Yeast Nuclear Gene Research Articles

Gene MRP-L4, encoding mitochondrial ribosomal protein YmL4, is indispensable for proper non-respiratory cell functions in yeast

In order to characterize individual protein components of the mitochondrial (mt) ribosome for regulatory, functional and evolutionary studies, the yeast nuclear gene MRP-L4 (accession No. Z30582), coding for the mt ribosomal protein ( MRP) YmL4, has been cloned using oligodeoxyribonucleotides (oligos) deduced from a partial amino acid (aa) sequence [Graack et al., FEBS Lett. 242 (1988) 4–8] as screening probes. MRP-L4 is located on chromosome XII and codes for a slightly basic protein of 319 aa. The first 14 aa have not been found in the mature protein, and putatively form a signal peptide that is cleaved off during or after mt import. YmL4 has an N terminus very rich in Pro residues, and at its C terminus contains four hydrophobic domains. YmL4 shows no significant sequence similarity to any other sequence from the databases. Gene disruption shows the MRP-L4 product to be indispensable for mt function in cells growing on non-fermentable carbon sources. In contrast to nearly all other MRPs investigated so far, gene disruption of MRP-L4 also affects growth of yeast cells on fermentable carbon sources, suggesting additional cytosolic and/or mt functions of YmL4 besides its involvement in mt protein biosynthesis.

YKE2, a yeast nuclear gene encoding a protein showing homology to mouse KE2 and containing a putative leucine-zipper motif

The nucleotide sequence of YKE2, a yeast nuclear gene, has been determined. The deduced YKE2 protein has 114 amino acids (12kDa), shows significant homology with the murine KE2 and contains a putative Leu-zipper motif characteristic of a group of DNA-binding proteins [Landschulz et al., Science 240 (1988) 1759–1764]. Strains in which YKE2 has been disrupted show normal cell growth in glucose and galactose media over the temperature range of 16 to 40°C. Disrupted strains also display normal mating and sporulation abilities. Northern analysis revealed that the transcription of YKE2 is unresponsive to catabolite repression by glucose.

Cloning of the yeast ATP3 gene coding for the gamma-subunit of F1 and characterization of atp3 mutants

Saccharomyces cerevisiae pet mutants, of complementation group G115, are deficient in mitochondrial ATPase and have properties indicative of defective F1. In this study we show that C287/LU1, a mutant belonging to group G115, is complemented by the yeast nuclear ATP3 gene coding for the gamma-subunit of the mitochondrial F1-ATPase. The amino-terminal sequence of the mature gamma-subunit matches the sequence encoded by ATP3 starting with the 34th amino acid confirming the identity of the gene, and earlier evidence indicating that this F1 component is synthesized as a precursor with a long amino-terminal extension. The properties of the mitochondrial ATPase have been studied in C287/LU1 with an Ala273-->Val substitution in the carboxyl-terminal region of the gamma-subunit and in W303 delta ATP3, a mutant lacking the gamma-subunit as a result of a deletion in ATP3. Both strains have negligible ATPase activity but near normal concentrations of the alpha- and beta-subunits of F1. In W303 delta ATP3, the subunits do not form a stable F1 oligomer nor are they firmly associated with F0. This is not true of C287/LU1, which was found to assemble an F1-F0 complex. These data indicate that the yeast gamma-subunit has dual functions, one in catalysis of ATP hydrolysis/synthesis and the second in assembly/stability of F1.

The yeast nuclear gene MRP-L13 codes for a protein of the large subunit of the mitochondrial ribosome.

The nuclear gene MRP-L13 of Saccharomyces cerevisiae, which codes for the mitochondrial ribosomal protein YmL13, has been cloned and characterized. It is a single-copy gene residing on chromosome XI. Its nucleotide sequence was found to be identical to that of the previously reported ORF YK105. A comparison of the predicted protein sequence of the MRP-L13 gene product and the actual N-terminal amino-acid sequence of the isolated YmL13 protein indicated that the mature protein is preceded by a mitochondrial signal peptide of 86 amino-acid residues, which is the longest among all known mitochondrial ribosomal proteins of S. cerevisiae. No sequence similarity was found to any other ribosomal protein in the current databases. The transcription of MRP-L13 was found to be repressed in the presence of glucose. Its protein product is not strictly essential for mitochondrial functions, but disruption of the gene by insertion of LEU2 noticeably affected cellular growth on non-fermentable carbon sources.

The HAP2,3,4 transcriptional activator is required for derepression of the yeast citrate synthase gene, CIT1.

The yeast nuclear gene CIT1 encodes mitochondrial citrate synthase, which catalyses the first and rate-limiting step of the tricarboxylic acid (TCA) cycle. Transcription of CIT1 is subject to glucose repression. Mutations in HAP2, HAP3 or HAP4 block derepression of a CIT1-lacZ gene fusion. The HAP2,3,4 transcriptional activator also activates nuclear genes encoding components of the mitochondrial electron transport chain, and thus it co-ordinates derepression of two major mitochondrial functions. Two DNA sequences resembling the consensus HAP2,3,4-binding site (ACCAATNA) are located at approximately -310 and -290, upstream of the CIT1 coding sequence. Deletion and mutation analysis indicates that the -290 element is critical for activation by HAP2,3,4. Glucose-repressed expression of CIT1 is largely independent of HAP2,3,4, is repressed by glutamate, and requires a DNA sequence between -367 and -348. Evidence is presented for a second HAP2,3,4-independent activation element located just upstream and overlapping the -290 HAP2,3,4 element.

Interaction between the first and last nucleotides of pre-mRNA introns is a determinant of 3' splice site selection in S. cerevisiae.

The splicing of group II and nuclear pre-mRNAs introns occurs via a similar splicing pathway and some of the RNA-RNA interactions involved in these splicing reactions show structural similarities. Recently, genetic analyses performed in a group II intron and the yeast nuclear actin gene suggested that non Watson-Crick interactions between intron boundaries are important for the second splicing step efficiency in both classes of introns. We here show that, in the yeast nuclear rp51A intron, a G to A mutation at the first position activates cryptic 3' splice sites with the sequences UAC/ or UAA/. Moreover, the natural 3' splice site could be reactivated by a G to C substitution of the last intron nucleotide. These results demonstrate that the interaction between the first and last intron nucleotides is a conserved feature of nuclear pre-mRNA splicing in yeast and is involved in the mechanism of 3' splice site selection.

Inactivation of YME1, a member of the ftsH-SEC18-PAS1-CDC48 family of putative ATPase-encoding genes, causes increased escape of DNA from mitochondria in Saccharomyces cerevisiae.

The yeast nuclear gene YME1 was one of six genes recently identified in a screen for mutations that elevate the rate at which DNA escapes from mitochondria and migrates to the nucleus. yme1 mutations, including a deletion, cause four known recessive phenotypes: an elevation in the rate at which copies of TRP1 and ARS1, integrated into the mitochondrial genome, escape to the nucleus; a heat-sensitive respiratory-growth defect; a cold-sensitive growth defect on rich glucose medium; and synthetic lethality in rho- (cytoplasmic petite) cells. The cloned YME1 gene complements all of these phenotypes. The gene product, Yme1p, is immunologically detectable as an 82-kDa protein present in mitochondria. Yme1p is a member of a family of homologous putative ATPases, including Sec18p, Pas1p, Cdc48p, TBP-1, and the FtsH protein. Yme1p is most similar to the Escherichia coli FtsH protein, an essential protein involved in septum formation during cell division. This observation suggests the hypothesis that Yme1p may play a role in mitochondrial fusion and/or division.

PET111 acts in the 5'-leader of the Saccharomyces cerevisiae mitochondrial COX2 mRNA to promote its translation.

PET111 is a yeast nuclear gene specifically required for the expression of the mitochondrial gene COX2, encoding cytochrome c oxidase subunit II (coxII). Previous studies have shown that PET111 activates translation of the COX2 mRNA. To map the site of PET111 action we have constructed, in vitro, genes coding for chimeric mRNAs, introduced them into mitochondria by transformation and studied their expression. Translation of a chimeric mRNA with the 612-base 5'-untranslated leader of the COX3 mRNA fused precisely to the structural gene for the coxII-precursor protein is independent of PET111, but does require a COX3 mRNA-specific translational activator known to work on the COX3 5'-leader. This result demonstrates that PET111 is not required for any post-translational step. Translation of a chimeric mRNA with the 54-base 5'-leader of the COX2 mRNA fused precisely to the structural gene for cytochrome c oxidase subunit III was dependent on PET111 activity. These results demonstrate that PET111 acts specifically at a site in the short COX2 5'-leader to activate translation of downstream coding sequences.

NAM7 nuclear gene encodes a novel member of a family of helicases with a Zn-ligand motif and is involved in mitochondrial functions in Saccharomyces cerevisiae

The yeast nuclear gene NAM7 was previously isolated within a genomic fragment of 7.7 kb (1 kb = 10 3 bases or base-pairs), having the ability to suppress mitochondrial intronic mutations defective in RNA splicing. We have identified and sequenced the region on the insert corresponding to the NAM7 gene. A long open reading frame has been revealed which could code for a protein of 971 amino acids. Comparison of the NAM7 putative protein with data libraries did not reveal any strong similarity with known proteins. However, the NAM7 protein contains several motifs typical for proteins interacting with nucleic acids: (1) five motifs diagnostic for a superfamily of helicases appear in the same order and with similar distances; (2) the N-terminal portion possesses potential Zn-ligand structures belonging to the C x superfamily. Deletion of the chromosomal copy of NAM7 gene leads to a partial impairment in respiratory growth that is particularly striking at low temperature. Southern blot analysis of DNA extracted from a nam7: URA3 deleted strain revealed the presence of a second gene whose sequence is related to that of the NAM7 gene and which could participate in the same process.

The yeast nuclear gene MRF1 encodes a mitochondrial peptide chain release factor and cures several mitochondrial RNA splicing defects.

We report the molecular cloning, sequencing and genetic characterization of the first gene encoding an organellar polypeptide chain release factor, the MRF1 gene of the yeast Saccharomyces cerevisiae. The MRF1 gene was cloned by genetic complementation of a respiratory deficient mutant disturbed in the expression of the mitochondrial genes encoding cytochrome c oxidase subunit 1 and 2, COX1 and COX2. For COX1 this defect has been attributed to an impaired processing of several introns. Sequence analysis of the MRF1 gene revealed that it encodes a protein highly similar to prokaryotic peptide chain release factors, especially RF-1. Disruption of the gene results in a high instability of the mitochondrial genome, a hallmark for a strict lesion in mitochondrial protein synthesis. The respiratory negative phenotype of mrf1 mutants lacking all known mitochondrial introns and the reduced synthesis of mitochondrial translation products encoded by unsplit genes confirm a primary defect in mitochondrial protein synthesis. Over-expression of the MRF1 gene in a mitochondrial nonsense suppressor strain reduces suppression in a dosage-dependent manner, shedding new light on the role of the '530 region' of 16S-like ribosomal RNA in translational fidelity.

Characterization of Satellite DNA from Three Marine Dinoflagellates (Dinophyceae): Glenodinium sp. and Two Members of the Toxic Genus, Protogonyaulax

Using CsCl-Hoechst dye or CsCl-ethidium bromide gradients, satellite and nuclear DNAs were separated and characterized in three marine dinoflagellates: Glenodinium sp., and two toxic dinoflagellates, Protogonyaulax tamarensis and Protogonyaulax catenella. In all three dinoflagellates, the lowest density fraction, satellite DNA(1), hybridized to chloroplast genes derived from terrestrial plants and/or other algae. Dinoflagellate chloroplast DNAs exhibited molecular sizes of 114 to 125 kilobase pairs, which is consistent with plastid sizes determined for other chromophytic algae (120-150 kilobase pairs). Mitochondrial DNA was not resolved from nuclear DNA in this system. Two additional satellite DNAs, satellite DNA(2) and satellite DNA(3), recovered from P. tamarensis and P. catenella were similar to one another, both within and between species, when characterized by restriction enzyme analysis. These satellites were 85 to 95 kilobase pairs in size, and exhibited restriction fragments that hybridized to yeast nuclear ribosomal RNA genes. Restriction enzyme analyses and DNA hybridization studies of cpDNA document that the two Protogonyaulax isolates are not evolutionarily identical.

ABC1, a novel yeast nuclear gene has a dual function in mitochondria: it suppresses a cytochrome b mRNA translation defect and is essential for the electron transfer in the bc 1 complex.

We have cloned and sequenced a novel yeast nuclear gene ABC1 which suppresses, in multicopy, the cytochrome b mRNA translation defect due to the nuclear mutation cbs2-223. Analysis of the ABC1 gene shows that it is weakly expressed, it could code for a protein of 501 amino acids which has a typical presequence of a protein imported into mitochondria and which does not display a strong similarity to any known protein. Inactivation of the ABC1 gene is not lethal to the cell but leads to a respiratory defect: no oxygen uptake and no growth on non-fermentable media. A total absence of NADH-cytochrome c oxidoreductase and succinate-cytochrome c oxidoreductase activities concomitant with the presence of specific dehydrogenases, suggests a block in the bc 1 segment of the respiratory chain. However, all the cytochromes are spectrally detectable. Cytochrome b is quite efficiently reduced while cytochromes c1 and c are not. The function of ABC1 in the suppression of a defect in apocytochrome b mRNA translation and in the activity of the bc1 complex suggests that the ABC1 protein would be a novel chaperonin involved both in biogenesis and bioenergetics of mitochondria.

Characterization of a yeast nuclear gene, AEP2, required for accumulation of mitochondrial mRNA encoding subunit 9 of the ATP synthase

The temperature-conditional pet mutant, ts379, of Saccharomyces cerevisiae fails to synthesize mitochondrial ATP synthase subunit 9 at the restrictive temperature due to mutation of a single nuclear locus, AEP2. The inability to synthesize subunit 9 correlates with a lowered accumulation of the cognate oli1 mRNA indicating that the AEP2 product is involved in oli1 transcript maturation or stabilization. The AEP2 gene has been isolated in this study from a wild-type yeast genomic library by genetic complementation of ts379 at the restrictive temperature. A 1,740 nucleotide open-reading frame was observed that encodes a basic, hydrophilic protein of 67,534 Da which possesses a putative mitochondrial address signal. Disruption of chromosomal DNA within this reading frame produced a non-conditional respiratory mutant unable to synthesize subunit 9, identifying the AEP2 gene. Hybridization analyses indicate that AEP2 is located on chromosome XIII and produces a 2.1 kb poly(A)+ transcript. Two additional open-reading frames were found in close proximity to that of AEP2. The three open-reading frames shared no significant homology with entries in several data bases.

Extended N-terminal sequencing of proteins of the large ribosomal subunit from yeast mitochondria

We have determined the N-termini of 26 proteins of the large ribosomal subunit from yeast mitochondria by direct amino acid micro-sequencing. The N-terminal sequences of proteins YmL33 and YmL38 showed a significant similarity to eubacterial ribosomal (r-) proteins L30 and L14, respectively. In addition, several proteins could be assigned to their corresponding yeast nuclear genes. Based on a comparison of the protein sequences deduced from the corresponding DNA regions with the N-termini of the mature proteins, the putative leader peptides responsible for mitochondrial matrix-targeting were compiled. In most leader sequences a relative abundance of aromatic amino acids, preferentially phenylalanine, was found.

Structure and evolution of a group of related aminoacyl-tRNA synthetases

A yeast nuclear gene, designated MSK1, has been selected from a yeast genomic library by transformation of a respiratory deficient mutant impaire lysine tRNA. This gene confers a respiratory competent phenotype and restores the mutant's ability to acylate the mitochondrial lysine tRNA. The amino acid sequence of the protein encoded by MSK1 is homologous to yeast cytoplasmic lysyl-tRNA synthetase and to the product of the herC gene, which has recently been suggested to code for the Escherichia coli enzyme. These observations indicate that MSK1 codes for the lysyl-tRNA synthetase of yeast mitochondria. Several regions of high primary sequence conservation have been identified in the bacterial and yeast lysyl-tRNA synthetases. These domains are also present in the aspartyl-and asparaginyl-tRNA synthetases, further confirming the notion that all three present-day enzymes originated from a common ancestral gene. The most conserved domain, located near the carboxyl terminal ends of this group of synthetases is characterized by a cluster of glycines and is also highly homologous to the carboxyl-terminal region of the E. coli ammonia-dependent asparagine synthetase. A catalytic function of the carboxyl terminal domain is indicated by in vitro mutagenesis of the yeast mitochondrial lysyl-tRNA synthetase. Replacement of any one of three glycine residues by alanine and in one case by aspartic acid completely suppresses the activity of the enzymes, as evidenced by the inability of the mutant genes to complement an mski mutant, even when present in high copy. Other mutations result in partial loss of activity. Only one glycine replacement affects the stability of the protein in vivo. The observed presence of a homologous domain in asparagine synthetase, which, like the aminoacyl-tRNA synthetases, catalyzes the formation of an aminoacyladenylate, suggests that the glycine-rich sequence is part of a catalytic site involved in binding of ATP and of the aminoacyladenylate intermediate.

Characterization of ATP12, a yeast nuclear gene required for the assembly of the mitochondrial F1-ATPase.

Mitochondrial F1-ATPase is an oligomeric enzyme composed of five distinct subunit polypeptides. The alpha and beta subunits make up the bulk of protein mass of F1. In Saccharomyces cerevisiae both subunits are synthesized as precursors with amino-terminal targeting signals that are removed upon translocation of the proteins to the matrix compartment. Recently, two different complementation groups (G13, G57), consisting of yeast nuclear mutants with defective F1, have been described. Biochemical analyses indicate that the mutational block in both groups of mutants affects a critical step needed for the assembly of the alpha and beta subunits into the F1 oligomer after their transport into mitochondria. In this study the ATP12 gene representative of the nuclear respiratory-deficient mutant of S. cerevisiae (pet) complementation group G57 has been cloned and the encoded product partially characterized. The ATP12 reading frame is 975 base pairs long and codes for a protein of Mr = 36,587. The ATP12 protein is not homologous to the subunits of F1 whose sequences are known, nor does it exhibit significant primary structure similarity to any known protein. In vitro import assays indicate that ATP12 protein is synthesized as a precursor approximately 3 kDa larger than the mature protein. The mitochondrial localization of the protein has been confirmed by Western blot analysis of mitochondrial proteins with an antibody against a hybrid protein expressed from a trpE-ATP12 fusion. Fractionation of mitochondria indicates further that the ATP12 protein is either a minor component of the matrix compartment or is weakly bound to the matrix side of the inner membrane. The molecular weight of the native protein, estimated from its sedimentation properties in sucrose gradients, is at least two times larger than the monomer. This suggests that the ATP12 protein is probably part of a larger complex.

ATP10, a yeast nuclear gene required for the assembly of the mitochondrial F1-F0 complex.

A yeast nuclear gene (ATP10) is reported whose product is essential for the assembly of a functional mitochondrial ATPase complex. Mutations in ATP10 induce a loss of rutamycin sensitivity in the mitochondrial ATPase but do not affect respiratory enzymes. This phenotype has been correlated with a defect in the F0 sector of the ATPase. The wild type ATP10 gene has been cloned by transformation of an atp 10 mutant with a yeast genomic library. The gene codes for a protein of Mr = 30,293. The primary structure of the ATP10 product is not related to any known subunit of the yeast or mammalian mitochondrial ATPase complexes. To further clarify the role of this new protein in the assembly of the ATPase, an antibody was prepared against a hybrid protein expressed from a trpE/ATP 10 fusion gene. The antibody recognizes a 30-kDa protein present in wild type mitochondria. The protein is associated with the mitochondrial membrane but does not co-fractionate either with F1 or with the rutamycin-sensitive F1-F0 complex. These data suggest that the ATP10 product is not a subunit of the ATPase complex but rather is required for the assembly of the F0 sector of the complex.

Common protein-binding sites in the 5'-flanking regions of human genes for cytochrome c1 and ubiquinone-binding protein.

The single nuclear gene encoding human ubiquinone-binding protein, a subunit of the cytochrome bc1 complex, was isolated previously (Suzuki, H., Hosokawa, Y., Toda, H., Nishikimi, M., and Ozawa, T. (1989) Biochem. Biophys. Res. Commun. 161, 371-378). The 5'-flanking region contains four putative CCAAT boxes, one putative NF-Y-binding site, and one sequence homologous to the AP-1 recognition site, but lacks typical TATA and GC boxes. Three common nucleotide sequences (5'-TATTCAGGT-3', 5'-ATCTGGCT-3', and 5'-TGGTGA(T/G)AG-3', designated Mt1, Mt3, and Mt4, respectively) were found in the 5'-flanking regions of the genes for human ubiquinone-binding protein and for human cytochrome c1, another subunit of the cytochrome bc1 complex. All three sequences are located in the 280-base pair BamHI-HindIII fragment of the ubiquinone-binding protein gene and in the 154-base pair SalI-PstI fragment of the cytochrome c1 gene. Interestingly, a computer-assisted homology search revealed that the SalI-PstI fragment of the cytochrome c1 gene is related to the long terminal repeat of an endogenous retrovirus. Sequences highly homologous to the three Mt-sequences were also found in the 5'-flanking regions of the genes for the beta subunit of human F0F1-ATPase and rat somatic cytochrome c. The Mt1 sequence (TATT-CAGGT) is similar to the GFII recognition site (RTCACGTG) found in the 5'-flanking regions of the yeast nuclear genes for three cytochrome bc1 complex subunits. Gel retardation assay demonstrated that protein factors in a whole HeLa cell extract bound to both of those fragments. At least three different specific binding sites were thought to exist in the fragments. Specific binding of the protein factors to the sites, possibly to the three Mt sequences, may play an important role in the coordinate regulation of the transcription of nuclear genes encoding subunits responsible for mitochondrial oxidative phosphorylation.

Disruption of the yeast nuclear PET54 gene blocks excision of mitochondrial intron aI5 beta from pre-mRNA for cytochrome c oxidase subunit I.

The nuclear PET54 gene of Saccharomyces cerevisiae was cloned and a pet54::LEU2 gene disruption strain was constructed. Analysis of the phenotype of this strain revealed a defect in expression of two mitochondrial genes: COX1, which encodes cytochrome c oxidase subunit I, and COX3, which encodes cytochrome c oxidase subunit III. The defect in COX1 gene expression in the pet54 mutant was shown to be the result of inefficient excision of COX1 intron aI5 beta. Two lines of evidence indicate that inefficient excision of intron aI5 beta is the sole defect in COX1 gene expression. First, a pet54::LEU2 cytoductant bearing the 'short' mitochondrial genome that lacks both COX1 introns aI5 alpha and aI5 beta is defective only in COX3 gene expression and not in COX1 mRNA splicing or mRNA translation. Second, Northern analysis of COX1 transcipts from the pet54 mutant showed that a 3.8 kb COX1 transcript containing unexcised intron aI5 beta and lacking intron aI5 alpha is accumulated while the amount of 2.2 kb mature COX1 mRNA is diminished. In an effort to relate the role of the PET54 gene product in splicing of COX1 pre-mRNA to the previously characterized role for PET54 in translation of mitochondrial COX3 mRNA, the sequence of the PET54-responsive portion of the COX3 5' untranslated leader region was compared to the COX1 intron aI5 beta sequence. Two blocks of RNA sequence present in COX3 have similar counterparts within intron aI5 beta of COX1. The possibility that the PET54 protein binds to one or the other of these blocks of RNA sequence and the potential consequences of this interaction are discussed.

The Yeast CBP1 Gene Produces Two Differentially Regulated Transcripts by Alternative 3′-End Formation

CBP1 is a yeast nuclear gene encoding a mitochondrial protein that stabilizes the 5' end of cytochrome b (cob) pre-mRNA. Cytochrome b is the only mitochondrially synthesized component of the respiratory chain complex III. Since the nuclearly encoded subunits of this complex are regulated at the transcriptional level by catabolite repression, we hypothesized that CBP1 might be similarly regulated. To test the idea that transcriptional regulation of CBP1 could coordinate an increase in cytochrome b mRNA stability with an increase in nuclearly encoded complex III subunit production, we characterized the change in abundance of CBP1 mRNA during derepression on a nonfermentable carbon source. Poly(A)+ RNA from derepressed yeast cells was examined by Northern (RNA) analyses with cRNA probes from CBP1. Both 2.2- and 1.3-kilobase (kb) transcripts were detected. The 1.3-kb mRNA lacked approximately 900 nucleotides of the 3' end of the 2.2-kb mRNA, which encodes the carboxyl-terminal 250 amino acid residues of the CBP1 coding sequence. Northern analyses of RNA isolated from deletion-insertion mutants of CBP1 and from strains that overexpress CBP1 mRNA demonstrated that both mRNAs were transcribed from the CBP1 gene. Furthermore, we demonstrated that the levels of the two CBP1 mRNAs were reciprocally regulated by the carbon source in the growth medium. This is the first description of a yeast gene from which two transcripts that can encode proteins with distinctly different coding properties are generated by alternative 3'-end formation.