ABSTRACT
We previously reported the identification of an intron (CaLSU) in the 25S
ribosomal RNA of some
Candida albicans
yeast strains. CaLSU was shown to self-splice and has the potential to adopt a secondary structure typical of
group I introns. The presence of CaLSU in
C.albicans
strains correlates with a high degree of susceptibility to base analog antifungal agents, 5-fluorocytosine (5-FC) or 5-fluorouracil (5-FU). Cell death, resulting from addition of base analogs
to growing cultures, precluded demonstration of a causal relationship between
CaLSU presence and susceptibility to base analogs. In the present study, CaLSU
was inserted in a non-essential
lac
Z reporter gene and expression was examined in
Saccharomyces cerevisiae
. Different mutations affecting
in vitro
self-splicing also had similar effects on reporter gene expression
in vivo
. This indicates that
in vivo
removal of CaLSU from the reporter gene occurs through the typical self-splicing mechanism of group I introns. Base analogs inhibited expression of the reporter gene product in a concentration-dependent manner upon their addition to the cultures. This supports
a model in which disruption of intron secondary structure, consecutive to the
incorporation of nucleotide analogs, is a major factor determining the
susceptibility of
C.albicans
cells to base analogs.
The discovery of catalytic RNA molecules has deeply modified our understanding
of biological systems, genetic organisation and evolution (for reviews see
among others:
1
-
6
). Among catalytic RNAs, group I introns possess the capability of promoting
their own removal from precursor RNA molecules to generate mature transcripts;
this process is referred to as self-splicing. Despite the fact that we know >200 examples of group I introns,
relatively few consequences of their presence has been observed at the
phenotypic level. As an example, one `petite' mutation in
Saccharomyces cerevisiae
has been mapped to a splicing-defective intron in the cytochrome b gene of mitochondria (
7
).
We have recently reported the discovery of a group I intron (CaLSU
1
) in the ribosomal 25S rRNA precursor molecule of
Candida albicans
(
8
), an important opportunistic fungal pathogen of humans. The 379 base long sequence of CaLSU shares the usual putative
secondary structure of group I introns, possesses the adequate nucleotides at
specific positions known to be crucial for self-splicing, and was shown to catalyze its own excision in a standard
in vitro
self-splicing assay. However, CaLSU primary sequence is among the most divergent in the usual consensus sequences established for group I introns (see among others:
4
,
9
-
12
).
In a previous study, phenotypes and genotypes of 120 clinical isolates of
C.albicans
were analyzed. In that regard, ~40% of the strains studied harbored CaLSU in their 25S rRNA coding
sequences. In those strains with CaLSU, Southern blots and PCR analysis failed
to reveal the presence of the rRNA-encoding gene devoid of CaLSU (unpublished data) in any of the 100-200 tandemly-arrayed copies encoding rRNA (rDNA) in the nucleus. This
presence of CaLSU, in apparently every copy of the 25S rRNA-coding gene, was shown to correlate with susceptibility to base analogs
exerting an antifungal activity, 5-fluorocytosine (5-FC) or 5-fluorouracil (5-FU). Intron-bearing
C.albicans
strains all exhibited a high degree of susceptibility to base analogs while
strains lacking the intron differed widely in their susceptibility (
8
). Upon addition of 5-FC or 5-FU base analogs, these will be metabolically used to generate fluoro-substituted nucleotides that will then be incorporated into
newly-synthesized RNA molecules. It has been shown that self-splicing of a
Tetrahymena
group I intron can be drastically reduced by the substitution of
fluoronucleotides (
13
); this is likely due to disruption of intron secondary structure consecutive to
weakening of base pairing. We postulated that a similar effect of base analogs
can result from their addition to
C.albicans
cultures, disrupting the secondary structure essential for CaLSU splicing and
resulting in an aberrant 25S rRNA and lack of normal functional ribosomes.
Addition of base analogs to cultures of
C.albicans
yeast cells harboring CaLSU results in cell death, preventing further study of
their postulated effect on
in vivo
self-splicing. Another approach was required to allow a more direct
demonstration of base analogs effect on splicing. In the present study, CaLSU
was thus inserted in the coding sequence of a
lac
Z reporter gene and the resulting construct was expressed in
S.cerevisiae
. Site-directed mutagenesis was then used to further establish the importance of specific
residues or sequence elements essential for catalytic activity of group I
introns. The same sequences were shown to be important both for
in vitro
splicing of CaLSU and
in vivo
expression of active [beta]-galactosidase encoded by the
lac
Z gene. In addition to providing definitive evidence for the assignment of CaLSU
to group I introns, these results suggest that
in vivo
removal of CaLSU from
lac
Z occurs through a self-splicing mechanism. Addition of either 5-FC or 5-FU to the
S.cerevisiae
yeast cells cultures prevented expression of [beta]-galactosidase and this supports the idea that base analogs can
interfere with
in vivo
self-splicing of CaLSU. The importance of these observations, in the context of
using inhibitors of self-splicing as antimicrobial agents, is briefly discussed.
The Bluescript II SK+ plasmid harboring the CaLSU intron flanked by ribosomal
sequences has been previously described (
8
). The yeast plasmid pLGSD5 (2[mu],
URA3
,
GAL1
::
lac
Z) and the
S.cerevisiae
strain MGD353-46D (
MAT
a
,
ura3-52, trp1-289, leu2-3, 112 his3-
[Delta]
1, cyh
R
) were obtained from Dr Pierre Legrain (Institut Pasteur, Paris) and have been
described elsewhere (
14
,
15
). The yeast actin clone was obtained from Dr Reginald K. Storms (Concordia
University, Montréal).
Subcloning procedures and plasmid constructs were done essentially according to
standard procedures (
16
). Polymerase chain reaction was performed using the
Taq
DNA polymerase (Promega). Site-directed mutagenesis using the unique site elimination procedure (
17
) was performed according to the manufacturer's instructions (Pharmacia). All
mutants were sequenced directly onto denatured double-stranded DNA using Sequenase, as recommended by the manufacturer (USB).
The CaLSU intron was originally subcloned in pBluescript vector as a fragment
flanked by 25S rRNA coding sequences (
8
); this plasmid was used as a template for PCR amplification. Oligonucleotides were designed to amplify a fragment of DNA encompassing CaLSU, short flanking
rDNA sequences, and additional
Bam
HI sites created at both ends. The conservation of flanking regions was
essential since they participate in two separate pairings, P1 and P10, which
enable accurate splicing and ligation of the exons in the classical model of
group I intron splicing (
4
,
18
). The strategy for PCR amplification and sequences of the oligonucleotides used
are presented in Figure
1
A. The amplified fragment was digested with
Bam
HI and subcloned at the unique
Bam
HI site of pBluescript II KS+ vector (Stratagene). Site-directed mutagenesis of CaLSU was performed on this construct; oligonucleotides used were 5'-TTGCCTCC
RNA substrates for self-splicing were produced by
in vitro
transcription using T3 RNA polymerase. The Bluescript II KS+ plasmids with
CaLSU inserted at the
Bam
HI site of the polylinker (wild-type, inverted and mutants) were linearized at the
Pvu
II site, leaving flanking plasmid
lac
Z sequences on both sides of CaLSU. The original pBluescript clone of CaLSU
flanked by ribosomal RNA sequences was used as a control (
8
). As previously described, transcription of this control was performed with T7
RNA polymerase after linearization at the
Nar
I site of
C.albicans
25S rDNA sequence, leaving ribosomal flanking sequences on both sides of CaLSU
(
8
). All transcription reactions were performed in the presence of 0.1 mM of each
ribonucleotide triphosphates; 50 [mu]Ci of radioactive UTP (800 Ci/mmol; ICN Biomedicals) was added to the
reaction. The radioactive RNA products were analyzed onto denaturing 4%
sequencing urea-polyacrylamide gels followed by autoradiography. Quantitation
was performed using a Personal Laser Densitomer (Molecular Dynamics).
Strains of
C.albicans
harboring (strain 4F) or lacking (strain Lecocq) the CaLSU intron in their
ribosomal RNA were grown 30 h at 30oC under gentle agitation in SC-uracil/glycerol/lactate medium [6.7 g/l yeast nitrogen base w/o amino acids (Difco), 5 g/l casamino acids (Difco), 20 mg/l tryptophan, 30 ml/l glycerol, 20 g/l lactate, 0.5 g/l glucose]. Cells were then centrifuged, washed twice in saline and resuspended in fresh medium at a concentration corresponding to an optical density of 0.25 at 600 nm. Base analog, 5-FC, was then added at various concentrations and the cultures
incubated in the same conditions for 24 h before measuring the optical density;
a control culture was also incubated in the absence of 5-FC.
Saccharomyces cerevisiae
transformations were performed by the lithium acetate protocol (
19
). Cultures for [beta]-galactosidase assays were prepared as follows. Transformants were plated on SC-uracil [6.7 g/l yeast nitrogen base w/o amino acids (Difco), 5 g/l casamino
acids (Difco), 20 mg/l tryptophan] containing 20 g/l glucose and 20 g/l bacto-agar. Pools of at least 100 transformants were directly used to inoculate liquid SC-uracil medium with 3% glycerol/2% lactate/ 0.05% glucose. Cultures were centrifuged, washed and used to inoculate
fresh liquid SC-uracil media (starting OD
600
of 0.1) supplemented with 2% galactose instead as carbon source in order to
activate transcription of the
lac
Z gene under the control of the GAL1 promoter. Cells were recovered by
centrifugation after overnight growth at 30oC. When indicated, varying concentrations of base analogs were added to the
medium at the same time as the yeast inoculum. The composition of the SC-uracil medium was chosen since it proved to be optimal to show the effect
of base analogs.
The assay was performed as previously described (
20
). Determination of the enzymatic activity, detection of the protein by immuno- blotting and RNA levels analysis (Northern blots), were all performed from
the same cultures. The [beta]-galactosidase activity was expressed in standard Miller units.
Proteins were extracted from the yeast cells using the glass beads extraction
procedure (
21
). Aliquots of ~100 [mu]g protein were loaded onto SDS-PAGE gels; following electrophoretic separation, proteins were electrotransferred onto nitrocellulose filters
and analyzed by immunoblotting. A monoclonal anti-[beta]-galactosidase antibody (Boehringer Mannheim) was used at a
1/2000 dilution and the secondary antibody was a rabbit anti-mouse alkaline phosphatase conjugate (BRL) used at the same dilution.
Antigen-antibody complexes were detected using NBT-BCIP chromogenic substrates as recommended by the manufacturer (Gibco/BRL).
Total RNA extraction and Northern blots were performed as previously described (
22
). Polyadenylated RNA was isolated using Dynabeads mRNA Purification Kit according to the manufacturer's instructions (Dynal). Probes were prepared by random priming (Pharmacia). The
CaLSU probe (401 nt) was gel-purified following a
Bam
HI restriction endonuclease digest of the pBluescript construct (see Plasmid
constructs). The
lac
Z probe is a
Bam
HI-
Acc
I DNA fragment (2834 nt) from pLGSD5. Finally, the plasmid bearing the complete
yeast actin gene, including its intron, was used as a probe.
Three specific sequence elements or residues, known to promote accurate and
efficient self-splicing of group I introns, were chosen as targets for site-directed mutagenesis. As a first mutant, a conserved guanosine
residue in the P7 pairing of the catalytic core of group I introns was
substituted for a uridine (`G247U' mutant). This residue was shown to be
responsible for the binding of the GTP molecule required for catalytic cleavage of the 5' intron-exon boundary during self-splicing of group I introns (
23
). Sequence elements, apparently required to position the two exons in close
proximity before self-cleavage at the boundary between the intron and 3'-exon and ligation of the two exons (
18
), were our next targets. The internal guide sequence (IGS) lies in the P1 stem-loop and it can pair with a second element, consisting of the first few
nucleotides of the 3'-exon, to form P10. We separately mutated either the IGS sequence
(`5'-disrupted' mutant) or the complementary sequence in the flanking 3'-exon (`3'-disrupted' mutant). We also generated a
double mutant in which both sequences were mutated (`restored' mutant).
Although the new sequences are different from those of the wild-type intron, they potentially restore the P10 pairing. The nature and
position of the different mutations are schematized in Figure
1
B and C.
The different mutant introns with flanking
lac
Z sequences in pBluescript were transcribed
in vitro
and their capacity to self-splice was examined and compared with the same intron flanked by ribosomal
RNA sequences in the original plasmid construct (Fig.
2
, CaLSU/25S rRNA). RNA products, resulting from self-splicing reaction occurring during the transcription reaction containing adequate magnesium and GTP concentrations required for splicing, were analyzed; we have been unable to achieve splicing of
lac
Z-CaLSU RNA species after gel purification. It is likely that, in this case, splicing has to be initiated concomitantly with transcription; this point will be further discussed. Nevertheless, splicing products generated in the transcription reaction were easily analyzed. The wild-type intron in
lac
Z (Fig.
2
, CaLSU construct wild type) was able to self-splice, generating free intron (379 bases long) and religated exons (263
bases long) from the 642 bases long precursor. Other minor intermediates
products were detected; this was also previously observed in the reaction of
CaLSU self-splicing in its original ribosomal context (free 5'-exon, intron-3'-exon and free 3'-exon). The CaLSU intron has
thus kept the capability of self-excision even though it is inserted in a
lac
Z gene different from its initial location in ribosomal RNA. However, the
efficiency of self-splicing in the
lac
Z context is reduced compared to the efficiency observed in the original
ribosomal context, as shown by a decreased percentage of precursor spliced to
its products from 75% in the original ribosomal context down to 25% in the
lac
Z context. A more important accumulation of the intermediate `intron-3'-exon' species was also observed in the
lac
Z context (Fig.
2
).
Following the analysis of
in vitro
self-splicing, we proceeded to determine the
in vivo
activity of the same constructs (wild-type and mutants). CaLSU was inserted at the beginning of the
lac
Z gene in the yeast vector pLGSD5 (Fig.
1
A), and transformed in
S.cerevisiae
yeast cells. In the absence of CaLSU removal, its presence should prevent
production of active [beta]-galactosidase since there are multiple termination codons in all
three reading frames of CaLSU; levels of [beta]-galactosidase activity should thus be a good indication of the
efficiency of CaLSU splicing
in vivo
.
The construct harboring wild-type CaLSU allowed production of active [beta]-galactosidase while intron insertion in the inverted
orientation resulted in complete inactivation of the
lac
Z gene (Table
0
). However, the level of activity was much lower in the intron-containing construct (~15% compared to the original intronless pLGSD5 plasmid). The GTP
binding site mutant (G247U) was inactive while mutants disrupted in either 3'-flanking (3'-disrupted) or 5'-IGS region (5'-disrupted) exhibited barely detectable levels of
activity (<0.1% of wild-type). About 10% of activity, compared to the wild-type value, was observed with the double mutant restoring base
pairing between IGS and the 3'-flanking sequence (restored, Table
0
). Altogether, these data indicate that production of [beta]-galactosidase activity, when CaLSU is inserted in the
lac
Z gene, is largely dependent on the same sequences that affect
in vitro
self-splicing. This suggests that
in vivo
splicing of CaLSU occurs through autocatalytic activity of the intron since production of [beta]-galactosidase is affected by the same mutations as
in vitro
self-splicing.
Table 1
[beta]-galactosidase activity of various
lac
Z constructs in pLGSD5
We also verified that the amount of protein correlates with the results of enzymatic activity dosage. This is of importance because, in the strategy for plasmid construct, it was impossible to avoid addition of
a few extra amino acids; in the event of correct splicing, there will be
addition of six amino acids close to the amino-terminal end of the [beta]-galactosidase protein (Fig.
1
A). Immunoblotting analysis was performed with a monoclonal anti-[beta]-galactosidase antibody; a major protein species was detected
with smaller species likely corresponding to degradation products. The amount
of both the full-length species or total immunoreactive products was significantly decreased by insertion of CaLSU in the
lac
Z coding region (Fig.
4
A compare first lane to fourth lane). This experiment showed that the reduction
of enzyme activity does reflect a reduced amount of [beta]-galactosidase enzyme produced by the cells harboring the
lac
Z gene interrupted by CaLSU, compared to the original intronless plasmid. There
is thus no evidence of a detrimental effect on the enzymatic activity due to
the presence of the six additional amino acids.
We next applied the reporter gene system to study the
in vivo
effect of base analog antifungal agents. Increasing concentrations of either 5-FC or 5-FU were added to cultures of
S.cerevisiae
harboring the pLGSD5 plasmid with the
lac
Z gene interrupted by wild-type CaLSU. In the range of concentration used, the growth of
C.albicans
strains, harboring the intron in their ribosomal rRNA-coding genes, was strongly inhibited by base analogs (Fig.
3
, and data not shown); as discussed later, the susceptibility of strains lacking
the intron was much more variable but they tend to be significantly more
resistant. In the same concentration range, the base analogs do not affect the
growth of the
S.cerevisiae
strain used in this study (data not shown). Similarly, the level of [beta]-galactosidase activity encoded by the native pLGSD5 plasmid vector
was only barely affected, with 80% activity remaining at high concentration of either base analogs, 25 [mu]g/ml 5-FC or 250 [mu]g/ml 5-FU (Fig.
4
). In contrast, when we analyzed the production of active [beta]-galactosidase from the plasmid harboring CaLSU in the
lac
Z gene, we observed a rapid decrease at low doses of either 5-FC (2-5 [mu]g/ml) or 5-FU (50 [mu]g/ml) (Fig.
4
).
In the present study, we devised a system to examine
in vivo
splicing of CaLSU in an effort to gather more evidence supporting our hypothesis that base analogs can exert an antifungal activity through
inhibition of self-splicing. The rationale was to introduce CaLSU in a non-essential reporter
lac
Z gene, an approach used with success by previous investigators (
24
,
25
). These earlier studies used
E.coli
as a model microorganism and the extent of self-splicing was visually estimated using the blue phenotype conferred by
expression of the
lac
Z gene in bacterial colonies grown on adequate indicator plates. However, we
have failed to adapt this bacterial system to a quantitative assay of CaLSU
splicing (data not shown). We thus devised a similar system using
S.cerevisiae
; this new original system also proved adequate to study the role of base
analogs since cells from the chosen strain can incorporate and use base analogs
for nucleic acid synthesis while being resistant to their lethal effect.
The predicted secondary structure of CaLSU, the presence of consensus sequence
elements, and the occurrence of magnesium and GTP-dependent
in vitro
splicing, initially allowed us to assign CaLSU to group I introns. In addition
to these standard criteria,
in vitro
splicing analysis of CaLSU mutants, reported in the present study, further
supports the idea that this initial assignment to classical group I self-splicing introns was justified and that prediction of GTP binding site and
sequence elements involved in the P10 pairing were correct. As previously
reported, the accumulation of the `intron-3'-exon' intermediate in the single P10 mutants suggests that
disruption of this particular pairing results in a loss of 3' splice site selection ability (
18
). As could be expected, such an accumulation was eliminated after restoration
of the P10 pairing by complementary mutations in both the IGS and 3'-exon. The variation observed in the amount of ligated exons with
the different constructs is comparable to results obtained by Michel
et al
. (
23
) after similar mutagenesis of the
Tetrahymena
intron. Mutants exhibiting reduced self-splicing proved to be similarly affected
in vivo
. However,
in vivo
splicing of the CaLSU constructs in yeast was quite inefficient even for the
wild-type intron. The low efficiency of splicing is also consistent with RNA
analysis since, despite our repeated attempts, the mature spliced form of the
lac
Z mRNA resulting from CaLSU removal could not be detected on Northern blots. The
presence of intermediate or aberrant splicing products may be partly responsible for the difficulties in pursuing further characterization by either RNase
mapping or reverse transcriptase polymerase chain reaction (RT-PCR) procedures (data not shown).
Many explanations come to mind for the apparently very low splicing efficiency
observed with CaLSU in the reporter gene. First of all, the intron is inserted
in a sequence context differing from its original location, and there is
evidence that the presence of certain flanking sequences can affect self-splicing of group I introns (
26
). This is also consistent with the fact that all our efforts to achieve
in vitro
self-splicing of the gel-purified
lac
Z-CaLSU precursor were unsuccessful, despite the fact that this was readily
achieved when CaLSU was in its original ribosomal sequence context (
8
, data not shown). We suspect that a 3'-flanking region inhibits self-splicing of complete transcripts; the spliced products
observed in the transcription reaction may then result from splicing initiated
during transcription, and before synthesis of the transcript is completed. It
should also be noted that,
in vivo
,
lac
Z mRNA synthesis will be the result of RNA polymerase II activity while rRNA
precursors are polymerase I transcripts; this, as well as the normally
nucleolar location of the ribosomal transcript, may also affect the efficiency
of
in vivo
splicing. Another possible explanation is the participation of accessory proteins during
in vivo
splicing of so-called self-splicing introns; the
S.cerevisiae
CBP2 protein and
Neurospora
CYT-18 have been conclusively identified as such accessory proteins involved
in the splicing of some group I introns (
27
-
30
). The absence of the adequate protein in the heterologous
S.cerevisiae
host, or non-nucleolar localization of the transcript, may prevent access of CaLSU to
the adequate protein cofactors.
It is likely that the
lac
Z reporter gene system will be useful for the study of splicing inhibitors as
well as various studies dealing with
in vivo
self-splicing in the context of an eucaryotic cell. Despite the limitations due
to a low splicing efficiency, this system was still well-suited to study the effect of base analogs on CaLSU splicing. Addition of
increasing concentrations of base analogs to the yeast cultures progressively
reduced expression of [beta]-galactosidase; this effect was observed only when the
lac
Z gene was interrupted by CaLSU, indicating that the intron is responsible for the effect. The level of protein produced was reduced although
unspliced mRNA was present at a level similar to the
lac
Z mRNA normally formed in the absence of CaLSU. Altogether, these results
strongly suggest that base analogs can block
in vivo
self-splicing of CaLSU when used at doses similar to the ones that are toxic to
C.albicans
strains harboring CaLSU in their 25S rRNA. We believe that the inhibition is
due to perturbations of CaLSU secondary structure resulting from incorporation
of fluoro-substituted nucleotides derived from the base analogs. This is consistent
with the inhibition of
in vitro
self-splicing observed for the
Tetrahymena
group I intron upon substitution of uracil by 5-fluorouracil residues (
13
).
Candida albicans
strains devoid of CaLSU vary widely in susceptibility to base analogs. This is
due to the pleiotropic effect of base analogs resulting from incorporation of
fluoro-substituted nucleotides into various nucleic acids: chromosomal DNA, ribosomal RNA and messenger RNA (
31
,
32
). Levels of metabolic incorporation vary from strains to strains and is likely
responsible for variations in susceptibility. In contrast, yeast strains
harboring CaLSU in their rRNA are uniformly very sensitive to the effect of
base analogs (
8
, and data not shown). Altogether, these observations support a model in which
self-splicing relying on adoption of a precise secondary structure by the
intron, can be more easily affected by base analogs than are other molecular targets; as a result, intron presence is a
major base analogs susceptibility factor in
C.albicans
.
It is interesting to notice that some antibiotics and antifungal agents have
been previously shown to inhibit
in vitro
self-splicing of group I introns (
33
-
37
). It has even been suggested that susceptibility to antimicrobial agents could
be used to detect new group I introns (
34
); our discovery of CaLSU is particularly significant in this regard. Another
interesting observation is the presence of group I introns in the chromosomally-encoded rRNA of
Pneumocystis carinii
, a fungal pathogen of increasing medical importance. It has been recently
observed that pentamidine, an anti-
Pneumocystis
agent, can inhibit
in vitro
self-splicing (
37
); there is, however, no other evidence that this inhibitory effect is
responsible for the
in vivo
effect of the agent. The reporter gene system described in the present study
could be applied to the study of such other introns and antifungal agents,
ultimately leading to new antimicrobial therapeutic approaches directed against
self-splicing introns.
We thank Nicole Rougeau for technical support in the initial part of this
project. We thank Dr Louis DeRepentigny for numerous helpful discussions. We
thank Dr Pierre Legrain (Institut Pasteur, Paris) for his generous gift of the
pLGSD5 plasmid as well as
S.cerevisiae
strain MGD353-46D, and Dr Reginald K. Storms (Concordia University, Montreal) for the
yeast actin clone. This work was supported in part by a group grant `Équipe prioritaire' from the `Fonds de la Recherche en Santé du Québec'. S. Mercure was initially the recipient of a
studentship from the Medical Research Council of Canada. P.B. and G.L. were
both recipients of a `Chercheur-Boursier' award from the `Fonds de la Recherche en Santé du Québec'.
*To whom correspondence should be addressed. Tel: +1 514 343 2422; Fax: +1 514
343 5701; Email: lemayg@ere.umontreal.ca
Lac
Z Constructs
(in pLGSD5)
a
[beta]-galactosidase activity
b
(Miller units)EXPT #1EXPT #2Intronless411366Wild-type CaLSU 32 70Inverted CaLSU <1 <1Mutants of CaLSU G247U (GTP binding) <1 <1 5'-disrupted <1 <1 3'-disrupted <1 <1 Restored (5'-3') 3 6.7
REFERENCES
Return