Mutational Analyses of Yeast RNA Triphosphatases Highlight a Common Mechanism of Metal-dependent NTP Hydrolysis and a Means of Targeting Enzymes to Pre-mRNAs in Vivo by Fusion to the Guanylyltransferase Component of the Capping Apparatus*

Saccharomyces cerevisiae Cet1p is the prototype of a family of metal-dependent RNA 5′-triphosphatases/NTPases encoded by fungi and DNA viruses; the family is defined by conserved sequence motifs A, B, and C. We tested the effects of 12 alanine substitutions and 16 conservative modifications at 18 positions of the motifs. Eight residues were identified as important for triphosphatase activity. These were Glu-305, Glu-307, and Phe-310 in motif A (IELEMKF); Arg-454 and Lys-456 in motif B (RTK); Glu-492, Glu-494, and Glu-496 in motif C (EVELE). Four acidic residues, Glu-305, Glu-307, Glu-494, and Glu-496, may comprise the metal-binding site(s), insofar as their replacement by glutamine inactivated Cet1p. E492Q retained triphosphatase activity. Basic residues Arg-454 and Lys-456 in motif B are implicated in binding to the 5′-triphosphate. Changing Arg-454 to alanine or glutamine resulted in a 30-fold increase in the K m for ATP, whereas substitution with lysine increased K m 6-fold. Changing Lys-456 to alanine or glutamine increasedK m an order of magnitude; ATP binding was restored when arginine was introduced. Alanine in lieu of Phe-310 inactivated Cet1p, whereas Tyr or Leu restored function. Alanine mutations at aliphatic residues Leu-306, Val-493, and Leu-495 resulted in thermal instability in vivo and in vitro. A secondS. cerevisiae RNA triphosphatase/NTPase (named Cth1p) containing motifs A, B, and C was identified and characterized. Cth1p activity was abolished by E87A and E89A mutations in motif A. Cth1p is nonessential for yeast growth and, by itself, cannot fulfill the essential role played by Cet1p in vivo. Yet, fusion of Cth1p in cis to the guanylyltransferase domain of mammalian capping enzyme allowed Cth1p to complement growth ofcet1Δ yeast cells. This finding illustrates that mammalian guanylyltransferase can be used as a vehicle to deliver enzymes to nascent pre-mRNAs in vivo, most likely through its binding to the phosphorylated CTD of RNA polymerase II.

The Saccharomyces cerevisiae RNA 5Ј-triphosphatase Cet1p is an essential enzyme that catalyzes the first step of mRNA cap formation, the hydrolysis of the ␥-phosphate of triphosphate-terminated pre-mRNA to form a diphosphate end that then serves as the substrate for capping by the yeast RNA guanylyltransferase Ceg1p (1,2). The yeast triphosphatase and guanylyltransferase interact in vivo and in vitro to form a bifunctional heteromeric capping enzyme complex (1)(2)(3)(4)(5)(6). The triphosphatase activity of Cet1p and its interaction with Ceg1p are both important for Cet1p function in vivo (6,7).
Cet1p is the prototype of a newly identified family of divalent cation-dependent nucleoside triphosphatases that include the RNA 5Ј-triphosphatases encoded by Candida albicans, poxviruses, and baculoviruses (7). The fungal and viral enzymes display a characteristic requirement for manganese or cobalt as the cofactor for their NTPase activities (7)(8)(9)(10)(11). The enzyme family is defined by the presence of three conserved colinear motifs (A, B, and C) that include clusters of acidic and basic amino acids essential for triphosphatase activity (7,(11)(12)(13). Motifs A, B, and C of yeast Cet1p are located within the Cterminal half of the 549-amino acid protein (Fig. 1) and are presumed to comprise the triphosphatase-active site. Six essential amino acids within these motifs, Glu-305, Glu-307, Arg-454, Glu-492, Glu-494, and Glu-496, were identified by alaninescanning mutagenesis (7).
Cet1p consists of three domains as follows: (i) a 230-amino acid N-terminal segment that is dispensable for catalysis in vitro and for Cet1p function in vivo; (ii) a protease-sensitive segment from residues 230 to 275 that is dispensable for catalysis but essential for Cet1p function in vivo; and (iii) a catalytic domain from residues 275 to 539 (6). The catalytic domain includes motifs A, B, and C. The segment of Ceg1p from residues 230 to 275 regulates Cet1p self-association (6) and is implicated in the binding of Cet1p to Ceg1p (5,6).
Here we address three questions concerning Cet1p. (i) Does the interaction of Cet1p with Ceg1p require a functional triphosphatase active site in Cet1p? By using zonal velocity sedimentation as an assay for Cet1p-Ceg1p complex formation, we found that the interaction of these proteins is unaffected by mutations in motifs A, B, or C that abrogate RNA triphosphatase activity. (ii) What structural features of the amino acid side chain are functionally relevant at essential residues Glu-305, Glu-307, Arg-454, Glu-492, Glu-494, and Glu-496? Studies of the effects of conservative side chain substitutions on Cet1p function revealed underlying structure-activity relationships that have implications for the catalytic mechanism. (iii) Are other amino acids within motifs A, B, or C essential for Cet1p function? Expansion of the alanine scan defined two other residues required for Cet1p function (Phe-310 and Lys-456), as well as three residues (Leu-306, Val-493, and Leu-495) at which side chain removal results in thermal instability of Cet1p in vivo and in vitro.
In addition, we ask whether the presence of motifs A, B, and C in a gene product of unknown function has predictive value with respect to the biochemical properties of that gene product. We noted previously (2) that the amino acid sequence of the 320-amino acid polypeptide encoded by the S. cerevisiae YMR180C open reading frame displays local similarity to the sequence of Cet1p. The region of sequence similarity spans Cet1p residues 302-532 (the C-terminal catalytic domain) and includes motifs A, B, and C (Fig. 1). The function of YMR180C is unknown. Here we show that the protein encoded by this yeast gene (renamed CTH1, cap triphosphatase homolog) possesses magnesium-dependent RNA triphosphatase and manganese-or cobalt-dependent NTPase activities in vitro. Cth1p is nonessential for yeast growth and, by itself, cannot replace Cet1p in vivo. Yet, we find that fusion of Cth1p to the guanylyltransferase domain of mammalian capping enzyme allows Cth1p to complement growth of yeast cells deleted for CET1. These findings bear on the question of how the capping apparatus is specifically targeted to pre-mRNAs in vivo.

EXPERIMENTAL PROCEDURES
Mutational Analysis of Cet1p-Amino acid substitution mutations were introduced into the CET1(201-549) gene by PCR 1 as described previously (7). The mutated genes were inserted into the yeast CEN TRP1 plasmid pCET1-5Ј3Ј, where expression of the inserted gene is under the control of the natural CET1 promoter.
Expression of Recombinant Cet1(201-549)p-NdeI/BamHI fragments encoding mutated versions of Cet1(201-549)p were excised from the respective pCET1-5Ј3Ј plasmids and inserted into pET16b (Novagen). The pET-CET1(201-549) plasmids were transformed into Escherichia coli BL21(DE3). Single transformants were inoculated into 100 ml of LB medium containing 0.1 mg/ml ampicillin and grown at 37°C until the A 600 reached approximately 0.5. Recombinant protein expression was induced by adding isopropyl-1-thio-␤-D-galactopyranoside to 0.4 mM final concentration and continuing incubation for 3 h at 37°C. Alternatively, expression was performed by placing the culture on ice for 30 min, followed by addition of isopropyl-1-thio-␤-D-galactopyranoside to 0.4 mM and ethanol to 2% final concentration (v/v), and then continuing incubation at 18°C for 24 h. The cells were harvested by centrifugation, and all subsequent procedures were performed at 4°C. The His-tagged Cet1(201-549) proteins were purified from soluble bacterial lysates by Ni 2ϩ -agarose chromatography as described (2). The 0.2 M imidazole eluate fractions containing Cet1(201-549)p were dialyzed against buffer containing 50 mM Tris-HCl (pH 8.0), 50 mM NaCl, 2 mM DTT, 10% glycerol, and 0.05% Triton X-100. Protein concentration was determined by using the Bio-Rad dye reagent with bovine serum albumin as the standard.
Glycerol Gradient Sedimentation-Aliquots (50 g of protein) of the Ni 2ϩ -agarose preparations of wild-type Cet1(201-549)p and E305A, R454A, and E494A mutant proteins in 0.2 ml of buffer containing 50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 10% glycerol were layered onto 4.8-ml 15-30% glycerol gradients containing 50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 2 mM DTT, and 0.1% Triton X-100. Where indicated, the wild-type Cet1(201-549)p, E305A, R454A, or E494A mutant proteins were mixed with 50 g of recombinant Ceg1p (6) in 0.2 ml of 50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 10% glycerol, and the mixtures were incubated on ice for 20 min before being layered onto glycerol gradients. The gradients were centrifuged in a Beckman SW50 rotor at 50,000 rpm for 20 h at 4°C. A mixture of marker proteins, catalase, BSA, and cytochrome c, was sedimented in a separate glycerol gradient. Fractions (ϳ0.2 ml) were collected from the bottoms of the tubes. Aliquots (20 l) of even-numbered fractions were analyzed by SDS-PAGE along with samples of the input protein mixtures for each gradient. Polypeptides were visualized by staining with Coomassie Blue dye.
RNA Triphosphatase Assay-Standard reaction mixtures (10 l) containing 50 mM Tris-HCl (pH 7.5), 5 mM DTT, 1 mM MgCl 2 , 20 pmol (of triphosphate termini) of ␥-32 P-labeled poly(A), and either wild-type or mutant proteins as specified were incubated for 15 min at 30°C. The reactions were quenched by adding 2 l of 5 M formic acid. Aliquots of the mixtures were applied to a polyethyleneimine-cellulose TLC plate, which was developed with 0.75 M potassium phosphate (pH 4.3). The release of 32 P i from ␥-32 P-labeled poly(A) was quantitated by scanning the TLC plate with a Fujix BAS2000 PhosphorImager.
ATPase Assay-Standard reaction mixtures (10 l) containing 50 mM Tris-HCl (pH 7.0), 5 mM DTT, divalent cation, and [␥-32 P]ATP as specified and enzyme were incubated for 15 min at 30°C. The reactions were quenched by adding 2 l of 5 M formic acid. Aliquots of the mixtures were applied to a polyethyleneimine-cellulose TLC plate, which was developed with 1 M formic acid. 0.5 M LiCl.
Effects of Motif B Mutations on the K m of Cet1(201-549)p for ATP-Reaction mixtures (10 l) containing 50 mM Tris-HCl (pH 7.0), 5 mM DTT, 2 mM MnCl 2 , varying concentrations of [␥-32 P]ATP, and wild-type or mutant enzymes were incubated for 15 min at 30°C. The extent of P i release at the lowest ATP concentrations tested was less than 10% of the input substrate. K m values were determined from double-reciprocal plots of the data. CTH1 Plasmids-A 2.8-kilobase pair genomic fragment extending from 1.4 kilobase pairs upstream of the start codon to 410 base pairs downstream of the stop codon of the CTH1 open reading frame was inserted into the SphI and SacI sites of pUC18 to yield pUC-CTH1. The CTH1 coding region was amplified by PCR using a sense primer that introduced an NdeI restriction site at the ATG start codon and an antisense primer that introduced a BglII site immediately following the stop codon. The PCR product was digested with NdeI and BglII and inserted into the NdeI and BamHI sites of pET16b to generate pET-CTH1. Alanine mutations E87A and E89A were introduced into the CTH1 gene by PCR using the two-stage overlap extension method. The second-stage PCR product was digested with NdeI and XhoI and inserted into NdeI/XhoI-digested pET-CTH1 to yield pET-CTH1-E87A and pET-CTH1-E89A. The wild-type, E87A, and E89A inserts in the pET plasmids were sequenced completely to exclude the acquisition of unwanted mutations during amplification and cloning. The CTH1 coding sequence and 410 base pairs of 3Ј-flanking DNA was cloned into yeast vector pYX232 (2 TRP1) to generate p232-CTH1. In this plasmid, expression of CTH1 is driving the constitutive TPI1 promoter. A gene disruption cassette conferring kanamycin resistance (14) was inserted into pUC-CTH1 in lieu of the CTH1 coding sequence from ϩ96 (aa 32) to ϩ834 (aa 278); the resulting plasmid was named pUC-cth1::KAN.
Expression and Purification of Cth1p-A 500-ml culture of E. coli BL21(DE3)/pET-CTH1 was grown at 37°C in LB medium containing 0.1 mg/ml ampicillin until the A 600 reached 0.5. The culture was adjusted to 0.4 mM isopropyl-1-thio-␤-D-galactopyranoside, and incubation was continued at 17°C for 24 h. Cells were harvested by centrifugation, and the pellet was stored at Ϫ80°C. All subsequent procedures were performed at 4°C. Thawed bacteria were resuspended in 25 ml of lysis , and baculovirus (Lef4) are aligned. Cet1p residues conserved in at least two other family members are shaded. The numbers of amino acids separating the motifs are indicated. Cet1p residues that were identified previously by alanine scanning as essential for Cet1p function (7) are denoted by arrowheads. These are as follows: Glu-305 and Glu-307 (motif A), Arg-454 (motif B), and Glu-492, Glu-494, and Glu-496 (motif C). Cet1p residues that were mutated in the present study are indicated by dots.
buffer (50 mM Tris-HCl (pH 7.5), 250 mM NaCl, 10% sucrose). Lysozyme was added to a final concentration of 100 g/ml; the suspension was incubated on ice for 10 min and then sonicated for 30 s. Triton X-100 was added to a final concentration of 0.1%, and sonication was repeated to reduce the viscosity of the lysate. Insoluble material was removed by centrifugation for 45 min at 17,000 rpm in a Sorvall SS34 rotor. The soluble extract was applied to a 2.5-ml column of Ni 2ϩ -NTA-agarose that had been equilibrated with lysis buffer containing 0.1% Triton X-100. The column was washed with the same buffer and then eluted stepwise with buffer B (50 mM Tris-HCl, pH 8.0, 250 mM NaCl, 10% glycerol, 0.05% Triton X-100) containing 50, 100, 200, 500, and 1000 mM imidazole. The polypeptide compositions of the column fractions were monitored by SDS-PAGE. Recombinant Cth1p was retained on the column and recovered predominantly in the 200 mM imidazole eluate. Cth1p mutants E87A and E89A were expressed and purified using the same procedures described for wild-type Cth1p.
Disruption of the CTH1 Gene in Yeast-Yeast strain YBS60 (MATa leu2 ade2 trp1 his3 ura3 can1 cth1::KAN) deleted at the chromosomal CTH1 locus was derived by targeted gene replacement in the diploid strain W303, followed by tetrad dissection and genotyping of haploid progeny. Diploid strain W303 was transformed with linearized pUC-cth1::KAN, and drug-resistant integrants were selected on YPD plates containing 200 g/ml G418 (14). Sporulation and tetrad dissection showed 2:2 segregation of G418 resistance. Correct insertion of the resistance marker into the CTH1 locus of YBS60 was confirmed by Southern blotting.
Chimeric Yeast Cth1-Mammalian Capping Enzyme-A chimeric gene encoding yeast Cth1p fused to the guanylyltransferase domain of the mouse capping enzyme (Mce1(211-597)p) was constructed as follows. The CTH1 coding sequence was PCR-amplified using pET-CTH1 as template and an antisense primer that changed the CTH1 stop codon to His and introduced an NdeI restriction site at the C terminus. The PCR product was digested with NdeI and then inserted into the NdeI site of pYX1-MCE1(211-597) (CEN TRP1) to yield the fusion gene CTH1-MCE1(211-597). An NheI-KpnI fragment containing the CTH1-MCE1(211-597) fusion gene was excised from pYX1-CTH1-MCE1(211-597) and inserted into the yeast multicopy expression plasmid pYX232 (2 TRP1) to generate p232-CTH1-MCE1(211-597). Expression of the chimeric gene in the pYX1 and p232 plasmids is under the control of the TPI1 promoter.  7). The six CET1(201-549)-Ala alleles were also lethal in vivo (7). It is proposed that motifs A, B, and C comprise the active site of Cet1p. Nonetheless, it is possible that mutations in these motifs also affect the interaction of Cet1p with the yeast guanylyltransferase Ceg1p, said interaction being important for Cet1p function in vivo (6). In order to gauge whether mutations of the catalytic domain of Cet1p affect Ceg1p binding, three of the triphosphatase-defective mutants (E305A, R454A, and E494A; one mutant for each motif) and the wild-type Cet1(201-549)p were sedimented in 15-30% glycerol gradients, either alone or after preincubation with Ceg1p. Marker proteins, catalase, BSA, and cytochrome c, were sedimented in a parallel gradient. As noted previously (2, 6), wild-type Cet1(201-549)p alone sedimented as a discrete peak near the BSA marker ( Fig. 2A, fractions 16 -18), as did Ceg1p alone (not shown), whereas pre-mixing Cet1(201-549)p with Ceg1p resulted in the formation of a more rapidly sedi-menting Cet1(201-549)p-Ceg1p complex recovered in gradient fractions 8 -10 ( Fig. 2A). Mutant E305A by itself sedimented as a discrete peak in fraction 16, but its position and that of Ceg1p was shifted to fraction 10 when the two proteins were pre- mixed and sedimented together (Fig. 2B). Similar results were obtained for mutants R454A (Fig. 2C) and E494A (Fig. 2D). We conclude that the RNA triphosphatase activity of Cet1(201-549)p is not required for its binding to Ceg1p.

Essential Residues in Motifs
Identification of New Essential Residues by Alanine Scanning-By having previously identified by alanine scanning six essential amino acids in motifs A, B, and C, we sought to expand the alanine scan to other positions in and near the motifs. Twelve positions were chosen for alanine substitution as follows: Leu-306 and Phe-310 (motif A); Thr-455, Lys-456, Ser-460, and His-463 (motif B); Asn-481, Lys-483, Ser-484, Arg-485, Val-493, and Leu-495 (motif C). These residues of Cet1p are conserved in at least two other members of the triphosphatase family, including one or more of the virusencoded RNA triphosphatases (Fig. 1).
The 12 CET1(201-549)-Ala alleles were tested for their function in vivo using the plasmid shuffle assay described by Ho et al. (2). The mutated genes were cloned into a CEN TRP1 vector so as to place them under the transcriptional control of the natural CET1 promoter. The plasmids were transformed into the cet1⌬ strain YBS20, in which the chromosomal CET1 locus has been deleted and replaced by LEU2. Growth of YBS20 is contingent on maintenance of a wild-type CET1 allele on a CEN URA3 plasmid. Therefore, YBS20 is unable to grow on agar medium containing 5-FOA, which selects against the URA3 plasmid, unless it is transformed with a biologically active CET1 allele. We found that YBS20 yielded colonies on 5-FOA at 30°C after transformation by mutant alleles L306A, T455A, S460A, H463A, N481A, K483A, S484A, R485A, V493A, or L495A. The timing of the appearance of the 5-FOA-resistant colonies after plating on selective medium and the size of the colonies formed on the selective plates were similar to what was observed for YBS20 transformed with wild-type CET1(201-549) (not shown). Mutant K456A formed only tiny colonies after 5-7 days of 5-FOA selection at 30°C (not shown). Mutant F310A failed to support growth even after incubation for 7-10 days on 5-FOA plates; thus, we conclude that the F310A mutation was lethal in vivo (Table I).
The 11 viable CET1(201-549)-Ala strains were tested for growth at 30 and 37°C in rich medium. T455A, S460A, H463A, N481A, K483A, S484A, and R485A cells grew at both temper-atures, and their colony sizes were similar to those formed by wild-type cells ( Fig. 3 and other data not shown). Therefore, the growth of these mutants was scored as ϩϩϩ (Table I). L306A, V493A, and L495A cells displayed a temperature-sensitive (ts) growth phenotype, i.e. these mutants grew well at 30°C but formed only pinpoint colonies at 37°C (not shown). K456A cells formed pinpoint or microscopic colonies at all temperatures ( Fig. 3 and data not shown); the growth of this mutant was scored as ϩ ( Table I).
Effects of Ala Substitutions on RNA Triphosphatase Activity-The 12 Cet1(201-549)-Ala proteins were expressed in bacteria as His-tagged fusions and purified from soluble lysates by Ni 2ϩ -agarose column chromatography. K456A, S460A, N481A, K483A, and S484A were expressed at 37°C and purified in parallel with wild-type Cet1(201-549)p. SDS-PAGE analysis of the polypeptide compositions of the Ni 2ϩ -agarose protein preparations revealed similar extents of purification (Fig. 4). L306A, F310A, T455A, H463A, V493A, and L495A were expressed in bacteria at 18°C in parallel with wild-type Cet1(201-549)p. (Lower temperatures were deemed necessary for production of the L306A, V493A, and L495A mutants, which were functionally thermolabile in vivo in yeast. The other mutants were produced at lower temperature after empirical determination that their solubility in bacteria was improved at 18 versus 37°C.) SDS-PAGE analysis of the mutants produced at lower temperature is shown in Fig. 5. The ϳ44-kDa His-Cet1(201-549) protein was the predominant species in every enzyme preparation (Figs. 4 and 5).
The RNA triphosphatase activities of the wild-type and mutant Cet1(201-549)p proteins were assayed by the release of 32 P i from 2 M ␥-32 P-labeled poly(A). Specific enzyme activity was determined from the slopes of the protein titration curves in the linear range of enzyme dependence. The specific activity of the wild-type Cet1(201-549)p preparation expressed in bacteria at 37°C was 25 pmol of P i released per ng of protein in 15 min. The specific activity of the Cet1(201-549)p preparation expressed in bacteria at 18°C was 20 pmol of P i released per ng of protein in 15 min. The specific activities of the Ala mutants, TABLE I Effects of alanine substitutions on CET1(201-549) function in vivo YBS20 was transformed with CEN TRP1 plasmids containing the indicated mutant alleles. Trpϩ transformants were selected and then streaked on medium containing 5-FOA (0.75 mg/ml). The plates were incubated at 25 and 30°C. Lethal mutations were those that formed no colonies after 7 days at either temperature. Individual colonies were picked from the FOA plate and patched on YPD agar. Two isolates of each mutant were streaked on YPD agar at 25, 30, and 37°C. Growth was assessed as follows: ϩϩϩ indicates colony size indistinguishable from strains bearing wild-type CET1; ϩ indicates that only pinpoint colonies were detected. Temperature-sensitive (ts) mutants were those that grew at 25 and 30°C but formed pinpoint colonies (ϩ growth) at 37°C.

Motif
Mutation Growth normalized to the wild-type specific activity at the applicable expression temperature, are shown in Table II. Comparison of the mutational effects on yeast cell growth and RNA triphosphatase activity of the recombinant proteins reveals the following correlations. First, the following seven Ala mutations that had no apparent effect on Cet1p function in vivo had no effect or a modest effect (defined operationally as less than a 10-fold decrement) on RNA triphosphatase-specific activity: T455A (84%), S460A (18%), H463A (17%), N481A (100%), K483A (96%), S484A (87%), and R485A (110%). Second, the lethal F310A mutation abolished RNA triphosphatase activity in vitro (0.2% of wild-type activity). Thus, for these cases, there is a direct correlation between activity in vivo and retention of function in vitro. The results are less straightforward for K456A, which is defective in vivo (albeit not lethal), and has only 0.4% of the wild-type RNA triphosphatase activity in vitro. The specific activity of recombinant K456A relative to the wildtype enzyme was the same when K456A expression was performed at 37 or 18°C (not shown). We suspect that the observed activity of the recombinant K456A protein purified from bacteria may underestimate its level of function when expressed in yeast cells; alternatively, it is possible that the threshold for slow growth versus no growth of yeast cells is between 0.4 (K456A) and 0.2% (F310A) of the wild-type-specific RNA triphosphatase activity. (Subsequent mutational analyses argue against such a tight threshold for growth versus no growth; see below.) The three ts mutants displayed a range of RNA triphosphatase-specific activities as follows: L306A (5%), V493A (21%), and L495A (2%).
Effects of Ala substitutions on ATPase Activity-Mutational effects on manganese-dependent ATP hydrolysis were gauged by enzyme titration at saturating ATP concentration (1 mM). The specific activity of the wild-type Cet1(201-549)p preparation expressed in bacteria at 37°C was 0.27 nmol of P i released per ng of protein in 15 min. The specific activity of the Cet1(201-549)p preparation expressed in bacteria at 18°C was 0.21 nmol of P i released per ng of protein in 15 min. The specific activities of the Ala mutants, normalized to the wild-typespecific activity at the applicable expression temperature, are shown in Table II. The following seven "functional" Ala mutants had little or no decline in their ATPase-specific activity: T455A (93%), S460A (64%), H463A (49%), N481A (110%), K483A (88%), S484A (96%), and R485A (100%). The F310A and K456A mutants displayed the most severe ATPase defects (3 and 4% of wild-type ATPase activity, respectively). The normalized specific activities of the ts mutants were as follows: L306A (15%), V493A (51%), and L495A (10%).
Mutants L306A, V493A, and L495A Are Thermolabile in Vitro-The thermal stability of wild-type Cet1(201-549)p and the L306A, V493A, and L495A mutants was tested by preincubation of the purified enzyme preparations for 10 min at either 30, 35, 40, 45, or 50°C, followed by quenching on ice. The protein samples were then assayed for ATPase activity at 30°C. The level of input enzyme in the assay mixtures was adjusted to achieve approximately the same extent of ATP hydrolysis in the unheated control reactions. The data were expressed as the ratio of ATP hydrolysis by enzyme preincubated at a given test temperature to the activity of the unheated control. The thermal inactivation curves are plotted in Fig. 6. The ATPase activity of wild-type Cet1(201-549)p was stable to preincubation at 30°C and reduced only modestly by treatment at 35 and 40°C. The ATPase activity fell off more sharply at 45°C (to 35% of the unheated control value) and 50°C (to 10% of the control value). L306A, V493A, and L495A were clearly thermolabile. The inactivation curve for L495A was shifted more than 10°C to the left relative to the wild-type enzyme; a similar effect was noted for L306A. The inactivation curve for V493A was shifted at least 5°C to the left compared with the wild-type enzyme.
Structure-Activity Relationships at Essential Residues in Motifs A, B, and C-Conservative substitutions were introduced at the 8 residues defined by alanine scanning as essential or important for activity: Glu-305, Glu-307, and Phe-310 (motif A); Arg-454 and Lys-456 (motif B); Glu-492, Glu-494, and Glu-496 (motif C). The five glutamates were replaced by aspartate or glutamine, and the mutant alleles were tested for function in  vivo by plasmid shuffle (Table III). E305D, E305Q, E307D, and E307Q were lethal, implying that the glutamates are strictly essential in motif A. E494D and E494Q were also lethal, signifying that the middle acidic residue in motif C was also strictly essential. E492Q cells were viable but grew more slowly than wild-type cells (Fig. 3A, scored as ϩϩ in Table III), whereas E492D cells were even more defective, forming only pinpoint colonies on rich medium (Fig. 3A, scored as ϩ). We surmise that an acidic moiety is not strictly essential at the first of three alternating glutamates in motif C; that glutamine is functionally superior to aspartate here implies that a polar group is key and that the distance of the functional group from the main chain is also critical. A similar structure-activity relationship was observed for Glu-496, insofar as E496D was lethal, whereas E496Q cells formed pinpoint colonies on rich medium (Fig. 3A). The R454K and R454Q mutations in motif B were lethal, which argues that function requires the bidentate arginine side chain and not merely positive charge. Whereas the K456A mutant grew extremely poorly, the K456R cells grew as well as wild-type cells (Fig. 3B and Table III). K456Q colonies were much smaller than wild type, much like K456A cells (Fig. 3B, scored as ϩ), implying that a positive charge at this position provides optimal function. The essential Phe-310 in motif A was substituted by tyrosine and leucine. F310Y and F310L grew as well as wild-type cells at all temperatures (Table III), which indicates that an aromatic group is not critical at this position; rather, a bulky aliphatic side chain suffices for Cet1p function in vivo.
The 16 conservatively substituted Cet1(201-549)p mutants were expressed in bacteria and purified by Ni 2ϩ -agarose chromatography (Figs. 4 and 5). Their RNA triphosphatase and ATPase specific activities (expressed as the percent of the wild-type specific activity) are shown in Table IV. The in vitro activities of the motif A mutants were in accord with the in vivo phenotypes. E305D, E305Q, E307D, and E307Q were catalytically defective for both RNA triphosphatase and ATPase, consistent with their in vivo lethality. The defects of the E305Q and E307Q mutants (Table IV) were just as severe as those of the E305A and E307A mutants (7). The aspartate-substituted proteins, although marginally more active than the alanine and glutamine mutants, were still 2 orders of magnitude less active than wild-type Cet1(201-549)p. At the Phe-310 position, the tyrosine-substituted protein displayed near wild-type activity in vitro, whereas F310L was one-half to one-third as active. Retention of triphosphatase activity by F310Y and F310L correlated with their ability to support cell growth.
The severe decrements in RNA triphosphatase activity elicited by replacement of the motif B Arg-454 by lysine (0.5% of wild-type activity) or glutamine (Ͻ0.1%) (Table IV) were consistent with the lethality of these changes in vivo. The RNA triphosphatase activities of R454K and R454Q were similar to that of R454A (0.2% of wild-type activity) (7). As noted previously for the R454A mutant (7), there was a discordance between R454K and R454Q effects on RNA triphosphatase (loss of function) and ATPase (substantial retention of activity). R454Q had 14% of the wild-type ATPase specific activity, similar to the 15% activity of the R454A mutant (7), whereas R454K had half the ATPase activity of wild-type Cet1(201-549)p. Replacement of the motif B lysine by arginine reduced RNA triphosphatase activity to 8% and ATPase to 24% of the wild-type values (Table IV). The residual activity apparently suffices for cell growth. Glutamine substitution for Lys-456 reduced RNA triphosphatase to 0.4% and ATPase to 6%; the glutamine mutation had virtually the same impact as alanine substitution.
For the six conservative mutations of the motif C glutamates, there was a fair correlation between the in vivo phenotypes and the mutational effects on RNA triphosphatase activity in vitro. For example, E492Q displayed the highest residual RNA triphosphatase activity (11% of wild-type) and supported ϩϩ growth in vivo. Lethal mutations E494D, E494Q, and E496D

TABLE III
Effects of conservative substitutions on CET1(201-549) function in vivo YBS20 was transformed with CEN TRP1 plasmids containing the indicated mutant alleles. Trpϩ transformants were selected and then streaked on medium containing FOA (0.75 mg/ml). The plates were incubated at 25 and 30°C. Lethal mutations were those that formed no colonies after 7 days at either temperature. Individual colonies were picked from the FOA plate and patched on YPD agar. Two isolates of each mutant were streaked on YPD agar at 25, 30, and 37°C. Growth was assessed as follows: ϩϩϩ indicates colony size indistinguishable from strains bearing wild-type CET1; ϩϩ denotes slightly reduced colony size; ϩ indicates that only pinpoint colonies were detected.

Motif
Mutation Growth Lethal E496Q ϩ reduced RNA triphosphatase activity by 2 orders of magnitude or more (Table IV). The mutations with a tiny colony phenotype, E492D and E496Q, displayed 1 and 3% of the wild-type RNA triphosphatase activity, respectively. The effects of substitutions of the motif C glutamates on ATP hydrolysis generally paralleled the effects on RNA triphosphatase. An exception to this trend was E494D, which retained one-fourth the ATPase activity of wild-type Cet1(201-549)p but only 0.8% residual RNA triphosphatase activity (Table IV). Motif B Mutations Increase the K m for ATP-Kinetic parameters were determined for manganese-dependent ATP hydrolysis by wild-type Cet1(201-549)p and the motif B mutants by measuring activity as a function of ATP concentration (7). Recombinant wild-type enzyme that had been expressed in bacteria at 37°C had a K m for ATP of 3.5 M; the K m of wild-type enzyme expressed at 18°C was 4.1 M (data not shown). Thus, the conditions of bacterial expression did not influence the affinity for ATP. The R454Q mutant had a significantly higher K m for ATP (100 M); this value was identical to the K m determined previously for the R454A protein (7). The K m of R454K was 23 M ATP (data not shown). These results underscore the importance of the bidentate arginine group (and not merely positive charge) in nucleotide binding affinity. The apparent K m values for the motif B lysine mutants were as follows: K456A (55 M ATP), K456Q (40 M ATP), and K456R (6.1 M ATP). We infer that the positive charge at position 456 is important for nucleotide binding.
RNA Triphosphatase and NTPase Activities of Cth1p, a Yeast Homologue of Cet1p-Motifs A, B, and C of yeast Cth1p include all 8 of the residues defined by alanine scanning as essential for the RNA triphosphatase activity of Cet1p (Fig. 1). To evaluate whether Cth1p is a triphosphatase, we expressed His-tagged Cth1p in E. coli under the control of an inducible T7 RNA polymerase promoter. The protein was purified from a soluble lysate of induced bacteria by Ni 2ϩ -agarose chromatography. SDS-PAGE analysis of the 0.2 M imidazole eluate fraction revealed a predominant 40-kDa polypeptide corresponding to His-Cth1p (Fig. 7A).
The Cth1p fraction catalyzed the release of 32 P i from ␥ 32 P-labeled triphosphate-terminated poly(A). Activity was proportional to input enzyme and the reaction proceeded to near completion at saturating enzyme concentration (Fig. 7B). We calculated that recombinant Cth1p released 125 fmol of P i per ng of protein in 15 min. Activity was strictly dependent on inclusion of magnesium in the reaction mixture (not shown). Cth1p catalyzed the release of 32 P i from 0.2 mM [␥-32 P]ATP in the presence of 1 mM manganese as the divalent cation cofactor. ATPase activity was proportional to the amount of input Cth1p protein (Fig. 7C). There was no detectable ATP hydrolysis in the absence of a divalent cation. ATPase activity was tested with other divalent actions added at 1 mM concentration. Cobalt was 10% as effective as manganese, whereas magnesium, calcium, copper, and zinc were inactive (not shown). Hydrolysis of 0.2 mM ATP was optimal at 0.4 -2 mM MnCl 2 . The titration curve was sigmoidal at manganese concentrations below the level of input ATP (not shown). ATP hydrolysis in 50 mM Tris buffer was optimal from pH 6.5 to 7.0. Activity at pH 8.5 to 9.5 was ϳ20% that at pH 7.0 (not shown).
Cth1p Activity Is Abolished by Replacement of the Motif A Glutamates with Alanine-Cth1p residues Glu-87 and Glu-89 in motif A (Fig. 1) were replaced individually by alanine. The E87A and E89A proteins were expressed as His-tagged fusions and purified from soluble lysates by Ni 2ϩ -agarose column chromatography. The purity of the recombinant E87A and E89A proteins was comparable to that of wild-type Cth1p (Fig. 7A). The E87A and E89A mutants were unable to hydrolyze triphosphatase-terminated RNA or ATP at levels of input enzyme sufficient for near-quantitative release of 32 P i by wild-type Cth1p (Fig. 7, B and C). From these data, we calculated that the specific RNA triphosphatase and ATPase activities of E87A and E89A were Ͻ0.1% of the activity of wild-type enzyme. We conclude that motif A is essential for the phosphohydrolase activity of Cth1p.
Characterization or the NTPase Activity of Cth1p-Cth1p catalyzed the quantitative conversion of [␣-32 P]ATP to [␣-32 P]ADP (Fig. 8A). Cth1p also catalyzed manganese-dependent hydrolysis of [␣-32 P]dATP to [␣-32 P]dADP with nearly identical kinetics (Fig. 8A). Thus, the enzyme has no apparent specificity for ribose versus deoxyribose sugars. We detected no formation of [ 32 P]AMP from [␣-32 P]ATP or [ 32 P]dAMP from [␣-32 P]dATP, even after 20 -45 min of incubation, by which time all of the nucleotide had been converted to ADP or dADP. We conclude that Cth1p catalyzes the hydrolysis of ATP to ADP plus P i and is unable to hydrolyze further the ADP reaction product. Kinetic parameters were determined from the dependence of activity on input [␥-32 P]ATP concentration. From a double-reciprocal plot of the data (Fig. 8B), we calculated a K m of 75 M for ATP and a V max of 2 s Ϫ1 . The NTPase activity of Cth1p was not restricted to adenosine nucleotides; Cth1p hydrolyzed [␣-32 P]GTP to [␣-32 P]GDP and [␣-32 P]UTP to [␣-32 P]UDP (not shown).
The native size of Cth1p was analyzed by sedimentation of the Ni 2ϩ -agarose enzyme fraction through a 15-30% glycerol gradient containing 250 mM NaCl. Marker proteins, catalase, BSA, and cytochrome c, were included as internal standards in the same glycerol gradient. After centrifugation, the polypeptide compositions of the gradient fractions were analyzed by SDS-PAGE, and the fractions were assayed for manganese-dependent ATPase. The ATPase sedimented as a discrete peak in fraction 19 (Fig. 8C) along with the 40-kDa Cth1p protein. An apparent sedimentation coefficient of 3.7 S relative to internal standards suggested that recombinant Cth1p is a monomeric enzyme.
Genetic Analysis of CTH1-The CTH1 gene was replaced by insertion of a kanamycin resistance gene (14). The disruption was performed in a diploid strain such that marker insertion eliminated the CTH1 coding sequence from amino acid positions 32-278. Correct insertion into one CTH1 locus was confirmed by Southern blotting of kanamycin-resistant transformants. After sporulation and tetrad dissection, viable kanamycin-resistant haploids were recovered in a 2:2 segregation pattern (not shown). The size of colonies formed by the cth1⌬ null strains was indistinguishable from that of the parental CTH1 strain at either 16, 25, or 37°C. We conclude that CTH1 is nonessential. The CTH1 gene was cloned into a 2 plasmid under the control of the constitutive yeast TPI1 promoter. Introduction of this plasmid into YBS20 (cet1⌬ p360-CET1) did not allow for growth of the transformants on 5-FOA (not shown). We surmise that Cth1p by itself was unable to substitute for Cet1p in vivo, even when overexpressed.

Fusion of Cth1p to the Guanylyltransferase Domain of Mouse Capping Enzyme Confers the Ability to Complement Cet1p
Function in Vivo-Cth1p does not contain a segment homologous to the conserved portions of S. cerevisiae Cet1p and C. albicans CaCet1p flanking the catalytic core that are implicated in triphosphatase-guanylyltransferase complex formation (5, 6). A plausible scenario to explain why Cth1p is unable to complement Cet1p function is as follows: (i) Cet1p is nor-mally targeted to pre-mRNAs by virtue of its association with Ceg1p, and (ii) Cth1p has no chaperone to direct it to nascent pre-mRNAs. Lehman et al. (6) showed recently that the in vivo requirement for the putative Ceg1p-binding site of Cet1p can be bypassed by linking the Cet1p triphosphatase catalytic domain in cis to a heterologous guanylyltransferase, i.e. the Cterminal guanylyltransferase domain of the mammalian capping enzyme Mce1p. Mce1p is a bifunctional polypeptide composed of an N-terminal triphosphatase domain (aa 1-210) and a C-terminal guanylyltransferase domain (aa 211-597) (15,23). Mce1(211-597)p binds directly to the phosphorylated CTD of RNA polymerase II, whereas the isolated N-terminal mouse triphosphatase domain Mce1(1-210)p does not bind the CTD (15,16). In effect, the guanylyltransferase chaperones the triphosphatase to the transcription complex. Given that the mammalian guanylyltransferase can also act as chaperone for the catalytic domain of yeast triphosphatase Cet1p when the two are fused, we tested whether fusion of Cth1p to the mouse guanylyltransferase might target Cth1p to RNA polymerase II and thereby result in a gain-of-function in vivo.
The fusion gene CTH1-MCE1(211-597) was cloned into a 2 TRP1 vector such that it was under the control of the yeast TPI1 promoter. The CTH1-MCE1(211-597) plasmid was transformed into a cet1⌬ ceg1⌬ double-deletion strain (YBS50). Transformation of YBS50 with plasmids bearing MCE1 or MCE1(211-597) provided positive and negative controls, respectively. Growth of YBS50 depends on maintenance of a CEN URA3 CET1 CEG1 plasmid. Transformation of YBS50 with a TRP1 plasmid containing MCE1 (encoding the bifunctional mouse capping enzyme) under control of the yeast TPI1 promoter allowed growth of YBS50 on medium containing 5-FOA. Transformants bearing MCE1(211-597), which encodes only the guanylyltransferase domain of mouse capping enzyme, failed to give rise to FOA-resistant colonies even after prolonged incubation (up to 7 days). This was expected because a functional triphosphatase on the TRP1 plasmid is needed to complement the cet1⌬ ceg1⌬ double-deletion (17). The instructive finding was that YBS50 cells bearing the 2 CTH1-MCE1(211-597) plasmid did give rise to small FOA-resistant colonies after 3-5 days (not shown).
Two independent FOA-resistant isolates of S. cerevisiae YBS50 CTH1-MCE1(211-597) cells were tested for growth in rich medium in parallel with FOA-selected derivatives of YBS50 containing MCE1. Although the cells grew when the fusion protein Cth1p-Mce1(211-597)p was the only source of triphosphatase and guanylyltransferase activities, cell growth rate (gauged by colony size) was clearly slower than when native Mce1p was present (Fig. 9). The colony size of MCE1 cells is indistinguishable from CET1 CEG1 cells (15,17).
We draw the following two conclusions from the genetic analysis: (i) Cth1p is not essential for cell viability and is not normally involved in mRNA cap formation (that role being played by Cet1p); (ii) nonetheless, Cth1p can function in vivo in lieu of Cet1p when fused to a chaperone that delivers the enzyme to the transcription complex. DISCUSSION The present study of the yeast RNA triphosphatase Cet1p consolidates the hypothesis that motifs A, B, and C participate directly in phosphohydrolase reaction chemistry and substrate binding. Structure-activity relationships at essential side chains suggest a plausible catalytic mechanism, and the results of alanine scanning of the motifs lead to predictions for the secondary structures of motifs A and C. The identification of Cth1p as a second RNA triphosphatase in S. cerevisiae highlights that the occurrence of motifs A, B, and C has predictive value for the function of proteins identified by genomic sequencing, a point underscored by the demonstration that the motif A glutamates are required for Cth1p triphosphatase activity.
Insights from Mutagenesis into the Structure and Catalytic Mechanism of Cet1p-The striking feature of motifs A (IELEMKF) and C (EVELE) is that they consist of charged side chains at every other position interdigitated with alternating aliphatic/aromatic side chains. This arrangement, together with mutational data, suggests that motifs A and C are folded as ␤-strands so that the hydrophilic and hydrophobic functional groups are arrayed on opposite sides of the main chain. Alanine substitutions of Phe-310 in motif A result in loss of function; the fact that function is restored by introduction of a leucine argues that the hydrophobic character of the side chain is important. Alanine substitutions for three of the other alternating aliphatic side chains (Leu-306 in motif A, plus Val-493, and Leu-495 in motif C) elicit thermolability in vivo and vitro. This suggests to us that the "back" surface of the predicted ␤-strands engages in hydrophobic interactions with other structural elements of the protein core that stabilize the active conformation of the hydrophilic residues of the strands. We predict that the hydrophilic surfaces of motifs A and C, which contain 5 essential glutamates, comprise part of the triphosphatase-active site. We hypothesized previously that the glutamates in motifs A and C facilitate catalysis by coordinating the essential divalent cation(s) (7). The expectation is that an acidic side chain should be critical for metal binding, in which case replacement of a metal-binding glutamate by glutamine should abrogate activity. The effects of conservative substitutions on Cet1p triphosphatase activity point toward Glu-305 and Glu-307 in motif A and Glu-494 and Glu-496 in motif C as prime candidates for the metal-binding site(s), because their replacement by glutamine inactivates Cet1p. Glu-492 is probably not involved in metal binding, insofar as there is substantial residual triphosphatase activity when a glutamine occupies this position. The loss of activity when Glu-492 is replaced by alanine suggests that side chain polarity, with the capacity to hydrogen-bond, is the key property of this position of Cet1p. Note that the poxvirus triphosphatases have only two glutamates in motif C (Fig. 1). Replacement of either glutamate of vaccinia D1 by alanine abolished triphosphatase activity (13). Thus, one or two glutamates in vaccinia motif C may suffice for metal binding, and the role played by the proximal glutamate in Cet1p may be fulfilled by the polar serine in the vaccinia protein (Fig. 1).
Motif B of Cet1p includes two important basic residues, Arg-454 and Lys-456. The surrounding sequence context is quite hydrophilic (SERTKDR), and we make no presumptions from the available data about the secondary structure of motif B. The mutational data implicate motif B in substrate binding. Changing Arg-454 to alanine or glutamine does not abolish ATP hydrolysis but does result in a 30-fold increase in the K m for ATP. Even a conservative lysine replacement increases the K m for ATP by a factor of 6. We suggest that the arginine side chain makes a bidentate contact with the 5Ј-triphosphate of the substrate. Arg-454 mutations appear to have a much greater impact on RNA triphosphatase activity than on ATP hydrolysis. Note, however, that the ATPase assays are performed at 1 mM ATP, a concentration 300-fold in excess of the K m for ATP (7), whereas the RNA triphosphatase assays are performed at 2 M substrate, which is quite close to the K m value of 1 M for triphosphate-terminated poly(A) (7). We suspect that the very low specific RNA triphosphatase activity of the Cet1p mutants with low affinity for ATP can be attributed, at least in part, to similarly reduced affinity for RNA. (The yield of labeled RNA ends in the enzymatic synthesis of triphosphate-terminated poly(A) limits the substrate concentrations attainable in the RNA triphosphatase assay. It would be difficult in practice to conduct the assays at the higher substrate concentrations that are easily attained with ATP.) FIG. 9. Fusion of Cth1p to the guanylyltransferase domain of mouse capping enzyme results in complementation of a ⌬cet1 deletion. Yeast strain YBS50 (cet1⌬ ceg1⌬) was transformed with 2 TRP1 plasmids containing the CTH1-MCE1(211-597) fusion gene driven by the TPI1 promoter. A control transformation was performed with a CEN TRP1 plasmid containing MCE1 (15). Trpϩ isolates were streaked on agar plates containing 0.75 mg/ml 5-FOA. FOA-resistant cells from single colonies were patched onto agar medium lacking tryptophan and were incubated for 2 days at 30°C. Cells from single patches were then streaked on YPD agar. Two independent CTH1-MCE1(211-597) isolates were tested. The plates were photographed after incubation for 4 days at 30°C.
Replacing motif B Lys-456 by alanine or glutamine increases K m for ATP by at least an order of magnitude. ATP binding is restored when arginine is introduced at this position. We speculate that Lys-456 makes a monovalent contact with the 5Јtriphosphate of the substrate. The K456A and K456Q mutants have very low catalytic activity, even at 1 mM ATP. Thus, Lys-456 may also play a role in catalysis by Cet1p.
A Second Yeast RNA Triphosphatase Cth1p-We have shown that yeast Cth1p is both an RNA 5Ј-triphosphatase and an NTPase. Similar findings were reported by Rodriguez et al. (18) while this manuscript was in preparation. We find that the NTPase of Cet1p is activated by manganese and cobalt. This is a property shared with the triphosphatase components of the yeast (Cet1p), vaccinia (D1), and baculovirus (LEF-4) capping enzymes (7)(8)(9)(10)(11). The turnover number of the Cth1p in ATP hydrolysis (2 s Ϫ1 ) is lower than the values reported for Cet1p (25-33 s Ϫ1 ), baculovirus LEF-4 (30 s Ϫ1 ), and vaccinia virus D1 (10 s Ϫ1 ) (7,8,10,19). The K m value of Cth1p for ATP (75 M) is higher than that of either Cet1p (2.8 M) or LEF-4 (43 M) but lower than that of vaccinia capping enzyme (800 M) (7,8,10,19). The function of Cth1 in vivo is unknown; this enzyme may well catalyze phosphohydrolase reactions unrelated to mRNA capping. Indeed, it is not even clear that RNA 5Ј ends are the relevant substrates for Cth1p action in vivo. Nonetheless, we have shown that Cth1p can act as an RNA triphosphatase in the cap synthetic pathway in vivo, provided that it is fused in cis to mammalian guanylyltransferase.
Targeting Enzymes to Pre-mRNAs in Vivo-Targeting of the cellular capping apparatus to nascent RNA polymerase II transcripts is achieved via the binding of one or more components of the capping enzymes to the phosphorylated C-terminal domain (CTD) of elongating RNA polymerase II (20 -22). The CTD, consisting of tandem repeats of a heptapeptide of the consensus sequence YSPTSPS, is extensively phosphorylated in the context of the transcription elongation complex (24). The guanylyltransferase domain of mammalian capping enzyme Mce1p binds specifically to the phosphorylated CTD, but not to unmodified CTD (15,16,20,21). The triphosphatase domain of mammalian capping enzyme does not bind the CTD, but is normally brought along via its linkage in cis to the guanylyltransferase. In yeast, the guanylyltransferase Ceg1p binds to CTD-PO4, whereas the triphosphatase Cet1p does not (5,21). Formation of a Cet1p-Ceg1p complex in trans allows the yeast guanylyltransferase to chaperone the triphosphatase to the transcription complex. We showed previously that mammalian guanylyltransferase can act as chaperone in cis for a catalytic domain of yeast triphosphatase Cet1p that lacks the ability to bind to Ceg1p (6). Now, we find that mammalian guanylyltransferase can target Cth1p, an RNA triphosphatase not normally involved in capping, and thereby convert it into a capforming enzyme in vivo. This result suggests that the mammalian guanylyltransferase can be used as a vehicle to transiently deliver heterologous proteins to the RNA polymerase II transcription elongation complex in vivo. Such a vehicle may prove useful in designing strategies to alter the structural or functional properties of the elongating polymerase or the nascent RNA and thereby modulate gene expression.