (indicated with lower case letters). Boxed nucleotides are those conserved in nearly all tDNAs; grey background indicates nucleotides requested by the eukaryotic cloverleaf model used to search tRNA genes in the genomes (21) (for clarity, the T-loop is drawn twice). The details of this model are indicated in the surounding grey-backgrounded boxes (one for each stem and one at top left applying to all four stems together). In these boxes, the maximum number of GT or TG base pairs (GU or UG in mature tRNA) is indicated as ‘GT’ and the maximum number of non Watson–Crick base pairs (mismatched) are indicated as ‘mm’. Values in parenthesis were used only for tDNA search in Y.lipolytica (YALI). Remarkable features are indicated in the other boxes, sequence exceptions are the heavy lined boxes. More exceptions to conserved bases in the D- and T-loops, that may affect tRNA genes transcription, are shown below in Figure 5A.

Distance tree analysis of 603 tDNA isoacceptor sequences from nine hemiascomycetous genomes and S.pombe. (A) Examples of tDNA sequences prepared for the p-distance analysis. The intron and sequences between nt 46 and 48 were removed to obtain perfectly aligned sequences, all 75 nt long. The stems are symbolized with ‘{}’, acceptor stem; ‘<>’, D- and T-stems; ‘()’, anticodon stem; the anticodon is indicated with ‘###’. Stars indicate sequence variations. (B) The 603 different isoacceptor tDNA sequences analysed (see Materials and Methods) generate 181 503 pairwise p-distances. This histogram displays the number of p-distance values between two tDNA versus the value of the p-distance. (C) A p-distance unrooted tree was computed from the p-distance matrix. For the sake of clarity, this tree is presented vertically and the sub-trees in which neighbour tDNA encoding the same amino acid cluster together are symbolized by boxes. The actual branches inside the sub-trees extends rightwards far beyond the right edge of the boxes. The number of anticodons and codons specific to the amino acid are given inside each box (e.g. ‘2,3/4 Ala’ means 2 or 3 anticodons and 4 codons for alanine). The three types of boxes are as follows: (i) Heavy lined boxes: all isoacceptors for a given amino acid (whatever are the anticodons) from the ten genomes cluster together; the total number of corresponding sequence types are given outside the boxes at right. (ii) Light boxes: not all the tDNA isoacceptors from the ten genomes cluster together; numbers at right indicate the fraction of tDNA that cluster over the total number of sequences considered for the amino acid. (iii) Light boxes filled with grey: all isoacceptors from the nine hemiascomycetes cluster together, but not those from S.pombe (SCPO) which are indicated at the right side (long grey horizontal branches ending with a dot). The signs ‘+’ at right indicate extra tDNA sequences that cluster inside uncomplete sub-trees (light boxes). For clarity, some vertical spacing was introduced in the drawing of the tree but the length of the horizontal branches was not modified. Notes: (#1) tDNAs-Gly split into two clusters; the upper one contains most of the tDNAs-Gly (TTC). (#2) tDNAs-Asp and tDNAs-Glu do not form two separate clusters but a single one mixing these two neighbour isoacceptors. (#3) The gene of the tDNA-Arg (CCG) of S.cerevisiae [which is related to the tRNA-Asp (GTC) of the same organism (see text)] as well as those from five other genomes (over eight harbouring such as tDNA) define a special cluster close to the Asp/Glu one (see also note #6). (#4) in D.hansenii (DEHA) and C.albicans (CAAL), the CUG codon is used for serine instead of leucine and only five codons remain for leucine. (#5) This cluster includes the two special single copy tDNA-Ser (CAG) of D.hansenii (DEHA) and C.albicans (CAAL) (CUG is a 7th serine codon in these two genomes). (#6) This cluster gets together tDNA-Arg other than tDNA-Arg (CCG) (however that of D.hansenii clusters here and not in the extra Arg (CCG) cluster).

Conservation of the A- and B-boxes promoter sequences in four RNA Pol III genes. The schematic representation of the genes uses the same convention as in Figure 5B. Coordinates of the mature products are given with a letter indicating the chromosome (or the contig number) and the direction (‘>’,Watson strand; ‘<’, Crick strand). Positions of the first base of A- and B-boxes (conserved T for the A-box, conserved G for the B-box) are given with respect to the first nucleotide of the mature product (numbered +1). A positive coordinate indicates that the promoter sequence is located inside the mature product (as in SNR6 and SCR1); while a negative coordinate indicates that the promoter sequence is located in a leader sequence cleaved posttrancriptionally (shown as dashed lines in SNR52 and RPR1). ‘ΔA-B’ indicates the distance (nt) separating the A- and B- boxes. Additionally, for the SNR6 genes, the distance between the 3′ end of the gene and the external 3′ B-box is given (‘Δter/B’). Nucleotides corresponding to ‘n’ (any nucleotide) in the genomic sequences are written in lower case. Exceptions in the positions conserved or semi-conserved in the consensus are also written in lower case. The nucleotides preceding and following the A- and B-boxes in the genomic sequences are also reported (separated with a blank) to better enhance the actual boundaries of the sequences putatively recognized by TFIIIC. The ‘na’ indication stands for ‘not applicable’. Boxes hilite the peculiar organization of the S.pombe SNR6 gene (B-box inside a spliceosomal intron) and that of RPR1 gene of E.gossypii and D.hansenii (B-box overlapping the 5′ boundary of the mature product) and C.albicans and Y.lipolytica (B-box internal). Notes: (#1) In the SNR6 gene of S.pombe (black box), the B-box is not located beyond the gene but inside a 50 nt spliceosomal intron located at positions 51–100; boundaries and length reported here include the intron. (#2) The SNR52 and RPR1 genes of S.pombe are probably Pol II genes as no A/B-boxes nor poly-T terminators are present. (#3) Two 100% identical SCR1 genes are present in S.castellii. (#4) Two SCR1 genes (94% identity) are present in Y.lipolytica.

Localization of decoding changes for Leu CUN and Arg CGN in the phylogeny of hemiascomycetes. The schematic phylogeny of hemiascomycetes versus S.pombe is taken from (22,23). The two four-codon boxes at left show the regular decoding rules of Leu CUN and Arg CGN codons in eukaryotes. In both boxes, the tRNA with anticodon starting with C₃₄ (shown in grey) is dispensable. Grey arrows indicate which is the switch in decoding occurring at the targeted node. Node 1: switch in the decoding of arginine CGN codons from G34-sparing to U₃₄-sparing; node 2: change in the genetic code (CUG codon reassigned to serine) in the Candida genus and switch in the decoding of leucine CUN codons from G34-sparing to U₃₄-sparing; node 3: switch in the decoding of leucine CUN codons from G₃₄ to A₃₄-sparing; node 4: location of the ancestral whole genome duplication (WGD) (104); node 5: S.castellii reverts to the standard eukaryotic rule for the decoding of Leu CUN codons. Grey squares at left indicate which species have lost the tDNA-Leu (CAG) (C₃₄-sparing); Note the similarity of the decoding of Leu CUN in the Candida genus with that of Arg CGN in yeasts from S.cerevisiae to C.albicans.

tRNA/anticodon and tDNA usages in the ten genomes. The upper and lower panels correspond to the left and right part of the conventional genetic code tabulation (upper and lower insets at right, respectively). The ten genomes are designated by their acronyms; the ‘#’ signs following some acronyms denote genomes of low coverage or without full assembled chromosomes (uncomplete genomes). The 64 regular codons and anticodons, plus the initiator mehionine (iMet), are listed vertically. Anticodons which are never used in tRNAs (in any of the three domains of life) are replaced by ‘---’. Numbers in the main array indicate the number of tRNA genes present per genome for the given anticodon; ‘-’ signs stand for no gene; ‘/’ signs indicate that one copy (at least) of the tDNA is probably present in the actual genome but could not be identified in the available sequences. The ‘+’ signs report tRNA genes that harbour an intron and ‘±’ those in which the intron is absent from some copies. Grey background indicates tDNAs that exhibit deviation from the consensus sequences defined for the A-box at G₁₀; open boxes indicate other exceptions (see Figure 5C). The indications ‘ΔA’, ‘ΔG’, ‘ΔU’ and ‘ΔC’ emphasize the lack of tRNA bearing anticodon starting with the indicated nucleotide (first base of anticodon). The ‘A or G’ sparing rule (‘ΔG’ or ‘ΔA’) as well as ‘ΔU’ and ‘ΔC’ rules are summarized at right with ‘E’, ‘A’ and ‘B’ to indicate whether the rule applies to each of the three domains of life (Eukaryotes, Archaea and Bacteria, respectively). Black boxes emphasize the decoding of leucine and arginine (see Figure 3 for details) in which double arrows denote anticodons of variable usage among the ten genomes. In the grey rectangle at bottom are given the number of tRNA species per genome (number of differents anticodons with initiator and elongator tRNA-Met considered as different), the number of variant tRNA genes, the total number of tRNA genes per genome and the genome size expressed in Mb. For K.waltii (KLWA) and S.castellii (SACA), values of 41 different tRNA (in brackets) are underestimated due to the uncompleteness of sequence data. These species probably harbour 43 and 42 tRNA, respectively.

Various strategies used to decode the Leu UUR/CUN and Arg AGR/CGN codons. In addition to the ten genomes explored in this work, data of other eukaryotic genomes are reported for comparison: NECR, Neurospra crassa (99); MAGR, Magnaporthe grisea [Data Version 10/31/2003 (Release 2.3)]; COCI, Coprinus cinereus (Data Version 6/25/2003); FUGR; Fusarium graminearum (Data Version 3/11/2003); CAEL, C.elegans (100); DRME, D.melanogaster (101); ARTH, Arabidopsis thaliana (102) and ENCU, E.cuniculi (103). ‘#’ signs indicate genomes of low coverage or without full assembled chromosomes. (A) Numbers indicate how many genes encode tRNA reading the UUR and CUN leucine codons (with ‘-’ standing for no gene). Underline and italic numbers indicate the most and second most used codons [taken from (24), Supplementary Table S3]. All genomes harbours the two tRNA-Leu (UAA) and (CAA). The two heavy boxes emphasize the particular situation in D.hansenii (DEHA) and C.albicans (CAAL): i.e. the tRNA-Leu G₃₂ (AAG) reads the three codons CUU, CUC and CUA (top box) and the single copy tRNA-Ser G₃₃ (CAG) reads the CUG codon (bottom box). The dashed vertical bar separates the species using an eukaryotic-type of sparing (at right) from those using bacteial types (at left). (B) Schematic representation of the decoding of the Leu CUN codon in the different genomes indicated. The sequences of the anticodon stems and loops of the three remarkable tRNA-Leu (GAG) of S.cerevisiae and four related genomes, tRNA-Leu (AAG) and tRNA-Ser (CAG) of D.hansenii and C.albicans are shown to illustrate that all three harbour an unusual nucleotide (boxed) close to the anticodon (boxed): C₃₃, G₃₂ or G₃₃, respectively. (C and D) A similar presentation is used for the decoding of the six arginine codons. Darker grey background denote, among the tDNA-Arg (CCG), which ones are presumably derived from the tDNA-Asp (GTC) (see Figure 4B). In the uncomplete genome of K.waltii (KLWA), this tDNA was not found, but it is probably present in this organism.

tRNA genes promoter sequences and polymorphism of the A-box/D-loop and B-box/T-loop sequences. (A) The consensus sequence of the 274 tRNA genes from S.cerevisiae is used to emphasize the conserved elements recognized by the RNA polymerase III machinery at the DNA level. These elements concentrate in the two areas referred to as the A- and B-boxes (grey boxes) which include, at the RNA level, the forward strand of the D-stem plus the D-loop (A-box) and the T-loop plus the terminal base pair of the T-stem (B-box), respectively. Typographical symbols and nucleotide abbreviations are given in the legend to Figure 2. The arrows indicate the four variable positions of the D-loop (occupied, or not, indicated with lower case letters). Exceptions in the sequences of the A- and B- boxes of some tDNA are indicated in surounding boxes. Grey lines connecting nucleotides between the D- and T-loops represent tertiary base pairs. (B) Schematic representation of a tRNA gene without (left) or with (right) an intron. The solid black bar represents upstream and downstream DNA, the open rectangle(s) the mature product and the two grey rectangles the A- and B-boxes. Transcription starts about 20–25 bases upstream the A-box (vertical arrows) and terminates inside the poly-T track (T_n). (C) Consensus sequences observed in the A- and B-boxes for the ten genomes; the position 17a (second vertical arrow) is never occupied in the nuclear eukaryotic tDNA. Sequence deviations [detailed in (A)] at positions 10, 19, 53, 55 and 61 are boxed and the tDNA in which they occur are highlighted with grey background in Figure 1. (D) The six possible cases of A-box sequences (according to various occupancies at positions 17, 20a and 20b) written without gap. Various occupancies at positions 20a and 20b generate three possible patterns (no base, only 20a occupied, both 20a and 20b occupied). For each of these patterns, position 17a can be occupied or not, thus generating a total of six possible patterns. (E) Final consensus sequences (not taking into account the exceptions shown in (A and C)) of the A- and B-boxes used to search these elements in the other Pol III genes from the ten genomes (see Figure 6). Note that the fourth nucleotide downstream A₁₄ is always a G (either G₁₈ or G₁₉, twin horizontal arrows).

Structural alignment of the RPR1 and SCR1 RNAs from nine hemiascomycetes and S. pombe. Header lines indicates the A- and B-promoter elements (green and blue background, respectively; white letters for B) and some consensus sequence elements. On the next two lines are displayed the names of helices with the bracket notation: a dot indicates a single stranded nucleotide and brackets, open for the 5′ end and closed for the 3′ end, indicate helices. Regions for which the structure is not specified are represented as single strands (with dots). Sequences of each species are aligned in a phylogenetic order favouring closest homology between neighbour genomes. First column indicates species names. In each sequence, a dash sign (-) indicates a 1 nt gap, whereas the number of nucleotides (in brackets) indicates a longer gap. Underlined pairs of nucleotides in red color indicate that they do not form nor a Watson–Crick or a GU wobble base pair. Bulges are highlighted in light grey and terminal loops in dark grey. Some nucleotides are highlighted in light blue to emphasize the variations occurring only in one sequence. The lowercase letters, highlighted in a darker color, indicate the nucleotides which are different from the consensus sequence of each promoter. The boundaries of the mature products are indicated with ‘5′’ and ‘3′’. With the exception of S.pombe, the Pol III termination signals (poly-T, underlined with blue) are followed by A or G (underlined in pink color) indicating an efficient terminator. (A): RPR1 RNAs, the product of the S.pombe gene (transcribed by Pol II) is not shown. (B): SCR1 RNAs, the genes of the two RNAs from Y.lipolytica are located on chromosomes A and D. The nucleotide abbreviations are given in the legend to Figure 2.