Vol 13, Issue 4, 544-557, April 2003
Comparative Evolutionary Genomics Unveils the Molecular Mechanism of Reassignment of the CTG Codon in Candida spp.
Steven E. Massey1,
Gabriela Moura2,
Pedro Beltrão2,
Ricardo Almeida2,
James R. Garey1,
Mick F. Tuite3 and
Manuel A.S. Santos2,4
1Department of Biology, University of South Florida,
Tampa, Florida 33620, USA; 2Centre for Cell Biology,
Department of Biology, University of Aveiro, 3810-193 Aveiro, Portugal;3
Research School of Biosciences, University of Kent at
Canterbury, Canterbury CT2 7NJ, UK
 |
ABSTRACT
|
|---|
Using the (near) complete genome sequences of the yeasts
Candida albicans, Saccharomyces cerevisiae, and
Schizosaccharomyces pombe, we address the evolution of a
unique genetic code change, which involves decoding of the standard
leucine-CTG codon as serine in Candida spp. By using two
complementary comparative genomics approaches, we have been able to
shed new light on both the origin of the novel Candida spp.
Ser-tRNACAG, which has mediated CTG reassignment, and on the
evolution of the CTG codon in the genomes of C. albicans,
S. cerevisiae, and S. pombe. Sequence analyses of
newly identified tRNAs from the C. albicans genome demonstrate
that the Ser-tRNACAG is derived from a serine and not a
leucine tRNA in the ancestor yeast species and that this codon
reassignment occurred  170 million years ago, but the origin of the
Ser-tRNACAG is more ancient, implying that the ancestral
Leu-tRNA that decoded the CTG codon was lost after the appearance of
the Ser-tRNACAG. Ambiguous CTG decoding by the
Ser-tRNACAG combined with biased AT pressure forced the
evolution of CTG into TTR codons and have been major forces driving
evolution of the CTN codon family in C. albicans. Remarkably,
most of the CTG codons present in extant C. albicans genes are
encoded by serine and not leucine codons in homologous S.
cerevisiae and S. pombe genes, indicating that a
significant number of serine TCN and AGY codons evolved into CTG codons
either directly by simultaneous double mutations or indirectly through
an intermediary codon. In either case, CTG reassignment had a major
impact on the evolution of the coding component of the Candida
spp. genome.
[Supplemental material is available online at http://www.genome.org
and at http://www.bio.ua.pt/genomica/Lab/Genomedata.html. The
following individuals kindly provided reagents, samples, or unpublished
information as indicated in the paper: N. Federspiel.]
Alterations to the standard genetic code have been
found in several organisms and organelle genomes during the past 20
years (for review, see Osawa et al. 1992 ), resulting in intense debate
about their origin and evolution. In an attempt to explain their
occurrence, two mutually exclusive models, (1) the "Codon Capture"
theory (Osawa and Jukes 1989; Jukes and Osawa 1993) and (2) the
"Ambiguous Intermediate" theory (Schultz and Yarus 1994, 1996),
have been put forward in recent years. The Codon Capture theory is a
neutral theory, which postulates that the reassigned codon completely
disappears from the genome under AT or GC pressure. On its
reappearance, a tRNA with a different amino acid identity and an
anticodon complementary to the codon being reassigned takes over the
decoding of the codon. Any decoding ambiguity is excluded. The driving
force is the neutral effect of GC/AT pressure. Conversely, the
Ambiguous Intermediate theory has no requirement for the disappearance
of the codon from the genome. The codon undergoes a transitional stage
in which it is ambiguously decoded. Such ambiguity may be one of
aminoacylation identity (Schultz and Yarus 1994, 1996), anticodon
misreading (Yarus and Schultz 1997), or a competition between two
nonambiguous tRNAs (i.e., codon competition). The driving force in this
case is likely to be a positive benefit resulting from the
reassignment.
An interesting example of a genetic code deviation occurs in
Candida yeasts, in which the nuclear CTG codon codes for
serine rather than the "universal" leucine (Kawaguchi et al. 1989;
Ohama et al. 1993; Santos and Tuite 1995; Sugiyama et al. 1995; Sugita
and Nakase 1999). The reassignment of the CTG codon from leucine to
serine is mechanistically unusual in that it cannot be accomplished via
a single mutation in the anticodon of a serine tRNA (Ser-tRNA). All
other known codon reassignments, with the exception of the reassignment
of the CUN leucine codon box to threonine in yeast mitochondria (Osawa
et al. 1990), may be accomplished via a single mutation of the
anticodon of the tRNA concerned. In the Candida spp., a single
tRNA with a CAG anticodon, the Ser-tRNACAG, decodes the CTG
codon as serine (Yokogawa et al. 1992; Ohama et al. 1993; Santos et al.
1993; Suzuki et al. 1994; Ueda et al. 1994; Sugiyama et al. 1995).
Interestingly, this tRNA is charged with serine, but is also mischarged
with leucine at a 3% rate in vivo in Candida zeylanoides
(Suzuki et al. 1997). Therefore, the CTG codon has the unique property
of being "polysemous" (Suzuki et al. 1997), that is, it codes for
two amino acids, indicating that the CTG codon in some of the extant
Candida species is still ambiguous. This has been taken as
providing supporting evidence for an Ambiguous Intermediate mechanism
for CTG reassignment (Santos et al. 1996; Knight et al. 2001a,b).
However, it questions the origin of the hybrid Ser-tRNACAG.
Here we demonstrate that the ancestor of the Ser-tRNACAG was
a serine tRNA, based on an analysis of newly identified tRNAs from the
Candida albicans genome, tRNAs from other yeast species, and a
consideration of tRNA intron sequences. In addition, we date the codon
reassignment to 170 million years ago using the
Ser-tRNACAG sequences of the Candida spp., and show
that the ancestor of the Ser-tRNACAG originated some time
before the codon reassignment, indicating that CTG evolution should
have been driven by a combination of genome GC pressure and ambiguous
CTG decoding. In addition, a comparative genomics study was carried out
to trace the origin of the 17,000 CTG codons present in the C.
albicans genome and to evaluate both the impact of biased AT
pressure and serine-CTG misreading on the usage of the CTN codons in
C. albicans. Most of the original C. albicans CTG
codons mutated to TTA (27.8%) and TTG (25.3%) leucine codons.
Remarkably, CTG codons present in the C. albicans genome
evolved relatively recently from codons encoding serine or
conserved/semiconserved serines but not leucine codons. Only a minor
fraction (0.2%; total = 102) of the CTG codons present in C.
albicans exist in Saccharomyces cerevisiae, implying that
almost all the original CTG codons disappeared from the C.
albicans genome. In contradiction to the polysemous nature of the
CTG codon, this unexpected observation provides apparent support for a
Codon Capture mechanism for CTG reassignment. However, CTG elimination
cannot be explained by biased AT pressure, raising the possibility that
appearance of the Ser-tRNACAG and consequent serine-CTG
misreading was the determining factor driving CTG almost to extinction.
|
RESULTS
|
|---|
Identity of the Ancestor of the C. albicans Ser-tRNACAG
The tRNAs identified from the C. albicans genome are
displayed at http://www.bio.ua.pt/genomica/Lab/Genomedata.html.
Pairwise alignments were conducted between the Ser-tRNACAG of
C. albicans, Candida cylindracea, Candida
tropicalis, and Candida rugosa and the class II tRNAs of
S. cerevisiae, Schizosaccharomyces pombe, C.
cylindracea, and C. albicans. Without exception, the
Ser-tRNACAGs of C. albicans, C.
cylindracea, C. tropicalis, and C. rugosa possess
a higher mean identity with the serine tRNAs of S. cerevisiae,
S. pombe, C. cylindracea, and C. albicans
than with the leucine tRNAs, indicating that the ancestral tRNA was a
serine tRNA (data not shown). Holmquist et al. (1973) determined that
the average divergence for pairs of eukaryotic tRNAs coding for
different amino acids is 48%. A somewhat higher mean identity is
observed when the Ser-tRNACAG is compared with the leucine
tRNAs of S. pombe, S. cerevisiae, C.
cylindracea, and C. albicans (59%, 60%, 61%, and 59%,
respectively). This is likely indicative of a common origin for serine
and leucine tRNAs, consistent with the observation that the
isoacceptors are unique in eukaryotes for possessing an extra arm.
Using neighbor-joining (NJ) analysis, the C. albicans
Ser-tRNACAG was compared with the class II tRNAs of C.
albicans (Fig. 1A), S. cerevisiae(Fig. 1B), C. cylindracea (Fig. 1C), and S. pombe
(Fig. 1D). In all cases the C. albicans Ser-tRNACAG
grouped with serine tRNAs rather than leucine tRNAs, with varying
degrees of bootstrap support. Thus, the above data indicate that the
ancestor of the C. albicans Ser-tRNACAG was a serine
tRNA. The close relationship between C. albicans
Ser-tRNAAGA and Ser-tRNATGA1 and
Ser-tRNATGA2 (Fig. 1A) indicates that these tRNAs have
undergone a change in anticodon. Such events make it difficult to
predict the specific anticodon of the ancestor of the
Ser-tRNACAG.
Sequence Analysis of tRNA Introns
Sequence alignments reveal that the introns of C. albicans
Ser-tRNACGA and Candida guilliermondii
Ser-tRNACGA are similar to each other and to that of C.
cylindracea Ser-tRNACAG (Table
1). This is evidence for a common ancestry
of the Ser-tRNACAG and Ser-tRNACGA of the
Candida spp. Interestingly, an insertion of A into position 34
of the anticodon of the C. guilliermondii
Ser-tRNACGA produces a CAG anticodon and adds an A to the 5'
end of the intron, found in the C. cylindracea
Ser-tRNACAG intron. The presence of an intron may have been a
predisposing factor for the codon reassignment, because no serine tRNA
can change its anticodon to CAG via a single point mutation. An
insertion in the anticodon loop would produce a loop larger than the
canonical 7 nt, which would likely be detrimental given that the size
of the anticodon loop is highly conserved in tRNAs (Sprinzl et al.
1998). However, the presence of an intron would absorb such an
anticodon expansion, allowing the loop to maintain its size as the
sequence and structure of tRNA introns do not influence splice-site
selection by the S. cerevisiae tRNA splicing endonuclease
(Reyes and Abelson 1988). Hence, a nucleotide insertion (or deletion)
into the intron should not cause any perturbation to the splicing of
introns from the anticodon loop, as already noted by Yokogawa et al.
(1992).
Date of the CTG Codon Reassignment
Using the Ser-tRNACAG sequences to construct an NJ tree
(Fig. 2), a date of at least 171 ± 27
million years ago was obtained for the emergence of the codon
reassignment. Using SSU rRNA sequences to construct an NJ tree, a date
of 178 ± 19 million years ago was obtained for the codon
reassignment (Fig. 3), which agrees with
the date obtained from the Ser-tRNACAG tree. The date is also
in agreement with a figure of 150 million years ago, derived using the
maximum likelihood method and SSU rRNA sequences (Pesole et al. 1995).
The discrepancy is probably due to the use of a limited data set in the
previous analysis as C. cylindracea was omitted, which may
cause an underestimation of the divergence date. Overall, the evidence
points to the codon reassignment having occurred during the Jurassic
era. It should be noted that saturation is reached at the base of the
tree, and this would lead to an underestimation of the time of
divergence of the outgroup, that is, Homo sapiens
Ser-tRNAGCT.
A tree of the C. albicans serine tRNAs indicates that the
Ser-tRNACAG originated at least 272 ± 25 million years ago
(Fig. 4B), which is considerably more
ancient than the date for the codon reassignment. The probability that
the codon reassignment and appearance of the Ser-tRNACAG
occurred at approximately the same time, that is, within 10 million
years of each other, is low, with p < 0.0006. Therefore, a
major new feature of this work is that the appearance of the
Ser-tRNACAG predates the reassignment event by a significant
amount of time. Some caution should be expressed when assigning
absolute dates in the absence of a fossil record; however, the dates
are valuable in indicating the ancient origin of the tRNA compared with
the CTG reassignment event.
Apparently, all copies of the ancestral tRNA(s) were converted into
Ser-tRNACAG, or remaining copies were lost from the genome.
Therefore, a close relative of the tRNA is not present in the genomes
of C. albicans or C. cylindracea (Fig. 1A,C).
Interestingly, considerations of the class II tRNAs of S.
cerevisiae do not indicate a close relative either (Fig. 1B),
implying that the homolog of the ancestral tRNA was lost from this
genome also after the divergence of S. cerevisiae from C.
albicans.
The agreement of the SSU rRNA and Ser-tRNACAG phylogenetic
trees for the date of the codon reassignment indicates that the
Ser-tRNACAG is evolving at a rate comparable to that of the
tRNAs used to calibrate the molecular clock. Thus, the reassignment is
complete and the other CTN codons, especially the CTA codon, are
"safe" from further reassignments, that is, the evolutionary
imperative behind the CTG reassignment is not driving further change.
The Effect of Genome GC Pressure on Codon Usage in C. albicans
Ambiguous CTG decoding provides compelling evidence for its
reassignment through a misreading mechanism as proposed by the
Ambiguous Intermediate theory (Schultz and Yarus 1994). However, this
should have a negative impact on the organism's fitness and therefore
prompts the question of whether biased GC pressure reduced CTG usage to
a bearable minimum prior to reassignment (Jukes and Osawa 1993). To
shed new light on this important question, a comparative study of the
effect of GC pressure on the coding component of C. albicans,
S. cerevisiae, and S. pombe genomes was carried out.
The total GC content is similar in the three species (Table
2); however, C. albicans shows the
lowest GC content at the codon third position. A comparative analysis
of codon usage carried out for the three species shows that, with very
few exceptions, codon usage follows a similar trend, which is
represented by the usage of the serine codon family shown in Figure
5B. However, leucine codons do not follow
this overall trend, in particular, significant differences in frequency
are found for the CTA, CTC, and CTG codons, whose usage is clearly
repressed in C. albicans in relation to S. cerevisiae
and S. pombe (Fig. 5A). That CTG and CTA usage is repressed
whereas TTG usage is favored in C. albicans indicates that AT
pressure alone is not the main force driving the evolution of the
leucine codons. One alternative explanation is that tRNA selection is
responsible for this effect, as is the case in many organisms (Osawa
1995). This is also in line with the observation that S.
cerevisiae, S. pombe, and C. albicans decode the
CTN codon family with a rather different set of anticodons, that is,
GAG and TAG for S. cerevisiae; AAG, TAG, and CAG for S.
pombe; and AAG for C. albicans. Because the AAG anticodon
in C. albicans decodes the CTC, CTT, and CTA codons using
extended wobbling, it is likely that the weak interaction between the
AAG anticodon and the CTA/C/T codons, in particular with CTA, is an
important factor in reducing CTN usage in C. albicans, as is
observed for two-codon sets in highly expressed genes in most organisms
(Ikemura 1981a,b; Ohama et al. 1990; Osawa 1995).
View this table:
[in this window]
[in a new window]
|
Table 2. Decreased GC Content at the Third Codon Position in Candida
Species That Reassigned the CTG Codon From Leucine
to Serine
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Figure 5. Usage of leucine and serine codons in Candida albicans,
Saccharomyces cerevisiae, and Schizosaccharomyces
pombe. (A) Of the six leucine codons, TTA and TTG are the
most frequently used by the nuclear genomes of the three yeast species,
the exception being CTT in S. pombe. In C. albicans,
the usage of CTN codons, in particular CTC and CTA but also CTG codons,
is repressed in relation to the same codons in the other two yeasts,
whereas usage of the CTT codon is similar between C. albicans
and S. cerevisiae. The usage of the CTA codon is also
repressed in C. albicans, indicating that other forces apart
from genome AT pressure shape CTN usage in this species. (B)
The bias in C. albicans CTN usage is not observed for serine
codons, whose distribution follows the expected pattern for a genome
with low GC content in coding sequences (Table 2). The values indicated
are relative to the total synonymous codon count for each genome and
independent of amino acid frequency.
|
|
Considering that the three yeast species being studied have similar
total GC content and rather conserved tRNA populations (data not
shown), the exceptions being higher AT content at N3 and the Leu-tRNA
isoacceptors that decode the CTN codon family in C. albicans,
specific effects of biased AT pressure at N3 and tRNA selection on CTN
usage should be unveiled by determining the relative codon usage
frequency for genome pairs, that is, C. albicans versus
S. cerevisiae and C. albicans versus S.
pombe (Fig. 6A,B). The increase in AT
pressure at N3 in C. albicans is clearly visible for
frequently used codons in the C. albicans/S.
cerevisiae pair, where C. albicans clearly prefers A- and
T-ending codons. Interestingly, the Leu-TTG codon, the Val-GTG, and the
Gly-GGG codons are exceptions to this rule and imply that translational
selection is an important factor modulating usage of these codons in
C. albicans. When a similar analysis is carried out for the
C. albicans/S. pombe pair, the same trend is
observed, but the number of exceptions increases for C-ending and
rarely used codons (Fig. 6B), thus further supporting the hypothesis
that both AT pressure and tRNA selection modulate codon usage in
C. albicans (Fig. 6A,B).
View larger version (55K):
[in this window]
[in a new window]
|
Figure 6. Candida albicans prefers A- and T-ending codons. Codon usage
in C. albicans, Saccharomyces cerevisiae, and
Schizosaccharomyces pombe was analyzed by counting the total
number of codons for each species. To determine the codon preference
between two species, the relative frequency of usage of each codon was
divided by the corresponding value for the same codon in the other
species. The codon usage ratios obtained for C.
albicans/S. cerevisiae (A), C.
albicans/S. pombe (B), and S.
cerevisiae/S. pombe (C) were plotted as indicated
in the graph, with ratio values above 1 (upper part in the
graph) indicating the preferred codons in C. albicans in
relation to the other two species (A and B) and in
S. cerevisiae in relation to S. pombe (C).
The C. albicans/S. cerevisiae codon ratios indicate
that C. albicans prefers highly used codons (green bars)
ending with A and T, with the exception of Leu-TTG and Gly-GGG. For the
C. albicans/S. pombe pair, the preference for
frequently used A- and T-ending codons is maintained, but four
frequently used C-ending codons, Asn-AAC, Thr-ACC, Ile-ATC, Phe-TTC,
and also the G-ending codons Leu-TTG, Val-GTG, and Gly-GGG, are
preferred. Therefore, the data indicate that in C. albicans
the effect of AT pressure is more visible at the third codon position
for highly used codons. That the CTG codon is used at low frequency in
C. albicans implies that AT pressure was not the main force
driving its reassignment to serine. Green and brown bars indicate
highly and rarely used codons, respectively.
|
|
A deeper insight into the effect of GC pressure on codon usage can be
gained by aligning C. albicans and S. cerevisiae
homologous genes and determining the content of A and T at N3 for each
homologous codon. That is, for each S. cerevisiae codon in a
double alignment, the nucleotide present at the corresponding N3
position in C. albicans can be determined and compared with
all other codons for the complete set of homologous genes (Fig.
7). C. albicans prefers A- and
T-ending codons without exception, and G- and C-ending codons are
repressed. However, leucine codons show a significant deviation to this
trend in that there is a clear relative increase in G3. This is mainly
achieved by decreasing C3 and T3. A similar trend is observed for the
arginine codon family, but in this case there is a relative increase in
A3 and the G3 effect is not so visible (Fig. 7). Thus, leucine codons
have a slight bias favoring purine- and repressing pyrimidine-ending
codons instead of favoring A- and T-ending leucine codons, which
contradicts the general trend of an increased frequency of A3 and T3.
Considering that this is a rather localized effect, the most likely
explanation for it is that the weak interaction between the AAG
anticodon and CTA, CTC, and CTT codons and the disappearance of the CAG
or TAG anticodons from the C. albicans genome drove a massive
conversion of CTN into TTA and TTG codons, which are decoded by a
strong interaction with cognate TAA and CAA anticodons, respectively.
View larger version (31K):
[in this window]
[in a new window]
|
Figure 7. GC pressure alone at the third codon position does not explain the
evolution of leucine codons in Candida albicans. To elucidate
the role of GC pressure in the evolution of C. albicans ORFs,
an analysis of GC3 pressure (GC pressure at the third codon position)
was carried out by comparing the complete ORFs set of both genomes
using BLASTP at an E-value of 10–5. For each
Saccharomyces cerevisiae codon, the corresponding N3 codon
position was identified in C. albicans ortholog genes. The
data show that S. cerevisiae codons are represented in C.
albicans mainly by A- or T-terminating codons (yellow and light
blue bars), in agreement with high AT pressure at the third codon
position in the C. albicans genome (N3 = 71% AT; Table 2).
However, the leucine codons show a small deviation from the pattern
observed for all other codons. That is, they change more frequently
than expected into G-terminating codons (blue bar), thus contradicting
the high AT pressure at the N3 position observed in C.
albicans genes. This increase in G3 is matched by a slight increase
in A3, indicating an increase in purines at N3. This is achieved by
decreasing the usage frequency of C- and T-ending codons. That is, when
compared with all other codons, there is a relative increase in purines
at the N3 position in C. albicans genes instead of an increase
in AT-ending codons, as would be expected from the relative increase of
AT pressure at N3 in the C. albicans genome, thus indicating
that other forces apart from GC pressure shaped the evolution of
leucine codons in the latter.
|
|
Leucine-Encoding CTG Codons Have Disappeared From the C. albicans Genome
The observation that the Ser-tRNACAG appeared prior to CTG
reassignment combined with the unusual evolutionary pattern of the
leucine codons raises the interesting possibility that the
Ser-tRNACAG itself shaped the evolution of leucine codons,
and in particular the CTG codons. That is, the negative pressure
imposed on CTG codons by incorporation of serines instead of leucines
might have played a major role in the evolution of all six leucine
codons. To clarify this question, the complete set of C.
albicans and S. cerevisiae homologous genes were aligned
as described in Methods, and for each S. cerevisiae amino acid
the codons present in the homologous positions in C. albicans
genes were computed. By plotting the percentage of the amino acid found
in the S. cerevisiae genome for each codon present at
homologous positions in the C. albicans genome, an overall
picture of amino acid conservation between the two species is obtained.
As expected, the level of amino acid conservation between the two
species is very high (Fig. 8A,B). If a
similar analysis is carried out, but instead of comparing amino acids
in S. cerevisiae with codons at homologous positions in
C. albicans homologous genes, the CTG codons present in the
former are compared with codons present in the latter, an evolutionary
pattern for the CTG codon between the two species emerges (red line in
Fig. 8A,B). Remarkably, almost all CTG codons present in the S.
cerevisiae genome are represented by leucine-TTG and -TTA codons in
C. albicans. Only 0.2% of the total number of CTG codons
remain conserved at homologous positions (Fig. 8B).
View larger version (46K):
[in this window]
[in a new window]
|
Figure 8. The majority of the CTG codons present in the Saccharomyces
cerevisiae genome are encoded by TTG and TTA codons in the
Candida albicans genome. (A) To quantify the relative
conservation of amino acids and codons between C. albicans and
S. cerevisiae, the two genomes were aligned using BLASTP at an
E-value of 10–5. For each S. cerevisiae
amino acid, the corresponding codons present at homologous positions in
C. albicans orthologs were identified, thus providing overall
information about amino acid and codon conservation between the two
species. To elucidate the mutational pattern of the CTG codon between
the two species, its frequency of conversion at each position of the
alignment was computed independently of the other leucine codons (thick
red line in the graph). As would be expected, the major trend is
residue conservation between the two species at each position for each
respective codon family. For the CTG and the other leucine codons, two
important trends are observed: first, their conversion into leucine TTG
and TTA codons and also into conserved amino acids of leucine (Ile,
Met, and Phe); second, the avoidance of nonconserved leucine, namely,
serine codons. As observed in Figures 6 and 7, the preferential
conversion of CTN codons into TTG does not follow the rules imposed by
increased AT pressure in C. albicans as TTG is a G-ending
codon, thus indicating that translational selection is a strong driving
force in the evolution of the CTN codons in C. albicans.
(B) Magnification of the previous graph in the regions of
serine and leucine codons (boxes A and B, respectively).
|
|
The above results only provide a partial picture for the evolution of
the CTG codon in that the alignments are not reciprocal and
consequently do not show how the 17,000 CTG codons present in C.
albicans are represented in the S. cerevisiae genome. For
this, the converse comparative analysis was carried out. C.
albicans residues in the aligned gene set were fixed, and the
corresponding codons present at homologous positions in the S.
cerevisiae genome were computed (Fig.
9A,B). The C. albicans CTG codons
were then fixed in the alignment, and the codons present in the S.
cerevisiae genome were identified. Most strikingly, almost none of
the C. albicans CTG codons are represented by leucines in
the S. cerevisiae genome; instead, serines appear at
homologous positions, indicating that almost all C. albicans
CTG codons evolved recently from serine or conserved serine codons
(Fig. 9B).
Considering that evolution of the CTG codon from serine codons is
rather surprising, because of the need for two mutations to convert a
serine or conserved serine codon into a CTG codon, a triple alignment
between C. albicans, S. cerevisiae, and S.
pombe genomes was carried out to remove the background noise
inherent in a double genome alignment and consequently improve data
quality. The results of the triple alignment are a clear-cut
confirmation of the double alignment (Fig.
10). That is, C. albicans CTG
codons are represented in the S. cerevisiae and S. pombegenomes by serine and not leucine codons. There is a very low level
of conserved leucine codons and an even lower number of conserved CTG
codons in the three genomes. That none of the other leucine codons
shows this trend highlights the unique evolutionary pathway for CTG
codons in the C. albicans genome (Fig. 10). The residual
number of conserved CTGs present in the three genomes indicates that
they are likely to be located at positions that tolerate leucine or
serine residues by C. albicans proteins.
View larger version (31K):
[in this window]
[in a new window]
|
Figure 10. A high percentage of Candida albicans CTG codons are
represented by serine residues in the Saccharomyces cerevisiae
and Schizosaccharomyces pombe genomes. To determine the origin
of leucine codons present in the C. albicans genome, the
percentage of serine and leucine residues present simultaneously in the
S. cerevisiae and S. pombe genomes was determined for
each of the six leucine codons present in the C. albicans
genome. For this, the three genomes were compared using the BLASTP
program as in Figures 8 and 9. The conservation of CTN codons between
C. albicans and S. cerevisiae/S. pombe is
very low and reaches the lowest value for CTG codons (0%).
Interestingly, 14% of the C. albicans CTG codons encode
serine residues and only 0.7% encode leucine residues in S.
cerevisiae and S. pombe. This trend is not observed for
any other leucine codon; leucine is always highly conserved in
homologous S. cerevisiae/S. pombe genes.
|
|
|
DISCUSSION
|
|---|
The occurrence of genetic code changes raises two important
evolutionary questions, namely: How do genetic code alterations evolve?
and, Why do they evolve? The reassignment of the leucine-CTG codon to
serine in Candida spp. is a paradigm genetic code change whose
study has already unveiled several intricate and subtle evolutionary
forces involved in the evolution of genetic code changes (Santos et al.
1993, 1996, 1999; Perreau et al. 1999). The present evolutionary study
sheds important new light on the evolution of CTG reassignment by
showing that a combination of the evolutionary mechanisms postulated by
both the Codon Disappearance and Ambiguous Intermediate theories has
driven CTG reassignment. That is, increased AT pressure at the third
codon position increases usage of A- and T-ending codons; however, the
appearance of the Ser-tRNACAG, which decodes the CTG codon as
serine, has been the major force repressing CTG, and perhaps
CTA/CTC/CTT, and increasing TTR codon usage.
The Evolution of the Ser-tRNACAG
A serine ancestry for the Ser-tRNACAG is consistent with
structural probing analyses that demonstrate that the structure of
C. albicans Ser-tRNACAG is that of a typical serine
tRNA (Perreau et al. 1999). A serine ancestry for the
Ser-tRNACAG indicates that the suggested mechanism of tRNA
mis-aminoacylation (Schultz and Yarus 1994, 1996) has not occurred in
this case, despite the polysemous nature of the CTG codon (discussed
further below). Therefore, if codon ambiguity were a part of the
mechanism of reassignment, then it was the result of either the
decoding ambiguity of the Ser-tRNACAG, mediated via an
ambiguous codon–anticodon interaction, or "codon competition"
between a cognate leucine and a cognate serine tRNA. Therefore, the
ancestor of the Ser-tRNACAG may have initially decoded a
serine codon, but also the CTG codon, to a minor extent, via an
ambiguously decoding anticodon. The leucine tRNA that originally
decoded the CTG codon would have been lost from the genome of C.
albicans, consistent with the observation that the S.
pombe Leu-tRNACAG bears no similarity to the C.
albicans Ser-tRNACAG (Fig. 1D).
The CTG Codon is Polysemous by Default
A problem with proposing a serine ancestry for the
Ser-tRNACAG is to explain the presence of m1G37 in
the Ser-tRNACAG, which in Candida zeylanoides is
responsible for mischarging of the tRNA with leucine (Suzuki et al.
1997), resulting in the polysemous nature of the CTG codon. The
proposal has been made that the m1G37 evolved to ameliorate
the detrimental effects of reassigning the CTG codon from leucine to
serine (Suzuki et al. 1997). In addition, m1G37 could be
viewed as a remnant of a putative leucine ancestry of the tRNA. The
presence of m1G37 in the Ser-tRNACAG of the
Candida spp. is likely an adaptive mutation that occurred
after the anticodon of the tRNA had mutated to CAG, because
m1G37 is found in tRNAs reading C Y/G N codons from
Eubacteria, Archaea, and Eukaryotes (Bjork 1986; Bjork et al. 1987),
whereas serine tRNAs possess a modified A37. The Ser-tRNACAG
reads a codon (CTG) that belongs to the C Y/G N group. m1G37
appears to have a role in maintaining the fidelity of codon–anticodon
interaction (Bjork et al. 1989, 2001; Hagervall et al. 1990, 1993; Li
and Bjork 1995; Li et al. 1997; Urbonavicius et al. 2001). We suggest,
therefore, that m1G37 is required for the efficient decoding
of the CTG codon and that the low level of leucylation is a by-product,
not an imperative, for the mutation. m1G37 therefore arose as
an adaptive mutation after the reassignment occurred, rather than as a
facilitative mutation that allowed the reassignment to occur. The
presence of m1G37 is probably mildly detrimental to the
yeast, but less detrimental than retaining A37. Further evidence for
the above hypothesis is provided by the reassignment of the entire
leucine CTN codon box to threonine in yeast mitochondria. This is the
only other codon reassignment, apart from the Candida codon
reassignment, that involves a change in identity of the 37 nucleotide
(N37) in the anticodon loop of the associated tRNA. During the course
of the reassignment, A37, typical of threonine tRNAs (Sprinzl et al.
1998), has mutated to m1G37 in the threonine tRNA responsible
for decoding the CTN codons.
An explanation for the misacylation of the C. zeylanoides
Ser-tRNACAG with leucine, rather than with any of the other
18 amino acids, lies in the observation that, as class II tRNAs,
Ser-tRNA and Leu-tRNA are structurally similar. Therefore, the
Ser-tRNACAG is most likely to be misrecognized by leucyl-tRNA
synthetase, rather than the other 18 aminoacyl-tRNA synthetases that
charge class I tRNAs. Experimental evidence supports this assertion;
mutations in S. cerevisiae (Soma et al. 1996) and human
(Breitschopf et al. 1995) Ser-tRNAs result in misacylation with
leucine. Likewise, mutations in S. cerevisiae (Himeno et al.
1997) and human (Breitschopf and Gross 1994) Leu-tRNAs result in
misacylation with serine. Hence, we propose that the nature of the
polysemous codon results from the process of reassignment, rather than
being an integral part of the mechanism.
The Evolution of the CTG Codon
The discovery that the Ser-tRNACAG appeared before CTG
reassignment is in agreement with the finding that GC pressure alone
cannot explain the evolutionary pattern of the CTN codon family in
C. albicans and most likely in the other Candida
species that reassigned the CTG codon. This is particularly clear if
one considers the relative preference for G- and A-ending codons (N3
position) in C. albicans in relation to S. cerevisiae
leucine codons (Fig. 7). That is, C. albicans has a relative
preference for purine-ending rather than pyrimidine-ending codons in
the leucine codon family. The A3/G3 preference can be explained by the
strong directional mutational bias of CTN codons into TTA and TTG
codons, and the G3 preference is mainly due to conversion of CTN codons
into TTG codons. However, if one considers the higher AT content at N3
in C. albicans codons, one would expect a biased preference
for TTA and not TTG as is the case for the six arginine codons (Fig.
7).
Another important discovery emerging from the study presented herein
relates to the nature of the 17,000 CTG codons present in extant
C. albicans. It is remarkable that almost all CTG codons
present in the ancestral yeast species disappeared from the C.
albicans genome and that the new codons evolved from serine or
conserved serine codons whose conversion into CTG codons require at
least two mutations. Alternatively, the conversion of serine codons
into CTG codons may have occurred via doublet mutations, such as
described for the conversion of serine-AGY codons into serine-TCN
codons (Averof et al. 2000). If taken separately, this could be
considered as providing strong support for the "Codon
Disappearance" theory proposed by Jukes and Osawa (1993); however,
the unique evolutionary pattern of the CTN codons in C.
albicans and the early appearance of the Ser-tRNACAG in
the Candida spp. imply a novel evolutionary mechanism in which
the ambiguous decoding of the CTG codon by the newly created
Ser-tRNACAG was the main force driving CTG codons to
extinction. It is likely that increased AT pressure at the N3 codon
position also played an important role. If so, CTG (CTN) evolution has
been driven by a combination of the negative pressure imposed by CTG
misreading and biased AT pressure at the N3 codon position.
Should the Ambiguous Intermediate Theory Be Reformulated?
The Ambiguous Intermediate theory originally proposed by Schultz and
Yarus (1994) postulates that codon reassignment is driven by a
mechanism, which requires ambiguous decoding of the codon being
reassigned by tRNAs with expanded decoding properties. Briefly, the
theory postulates that a mutant tRNA with expanded decoding properties
starts decoding a codon belonging to a noncognate amino acid codon
family, making it ambiguous. The theory proposes that this ambiguity
increases because of additional mutations in the newly created tRNA up
to a point at which the mutant tRNA efficiently decodes the ambiguous
codon, which can then acquire a new meaning. The implication of this
theory is that, contrary to expectation, codon ambiguity provides some
sort of selective advantage to drive codon reassignment to completion.
That the CTG and TGA codons are ambiguous in C. albicans and
Bacillus subtilis, respectively (Lovett et al. 1991; Suzuki et
al. 1997) and that reconstruction of the C. albicans CTG
reassignment in S. cerevisiae showed that the latter has a
selective advantage under some stress conditions provide support for a
reassignment mechanism through ambiguous decoding in which the
environment may play a significant role (Santos et al. 1999).
A recent study carried out in mitochondria showed a good correlation
between codon reassignment and codon disappearance; that is, of the 11
codons that have been reassigned in some lineages, 8 have disappeared
in some other lineage (Knight et al. 2001a). However, no significant
association between codons predicted to disappear by mutation pressure
and reassigned codons was found. Rather, some codons apparently
disappear for reasons unrelated to mutational bias, and these are the
ones that are more like to be reassigned (Knight et al. 2001a). Our
data provide the missing link between codon reassignment and
disappearance by showing that codon misreading by mutant tRNAs is an
important evolutionary force driving codons to extinction. That most
codon reassignments are associated with tRNA mutations, editing,
altered tRNA modification, and release factor mutations (Knight et al.
2001a,b) provides strong support for this hypothesis, which
unveils a novel mechanism for codon reassignment through
codon ambiguity and supports a reformulation of the original Ambiguous
Intermediate theory, as follows:
Mutant tRNAs with expanded decoding properties, or mutations in the
translational machinery, which create decoding ambiguity, in
conjunction with biased GC pressure, impose a negative selective
pressure on particular codons, decreasing their usage to very low
levels or forcing them to disappear from the genome. Some overall
positive advantage must be present in order for this process to occur.
These codons can be gradually reassigned through structural change of
the translational machinery and loss of the ancestral cognate
tRNAs.
The novelty of the reformulated theory relies on the disappearance
or reduction of codon usage to a tolerable minimum caused by ambiguous
decoding and not GC pressure only. Despite introducing a new mechanism
for codon disappearance, the reformulated theory still relies on
ambiguous decoding as a mechanism for codon reassignment and,
therefore, we consider that the Ambiguous Intermediate theory should
maintain its original name.
Conclusions
This study shows that the Ser-tRNACAG evolved from a
serine and not a leucine tRNA and that the CTG codons present in extant
Candida species are new codons, which have evolved recently.
This study also unveils a hidden aspect of Candida biology,
that of having a very unstable proteome during the last 272 million
years or so. This instability clearly has important consequences at the
phenotypic level, and one is urged to unravel its physiological and
evolutionary meaning. This instability arises from three distinct
mechanisms: (1) initial ambiguous decoding of the CTG codon by the
newly created Ser-tRNACAG, forcing its disappearance; (2)
evolution of serine or conserved serine codons into CTG codons through
an intermediary codon; and (3) ambiguous charging of the
Ser-tRNACAG by both the seryl- and leucyl-tRNA synthetases.
|
METHODS
|
|---|
Genome Data Retrieval and Selection
Sequence data for C. albicans was obtained from the
Stanford DNA Sequencing and Technology Centre Web site at
http://www-sequence.stanford.edu/group/candida. Sequencing of the
C. albicans genome was accomplished with support from the
National Institute of Dental and Craniofacial Research and the
Burroughs Wellcome Fund.
The sequences are of 10x mean coverage. Complete genome sequence data
builds for S. cerevisiae (May 01, 2002) and S. pombe(May 16, 2002) were downloaded from GenBank
(ftp://ftp.ncbi.nih.gov).
Sequence data were filtered before any analysis was conducted. All open
reading frames (ORFs) were scanned for irregularities in sequence size,
start and stop codon positions, and invalid character usage. The total
of validated ORFs for each genome is 8467, 6313, and 4659 for C.
albicans, S. cerevisiae, and S. pombe,
respectively. The C. albicans serine and leucine tRNA genes
were identified by homology searching of the database, using the BLASTN
program and the S. cerevisiae serine and leucine tRNA gene
sequences. Introns were assigned using the tRNAscan-SE 1.1 program
(Lowe and Eddy 1997). S. cerevisiae, S. pombe, and
C. cylindracea tRNA gene sequences were obtained from GenBank.
RNA Sequence Alignments
The tRNA sequences were manually aligned, excluding introns and
ensuring that the stems, anticodons, and invariant nucleotides (U8,
A14, G18, G19, A21, G53, U54, U55, and C61) were aligned. The
percentage identity of two tRNAs was calculated as the number of
identical nucleotides divided by the nucleotide length of the longest
of the two tRNA molecules. The 5'-CCA terminus of mature tRNAs was not
included in any of the sequence analyses. Small subunit (SSU) rRNA
sequences were aligned using the DCSE program (De Rijk and De Wachter
1993), taking into account secondary structure information. The
alignments used in this publication are available at
http://chuma.cas.usf.edu/garey/alignments/alignment.html.
Tree Construction
NJ trees using Kimura 2-parameter distances and  corrections for
site-to-site variation were constructed using the MEGA 2.1 program
(Kimura 1980; Kumar et al. 2001). The PAUP 4.0 program (Swofford 1993)
was used to calculate the shape parameter using maximum likelihood.
Tree Calibration and Estimates of Divergence Times
The trees were linearized under the molecular clock assumption
using the MEGA 2.1 program and the method of Nei and Kumar (2000) to
obtain estimates of divergence times. The tRNA nucleotide substitution
rate was estimated by considering tRNAs from C. albicans,
S. cerevisiae, and S. pombe, using the respective
tRNAs from H. sapiens as out-groups. The tRNAs considered were
aspartate, asparagine, cysteine, phenylalanine, histidine,
tryptophan, and tyrosine. These tRNAs were chosen because each of the
tRNAs specific for these amino acids possesses a single isoacceptor in
each genome; thus, the common ancestry of each set of tRNAs was
unambiguous. The C. albicans tRNA sequences used are listed at
the Web site http://www.ukc.ac.uk/bio/tuite/research/tRNA.htm. The
S. cerevisiae, S. pombe, and H. sapiens
tRNAs used are listed at the Web site http://rna.wustl.edu/GtRDB/. The
tRNA sequences were concatenated into a single sequence for each
species. A shape parameter of 0.578 was estimated using the maximum
likelihood method in PAUP, and a linearized NJ tree was constructed.
A nucleotide substitution rate of 1.156 x 10–9
substitutions per site per year was estimated from the NJ tree using
420 million years ago for the divergence of S. cerevisiae
from S. pombe (Lum et al. 1996). This tRNA nucleotide
substitution rate was used to estimate the divergence times of key
genes within the linearized tRNA trees displayed in Figures 2 and 4.
The divergence times of key genes within the linearized ribosomal RNA
tree (Fig. 3) were estimated by calibrating the rRNA tree using 420
million years ago for the divergence of S. cerevisiae and
S. pombe.
To test the statistical significance of the key divergence times, we
calculated the standard error of branch lengths of the genes that were
used to estimate the respective divergence times. These standard errors
were used to estimate the probability that the codon reassignment and
origin of the Ser-tRNACAG did not occur at the same time
(i.e., within 10 million years of each other), using approximations to
the normal distribution.
Codon Usage Analysis
A platform for protein sequence data manipulation and analysis was
built on Perl (ActivePerl 5.6.1; http://www.activeperl.com) using
BioPerl 1.0 (http://www.bioperl.org) and consists of a series of
programs and CGI scripts. The three groups of validated sequences were
introduced separately, and each ORF was analyzed according to its codon
usage. A range of several statistical values was generated at three
different levels of genome comparison: (1) distribution of individual
codons, (2) dependencies between groups of synonymous codons, and (3)
differences in amino acid usage.
Homolog Alignments
A stand-alone BLAST (Altschul et al. 1997; version 2.2.2,
http://www.ncbi.nlm.nih.gov/BLAST/) was used to find S.
cerevisiae–C. albicans homologous proteins. All protein
pairs that were reciprocal best hits (Rivera et al. 1998) with a BLAST
expected value lower than 10–5 were included in this study.
In all, 2899 pairs of homologous proteins that matched these criteria
were found. The C. albicans genes in this homolog set have a
total of 6084 CTGs (40% of the genome total), an average size of
533.68 codons, and an average of 3.52 CTGs per thousand (the genome
average is 6.97 CTGs per thousand). The protein homologs were aligned
with T-Coffee (Notredame et al. 2000;
http://igs-server.cnrs-mrs.fr/cnotred/Projects-home-page/t-coffee-home-page.html).
N- or C-terminal extensions or insertions in the protein sequence of
either species that lacked a mirror residue in the homologous sequence
(gaps) were not included in the analysis. The alignments around gaps
were analyzed to reduce possible ambiguity. Based on a manual analysis
of 30 alignments, a set of sequential rules was determined to formalize
our definition of ambiguous regions. We consider ambiguous: (1) all the
aligned portions between a gap and a perfect match; (2) all aligned
segments of at least five amino acids that surrounded ambiguous regions
with at least 20% negative score or; (3) any window of 15 amino acids
with one third negative alignment.
All amino acids and their respective codons within S.
cerevisiae proteins and genes were compared with their aligned
residues in the homologous C. albicans protein and DNA
sequences. Codon conservation for each species was determined from the
alignments and scored according to the codon or amino acid residue
present at the same position in the other species and its respective
third-position nucleotide (N3). A triple alignment was also performed
between C. albicans, S. cerevisiae, and S. pombeproteins, using the same method already described, to determine the
origin of the leucine codons present in modern day C. albicans
strains.
|
WEB SITE REFERENCES
|
|---|
ftp://ftp.ncbi.nih.gov; GenBank.
http://chuma.cas.usf.edu/garey/alignments/alignment.html; small
subunit rRNA sequences aligned in this paper using the DCSE program.
http://igs-server.cnrs-mrs.fr/cnotred/Projects-home-page/t-coffee-home-page.html;
T-Coffee.
http://rna.wustl.edu/GtRDB/; the S. cerevisiae, S.
pombe, and H. sapiens tRNAs used in this paper.
http://www.activeperl.com; ActivePerl 5.6.1.
http://www.bioperl.org; BioPerl 1.0.
http://www.bio.ua.pt/genomica/Lab/Genomedata.html; tRNAs identified
from the C. albicans genome.
http://www.ncbi.nlm.nih.gov/BLAST/; BLAST.
http://www-sequence.stanford.edu/group/candida; Stanford DNA Sequencing
and Technology Centre.
http://www.ukc.ac.uk/bio/tuite/research/tRNA.htm; C. albicans
tRNA sequences used in this paper.
|
Acknowledgements
|
|---|
The research reported here was supported by project grants to
M.A.S.S. from the Portuguese Foundation for Science and Technology
(FCT; Projects POCTI REFs: BME/32938, 32942/99 and BME/39030/01) and
the EMBO YIP Programme, and to M.F.T. and M.A.S.S. from the
Biotechnological Biological Research Council (BBSRC). G.M. is supported
by a postdoctoral research fellowship (SFRH/BPD/7195/2001) from FCT. We
thank Nancy Federspiel for her kind permission to use the C.
albicans genome data. Finally, we are most thankful to Justin
O'Sullivan for his helpful discussions and input into this
study.
The publication costs of this article were defrayed
in part by payment of page charges. This article must therefore be
hereby marked "advertisement" in accordance with 18 USC section
1734 solely to indicate this fact.
|
Footnotes
|
|---|
4 Corresponding author. 
E-MAIL msantos{at}bio.ua.pt; FAX 351234426408.
Article and publication are at
http://www.genome.org/cgi/doi/10.1101/gr.811003.
|
REFERENCES
|
|---|
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.J. 1997. Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res. 25: 3389-3402.[Abstract/Free Full Text]
Averof, M., Rokas, A., Wolfe, K.H., and Sharp, P.M. 2000. Evidence for a high frequency of simultaneous double-nucleotide substitutions. Science 287: 1283-1286.[Abstract/Free Full Text]
Bjork, G.R. 1986. Transfer RNA modification in different organisms. Chem. Scripta 26: 91-95.
Bjork, G.R., Ericson, J.U., Gustafsson, C.E., Hagervall, T.G., Jonsson, Y.H., and Wikstrom, P.M. 1987. Transfer RNA modification. Annu. Rev. Biochem. 56: 263-287.[CrossRef][Medline]
Bjork, G.R., Wikstrom, P.M., and Bystrom, A.S. 1989. Prevention of translational frameshifting by the modified nucleoside 1-methylguanosine. Science 244: 986-989.[Abstract/Free Full Text]
Bjork, G.R., Jacobsson, K., Nilsson, K., Johansson, M.J., Bystrom, A.S., and Persson, O.P. 2001. A primordial tRNA modification required for the evolution of life? EMBO J. 20: 231-239.[CrossRef][Medline]
Breitschopf, K. and Gross, H.J. 1994. The exchange of the discriminator base A73 for G is alone sufficient to convert human tRNALeu into a serine-acceptor in vitro. EMBO J. 13: 3166-3167.[Medline]
Breitschopf, K., Achsel, T., Busch, K., and Gross, H.J. 1995. Identity elements of human tRNALeu: Structural requirements for converting human tRNASer into a leucine acceptor in vitro. Nucleic Acids Res. 23: 3633-3637.[Abstract/Free Full Text]
De Rijk, P. and De Wachter, R. 1993. DCSE, an interactive tool for sequence alignment and secondary structure research. Comput. Appl. Biosci. 9: 735-740.[Abstract/Free Full Text]
Hagervall, T.G., Ericson, J.U., Esberg, K.B., Li, J.M., and Bjork, G.R. 1990. Role of tRNA modification in translational fidelity. Biochim. Biophys. Acta 1050: 263-266.[Medline]
Hagervall, T.G., Tuohy, T.M., Atkins, J.F., and Bjork, G.R. 1993. Deficiency of 1-methylguanosine in tRNA from Salmonella typhimurium induces frameshifting by quadruplet translocation. J. Mol. Biol. 232: 756-765.[CrossRef][Medline]
Himeno, H., Yoshida, S., Soma, A., and Nishikawa, K. 1997. Only one nucleotide insertion to the long variable arm confers an efficient serine acceptor activity upon Saccharomyces cerevisiae tRNALeu in vitro. J. Mol. Biol. 268: 704-711.[CrossRef][Medline]
Holmquist, R., Jukes, T.H., and Pangburn, S. 1973. Evolution of transfer RNA. J. Mol. Biol. 78: 91-116.[CrossRef][Medline]
Ikemura, T. 1981a. Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes. J. Mol. Biol. 146: 1-21.[CrossRef][Medline]
___, 1981b. Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes: A proposal for a synonymous codon choice that is optimal for the E. coli translational system. J. Mol. Biol. 151: 389-409.[CrossRef][Medline]
Jukes, T.H. and Osawa, S. 1993. Evolutionary changes in the genetic code. Comp. Biochem. Physiol. 106: 489-494.[CrossRef]
Kawaguchi, Y., Honda, H., Taniguchi-Morimura, J., and Iwasaki, S. 1989. The codon CUG is read as serine in an asporogenic yeast Candida cylindracea. Nature 341: 164-166.[CrossRef][Medline]
Kimura, M. 1980. A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16: 111-120.[CrossRef][Medline]
Knight, R.D., Landweber, L.F., and Yarus, M. 2001a. How mitochondria redefine the code. J. Mol. Evol. 53: 299-313.[CrossRef][Medline]
Knight, R.D., Freeland, S.J., and Landweber, L.F. 2001b. Rewriting the keyboard: Evolvability of the genetic code. Nat. Rev. Genet. 2: 49-58.[CrossRef][Medline]
Kumar, S., Tamura, K., Jakobsen, I.B., and Nei, M. 2001. MEGA2: Molecular Evolutionary Genetics Analysis software. Bioinformatics 17: 1244-1245.[Abstract/Free Full Text]
|
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
ABSTRACT
RESULTS
