|
|
|
|||||||
|
|||||||||
From the
Three mitochondrial mutants were characterized that block the
splicing of aI3
Many group I introns are mobile elements. Each such intron
contains a reading frame that encodes a site-specific endonuclease
required for the mobility. A number of families of intron-encoded
endonucleases have been identified, the largest of which has the
sequence motif LAGLIDADG (reviewed in Ref. 1). Nearly all such
proteins, including a number that are not encoded by group I introns (e.g. Refs. 2-4), have two copies of that motif, often
referred to as P1 and P2(1) . In addition to encoding
endonucleases, some group I introns, including three in the COB
Intron 4
Four other mitochondrial
group I introns in S. cerevisiae (
Here we report
studies of aI3
The fusion protein was purified from
extracts of induced E. coli cells by passage over an amylose
column and used to immunize two rabbits. Samples of immune serum were
partially purified by passage over a MalE affinity column. The
resulting flow-through sample was adsorbed on a column containing the
fusion protein antigen and eluted in the cold with a solution
containing 0.1 M glycine HCl, pH 2.3, and collected as 1-ml
fractions in tubes containing 0.25 ml of 1 M Tris, pH 8.0.
Pooled fractions containing the highest amount of protein were then
titered against mitochondrial protein samples and used in Western blot
experiments ( Fig. 6and Fig. 8) using chemiluminescence
(ECL kit from Amersham Corp.).
Next, the cDNAs from both mutants
and the control were sequenced. As shown in Fig. 4A,
mutant C1072 has a major cryptic 5`-splice site in exon 1 (Fig. 4B). Splicing occurs within a sequence (TGGTAA)
that is identical to the sequence at the wild-type splice junction,
resulting in a COXI mRNA deleted for 90 nucleotides of exon
sequence. The shortened mRNA retains the reading frame so that a small
amount of a CoxI protein deleted for 30 amino acids can be made. Based
on the RT-PCR results, it appears that the cryptic site is not used in
the wild-type strain (Fig. 3, lane2). The
227-bp minor RT-PCR product obtained from both mutants was also
sequenced; as shown in Fig. 4A for strain C1072, it
results from splicing at a site in exon 2 where the sequence TGGTCA is
present (Fig. 4B). Splicing there forms an mRNA 57
nucleotides shorter than the wild type that encodes a protein deleted
for 19 amino acids. The same results were obtained when the 186- and
227-bp cDNAs from mutant M44 were sequenced (data not shown).
We have tested the self-splicing
activity of all three mutant alleles (data not shown; see Ref. 42).
While wild-type aI3
For this experiment, a petite mutant containing the entire COXI gene was isolated from strain 161
As
shown in Fig. 5, mutant C1085 contains seven mutations in the
closed reading frame of aI3
The aI3
Although we have not determined whether
p44 has or lacks endonuclease activity, the following experiment shows
that the polypeptide is probably a precursor of p35. Samples of
[
This work characterizes the splicing and expression of the
protein encoded by the mobile group I intron of yeast mtDNA, aI3
The M44 mutation
directly affects the formation or stability of the P1 helix of
aI3
Mutants of
the exon portion of the P1 helix of other introns can lead to the use
of cryptic 5`-splice sites (e.g. Refs. 48 and 59). Winter et al.(60) reported self-splicing experiments showing
that a mutation of the internal guide sequence of aI3
In vivo, the M44 mutation
completely blocks accurate splicing of aI3
In most respects, the
phenotype of mutant C1072 resembles that of the M44 mutant. Since the
C1072 mutation lies upstream of the exon sequence predicted to be
needed for P1 formation, its effect on splicing of aI3
Mutant C1072 has a novel exon mutation blocking splicing of a group
I intron because it does not alter a nucleotide of the P1 helix. As
noted above, that mutation has little effect on in vitro self-splicing. Since both C1085 and M44 are mutant for known
splicing signals and are splicing-deficient both in vivo and in vitro, it appears that the C1072 mutation reveals a feature
of aI3
In agreement with the findings of
Szczepanek et al.(21) using another yeast strain, the
allele of aI3
Our finding that aI3
The mutants accumulate two proteins in
common, p35 and p44; C1072 has an additional protein, p38, that is also
detected by our antibody. The S. aureus V8 protease analysis
of p35 and p44 shows that these polypeptides are closely related to
each other, presumably via post-translational processing. It is
possible that the C1072 mutation alters the processing of the precursor
protein, yielding the novel protein, p38. If so, then it will be
interesting to learn whether p38 can be processed to yield active p35.
Schapira et al.(26) expressed in E. coli a
universal code equivalent of the aI3
Recently, protein introns (inteins) were discovered in several
organisms, including yeast (reviewed in Refs. 67 and 68). These
insertions contain long reading frames that encode polypeptides clearly
related to the family of group I intron-encoded proteins that have P1
and P2 motifs. Precursor polypeptides containing the insertions are
rapidly processed via transpeptidation reactions, resulting in the
``spliced'' product of the host gene and excision of the
intein protein(50, 52) . In several cases, the excised
protein has been shown to have endonuclease activity (e.g. Refs. 3 and 4) resembling, in many ways, the major family of
endonucleases encoded by mobile group I introns(67) . Since I-SceIII is a member of the LAGLIDADG family and is made as a
precursor (p44) that is processed to a mature form (p35), it is an
intriguing possibility that its processing may resemble that of protein
introns. Further studies are clearly needed to test this possibility.
Mutants C1072 and C1085 were initially characterized
by Kirk Mecklenburg (Ohio State University) and by Philip Sass and
Henry Mahler (Indiana University).
Volume 270,
Number 26,
Issue of June 30, pp. 15563-15570, 1995
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
of the Yeast Mitochondrial COXI Gene Encodes a 35-kDa Processed Protein That Is an Endonuclease
but Not a Maturase (*)
ABSTRACT
INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
, a mobile group I intron of the COXI gene
of yeast mtDNA. Mutant C1085 alters helical structures known to be
important for splicing of group I introns. M44 and C1072 are point
mutants in exon 3 that block correct splicing but allow some splicing
at cryptic 5`-splice sites. M44 alters the P1 helix needed for
5`-splice site definition, while the mutation in C1072 is a new kind of
mutation because it is located upstream of the exon sequence involved
in the P1 helix. All three mutants accumulate novel proteins of 35 and
44 kDa (p35 and p44, respectively) detected both by labeling of
mitochondrial translation products and by Western blotting. Partial
protease digestions indicate that p44 and p35 are closely related,
probably as precursor and processed protein. The level of the
intron-encoded endonuclease activity, I-SceIII, is elevated
10-fold in the mutants. Partial purification of I-SceIII
from the mutants showed that most, if not all, of the activity is
associated with p35. Finally, because aI3
splices accurately in a
petite mutant, we conclude that aI3 splicing does not depend on a
mtDNA-encoded maturase.
()gene, encode maturases required
for splicing in vivo (reviewed in Ref. 1).
of
the COXI gene (aI4) encodes an endonuclease (I-SceII) (5, 6) and, in wild-type strains,
requires the maturase encoded by COB intron 4 (bI4) for
splicing(7, 8, 9) . The protein encoded by
aI4 is bifunctional because it is also a latent maturase that is
activated either by a particular missense mutation in the aI4
reading frame (10) or by a missense mutation in the nuclear NAM2 gene encoding the mitochondrial leucyl-tRNA
synthetase(11, 12) . The protein encoded by COB intron 2 of Saccharomyces capensis is another example of
a bifunctional LAGLIDADG protein(13) ; however, in standard
laboratory strains of Saccharomyces cerevisiae, that intron is
not mobile and has only maturase activity.
, aI3
, aI5,
and aI5
) have reading frames with the P1 and P2 motifs. With the
exception of aI5, which lacks an obvious means of
expression(14) , all encode endonucleases (I-SceI, I-SceIII, and I-SceIV, respectively) and are
mobile(15, 16, 17, 18, 19) . Two
more group I introns with reading frames with P1 and P2 motifs have
been found in mtDNAs of several Saccharomyces species (6, 20, 21) and do not appear to be
mobile(21) . Interestingly, the three group I introns known to
encode maturases do not self-splice. The mobile intron of the 21
S rRNA gene self-splices (22) and does not encode or require a
maturase activity(23) . The other mobile introns, aI3
and
aI5, both self-splice (22, 24) but have not yet
been shown to encode or require a maturase activity.
in S. cerevisiae strain ID41-6/161,
including characterization of mutations that block in vivo splicing of the intron. We show that the protein encoded by
aI3 is synthesized as a 44-kDa precursor and that most or all of
the I-SceIII endonuclease activity present in mutant strains
that overproduce the intron-encoded protein is associated with a 35-kDa
processed form of the precursor. Although there have been suggestions
that aI3 encodes a maturase(21, 25, 26) ,
we show that the intron neither encodes a maturase nor depends on one
encoded by another mitochondrial intron.
Yeast Strains and Mitochondrial Mutants
The
strains employed in this study are described in Fig. 1. Petite
strains DS6/A401 and DS6/A407 are derived from strain D273-10B (MAT met) and were gifts of Dr. Alexander
Tzagoloff(27) . Petite strains DS6/A422/C2 and DS6/A422/N are
derivatives of petite mutant DS6/A422 (27) containing only the portion
of the COXI gene shown in Fig. 1(28) . Strain AD1
is a mit deletion mutant (29) derived from
strain ID41-6/161 (MATa ade1 lys1
);
this deletion was used here in the nuclear background of strain JC9/55 (MAT
leu1 kar1). The mit mutants M44,
C1072, and C1085 were induced by treatment of strain ID41-6/161 with
MnCl
(30, 31) and isolated using methods
reviewed in Refs. 32 and 33. A derivative of strain ID41-6/161 deleted
for introns 1 and 2 of the COXI gene (C1036
1,2) (34) was treated with ethidium bromide to induce petite mutants.
The mutants were screened using a test cross to mutant C245 (a
mit mutant lacking most of the COXI gene) (35) until one carrying the COXI gene was found
(COXI
1,2).
Figure 1:
Physical map of mutations in
the COXI gene. The COXI gene of mtDNA of yeast strain
ID41-6/161 is diagramed at the top. The portion shown between the BglII and BclI sites was cloned from the wild-type
and mutant strains for sequencing experiments; the PvuII site
shown was placed in the wild-type clone to permit an in-frame fusion of
the intron reading frame to an E. coliMalE gene
expression plasmid. The shaded part of the intron reading
frame was expressed in E. coli. Five mutant genomes used to
map the point mutations in this region of the gene are diagramed. The thinline indicates the portion of the gene present
in each mutant genome; the diagonallines indicate
which genomes extend beyond COXI. The breakpoint of each
deletion is denoted by a verticalline when it is
known with precision. The regions of mtDNA within the unfilledbox or outside of the vertical or diagonallines are deleted from the genome. Where some uncertainty
remains, a filledbox is show spanning the region
that may be present in the mutant. Petite mutants DS6/A422/N and
DS6/A407 are derived from strain D273-10B, which lacks introns 5
and 5; the missing introns are indicated by inverted Vs.
The three mit mutants reported here were mated to
each tester strain, and the results of those crosses defined a discrete
interval of the gene where the mutation is located, summarized at the
bottom.
Genetic Methods
Marker rescue with genetically
defined, physically mapped petite mutants and mit deletions (reviewed in Ref. 32) was used to map the new
mit
mutations of this study. Standard media were used
throughout. Spontaneous revertants of C1072 and M44 were isolated as
red papillae on colonies on 2% yeast extract, 2% bactopeptone, 2%
dextrose dishes. Representative revertant strains were subcloned,
aI3
plus flanking sequences were cloned, and exon 3 was sequenced.
Protein Labeling and Fingerprinting
Methods
Mitochondrial proteins were labeled with
[S]SO
in the
presence of cycloheximide, and proteins were separated by
polyacrylamide gel electrophoresis as described(36) .
One-dimensional fingerprinting of proteins excised from
SDS-polyacrylamide gels was carried out as described(37) .
RNA Blots
RNA was prepared from spheroplasts using
the extraction procedure of Ref. 38. RNA was denatured at 75 °C and
separated on 1.2% agarose gels containing 6% formaldehyde. The RNA was
then transferred to Hybond-N filters (Amersham Corp.) and probed.
Oligonucleotide probes specific for COXI exon 4 (containing
nucleotides 6754-6771 of Ref. 27), aI3 (complementary to
nucleotides 5303-5320), and aI5 (complementary to
nucleotides 148-171 of aI5) (14) were used. Each
probe was 5`-end-labeled using T4 polynucleotide kinase and
[-
P]ATP (Amersham Corp.), and blots of gels
were hybridized in Rapid Hyb buffer (Amersham Life Science) at 42
°C using between 10
and 10
cpm of
probe/filter. Filters were washed three times at room temperature in 5
SSC (0.75 M NaCl plus 0.075 M trisodium
citrate, pH 7.0) containing 0.1% SDS and then one time at 42 °C in
5
SSC containing 0.1% SDS. The filters were exposed to Kodak
X-AR5 medical x-ray film with a DuPont NEN Cronex intensifying screen.
In some cases, blots were quantitated using a Molecular Dynamics
PhosphorImager.
DNA Sequencing
Restriction fragments of the COXI gene of mutant and wild-type strains were cloned in
M13mp18, M13mp19, or pBLSKS and sequenced using
synthetic oligonucleotide primers. For all mutants, we sequenced the
entire portion of the gene to which the mutation was mapped (based on
the data of Fig. 1) and compared the result with the sequence
from the wild-type parent.
PCR Experiments
cDNAs containing splice sites in COXI mRNA present in various strains were obtained as follows.
The first strand of cDNA was made using an RT-PCR consisting of 1
avian myeloblastosis virus reverse transcriptase reaction
buffer (1 mM dNTPs, a 1 µM concentration of an
oligonucleoide complementary to nucleotides +6847 to +6869 of
exon 4(27) , 2 µg of whole cell RNA, and 16 units of avian
myeloblastosis virus reverse transcriptase (Promega)) in a final volume
of 20 µl. The reaction was incubated for 1 h at 45 °C and then
terminated by incubation for 15 min at 80 °C. Next, 80 µl of
PCR master mixture (1
reaction buffer (Boehringer Mannheim),
200 mM dNTPs, a 1 mM concentration of an
oligonucleotide containing nucleotides +110 to +131 of exon
1(27) , and 8 units of Taq polymerase (Boehringer
Mannheim)) was added and incubated in a Thermocycler (initial 3 min of
denaturation at 95 °C, followed by 30 cycles of 30 s for
denaturation at 95 °C, 30 s for annealing at 45 °C, and 30 s
for extension at 72 °C). The PCR products were fractionated on 1%
SeaPlaque GTG agarose gel, and the main products were isolated and
sequenced as described(39) . The PCR products shown in Fig. 3were fractionated on a 4% Metaphor agarose gel, stained
with Cyber green I, and scanned in a FluorImager (Molecular Dynamics,
Inc.).
Figure 3:
RT-PCR analysis of spliced transcripts
present in wild-type and mutant strains. RNAs analyzed in Fig. 2 were
used as templates in RT-PCR experiments using the strategy outlined
under ``Experimental Procedures.'' The RNA used in each
amplification is shown above lanes 2-6, and the sizes of
the main cDNA products are shown beside lane6. Lane1 contains a 123-bp DNA ladder (Life
Technologies, Inc.) used as a size standard.
Antibody Preparation and Western Blot Methods
Part
of the reading frame of aI3 was expressed in Escherichia coli as an extension of the E. coli MalE protein using the
pMAL-C system (New England Biolabs Inc.). The intron and flanking exons
were cloned in BamHI-linearized pBS+ (Stratagene) as a BglII-BclI fragment from mtDNA of strain 161 (see Fig. 1). A Muta-Gene kit (Bio-Rad) was used to change nucleotides
5690-5695 (27) from GCT TGA to CAG CTG, eliminating a TGA
stop codon and creating a PvuII site. This plasmid was cleaved
with PvuII, and the 1.7-kb fragment (from the new site through
the end of aI3 to a PvuII site in the vector) was cloned
into pMAL-C that was linearized by cleavage with StuI. A
plasmid containing the insert in the correct orientation was confirmed
by DNA sequencing. When expressed in E. coli cells, a 56-kDa
fusion protein containing 420 amino acids of MalE and 120 amino acids
of the aI3 reading frame accumulated. The intron-encoded portion
begins in the P2 motif and ends at a TGA codon 51 triplets from the end
of the intron reading frame; due to genetic code differences, the
protein translated in E. coli contains seven missense changes
(3 Ile residues that are Met in mitochondria and 4 Thr residues that
are Leu in mitochondria).
Figure 6:
Mitochondrial translation products
accumulated in mutant strains. Lanes 1-4 show
mitochondrial proteins labeled in vivo with
[S]sulfate as described under
``Experimental Procedures.'' The protein samples from the
strains shown were fractionated on a 10-15% polyacrylamide gel
and visualized by autoradiography. The major proteins present in the
wild-type sample are labeled beside lane1, and the
estimated sizes of prominent novel proteins present in some mutants are
indicated beside lane8. The samples shown in lanes 1-4 were transferred to nitrocellulose paper and
probed with a rabbit antiserum raised against a fragment of the
aI3
reading frame expressed in E. coli (see
``Experimental Procedures''), and the image was obtained
using enhanced chemiluminescence.
Figure 8:
p35
copurifies with the I-SceIII activity. Mutant C1085
(ID41-6/161 nuclear background) was grown on YEP/galactose (2%) medium,
and a mitochondrial fraction was isolated according to a scaled-up
version of the cell breakage/fractionation protocol (36). A is
a diagram of our standard procedure for enriching I-SceIII
activity from mitochondrial fractions (see ``Experimental
Procedures''). B shows a Western blot of proteins from
isolated mitochondria and the fraction 1 supernatant and pellet
fractions (analyzed as described in the legend of Fig. 6). Here equal
amounts of I-SceIII activity were loaded in each lane. Like
the fraction 1 supernatant, more highly purified samples of I-SceIII activity (both fraction 1a and peak fractions from
phosphocellulose chromatography; see A) also have a prominent
p35 band but no p44 (data not shown).
I-SceIII Extraction and Assay
Methods
Mitochondrial fractions were isolated from cells grown
on 2% yeast extract, 2% bactopeptone/galactose (2%) medium and
fractionated essentially as described(40) . I-SceIII
activity could be detected at all stages in the purification, including
lysed mitochondria, the supernatant of a 1 M KCl extraction
(fraction 1), redissolved 25-55% ammonium sulfate precipitate
(fraction 1a), and peak fractions from a phosphocellulose column
(fraction 2). The activity was assayed using buffer containing 50
mM Tris-HCl, pH 7.5, 10 mM MgCl, 50
mM KCl, 2 mM dithiothreitol, and 0.4 mg/ml bovine
serum albumin essentially as described for I-SceII(40) . The I-SceIII substrate is
plasmid p1-3; it is pEMBL18 containing a 1.2-kb HpaII-HindIII fragment of mtDNA that includes exons
1-4 plus most of intron 4 from strain 1611-3, a
derivative of strain ID41-6/161 lacking introns 1-3(6) .
From preliminary experiments, these I-SceIII preparations have
optimal activity at 30 °C and pH 6.5-8.5 in buffer containing
200 mM KCl. In some experiments, we used strains disrupted
for the NUC1 gene that encodes a mitochondrial nuclease
(strain WA12, MATa ade2 ura3 trp1 his3 leu2 nuc1::LEU2
IMP1)(40, 41) ; it was found that detection of I-SceIII activity is not influenced by the NUC1-encoded enzyme, so in most experiments, strains having
the nuclear background of strain ID41-6/161 (NUC1) were used.
Physical Mapping of Mutations in or near
aI3
Preliminary mapping experiments of sets of
mitochondrial respiration-deficient mutants revealed three mutants
(M44, C1072, and C1085) in or near aI3. Each mutant was mated to
the tester strains shown in Fig. 1, and the crosses were scored
for recombinant progeny capable of growth on glycerol medium
(Gly). These crosses define the physical location of
each mutation as summarized in Fig. 1. C1072 and M44 map to a
portion of the COXI gene containing the 3`-end of intron 2,
exon 3, and the beginning of intron 3. C1085 maps to the 3`-end of
intron 3 or part of exon 4. Mutants C1072 and M44 revert spontaneously,
suggesting that each owes its mutant phenotype to a single mutation;
however, C1085 does not revert, indicating that it contains at least
two mutations, each sufficient to block expression of the COXI gene (see below). Transient complementation experiments (see Ref.
30) showed that each strain has a cis-dominant defect of the COXI gene, indicating that none of the mutations inactivates a trans-acting splicing factor such as a maturase.
Splicing Phenotypes of the Mutants
Mitochondrial
RNA was extracted from cultures of the three mutant strains,
fractionated on agarose gels, and analyzed by Northern blot
hybridization. In mutant C1085, the COXI exon 4 probe detects
a 4.8-kb precursor RNA instead of the 1.9-kb COXI mRNA (Fig. 2, lane5). Hybridization using
intron-specific probes shows that the 4.8-kb RNA species contains
unspliced aI3 and aI5 (data not shown). The mapping data of Fig. 1do not rule out a second mutation downstream of aI3.
In mutants C1072 and M44, a major 3.4-kb COXI precursor RNA is
detected with the exon-specific probe (Fig. 2, lanes2 and 3) and with an aI3-specific probe,
showing that it contains unspliced aI3 plus exon sequences. Thus,
all three mutant strains are defective for aI3 splicing. Finally,
a spontaneous revertant of C1072 was isolated and found to restore a
wild-type level of splicing (Fig. 2, lane4)
(see below).
Figure 2:
COXI transcripts accumulate in
mutant strains. The strains shown were grown to late logarithmic phase
in YEP medium with 2% raffinose, and RNA was extracted. The RNA samples
were fractionated on gels as described under ``Experimental
Procedures,'' and the blotted gel was hybridized with a
5`-end-labeled oligonucleotide probe complementary to COXI exon 4. The sizes of the major bands detected are indicated on the
left. The 1.9-kb RNA is COXI mRNA; the 3.4-kb RNA is mRNA
spliced for all introns except aI3; and the 4.8-kb RNA is mRNA
spliced for all introns except aI3 and
aI5.
The Northern blots indicate that 3% of the COXI transcript in strains M44 and C1072 is apparently
spliced; however, the putative spliced RNA appears to be somewhat
smaller than the mRNA in the control sample (Fig. 2, compare lane1 with lanes2 and 3). To determine whether these mutant RNAs are accurately
spliced, cDNAs were amplified by an RT-PCR strategy and characterized.
The amplified material was found to be of the expected size (281 bp)
from the wild-type strain (Fig. 3, lane 2), but was
clearly smaller from both mutant strains (Fig. 3, lanes4 and 5). Most of the amplified cDNA from the
mutants was
186 bp long, but there were several other minor
products of intermediate sizes.
Figure 4:
aI3 splices inaccurately in mutant
C1072 but accurately in a petite mutant. A, sequence of cDNAs
from wild-type and mutant strains. cDNAs from an RT-PCR experiment like
the one shown in Fig. 3 were excised from a gel and sequenced using a
primer in exon 4. The sequences across the splice junction in the
281-bp cDNAs from wild-type and a mutant are
shown in the firsttwopanels. The sequences
across the splice junctions in the 186- and 227-bp cDNAs from mutant
C1072 are shown in the secondtwopanels.
The 186- and 227-bp cDNAs from mutant M44 were also found to result
from splicing at the same cryptic splice sites as in C1072 (data not
shown). Beside the gel lanes are shown the sequence of the antisense
strand of exon 4 (upper-case letters) and the 5`-exon sequence (lower-case letters). B, summary of accurate and
cryptic splice sites. A diagram of the first four exons of the COXI gene plus aI3
is shown. The sequence around each splice site
defined in A is shown. The splice sites used in the wild-type
and petite mutant strains are indicated by thickarrows. The major and minor cryptic splice sites used in
C1072 (and M44) are indicated by the tall and shortthinarrows, respectively. C, putative
P1 helices involved in accurate and cryptic splicing events. The three
pairings between the exon sequence preceding the splice site and the
internal guide sequence (P1 3`-element) of the intron are shown. In
each case, the splice site is shown by an arrow. The first is
the wild-type P1 helix; the second (P1`) is the helix formed by the
most abundant cryptic splice site in exon 1; and the third (P1") is the
helix formed by the cryptic splice site in exon 2. Exon nucleotides are
shown as upper-case letters, and intron nucleotides are shown
as italic lower-case letters.
We
note that the sequence of the wild-type form of aI3 in strain
ID41-6/161 used here differs at a number of sites from the published
sequence of strain D273-10B (see below). One of the sequence
differences alters the internal guide sequence of the intron in
ID41-6/161, substantially improving the P1 helix involving the internal
guide sequence. The P1 helix of aI3 of strain ID41-6/161 is shown
in Fig. 4C. The major cryptic splice site is capable of
forming a P1 helix that is nearly as stable as the one involved in
correct splicing (Fig. 4C), while the minor cryptic
splice site can make only a very short P1 helix. We conclude that
mutants M44 and C1072 are unable to form the wild-type P1 structure
(see below) and that nearby sequences that resemble the 5`-arm of P1
are used (inefficiently) instead.
from strain ID41-6/161 self-spliced under
conditions defined by Tabak et al.(25) , an in
vitro precursor RNA containing the C1085 mutations was completely
inactive. RNA containing the M44 allele was defective for reactions at
the 5`-splice site, but still carried out reactions at the 3`-splice
site and at an internal site, both of which occur with the wild-type
intron(25, 43) . Finally, RNA from strain C1072
self-spliced efficiently, although it has a partial deficiency in the
second step of splicing. The in vitro precursor RNA used in
those experiments lacks E1 and E2, so the cryptic splice sites detected in vivo were excluded from those self-splicing experiments.
aI3
Previous studies do not establish conclusively whether
aI3 Does Not Encode or Require a Maturase for
Splicing encodes a maturase(25, 44) . Our recent studies
of the phenotypes of mutants blocked for aI1 and aI2 splicing suggest
that aI3 does not encode a maturase since upstream splicing
defects, which would prevent expression of the aI3 reading frame,
do not block aI3 splicing(45) . We have investigated this
point explicitly by testing whether aI3 splices in petite mutants
that contain the COXI gene. Because petite mutants are
incapable of mitochondrial protein synthesis, any intron that splices
in petites (for example, , bI1, bI5, and aI5
) (23, 44, 46, 47) cannot encode a
maturase nor can its splicing depend on one encoded by another intron.
1,2;
this COXI allele contains introns 3
through 5 but
lacks aI1 and aI2. This intron configuration was chosen because aI1 and
aI2 do not splice in petite strains(44) , so their inclusion
would require the analysis of very large precursor RNAs. Northern blots
of RNA from strain 161
1,2 indicated that
the most abundant COXI transcript present is spliced for
aI3
(data not shown). We carried out the same RT-PCR experiment
used above with the mutant strains to determine whether the efficient
splicing of aI3 in this petite mutant is accurate. As shown in Fig. 3(lane3), the petite mutant yields the
same amplified cDNA as was obtained using RNA from a control
strain. The cDNA sequence showed that the intron
is spliced accurately in this petite. This finding establishes that
aI3
does not encode a maturase needed for its splicing and also
rules out the possibility that it depends on a maturase encoded by
another intron.
DNA Sequence Analysis of Mutant Strains
Fig. 5shows the core secondary structure of aI3 from strain
ID41-6/161 and summarizes the results of DNA sequencing experiments
that define the mutations in and around aI3 of the mutant strains.
These sequence data are reported using the published sequence
coordinates of the COXI gene of strain D273-10B (27) (GenBank accession number JO1481). The
sequence of aI3
from strain ID41-6/161 differs from that of strain
D273-10B at a number of positions, both within the intron reading frame
and in the closed reading frame that contains nearly all of the intron
core structure. Four sequence differences in the closed reading frame
are shown in Fig. 5, three of which influence the core secondary
structure of the intron. Finally, we note that the sequence in our
strain is identical to that of strain KL14-4A (25) and to the S. capensis strain analyzed by Szczepanek et al.(21) (GenBank accession number U00801), except
that it lacks the two stop codons present in the aI3
reading frame
of S. capensis.
Figure 5:
Sequence analysis of COXI gene
mutations. A secondary structure of aI3 of strain 161 is shown;
all helical (paired) regions of the intron are labeled (P1, P2, etc.).
The entire sequence of the intron is shown except for the 1099 bp of
loop 1 (L1), which includes nearly all of the intron reading
frame. Several sequence coordinates are provided using the published
sequence of this intron from strain D273-10B (27), and individual
nucleotides are marked with a filledcircle to
designate each span of 50 nucleotides (nt). Nucleotides in the
intron sequence from our strain that are reported as absent in strain
D273-10B are boxed. All nucleotide changes present in the core
structure of the intron of each mutant are indicated. M44 has just one
change in the part of exon 3 that is present in P1. C1072 has one
change in exon 3, just upstream of P1, and a second change in P6b.
C1085 has point mutations altering P1, P3, P4, P5b, P5c, P5e, and P6a;
it also has an insertion of an Ade residue after nucleotide 6218 in
loop 1.
Mutants C1072 and M44 were found to contain
different missense mutations in exon 3. In C1072, Gua is
mutated to Ade, changing a glycine to aspartate. There is also a T to C
transition in P6b that is outside of the region of the gene where the
phenotypic defect maps (compare Fig. 1and Fig. 5). The
spontaneous revertant of C1072 noted above in Fig. 2regained the
wild-type sequence in the exon but not in P6b, showing that the exon
mutation is responsible for the phenotype. It is surprising that an
exon mutation there would affect splicing since G
is
outside of the P1 helix of the intron, and no other splicing signal
upstream of P1 is known. Only one mutation, G
to T, was
found in mutant M44, changing a glycine to valine. As shown in Fig. 5, the mutated nucleotide is in the P1 helix of aI3
.
Several spontaneous revertants were sequenced, and each had the
wild-type exon sequence and spliced efficiently (data not shown).
Similar mutations of the P1 structure of other introns are known to
block splicing in vivo (e.g. Refs. 48 and 49).
. This part of the intron contains most
of the known cis-acting splicing signals, some of which (P1,
P3, P4, P5, and P6) are altered in this mutant. As shown in Fig. 2, this mutant is blocked for splicing of aI5, so it is
not surprising that it does not revert. The open reading frame of
mutant C1085 contains no mutations; thus, it is the only one of the
three mutants studied here that makes a completely wild-type precursor
protein encoded by the aI3 reading frame plus upstream exons.
Novel Mitochondrial Proteins Accumulate in the Mutant
Strains
Most splicing-deficient mutants of yeast mtDNA
accumulate pre-mRNAs, which often lead to elevated levels of
intron-encoded proteins(6) . Analysis of the in vivo mitochondrial translation products revealed that these three
mutants lack CoxI protein (Fig. 6, lanes 2-4).
Mutants M44 and C1085 accumulate 35- and 44-kDa proteins (p35 and p44,
respectively) not apparent in wild-type cells. In addition, mutant
C1072 accumulates a third protein of 38 kDa (p38).
open reading frame contains 335 codons. If translation begins at the
AUG codon of exon 1, then the predicted product would be 416 amino
acids long, roughly the size of the largest protein (p44) observed in
the mutants. At least two other group I intron reading frames (bI4 and
aI4) are translated as precursor proteins, presumably initiated
from the AUG start codon of exon 1; in those instances, the functional
intron-encoded protein is a large C-terminal fragment processed from
the precursor(7, 40, 51) . The p35 and p38
species present in these aI3 mutants probably result from such
processing of p44 (see below).
The aI3
Recently, Szczepanek et al.(21) showed
that the aI3 Mutant Strains Overproduce I-SceIII
Activity allele of S. cerevisiae strain 777-3A is
mobile in crosses. Mitochondria from this donor strain contain an
endonuclease activity (I-SceIII) encoded by the aI3
reading frame that cleaves the exon 3-exon 4
junction(19, 26, 53) . For our strain, we have
confirmed aI3 mobility and the exact cleavage site of the
endonuclease (data not shown). Using essentially the same preparative
methods used previously to study I-SceII(40) , we found
that mutant C1085 has a roughly 10-fold elevated level of I-SceIII activity (Fig. 7) and obtained a similar result
for mutant M44 (data not shown). The I-SceIII activity of
these strains is readily released from broken mitochondria by
extraction with 1 M KCl. Analysis of I-SceIII
activity following partial purification by ammonium sulfate
precipitation and chromatography on a phosphocellulose column (see
``Experimental Procedures'') revealed the same preferences
for reaction conditions reported by Schapira et al.(26) for an enzyme preparation of I-SceIII
expressed in E. coli from a universal code equivalent of most
of the aI3 reading frame.
Figure 7:
Mutant C1085 accumulates 10 times
more I-SceIII activity than does its wild-type parent strain.
Strains 161 (
) and C1085 (in the nuclear
background of strain ID41-6/161) were grown on 2% galactose containing
YEP medium, and mitochondrial fractions were isolated as described
under ``Experimental Procedures.'' Fraction 1a from the
standard purification protocol (see ``Experimental
Procedures'' and Fig. 8A) from each strain was used here.
Three different amounts of protein were used, and the fraction of an
excess of end-labeled substrate plasmid DNA converted to product was
measured by PhosphorImager scanning of agarose gels. WT, wild
type.
The 35-kDa Protein Is the Most Abundant Active Form of
I-SceIII
As shown above, these mutant strains contain several
proteins likely to be encoded by aI3. We used affinity-purified
polyclonal antibodies raised against a portion of the aI3 reading
frame (shaded region in Fig. 1; see ``Experimental
Procedures'') to characterize the aI3-encoded proteins
present in isolated mitochondria and in various fractions enriched for I-SceIII activity. As shown in Fig. 6(lanes
5-8), the antibody detects all of the novel species seen by
[S]SO
labeling
(compare with lanes 1-4). A low level of both p35 and
p44 is evident in the mitochondrial fraction from the wild-type strain.
The distribution of signal between the two bands is about the same in
the wild-type strain as in the mutants, suggesting that the presence of
the larger protein is not a consequence of the mutations. As
illustrated in Fig. 8B, however, only p35 is detected in
partially purified samples of I-SceIII. Since activity was
recovered efficiently in this step but only p35 was extracted, we
conclude that p35 accounts for most, if not all, of the I-SceIII activity.
S]SO
-labeled p35
and p44 were excised from a 10-15% acrylamide gradient gel
containing SDS and urea and digested with Staphylococcus aureus V8 protease prior to separation of the products on a 10%
polyacrylamide gel. As shown in Fig. 9, the undigested samples
have the expected mobility, and the digested samples yield very similar
protein fragment patterns. While it is generally accepted that some
intron-encoded proteins accumulate as large C-terminal fragments of a
precursor protein, this study contains the most clear-cut demonstration
of this point to date.
Figure 9:
p35 and p44 are closely related proteins.
Samples of
[S]SO
-labeled p35
and p44 from strain C1085 (161 nuclear background) were excised from a
gel like the one shown in Fig. 6 (lane4), digested
with S. aureus V8 protease, and analyzed on a 10%
polyacrylamide gel (lanes3 and 4,
respectively). Lanes1 and 2 show the
undigested starting material for p35 and p44,
respectively.
.
Strains C1072, M44, and C1085 are the first in vivo mutants to
be described affecting the splicing of aI3. While C1085 has the
most complex set of mutations, it is the only mutant that is not
mutated in the COXI exons and in the aI3 open reading
frame. The mutations responsible for the splicing defect in the other
two mutants are shown conclusively to be in the 35-bp exon 3. Each
mutation blocks the splicing of aI3 and has no effect on the
splicing of the adjacent group II intron, aI2.
. P1 is a secondary structure that contains the internal guide
sequence crucial for defining the 5`-splice site of all group I
introns(54, 55, 56) . Overall, the phenotype
caused by this mutation is similar to that of P1 mutants characterized
in other introns (e.g. Refs. 48 and 57). Some second-site
mutations that restore base pairing in P1 helices of other introns
restore efficient splicing(55, 58) , so our failure to
find suppressors in the internal guide sequence of this intron suggests
that either the glycine altered by the mutation is essential for COXI function or that the aspartic acid present there in the
mutant may be incompatible with COXI function.
nearly
blocks the use of the wild-type 5`-splice site and activates the use of
a cryptic 5`-splice site 6 nucleotides upstream. As noted above, the
M44 mutation blocks the use of the 5`-splice junction under
self-splicing conditions, but we have no indication that it activates
the cryptic splice site seen by Winter et al. (60).
Interestingly, the M44 mutant intron is quite reactive at the 3`-splice
junction and at a site within the intron that resembles the natural
5`-splice junction(42) , carrying out in vitro reactions at those sites typical of group I
introns(25, 43) .
, but permits a low
level of splicing using cryptic 5`-splice sites. Sequencing of several
cDNAs obtained from RT-PCR experiments defines a major cryptic splice
site in exon 1 and a minor one in exon 2; both sites resemble the
natural splice site, and each is capable of forming an alternative P1
helix. Strain G2457 has a similar mutation of the P1 helix of the
fourth intron of the COB gene and causes a similar splicing
defect, including missplicing(58) .
is not
explained by any feature of the current secondary structure model for
group I introns. Several studies have shown that the sequences flanking
a group I intron can influence splicing. Alternative secondary
structures associated with mutations in the 5`-exon can affect the
splicing of the Tetrahymena group I
intron(61, 62) . Seraphin et al.(63) showed that the MSS18 gene of yeast encodes a
protein needed for splicing of the group I intron, aI5. Point
mutations that suppress the splicing defect were found in the exon
preceding aI5, and these mutations were, likewise, in sequences
not already associated with the secondary structure of the intron.
splicing that is limited to the in vivo state.
Since mutant C1072 carries out the same cryptic splices as were found
for the authentic P1 mutant, M44, we conclude that the C1072 mutation
probably also interferes with the formation or stability of P1. While
the C1072 mutation does not appear to create a sequence that can pair
with either element of the P1 helix, it is possible that it alters some
other secondary structure, perhaps involving just exon sequences, that
now can base pair with the 3`-end of exon 3 and block the formation of
the natural P1 structure.
present in our wild-type strain ID41-6/161 is a
mobile group I intron. We have characterized the I-SceIII
endonuclease activity present in strain 161 and several of these mutant
derivatives and confirmed the cleavage site reported by Schapira et
al.(26) . Our studies of the intron-encoded protein
establish that the active enzyme is the 35-kDa C-terminal portion of a
44-kDa precursor protein. Our experiments define the molecular mass of
the active endonuclease (p35) and identify a specific precursor to p35.
Because of the splicing defects, all three mutants accumulate chiefly a
precursor to COXI mRNA that probably serves as the mRNA for
the 44-kDa precursor protein. We showed that blocking splicing leads to
an 10-fold overproduction of the intron-encoded protein based on
both I-SceIII activity and the level of p35 + p44
detected in Western blots of lysed mitochondria. RNA blots, however,
indicate that the level of 3.4-kb pre-mRNA is increased much more than
10 times (Fig. 2). This discrepancy suggests that the
accumulation of p35 may be controlled at a post-transcriptional level.
splices efficiently and accurately in a
petite mutant shows that this intron neither encodes a maturase nor
depends on one encoded by another mitochondrial reading frame. A number
of proteins encoded by nuclear genes participate in the splicing of
group I introns in yeast mitochondria(64) . So far, genes needed
for splicing of four introns have been reported, although none affects
aI3. The only nuclear gene reported that is known to interact with
this intron is suv3, a putative RNA helicase(65) . A
dominant mutant allele of this gene (SUV3-1) causes most
excised group I intron RNAs, including aI3, to accumulate in
mitochondria(66) .
reading frame containing 321
of the 335 codons of the intron reading frame and reported that the
active protein that accumulates is 35 kDa. Since we found an
active p35 species in vivo, we conclude that it probably does
not contain more than the 321 amino acids present in the gene that was
expressed in E. coli. However, Schapira et al.(26) did not determine whether the active protein obtained
from E. coli cells is full-length or, perhaps, processed.
,
aI4, aI5, aI5, and aI5, introns 3, 4, 5, 6, and 7,
respectively, of the long form of the COXI gene of yeast
mtDNA; bI1, intron 1 of the COB gene; I-SceI, I-SceII, I-SceIII, and I-SceIV,
endonucleases encoded by the reading frames of introns
,
aI4
, aI3, and aI5, respectively; kb, kilobase(s); bp,
base pair(s); mit, generic term for
respiration-deficient mutants of yeast due to base substitutions or
small deletions in mtDNA; RT-PCR, reverse-transcription polymerase
chain reaction;
, genotypic designation for
wild-type mtDNA;
, genotypic designation for
mtDNA in a respiration-deficient mutant due to a large deletion of
mtDNA.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. H. Barros, A. M. Myers, S. Van Driesche, and A. Tzagoloff COX24 Codes for a Mitochondrial Protein Required for Processing of the COX1 Transcript J. Biol. Chem., February 10, 2006; 281(6): 3743 - 3751. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Lin and V. M. Vogt I-PpoI, the Endonuclease Encoded by the Group I Intron PpLSU3, Is Expressed from an RNA Polymerase I Transcript Mol. Cell. Biol., October 1, 1998; 18(10): 5809 - 5817. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |