The Nuclear Gene MRSB Is Essential for the Excision of Group I1 Introns from Yeast Mitochondrial Transcripts in Vivo*

RNA splicing defects in mitochondrial intron mutants can be suppressed by a high dosage of several proteins encoded by nuclear genes. In this study we report on the isolation, nucleotide sequence, and pos- sible functions of the nuclear MRS2 gene. When present on high copy number plasmids, the MRSB gene acts as a suppressor of various mitochondrial intron muta- tions, suggesting that the MRS2 protein functions as a splicing factor. This notion is supported by the observations that disruption of the single chromosomal copy of the MRSB gene causes (i) a pet- phenotype and (ii) a block in mitochondrial RNA splicing of all four mito- chondrial group I1 introns, some of which are effi- ciently self-splicing in uitro. In contrast, the five group I introns monitored here are excised from pre-mRNA in a MRS2-disrupted background although at reduced rates. So far the MRS2 gene product is unique in that it is essential for splicing of all four group I1 introns, but relatively unimportant for splicing of group I in- trons. In strains devoid of any mitochondrial introns the MRSS gene disruption still causes apet- phenotype and cytochrome deficiency, although the standard pattern of mitochondrial translation products is produced. Therefore, apart from RNA splicing, the absence of the MRSS protein may disturb the assembly of mitochon- drial membrane complexes.


The Nuclear Gene MRSB Is Essential for the Excision of Group I1 Introns from Yeast Mitochondrial Transcripts in Vivo*
(Received for publication, June 3, 1991) Gerlinde WiesenbergerS, Martin Waldherr, and Rudolf J. SchweyenQ From the Institut fiir Mikrobiologie und Genetik, Universitat Wien, A-1090 Vienna, Austria RNA splicing defects in mitochondrial intron mutants can be suppressed by a high dosage of several proteins encoded by nuclear genes. In this study we report on the isolation, nucleotide sequence, and possible functions of the nuclear MRS2 gene. When present on high copy number plasmids, the MRSB gene acts as a suppressor of various mitochondrial intron mutations, suggesting that the MRS2 protein functions as a splicing factor. This notion is supported by the observations that disruption of the single chromosomal copy of the MRSB gene causes (i) a petphenotype and (ii) a block in mitochondrial RNA splicing of all four mitochondrial group I1 introns, some of which are efficiently self-splicing in uitro. In contrast, the five group I introns monitored here are excised from pre-mRNA in a MRS2-disrupted background although at reduced rates. So far the MRS2 gene product is unique in that it is essential for splicing of all four group I1 introns, but relatively unimportant for splicing of group I introns. In strains devoid of any mitochondrial introns the MRSS gene disruption still causes apet-phenotype and cytochrome deficiency, although the standard pattern of mitochondrial translation products is produced. Therefore, apart from RNA splicing, the absence of the MRSS protein may disturb the assembly of mitochondrial membrane complexes.
Organelle introns have been classified as group I and group 11 introns according to conserved secondary structure features and group specific excision pathways (1,2). Group I1 introns have been detected so far in mitochondrial and chloroplast genomes whereas group I introns are widespread in other genomes, namely in the nuclear rRNA genes of Tetrahymena and in genes of bacteriophages and cyanobacteria (1)(2)(3)(4)(5)(6).
Some members of each group have been shown to be catalytic in vitro; they excise themselves from pre-mRNA and ligate the exons in the absence of any protein (7-9). The splicing pathway and conserved sequence elements at the splice junctions relate group I1 introns with the nuclear mRNA introns; both classes of introns are excised as branched * This work was supported by the Austrian Fonds zur Forderung der wissenschaftlichen Forschung. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in thispaper has been  circular RNAs, the so-called lariats, with a most intriguing 2'-5' phosphodiester bond. This relatedness has led to the speculation that the autocatalytic group I1 introns are the evolutionary ancestors of the nuclear (non-autocatalytic) mRNA introns (10).
The finding of self-catalysis of mitochondrial introns in vitro contrasts a number of observations which have revealed that excision of group I and group I1 introns in vivo involves proteins (see Refs. 11 and 12 for review), encoded either within the introns themselves, the so-called maturases, or by nuclear genes. Two of these nuclear genes in yeast, MRSl and CBP2, encode products that are needed only for the excision of group I introns (b13 and aI5b b15, respectively); yeasts lacking the introns in question are unaffected by disruption of these nuclear genes (13)(14)(15). Others encode mitochondrial proteins of dual function, like CYT18 in Neurospora crassa and NAMZ in yeast which serve as tRNA synthetases and are in addition essential splicing factors (16,17). Finally, the yeast MSS116 gene encodes a protein with similarity to a helicase; its absence affects the excision of both group I and group I1 introns and other, so far unknown functions (18).
We previously described the isolation of three nuclear genes, MRS2, MRS3, and MRS4, whose products appear to be involved in the excision of group I1 introns from mitochondrial pre-mRNA (19). They have been detected by virtue of their ability to suppress a splice defect exerted by a group I1 intron mutation, when present in high copy number.
MRS3 and MRS4 were found to be closely related genes; both can be disrupted without causing any phenotypically detectable effect on splicing or any other function (20,21). We show here that the MRS2 gene, unlike the MRSB and MRS4 gene, is essential for the excision of all four group I1 introns present in yeast mitochondrial RNAs, whereas it hardly affects the excision of group I introns. MRSB has additional, so far unknown functions since its disruption in yeast strains lacking all known introns still leads to respiratory deficiency.

MATERIALS AND METHODS
Strains and Plasmids-The genotypes and origins of the yeast strains used in this study are described in Table I. Escherichia coli strain XL1-Blue (Stratagene) and the following plasmids were used for subcloning: YEp351 (24), Bluescript (Stratagene).
Rho" derivatives of yeast strains were obtained by growing cultures for about 30 generations in YPD medium containing 50 pg/ml ethidium bromide.

6963
DNA Subcloning and Sequencing-Appropriate DNA restriction fragments were subcloned into Bluescript vectors and sequencing was done with the didesoxy chain termination method using the T7 sequencing kit (Pharmacia LKB Biotechnology Inc.). Restriction enzyme sites used for subcloning are shown in Fig. 1. For sequencing of some larger fragments we used specific primers. Nucleotide sewe used the MICROGENIE program (Beckman). The SWISSPROT quences of both strands were determined. For DNA sequence analysis Protein Sequence Library (release 18, May 1991) and the EMBL Nucleotide Sequence Library (release 27, May 1991) were searched by use of the FASTP program (26).
Transformation of Yeast and E. coli-E. coli strains were transformed by the CaCl, procedure (25). Yeast transformation was done using the lithium acetate method (27).
Southern Blot Analyses-DNA preparation from yeast was performed following the method described in Ref. 28, whereas blotting and hybridization was done as described in Ref. 25.
Northern Blot Analyses-Northern blot analyses were done as described previously (21).
Analyses of Mitochondrial Translation Products-Labeling of mitochondrial proteins was done essentially as described by Haid et al. (29) except that cells were cultivated in YPR medium and starved for methionine prior to labeling in SR-meth medium (synthetic medium with raffinose as carbon source, lacking methionine). Cells were labeled with [35S]methionine (Amersham Corp.; 25 pCi/ml culture) in the presence of cycloheximide (100 pglml). Small scale isolation of mitochondria was done according to the method described by Needleman and Tzagoloff (30). Proteins were separated on 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and after fluorography the dried gels were autoradiographed (31).
Cytochrome Spectra-Cells were grown on YPR-plates for 2 days, suspended in water, and washed twice with water. After treatment of the cells with an excess of sodium dithionite in suspension, the pellets were applied to the window of a home-made cuvette and frozen in liquid nitrogen.
Spectra were recorded against several layers of Parafilm in a Hitachi 150-20 spectrophotometer. polarity phenomenon in COB RNA splicing, of all downstream introns (32,33).

Multi-copy Suppression-In a previous communication we
By Northern blot analyses we could demonstrate that the MRS2 gene product is in fact involved in mitochondrial RNA splicing: in transformants of the mutant (strain DBY747/ M1301) with the MRSS gene on the multi-copy plasmid YEp351 we could detect a band that corresponded to the mature transcript of the COB gene and which is not present in the same strain transformed with YEp351 alone (data not shown).
Subcloning and Expression of the MRS2 Gene-Subcloning of the MRSZ gene was done by partial digestion of the original 4.8-kb' BamHI insert (19) with the restriction endonuclease S a d A and cloning of fragments with sizes of 1.5-4 kb into a YEp351 vector linearized with BamHI. The suppressing activity could be allocated to a 2.2-kb fragment (Fig. 1). Suppressor activity of this fragment was lost when the 0.5-kb SauI fragment was replaced by the HIS3 gene (Fig. 1).
By use of the OFAGE technique (35) and the 1.1-kb HindIII restriction fragment as a hybridization probe we were able to locate the MRSZ gene on chromosome XV.
Northern blot analyses of poly(A)+ RNA revealed that MRS2 is weakly transcribed in vivo. A transcript of about 1.6 kb could be detected by hybridization with a 32P-labeled 1.1kb HindIII fragment of MRS2. Transcription of MRS2 appears not to be regulated by carbon source (data not shown).
Sequence Analyses of MRS2"The sequence of the 2.2-kb subclone is shown in Fig. 2. It contains one long open reading frame which can code for a protein of 470 amino acids with a calculated M, of 54,186.
The codon bias index (36) of the predicted MRS2 protein has been calculated to be 0.17 indicating that the gene is poorly expressed. The protein appears to be hydrophilic; about 24% of its amino acids are either positively or negatively charged, and the calculated polarity index (37) is 46.2%. The predicted amino acid sequence indicates that the protein is basic with an excess of 70 basic over 47 acidic residues.
Many of the nuclear encoded proteins that are directed into mitochondria have been shown to contain leader sequences. These leaders, which are cut off during import into mitochondria, are assumed to form amphiphilic helices, which are rich in arginine, leucine, and serine but do not contain lysine or any acidic residues (38). The N-terminal part of the predicted MRSB protein may form such a leader; this and the observation that MRSZ plays a role in mitochondrial RNA splicing indicate that MRSB is a mitochondrial protein. Poor in vitro translation efficiency of the MRS2 gene prevented us so far ' The abbreviation used is: kb, kilobase(s).
from performing import assays of the MRSB protein into isolated mitochondria as has been successfully done with the MRSB and MRS4 proteins (21).
The DNA sequence of the MRSB gene and the deduced amino acid sequence were compared with sequences of the available data libraries (EMBL and SWISSPROT). We did not find any convincing similarities of the MRSZ sequences with those in the libraries, nor did we detect any consensus sequence motifs which might represent RNA, DNA, or nucleotide binding elements.
Effects of the MRS2 Gene Disruption on Mitochondrial RNA Splicing-In order to inactivate the chromosomal copy of the MRSB gene, an internal 0.5-kb SauI fragment of this gene was replaced by a 1.7-kb fragment containing the HIS3 gene (Fig. 1). The haploid strain DBY747 rho+ was transformed with this construct following the one-step gene disruption method (39). Stable HIS' transformants were isolated, and the replacement of the wild-type MRSB copy by the disrupted one was confirmed by Southern hybridization (data not shown). The mrs2::HIS3 gene-disrupted strain (GW7/gd2-21.2) was unable to grow on non-fermentable substrates (cf. Fig. 6). The disruption thus causes a pet-phenotype. This indicates that the MRSB gene product is essential for the formation of a functional respiratory chain.
To determine a possible effect of the MRSZ gene disruption on mitochondrial RNA splicing, we have performed a series of Northern hybridization experiments using mtRNA of strain GW7/gd2-21.2 which contains the disrupted MRSZ gene. The patterns of the COB and COX1 transcripts of this strain are shown in Fig. 3 in parallel with those of the wildtype strain DBY747 and the intron bI1 mutant DBY747/ M1301, which are isogenic with strain GW7/gd2-21.2 except for the MRSB alleles and the bI1 mutation.
Probes specific for the COB exon B1 (including part of the leader) and for the COB intron bI1 detect the mature mRNA and the excised bI1 RNA as the major COB transcripts in the wild-type strain (Fig. 3). The strains with the nuclear MRSB gene disruption (GW7/gd2-21.2) or with the bI1 intron mutation (DBY747/M1301) lack these mature RNAs. Instead, they accumulate high molecular weight RNAs of similar sizes. These have been shown to comprise all COB exon and intron sequences, or all but the last COB intron (32).
The pleiotropic effect on intron excision may be due to a primary block in the excision of the group I1 intron bI1 which, as a secondary effect, prevents the expression of maturases in downstream introns and hence their excision from pre-mRNA (32,(40)(41)(42). Alternatively, disruption of MRS2 may equally block excision of group I and group I1 introns. This latter possibility was ruled out by the following experiment. Mitochondria of strain KGF177 (Saccharomyces uuarum) and of strain CK506 (Saccharomyces cerevisiae), containing only group I introns b14 and b15 or introns b13, b14 and b15, respectively (22, 14), were combined by cytoduction with the MRS2 disrupted nuclear background (strain GW7/gd2-21.2). The cytoductants accumulate mature COB mRNA (data not shown). The same result was obtained with a mrs2:: strain lacking all group I1 (and some of the group I) introns as shown in Fig. 5 . We conclude therefore that excision of the group I introns b13, bI4, and b15 from the COB precursor is only little effected by the MRSS disruption, whereas splicing of the group I1 intron bI1 is totally blocked in a mrd:: strain. Additionally, we were not able to investigate the excision of the group I intron b12, since there was no strain available so far that lacks only bI1; excision of bI1 is necessary for splicing of b12, because this step is dependent on a functional maturase (32).

~O S E S L I~D I K S T E E I V N I I L D A N R N S L M L L E L K V T I Y T L
TTAGAAAACGACGATGATTTAGCAAACATGTACTTGACAGTTAAGAAGTCTCCTAAGGACAATTTTTCGGACTTGGAAATGCTTATAGAGACGTACTACACCCAATGTGATGAATACGTT CAGCAATCAGAATCTTTGATTCAGGATATCAAATCTACTGAAGAAATTGTCAACATCATATTGGACGCAAATAGAAATTCCTTAATGTTGTTGGAGTTGAAAGTTACCATCTACACGTTG  MRS2 (right lanes, mrs2::) were hybridized with radioactivity labeled, in vitro generated RNA probes complementary to a part of the untranslated leader and the first exon (COB-1, panel A), to the first  intron ( b l f , panel B )  In order to test whether excision of group I1 introns in the COXl transcript is also affected, Northern hybridization experiments were performed with probes specific for exon A4 and for the three group I1 introns in the COXl gene, aI1, a12, and aI5c (Figs. 3 and 4). As shown in Fig. 3 (panel C) no mature mRNA can be detected in the mrs2:: strain.

K V K K F N M D N D I K N K~N R D M I W K W L I E D K K N
Excision of introns a11 and aI5c is independent of the presence or absence of other introns whereas excision of intron a12 requires that its open reading frame, possibly encoding a RNA maturase, is in frame with the upstream exons A1 and A2 (43). Therefore, we have chosen a strain lacking intron a11 (GW7/gd2-21.2/SU) to directly monitor excision or non-excision of intron a12.
The transcript patterns of the mrs2:: strains lack excised group I1 intron RNAs (aI1, aI5c; Fig. 4, panels B and D ) or show only traces of excised a12 RNA (Fig. 4C), whereas the corresponding bands are very dominant in the wild type. Instead, the disruption of MRS2 leads to accumulation of high molecular weight transcripts hybridizing with all group I1 introns. Taken together the results presented above show that the MRS2 gene product is an essential factor for the excision of all four group I1 introns from yeast mitochondrial COB and COXl pre-mRNAs. The COXl transcript pattern of the mrs2::HISB strains, however, differs from that of the bI1 mutant M1301 (Fig. 3,  probe A4)  aI2, or aI5c are marked by arrowheads. Note that strain GW7/gd2-21.2/SU bears mitochondria of S. uuarum which lack the upstream intron aI1; hence, accumulation of pre-RNAs containing intron a12 directly reflects reduced splicing efficiency of this intron and is not due to a polarity effect.
of COXl introns but it exerts only the so-called "box" effect, an interplay between the introns b14 and a14 (44). The COXl transcript pattern of the mrs2::HIS3 strain can be interpreted as containing various combinations of the three group I1 introns and the intron a14. This indicates that the other group I introns of the COXl pre-mRNA are excised in the absence of the MRS2 gene product. A polarity effect similar to that of the non-excised intron bI1 in COB pre-mRNA may not be detected with the COXl pre-mRNA since the effect of the MRS2 disruption on excision of the first two introns a11 and a12 is somewhat leaky and thus allows for sufficient expression of maturases encoded in the downstream introns. A more direct test for the effect of the MRS2 gene disruption on the excision of group I introns was performed with mitochondria lacking all group I1 introns but retaining some group I introns only. For this experiment mitochondria from strain GF132-10A (retaining group I introns b14, b15, a13, a14) were introduced into the nuclear background of the MRS2-disrupted strain GW7/gd2-21.2 via cytoduction. In Fig.  5 the effect of MRS2 disruption on processing of COB and COXl RNAs is compared in various mitochondrial backgrounds. In mrs2:: strains containing the "long" version of mitochondrial DNA (bI1-b15; aI1-aI5c) no mature mRNAs of both genes can be detected. In contrast, in mrs2:: cells lacking all group I1 and some group I introns ( A group 11) mature mRNAs are clearly present, although in a slightly lower amount than in the correspondent wild type. Further predominant bands in this figure correspond to COB mRNA+bI5 (panel A ) and COXl mRNA+aI4 (panel B ) .
These bands, which are also present in the wild type lanes, together with some other weaker precursor bands occurring only in the mrs2:: background might indicate that disruption of MRSB has also some effect on group I intron splicing. In the A group I1 strain (containing only introns a13 and a14) one would expect only three COXl precursor bands. The A4 probe we used contains also 870 bases of intron a14, and this intron is highly homologous to the COB intron b14 (1). Additional precursor bands in a blot hybridized with the A4 probe are caused by a crosshybridization between these two introns.
Functions of the MRS2 Gene Product aside from RNA Splicing-As shown above, deletion of group I1 introns or of all introns from mtDNA apparently cures MRS2-disrupted strains of their defect in RNA processing. If the only function of the MRS2 gene product were to participate in splicing, one would expect that this intron deletion would also restore growth of MRS2 disrupted cells on a non-fermentable substrate. This was not the case; the MRS2 gene disruption caused the same pet-phenotype irrespective of the mitochondrial genotype (Fig. 6). This suggests that the MRSZ gene product has another essential function in mitochondria. A comparison of cytochrome spectra of wild-type (DBY747/ wo; MRS2) and mrs2::HIS3 (GW7/gd2-21.2/wo; mrs2::) strains with intron-less mtDNAs reveals a striking absence of cytochrome a.a3 and a dramatic decrease of cytochrome b in the MRS2 disrupted cells compared with wild-type cells (Fig. 7); a cytochrome spectrum of the mrs2:: strain GW7/ gd2-21.2 containing all mitochondrial introns is identical to that of strain DBY747/M1301 (data not shown). However, it is not the absence of the mitochondrially made subunits of the cytochrome oxidase that cause this cytochrome a.a3 deficiency. As shown in Fig. 8, patterns of mitochondrial translation products are similar in MRS2 and mrs2::HISB strains lacking all introns whereas mrs2::HIS3 cells containing all introns lack apocytochrome b and subunit I of cytochrome oxidase, as expected from their lack of the respective mature mRNAs. The only effect which we can see is a reduced amount in the bands representing COX3 and subunit 6 of the ATPase. However, this effect is also seen in MRS2 cells with the cobmutation M1301 and therefore may be a secondary effect of their respiratory deficiency.

DISCUSSION
The MRS2 gene was found during a search for nuclear genes that suppress a mitochondrial mutation in a group I1 intron when this gene was present in high copy number (19). This suppressor activity suggested that the MRSB gene product might have a function in the excision of group I1 (and possibly also group I) introns from mitochondrial pre-mRNAs. Here we show that disruption of this gene indeed blocks the excision of all four known group I1 introns from the COB and COX1 transcripts and has little effect on the excision of some group I introns. The MRS2 gene product may have a weak effect on mitochondrial RNA splicing in general, but it is an essential factor only for the excision of group I1 introns.
In this respect the MRS2 gene is unique among the many known genes that affect mitochondrial RNA splicing. One of these genes, CBP2, is necessary for excision of only one single intron (bI5) (45), whereas another one, MRS1, is an indispensible factor for two group I introns (bI3 and aI5b) (13,14). Yeasts lacking the respective introns are unaffected by the disruption of these nuclear genes. On the other hand, most of the nuclear genes involved in mitochondrial RNA splicing described so far have other functions besides their involvement in the splicing process. CYT18 in N. crmsa and NAM2 in yeast code for proteins with dual function; they serve as tRNA synthetases and in addition are essential factors for group I intron excision (16,17); PET54 affects translation of the mitochondrially encoded COX3 gene and also splicing of aI5b (46); NAMl is a mitochondrial transcription factor and also seems to be involved in mitochondrial splicing (47,48). The MSS116 gene, which turned out to encode a RNA helicase type protein, and the MSS18 gene both have a second, still unknown function besides their action on mitochondrial RNA splicing (18,49).
Apparently the involvement in mitochondrial RNA splicing is also not the only function of the MRSZ gene product. This notion rests on the fact that disruption of the MRS2 gene causes a petphenotype, even when combined with a mitochondrial genome lacking the four known group I1 introns or all known introns (cf. Fig. 6). Furthermore, the intron-less mitochondria appear to synthesize the standard set of major mitochondrial proteins and yet lack the cytochrome a.a3 spectral bands (cf. Figs. 7 and 8). This parallels a study on the MSS116 gene (18), which is involved in both group I and group I1 intron excision; its disruption also resulted in a petphenotype, irrespective of the presence or absence of the known introns in mtDNA. Both findings could easily be explained by the assumption that not all introns in yeast mtDNA have been identified, but this seems unlikely since this genome has been sequenced almost completely. Alternatively the products of both genes MSS116 and MRS2 might be involved in other processing or modification events of mitochondrial transcripts.
Both of these explanations attribute to the MRSZ gene product the primary (and only) function in RNA processing. However, this does not easily explain the fact that cytochrome a.a3 is absent and cytochrome b is reduced in a MRS2 disrupted strain lacking all introns although the major, mitochondrially encoded proteins are synthesized. These data rather hint at the involvement of the MRS2 gene product in the formation of functional cytochrome complexes. It may fulfill a basic function in mitochondrial biogenesis, e.g. in the assembly of cytochromes as it has also been suggested for genes COX10 and COX11 (50,51) or it may be involved in the expression or the import of some nuclear encoded subunits of the cytochrome complexes.
We conclude from our observation that the MRS2 gene product might serve other functions besides RNA splicing and thus behave like several of the other nuclear genes involved in mitochondrial RNA splicing (12).
Yet, from our results we cannot exclude that MRSZ affects respiration and splicing fairly indirect. We are currently investigating spontaneous suppressor mutations of MRSZ that restore growth on nonfermentable carbon sources. Preliminary results reveal the existence of a class of dominant nuclear suppressors that restore growth on glycerol of mrs2:: strains lacking all introns in mtDNA, but not of mrs2:: strains harboring the full set of mitochondrial introns (long version).' Thus, the suppressor mutants are independent of the MRSZ gene product as far as basic functions in mitochondrial biogenesis are concerned. However group I1 intron splicing remains dependent on this product. We conclude therefore that the gene product of MRS2 is indeed involved in two different processes of mitochondrial biogenesis, directly or indirectly, and that the splicing deficiency is not simply the consequence of some yet unknown process also leading to respiratory deficiency.