Homodimeric quaternary structure is required for the in vivo function and thermal stability of Saccharomyces cerevisiae and Schizosaccharomyces pombe RNA triphosphatases.

Saccharomyces cerevisiae Cet1 and Schizosaccharomyces pombe Pct1 are the essential RNA triphosphatase components of the mRNA capping apparatus of budding and fission yeast, respectively. Cet1 and Pct1 share a baroque active site architecture and a homodimeric quaternary structure. The active site is located within a topologically closed hydrophilic beta-barrel (the triphosphate tunnel) that rests on a globular core domain (the pedestal) composed of elements from both protomers of the homodimer. Earlier studies of the effects of alanine cluster mutations at the crystallographic dimer interface of Cet1 suggested that homodimerization is important for triphosphatase function in vivo, albeit not for catalysis. Here, we studied the effects of 14 single-alanine mutations on Cet1 activity and thereby pinpointed Asp280 as a critical side chain required for dimer formation. We find that disruption of the dimer interface is lethal in vivo and renders Cet1 activity thermolabile at physiological temperatures in vitro. In addition, we identify individual residues within the pedestal domain (Ile470, Leu519, Ile520, Phe523, Leu524, and Ile530) that stabilize Cet1 in vivo and in vitro. In the case of Pct1, we show that dimerization depends on the peptide segment 41VPKIEMNFLN50 located immediately prior to the start of the Pct1 catalytic domain. Deletion of this peptide converts Pct1 into a catalytically active monomer that is defective in vivo in S. pombe and hypersensitive to thermal inactivation in vitro. Our findings suggest an explanation for the conservation of quaternary structure in fungal RNA triphosphatases, whereby the delicate tunnel architecture of the active site is stabilized by the homodimeric pedestal domain.

RNA triphosphatase catalyzes the first step in mRNA cap formation, the cleavage of the ␤-␥ phosphoanhydride bond of 5Ј-triphosphate RNA to yield a diphosphate end. In the second step of the pathway, the RNA diphosphate is capped with GMP by RNA guanylyltransferase to yield GpppRNA (1). The budding yeast Saccharomyces cerevisiae encodes separate triphosphatase (Cet1; 549 aa 1 ) and guanylyltransferase (Ceg1; 459 aa) proteins that interact in trans to form a stable capping enzyme complex consisting of one Ceg1 protomer bound to a dimer of Cet1 (2)(3)(4)(5). Although the fission yeast Schizosaccharomyces pombe also encodes separate triphosphatase (Pct1; 303 aa) and guanylyltransferase (Pce1; 402 aa) enzymes, they do not interact with each other (6 -8).
The Cet1-Ceg1 interaction stabilizes the intrinsically labile guanylyltransferase activity of Ceg1 against thermal inactivation at physiological temperatures (9). In addition, the physical tethering of Cet1 to Ceg1 facilitates recruitment of the triphosphatase to the RNA polymerase II elongation complex, via Ceg1 binding to the phosphorylated C-terminal domain (CTD) of the largest subunit of RNA polymerase II (10 -13). Cet1 by itself does not interact with the CTD. In contrast, the S. pombe guanylyltransferase Pce1 is inherently thermostable, and its stability is unaffected by the presence of the triphosphatase Pct1 (9). Also, S. pombe employs a distinctive strategy of targeting capping to polymerase II transcripts, whereby the Pct1 and Pce1 enzymes bind independently to the phosphorylated CTD (8). Thus, the fission yeast has elided on both counts the need for a triphosphatase-guanylyltransferase complex. It is therefore not surprising that Pct1 has no counterpart of the surface domain of Cet1 that mediates binding to the guanylyltransferase (7).
Although the fungal triphosphatase components display species-specific differences in their protein-protein interactions, they are nonetheless conserved with respect to their active site architecture, catalytic mechanism, and quaternary structure. The yeast triphosphatases belong to a family of metal-dependent phosphohydrolases that embraces the RNA triphosphatase components of the capping enzymes of unicellular eukaryotes and certain DNA viruses (7,14,15). The family is defined by the presence of two conserved glutamate-containing motifs (␤1 and ␤11 in Fig. 1) and the signature property of hydrolyzing NTPs to NDPs in the presence of manganese or cobalt (7,14). The crystal structure of the S. cerevisiae RNA triphosphatase Cet1 revealed that the enzyme is a homodimer with active sites located within parallel topologically closed tunnels composed of eight ␤ strands (16) (Fig. 1). The "triphosphate tunnel" architecture is supported by an intricate network of hydrogen bonds and electrostatic interactions within the cavity, most of which are required for catalytic activity (17,18). The tunnel floor rests on a globular "pedestal" domain. Amino acid sequence comparisons and mutational analyses of the RNA triphosphatases from other fungi (e.g. Candida albicans and S. pombe), microsporidia, protozoa, and Chlorella virus underscore the conservation of the ␤ strands that comprise the triphosphate tunnel (15). Mutational analyses of the C. albicans and S. pombe RNA triphosphatases indicate that their active sites and catalytic mechanism adhere closely to that of Cet1 (7,19).
The S. cerevisiae and S. pombe RNA triphosphatases are both homodimers (4,7,16). Available evidence indicates that homodimer formation is essential for Cet1 function in vivo but not for catalytic activity. Deletion analysis showed that the C-terminal domain Cet1(276 -549) has a monomeric quaternary structure and retains activity in vitro (4). However, the monomeric domain by itself cannot support yeast cell growth, even when it is overexpressed at high gene dosage under the control of a strong promoter. Interpretation of the deletion data is complicated by the fact that an N-terminal truncation to position 275 also removes the guanylyltransferase-binding site 247 WAQKW 251 , which is located on the protein surface (4,16,20,21) and is responsible for Cet1-mediated stabilization of the guanylyltransferase Ceg1 (9). The in vivo function of Cet1(276 -549) is restored when the monomeric triphosphatase is fused to either S. pombe guanylyltransferase (Pce1) or the guanylyltransferase domain of mammalian capping enzyme (4,9). The S. pombe and mammalian guanylyltransferases bind to the phosphorylated CTD (8,10,22,23) and can thereby act as vehicles to deliver the fused monomeric yeast RNA triphosphatase to the RNA polymerase II elongation complex (4,9). Also, because the S. pombe and mammalian guanylyltransferases are thermostable (unlike Ceg1), the chimeric capping enzymes bypass the need for the Ceg1-stabilization function of the 247 WAQKW 251 peptide of Cet1 (9). To focus specifically on the role of homodimerization in Cet1 function in vivo, we previously performed an alanine cluster mutational analysis guided by the Cet1 crystal structure (24). Double-alanine mutations at vicinal amino acids were introduced into the biologically active protein Cet1(201-549), which contains both the guanylyltransferase-binding and catalytic domains. A total of 42 residues were changed to alanine, 24 of which were constituents of the crystallographic dimer interface. Four of the Ala cluster alleles were lethal in vivo. Three other Ala cluster mutants displayed temperature-sensitive (ts) growth defects, even when the mutant alleles were present in high copy under the control of a strong promoter. Several of the lethal and ts mutations were suppressed by fusion of the Cet1-Ala/Ala protein to the mammalian guanylyltransferase. Moreover, two of the lethal mutant proteins were characterized in vitro and found to be catalytically active monomers (24). These results indicated that homodimerization of the budding yeast RNA triphosphatase is critical in vivo when Cet1 functions in concert with the endogenous yeast guanylyltransferase.
It remains unclear why homodimerization of yeast RNA triphosphatase is important in vivo. If the dimer is critical for the functional interactions of S. cerevisiae RNA triphosphatase and RNA guanylyltransferase, then it is not obvious why a homodimeric quaternary structure for RNA triphosphatase would be conserved in the fission yeast S. pombe, where the triphosphatase and guanylyltransferase components do not interact physically. On the other hand, homodimerization may confer added value to the triphosphatase in other ways that are independent of the guanylyltransferase component of the capping apparatus.
Here we conduct a series of experiments to define the individual essential constituents of the Cet1 homodimer interface and probe the role of quaternary structure in triphosphatase function in vivo and in vitro. Guided by the initial results of the Ala cluster mutagenesis, we tested the effects of 14 singlealanine mutations on Cet1 activity in vivo and thereby pinpointed Asp 280 as a critical side chain required for dimer formation. We find that disruption of the dimer interface renders Cet1 thermolabile in vitro. We engineered a catalytically active monomeric version of S. pombe Pct1 and show that it is thermolabile in vitro. Introduction of the monomeric triphosphatase into a S. pombe pct1⌬ strain confers a dosage-sup-pressible lethal phenotype in vivo. We propose a model whereby homodimerization of the globular pedestal domain is critical to stabilize the delicate active site tunnel architecture of the fungal RNA triphosphatases.

EXPERIMENTAL PROCEDURES
Mutagenesis of S. cerevisiae RNA Triphosphatase-Alanine mutations were introduced into the CET1(201-549) gene by PCR (25). 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 (26). The inserts were sequenced completely to exclude the acquisition of unwanted mutations during amplification and cloning. The in vivo activity of the mutated CET1 alleles was tested by plasmid shuffle. Yeast strain YBS20 (trp1 ura3 leu2 cet1::LEU2 p360-Cet1[CEN URA3 CET1]) was transformed with TRP1 plasmids containing the wild-type and mutant alleles of CET1(201-549). Trp ϩ isolates were selected and then streaked on agar plates containing 0.75 mg/ml 5-fluoroorotic acid (5-FOA). Growth was scored after 7 days of incubation at 18°, 25°, 30°, and 37°C. Lethal mutants were those that failed to form colonies on 5-FOA at any temperature. Individual colonies of the viable CET1 mutants were picked from the FOA plate at permissive temperature and patched to YPD agar. Two isolates of each mutant were tested for growth on YPD agar at 18°, 25°, 30°, and 37°C. Growth was assessed as follows: ϩϩϩ indicates colony size indistinguishable from strains bearing wild-type CET1(201-549); ϩϩ denotes slightly reduced colony size; ϩ indicates that only pinpoint colonies were formed.
Mutagenesis of S. pombe RNA Triphosphatase-Gene fragments encoding N-terminal-truncated versions of Pct1 were generated by PCR amplification using sense primers that introduced an NdeI site at the codons for amino acids 41 or 51. The antisense primers introduced a BamHI site immediately downstream of the stop codon. The PCR products were digested with NdeI and BamHI and then inserted into pET16b. Full-length Pct1 and the N⌬40 and N⌬50 mutants were produced in E. coli as N-terminal His 10 -tagged fusions and purified from soluble bacterial lysates by nickel-agarose chromatography as described previously (7,8).
Glycerol Gradient Sedimentation-Aliquots (45 g) of the nickelagarose preparations of wild-type Cet1(201-549) and the D280A mutant were mixed with BSA (40 g) and cytochrome c (40 g) in 0.2 ml of buffer G (50 mM Tris HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 2 mM DTT, 0.05% Triton X-100). Aliquots (50 g) of the nickel-agarose preparations of wild-type Pct1 and the N⌬40 and N⌬50 mutants were mixed with BSA (50 g), ovalbumin (50 g), and cytochrome c (50 g) in 0.2 ml of buffer G. The mixtures were layered onto 4.8 ml of 15-30% glycerol gradients containing buffer G. The gradients were centrifuged in a Beckman SW50 rotor at 50,000 rpm for 24 h at 4°C. Fractions (ϳ0.2 ml) were collected from the bottoms of the tubes. Aliquots (20 l) of oddnumbered fractions were analyzed by SDS-PAGE. Polypeptides were visualized by staining with Coomassie Blue dye.
Assay of Triphosphatase Activity-Reaction mixtures (10 l) containing 50 mM Tris HCl (pH 7.5), 5 mM DTT, 2 mM MnCl 2 , 1 mM [␥-32 P]ATP, and Cet1 or Pct1 as specified were incubated for 15 min at 30°C. The reactions were quenched by adding 2.5 l of 5 M formic acid. An aliquot of the mixture was applied to a polyethyleneimine-cellulose TLC plate, which was developed with 0.5 M LiCl and 1 M formic acid. The release of 32 P i from [␥-32 P]ATP was quantitated by scanning the TLC plate with a PhosphorImager.
Test of RNA Triphosphatase Function in Vivo in S. pombe-The full-length pct1 ϩ and truncated pct1-N⌬40 and pct1-N⌬50 cDNAs were cloned into the S. pombe expression vectors pREP81X, pREP41X, and pREP3X (LEU2 ars1 ϩ ) so as to place them under the control of the nmt1**, nmt1*, and nmt1 promoters, respectively (27,28). The plasmids were transformed into a heterozygous pct1 ϩ /pct1::kanMX diploid (29) using the lithium acetate method (30). The Leu ϩ diploid transformants were selected and then sporulated on ME plates at room tem-perature. A loopful of cells was inoculated into 500 l of sterile water, and the mixture was incubated overnight at 28°C with 10 l of ␤-glucuronidase (Sigma G7770). The spores were plated on EMM(-Leu) agar medium and incubated at 30°C. Individual colonies were then restreaked onto YES agar and on YES agar containing 200 g/ml G418. Growth was scored after incubation for 5-7 days at 30°and 37°C.

Structure-based Mutational Analysis of the Cet1 Homodimer
Interface-The Cet1 homodimer interface is extensive, with a buried surface area of 1860 Å 2 per protomer (16). Elements that comprise the dimer interface are strands ␤2 and ␤3, helices ␣1 and ␣4, the loop immediately preceding ␣1, the loop between ␤9 and ␤10, and the loop between ␣3 and ␣4 (see Fig.  1). The molecular contacts of the dimer interface entail multi-ple hydrophobic interactions and a network of side-chain and main-chain hydrogen bonds. We previously tested the effects of 21 double-alanine mutations of vicinal amino acids on Cet1 function in vivo (24). Twenty-four of the mutated residues were constituents of the crystallographic dimer interface. The Ala cluster mutations also targeted residues in helices ␣1 and ␣4 that compose the hydrophobic core of the pedestal upon which the triphosphate tunnel rests. Four of the Ala cluster alleles were lethal in vivo: D279A-D280A, C330A-V331A, L519A-I520A, and F523A-L524A. Three Ala cluster mutants displayed temperature-sensitive (ts) growth defects, even at high gene dosage: F272A-L273A, I470A-I472A, and I529A-I530A.
The cluster mutagenesis approach provided useful information in two respects: (i) it showed that 15 of the amino acids FIG. 1. Mutational analysis of the homodimer interface of yeast RNA triphosphatase. A, the Cet1 homodimer is shown as a ribbon diagram. The view is looking into the tunnel entrance. B, the secondary structure of Cet1 from residues 241 to 539 is shown above its amino acid sequence, which is aligned to the homologous segments of C. albicans CaCet1 and S. pombe Pct1. The C termini of the CaCet1 and Pct1 polypeptides are indicated by asterisks. The N-terminal 40-aa segment of Pct1, which has no apparent counterpart in Cet1 or CaCet1, is shown in italics. Gaps in the alignment are indicated by dashes. Individual amino acids in Cet1 that were mutated to alanine in the present study are denoted by dots above the sequence. Essential constituents of the Cet1 active site that are conserved in CaCet1 and Pct1 are denoted by the symbol . The margins of Pct1 deletions N⌬40 and N⌬50 are indicated by arrows below the sequence.
that comprise the crystallographic homodimer interface are not important for Cet1 function in vivo, and (ii) it narrowed down the critical constituents of the homodimer interface to one or both of the side chains of the clusters at which Ala-Ala mutations elicited severe growth defects. Here we set out to identify which individual components of the dimer interface are functionally relevant, by testing the effects of single-alanine mutations at each of the residues of those clusters. As in previous studies, the mutations were introduced into the biologically active Cet1(201-549) protein.
The CET1(201-549)-Ala genes were cloned into a CEN TRP1 vector under the control of the natural CET1 promoter and then tested by plasmid shuffle for their ability to complement a cet1⌬ strain of S. cerevisiae. Growth of cet1⌬ is contingent upon maintenance of a wild-type CET1 allele on a CEN URA3 plasmid. Therefore, the cet1⌬ strain is unable to grow on agar medium containing 5-FOA (5-fluoroorotic acid, a drug that selects against the URA3 plasmid) unless it is first transformed with a biologically active RNA triphosphatase gene on the TRP1 plasmid. Trp ϩ CET1(201-549)-Ala transformants were tested for growth on 5-FOA. The results are summarized in Table I. Triphosphatase mutations were judged to be lethal if they failed to support colony formation on 5-FOA after prolonged incubation at four different temperatures (18°, 25°, 30°, and 37°C). Two of the single-Ala mutations were lethal by this criterion: D280A and I520A.
The other 12 mutants gave rise to 5-FOA-resistant colonies at one or more of the selection temperatures. The viable CET1(201-549)-Ala strains were then tested for growth on rich medium (YPD agar) at 18°, 25°, 30°, and 37°C. A hierarchy of severity of mutational effects was thereby revealed ( Table I). Four of the mutants displayed a tight ts phenotype: I470A and L524A cells grew as well as wild-type yeast at 18°and 25°C (scored as ϩϩϩ based on colony size) but failed to grow at 30°a nd 37°C (Ϫgrowth); F523A and I530A cells grew normally at 18°C but formed small colonies at 25°C (scored as ϩϩ) and no colonies at 30°and 37°C. Four of the mutants displayed weaker ts phenotypes: the F272A, V331A, and I472A strains grew normally at 18°, 25°, and 30°C but formed small colonies at 37°C; L519A cells grew normally at 18°, 25°, and 30°C but formed only pinpoint colonies at 37°C. Four of the mutants, L273A, D279A, C330A, and I529A, grew as well as wild-type yeast at all temperatures tested.
Three of the residues at which single-alanine substitutions elicited significant growth defects are components of the crystallographic homodimer interface: Asp 280 , Ile 470 , and Ile 530 . The Asp 280 side chain is the only one of these that is strictly essential for cell viability. Thus, Asp 280 appears to be an important constituent of the homodimer interface. None of the other three amino acids (Ile 520 , Phe 523 , and Leu 524 ) at which alanine mutations resulted in significant growth defects are involved in cross-dimer contacts; rather, they are components of the hydrophobic core of the pedestal domains of the individual protomers.
Biochemical Characterization of Mutant Enzymes-We produced the 14 Cet1(201-549)-Ala proteins in bacteria as His 10tagged fusions and purified them from soluble bacterial lysates by nickel-agarose chromatography. Wild-type His 10 -Cet1(201-549) was purified in parallel. SDS-PAGE analysis showed that the ϳ44-kDa Cet1(201-549) protein was the predominant species in each enzyme preparation ( Fig. 2A). Phosphohydrolase activity was assayed by the release of 32 P i from [␥-32 P]ATP in the presence of manganese chloride (7). The extents of ATP hydrolysis increased as a function of input enzyme for each protein (Fig. 2B). A specific activity for the wild-type Cet1(201-549) of 0.31 nmol of ATP hydrolyzed per ng of protein in 15 min at 30°C was calculated from the slope of the titration curve in the linear range. (This value translates into a turnover number of 16 s Ϫ1 .) The specific activities of 11 of the Cet1(201-549)-Ala mutants were similar (within a factor of 2) to that of the wild-type enzyme, as follows: F272A (79% of wild-type), L273A (96%), D279A (79%), D280A (62%), C330A (100%), V331A (64%), I470A (57%), I472A (77%), L519A (86%), I529A (76%), and I530A (77%). The retention of triphosphatase activity by the F272A, L273A, D279A, C330A, V331A, I470A, I472A, L519A, I529A, and I530A proteins is consistent with the observations that each of these ten mutants supported yeast cell growth at two or more of the temperatures tested (Table I). The instructive finding was that the D280A mutation, which was lethal in vivo, had no significant impact on Cet1 phosphohydrolase activity in vitro.
Three of the Cet1(201-549)-Ala mutants were catalytically defective; I520A, F523A, and L524A were 2, 5, and 7% as active as wild-type, respectively (Fig. 2B, right panel). The hierarchy of mutational effects on triphosphatase activity in vitro paralleled the lethal (I520A) and severe ts (F523A and L524A) ef-  CET1(201-549) alleles in the yeast cet1⌬ strain as performed as described under "Experimental Procedures." Trp ϩ isolates were selected and then streaked on agar plates containing 0.75 mg/ml 5-FOA. Growth was scored after 7 days of incubation at 25, 30, and 37°C or 12 days at 18°C. Lethal mutants D280A and I520A failed to form colonies on 5-FOA at any temperature. Individual colonies of the viable CET1 mutants were picked from the 5-FOA plates at permissive temperature and patched to ϪTrp agar. Two isolates of each mutant were tested for growth on YPD agar at 18, 25, 30, and 37°C. Growth was assessed as follows: ϩϩϩ indicates colony size indistinguishable from strains bearing CET1; ϩϩ denotes slightly reduced colony size; ϩ indicates that only pinpoint colonies were formed; Ϫ denotes no growth. The cross-dimer and intra-protomer contacts of the 14 side chains seen in the Cet1 crystal structure are indicated in the middle and right columns, respectively. fects on Cet1 function in vivo (Table I), suggesting a causal relationship between the loss of phosphohydrolase activity and defective cell growth. Asp 280 Is Essential for Homodimerization-The native sizes of the wild-type and D280A proteins were investigated by zonal velocity sedimentation in a glycerol gradient. Marker proteins BSA and cytochrome c were included as internal standards. After centrifugation, the polypeptide compositions of the oddnumbered gradient fractions were analyzed by SDS-PAGE. The sedimentation profile for wild-type Cet1(201-549) is shown in Fig. 3 (top panel). The triphosphatase (44 kDa) sedimented as a discrete peak coincident with BSA (66 kDa), consistent with the wild-type enzyme being an asymmetric homodimer (4,16). The triphosphatase activity profile paralleled the distribution of the Cet1 polypeptide (Fig. 3, bottom panel).
D280A sedimented between BSA and cytochrome c, suggesting that D280A is a monomer (Fig. 3, middle panel). The triphosphatase activity profile of D280A paralleled the distribution of the Cet1 polypeptide in the gradient (Fig. 3, bottom  panel). The effects of the D280A single mutation of Cet1 quaternary structure (conversion to monomer) and Cet1 function in vivo (lethality) and in vitro (no significant effect on ATPase activity) are identical to those noted previously for the D279A-D280A cluster mutation (24). Given the present findings that the neighboring residue Asp 279 is not important for Cet1 activity, we conclude that the Asp 280 side chain per se is essential for Cet1 homodimerization and for Cet1 function in vivo. Control experiments showed that the sedimentation profiles of wildtype Cet1(201-549) and D280A were unaffected by inclusion of 0.2 mM ATP in the glycerol gradient (data not shown).
Mutational Effects on Cet1 Thermal Stability in Vitro-Sev-eral of the Cet1-Ala mutations studied here resulted in a ts growth defect in vivo (Table I). To evaluate the basis for the ts phenotype, we compared the thermal stability of wild-type Cet1(201-549) to that of the Cet1(201-549)-Ala mutants. The purified enzymes were preincubated for 10 min at 30°, 35°, 40°, 45°, or 50°C, followed by quenching on ice. The protein samples were then assayed for ATPase activity at 22°C. The data were expressed as the ratio of ATP hydrolysis by enzyme preincubated at a given test temperature to the activity of the respective unheated control. The thermal inactivation curves are plotted in Fig. 4. The activity of wild-type Cet1(201-549) was stable to preincubation at 30°C and reduced only 15% by treatment at 40°C. The activity declined sharply after heating at 45°C (to 55% of the unheated control value) and 50°C (to 10% of the control value). The I470A, L519A, F523A, L524A, and I530A proteins, which were temperature-sensitive in vivo, were clearly thermolabile in vitro (Fig. 4). The inactivation curves for I470A, F523A, L524A, and I530A were shifted ϳ15°C to the left relative to the wild-type enzyme. The L519A mutation, which elicited a less severe ts defect in vivo, shifted the thermal inactivation curve ϳ10°C to the left (Fig. 4). An instructive finding was that the D280A change, which disrupted homodimerization, rendered yeast RNA triphosphatase thermolabile in vitro. Heating for 10 min at 40°and 45°C reduced D280A phosphohydrolase activity by 82 and 88%, respectively, Thus, homodimerization enhances the stability of yeast RNA triphosphatase at physiological temperatures. Control experiments showed that the heat inactivation profiles of wild-type Cet1(201-549) and D280A were unaffected by inclusion of 0.2 mM ATP in the buffer during the preincubation step (data not shown). Deletion Analysis of S. pombe RNA Triphosphatase Identifies a Monomeric Catalytic Domain-The primary structure of S. pombe RNA triphosphatase Pct1 resembles the catalytic domains of budding yeast RNA triphosphatases Cet1 and CaCet1 across the segment extending from strand ␤1 to the C terminus (7). Reference to the crystal structure of Cet1 indicates that the essential active site residues in the ␤ strands that compose the triphosphate tunnel are strictly conserved (denoted by the symbol^in Fig. 1). Pct1 (303 aa) is considerably smaller than Cet1(549 aa) or CaCet1 (520 aa), because Cet1 and CaCet1 contain nonessential N-terminal extensions that are missing from Pct1.
To evaluate the contributions of the N-terminal segment to Pct1 structure and function, we constructed two N-terminal deletions, Pct1(41-303) [N⌬40] and Pct1(51-303) [N⌬50]. (The N termini of the N⌬40 and N⌬50 polypeptides are denoted by arrowheads below the Pct1 sequence in Fig. 1.) The N⌬40 and N⌬50 proteins were produced in bacteria as His 10 fusions and purified from soluble bacterial lysates by nickel-agarose chromatography. SDS-PAGE analysis of the imidazole eluate fractions of the full-length and truncated Pct1 proteins revealed similar extents of purification and the expected increments in electrophoretic mobility (Fig. 5A). The phosphohydrolase specific activity of N⌬40 was identical to that of wild-type Pct1, whereas N⌬50 was 61% as active as the full-length enzyme (Fig. 5B). Note that the N terminus of the catalytically active Pct1(51-303) protein corresponds to the proximal margin of the monomeric catalytic domain of S. cerevisiae Cet1.
The native sizes of the Pct1 N⌬40 and N⌬50 proteins were analyzed by glycerol gradient sedimentation (Fig. 6). Marker proteins BSA, ovalbumin, and cytochrome c were included as internal standards. Wild-type Pct1 and N⌬40 sedimented faster than ovalbumin and just slightly ahead of BSA, consistent with both proteins being homodimers. In contrast, N⌬50 sedimented between ovalbumin and cytochrome c, indicating that the deletion of the segment from amino acids 41-50 converted Pct1 into a monomeric enzyme.
Monomeric Pct1 Is Thermolabile in Vitro-The wild-type, N⌬40 and N⌬50 Pct1 preparations were treated for 10 min at either 30, 35, 40, 45, or 50°C, then quenched on ice and assayed for ATP hydrolysis at 30°C (Fig. 5C). The phosphohydrolase activities of wild-type Pct1 and the N⌬40 deletion mutant were unaffected by preheating at 30 -50°C. In marked contrast, the N⌬50 enzyme was thermolabile. Brief preincubations at 35°, 40°, and 45°C reduced enzyme activity by 42, 78, and 84%, respectively (Fig. 5C). The stability or lability of the Pct1 mutants correlated with their homodimeric versus monomeric quaternary structures. Thus, we surmise that homodimerization stabilizes S. pombe RNA triphosphatase at physiological temperatures, a feature shared with S. cerevisiae Cet1.
Homodimerization Is Important for Pct1 Function in Vivo in S. pombe-The full-length pct1 ϩ and truncated pct1N⌬ cDNAs

Fungal RNA Triphosphatases
were cloned into the S. pombe expression vectors pREP81X, pREP41X, and pREP3x (LEU2 ars1 ϩ ) so as to place them under the control of the nmt1** (low strength), nmt1* (medium strength), and nmt1 (full strength) promoters, respectively. The normalized constitutive expression levels provided by these three promoters are 1ϫ (nmt1**), 3.5ϫ (nmt1*), and 42ϫ (nmt1) (28). The expression plasmids were introduced into a heterozygous S. pombe pct1 ϩ /pct1::kanMX diploid strain in which one of the pct1 ϩ alleles was replaced with a gene conferring G418 resistance (29). The Leu ϩ diploid transformants were selected and then sporulated. Random populations of 32 individual Leu ϩ haploids were tested for G418 resistance or sensitivity. We found that half of the Leu ϩ haploids derived from a pct1 ϩ /pct1::kanMX strain containing a pREP81X plasmid with the pct1 ϩ cDNA under the control of the weakest promoter also contained the pct1::kanMX chromosomal allele and were resistant to G418 (Table II). In contrast, none of the Leu ϩ haploids obtained by sporulating the pct1 ϩ /pct1::kanMX diploid containing an empty pREP vector were resistant to G418 (29). These results show that the pct1⌬ strain is viable if the chromosomal deletion is complemented by an extrachromosomal triphosphatase gene driven by a weak promoter. Complementation was also observed when the wild-type pct1 ϩ cDNA was under the control of the intermediate-strength nmt1* promoter (29).
The pct1-N⌬40 allele complemented the pct1⌬ null mutation when expression of N⌬40 was driven by the weakest nmt1** promoter (15/32 G418 R haploids) or the intermediate-strength nmt1* promoter (16/32 G418 R haploids). In contrast, the pct1-N⌬50 mutant was unable to support growth of S. pombe when its expression was driven by the intermediate-strength nmt1* promoter (0/32 G418 R haploids) (Table II). However, the lethality of the N⌬50 mutant was suppressed by overexpression under the control of the strong nmt1 promoter (15/32 G418 R haploids) (Table II). Although the nmt1** promoter can be repressed ϳ6-fold by inclusion of 5 g/ml thiamine in the growth medium (28), we observed that the growth of the plasmiddependent nmt1**-pct1 ϩ and nmt1**-pct1-N⌬40 strains was not affected by exogenous thiamine (Fig 7B). We infer that the expression levels of the full-length Pct1 and truncated N⌬40 enzymes in these strains exceeded a threshold required for cell viability. In contrast, the growth of the nmt1-pct1-N⌬50 strain was inhibited by thiamine (Fig. 7B). These findings (Table II and Fig. 7B) provide two lines of evidence that the monomeric N⌬50 mutant, although catalytically active in vitro, is unable to perform all of the requisite functions of Pct1 in vivo.

Importance of Triphosphatase Homodimerization in Fungi-
The present study shows that a homodimeric quaternary structure is critical for the in vivo function of RNA triphosphatases in two highly divergent species of fungi: the budding yeast S. cerevisiae and the fission yeast S. pombe. The finding that homodimerization ensures the thermal stability of S. cerevisiae and S. pombe triphosphatase activity in vitro at physiological temperatures engenders an explanation for the conservation of quaternary structure, whereby the delicate tunnel architecture of the active site is stabilized by the homodimeric pedestal domain upon which it rests.
The fold of the pedestal domain of Cet1 is composed of multiple secondary structure elements and connecting loops that interdigitate across the homodimer interface ( Fig. 1A) (16). The ␤ strands that comprise the triphosphate tunnel project into the pedestal, where they or their interconnecting loops make cross dimer contacts to the partner protomer. Most of the side chains at the dimerization surface of Cet1 have now been subjected to alanine substitution, either in clusters or individually. The present alanine scan highlights the dominant contributions of Asp 280 to the dimer interface of Cet1. No other single alanine mutation tested here resulted in unconditional lethality in vivo, although single mutations of several other side chains at the dimer interface resulted in conditional phenotypes in vivo and thermolability in vitro. These effects are discussed in detail below.
In the case of fission yeast triphosphatase Pct1, we show that dimerization, thermal stability, and in vivo activity depend on the peptide segment 41 VPKIEMNFLN 50 located immediately prior to the start of the Pct1 catalytic domain. The 48 FL 49 dipeptide within this segment of Pct1 corresponds to the 272 FL 273 dipeptide of Cet1, which is an essential component of the Cet1 dimer interface, i.e. an F272A/L272A double mutation of Cet1 converts it into a catalytically active monomer that is thermosensitive in vivo (24) We suspect therefore that the homodimer interface of S. pombe Pct1 is at least partially similar to that of S. cerevisiae Cet1.
Initial speculations as to why S. cerevisiae RNA triphosphatase is a homodimer focused on a possible role for dimer-ization in the context of the heterotrimeric triphosphataseguanylyltransferase complex, whereby the guanylyltransferase Ceg1 bound to the surface peptide 247 WAQKW 251 in one protomer of Cet1 might make interactions with the other Cet1 protomer that are relevant to cap formation (24). Although not excluding this idea, the present finding that triphosphatase homodimerization is conserved and essential in S. pombe, where the triphosphatase and guanylyltransferase components do not interact physically, suggests either that: (i) homodimerization confers added value to the fungal triphosphatases in a common manner that is independent of the guanylyltransferases, or (ii) triphosphatase homodimerization is required for different reasons in budding yeast and fission yeast. For reasons of parsimony, we invoke the former model and posit, based on concordant biochemical properties of the monomeric versions of Cet1 and Pct1, that stabilization of the triphosphatase fold is the principal value added by dimerization.
This is not to say that stabilization is the only benefit of homodimerization. For example, the S. pombe triphosphatase Pct1 interacts directly with several protein components of the transcription elongation complex, including: (i) the phosphorylated CTD of S. pombe polymerase II (8); (ii) the S. pombe ortholog of transcription elongation factor Spt5 (31); and (iii) S. pombe Cdk9, a cyclin-dependent protein kinase that phosphorylates S. pombe Spt5 and the polymerase II CTD (32). Any one or several of these protein-protein interactions of Pct1 might be affected by the dissociation of the Pct1 homodimer, either because the interactions entail contacts with both protomers of the dimer or Pct1 needs to interact simultaneously in vivo with more than one of its binding partners, each one being tethered to a different Pct1 protomer. Recent studies in S. cerevisiae have imputed new functions to the Cet1 in transcriptional repression in vitro (33) and in vivo (34), and it is conceivable that Cet1 dimerization impacts on those functions. Nonetheless, it is improbable that the adverse effects of monomerization of Cet1 on S. cerevisiae cell growth can be attributed to ancillary functions other than cap formation, because the lethal cet1⌬ null mutation is rescued completely by the mammalian capping enzyme, whose triphosphatase component has no structural or mechanistic similarity whatsoever to yeast Cet1 (26,35,36).
In summary, the available evidence points to protein stabilization as a force behind the conservation of homodimeric quaternary structure among known fungal RNA triphosphatases involved in mRNA cap formation. It will be of interest to determine if nonfungal members of the tunnel family of RNA triphosphatases also rely on dimerization for stability. To date, we know that the RNA triphosphatase of the microsporidian parasite Encephalotozoon cuniculi is also a homodimeric tunnel family enzyme (37), but its dimer interface is undefined. Microsporidia are believed to be phylogenetically close to fungi. Tunnel family RNA triphosphatases are also found in protozoan parasites, including Plasmodium and Trypanosoma (38 -40), but their quaternary structures have not been determined.
Structural Interpretations of the Cet1 Mutational Effects-Cluster mutagenesis results had located two functionally important facets of the Cet1 homodimer interface (24). One of these entails hydrophobic interactions between the side chains of strand ␤2 of one protomer and strand ␤3 in the other protomer. This facet of the interface also embraces Asp 280 (in ␣1), which engages in a cross-dimer hydrogen bond with the backbone amide of Gln 329 in the turn connecting ␤2 and ␤3 (Fig. 8). We showed previously that two Ala cluster mutations involving or flanking these residues (D279A-D280A and C330A-V331A) elicited lethal phenotypes in vivo (24). Here we find that singlealanine mutations of Cys 330 , Val 331 , and Asp 279 did not elicit  15 17 significant effects on cell growth or triphosphatase activity in vitro, implying that their individual contributions to Cet1 function are either negligible or subtle at best. The single D280A mutation was lethal in vivo. Biochemical analysis of the recombinant D280A protein confirmed that its phosphohydrolase activity was intact and that the mutant protein sedimented as a monomer. Thus, we attribute the lethal in vivo phenotype of D280A to an isolated defect in homodimerization. The Asp 280 side chain emerges from this analysis a critical determinant of Cet1 homodimerization. The Asp 280 carboxylate coordinates two separate structural elements on the partner protomer: (i) O␦1 accepts a hydrogen bond from the backbone amide of Gln 329 at the top of the loop connecting the ␤2 and ␤3 strands, and (ii) O␦2 forms a bifurcated salt bridge with two arginine side chains (Arg 531 and Arg 532 ) in helix ␣4 of the partner protomer (Fig. 8).
A second functionally relevant dimer interface involves hydrophobic side-chain interactions between ␣4 and residues Phe 272 and Leu 273 in the loop preceding ␣1. We showed previously that simultaneous replacement of the Phe 272 and Leu 273 side chains by alanine resulted in a catalytically active monomeric enzyme and a severe ts growth phenotype in vivo. In the current study, we find that the single L273A mutation had no effect on activity in vivo or in vitro, whereas the single F272A mutation conferred a weak ts phenotype in vivo without affecting catalysis in vitro. Thus, neither Phe 272 nor Leu 273 per se is essential for function.
Phe 272 and Leu 273 make van der Waals contacts with Ile 470 , Ile 529 , and Ile 530 in the partner protomer. We found earlier that cluster mutants I470A-I472A and I529A-I530A were unable to grow at 25, 30, or 37°C and barely grew at 14°C (24). Here we see that the single I529A mutation had no impact on Cet1 function and the I472A change conferred only a weak ts phenotype. In contrast, the single I470A and I530A mutations resulted in tight ts growth defects that correlated with the thermolability of the I470A and I530A proteins in vitro. The thermolability of I470A is likely to result from a combination of factors, including: (i) effects on dimerization caused by the loss of contact between Ile 470 and Phe 272 on the partner protomer and (ii) effects of the I470A mutation on the conformation of the ␤11 strand caused by loss of the intra-protomer contact between Ile 470 and Leu 495 (see below).
Similarly, the I530A temperature sensitivity may also be multifactorial, insofar as the Ile 530 side chain, which contacts Phe 272 in the partner protomer, makes additional intramolecular contacts within the hydrophobic core of the pedestal (involving Phe 310 in ␤1 and Val 493 and Leu 495 in ␤11) that support the floor of the triphosphate tunnel (Table I). Neighboring residues Glu 305 and Glu 307 in ␤1 and Glu 494 and Glu 496 in ␤11 bind the essential metal cofactor and are required for catalysis FIG. 7. Effects of N-terminal deletions on Pct1 function in vivo. pct1⌬ haploid strains containing plasmid-borne nmt1**-pct1 ϩ (WT), nmt1**-pct1-N⌬40, or nmt1-pct1-N⌬50 alleles were patched on EMM(-Leu) agar medium and then streaked: A, on YES agar medium at 30°C (top) or 37°C (bottom); B, at 30°C on EMM(-Leu) agar medium containing no thiamine (top) or 5 g/ml thiamine (bottom). The plates were photographed after 5 days. Fungal RNA Triphosphatases (16). Because the I530A phenotype is more severe than the F272A mutation in its cross-dimer contact (Table I), we surmise that the Ile 530 change may directly destabilize the structure of the tunnel. This hypothesis is consistent with the previous findings that a single alanine mutation of Phe 310 in ␤1 was lethal and single alanine mutations of Val 493 or Leu 495 in ␤11 (the other residues contacted intramolecularly by Ile 530 ) resulted in ts growth defects at 37°C and thermolability of Cet1 triphosphatase activity in vitro (17).
The L519A mutation resulted in a weak ts growth phenotype and had no effect on catalysis in vitro, although the L519A protein was sensitized to thermal inactivation. Although Leu 519 makes a cross-dimer van der Waals contact with Ile 268 , we suspect that the ts defects do not reflect the simple loss of this contact, because mutation of Ile 268 to alanine had no effect on growth (26). Leu 519 makes other intramolecular contacts to Cys 467 , Ile 497 , and Leu 502 that tether the floor of the tunnel to the pedestal.
Three of the single alanine mutations studied here resulted in lethal (I520A) or tight ts (F523A and L524A) growth phenotypes that correlated with major defects in catalysis of ␥ phosphate hydrolysis. These phenotypes are likely caused by effects on active site architecture rather than quaternary structure, because Ile 520 , Phe 523 , and Leu 534 are oriented toward the hydrophobic core of the Cet1 protomer, and they make no contributions to the crystallographic dimer interface, although they make extensive contacts within the Cet1 protomer (Table  I). Ile 520 makes van der Waals contacts to Val 285 and Val 289 in helix ␣1 and also to Tyr 516 and Leu 524 in helix ␣4. This network of hydrophobic interactions stabilizes the helix packing within the pedestal, and its disruption by mutation would likely have global effects on Cet1 folding. Leu 524 interacts with Val 285 , Ile 428 , and Ile 520 . Phe 523 is situated in a hydrophobic-aromatichydrophobic sandwich between Met 308 (in ␤1), Ile 428 (in ␤8), and Ile 497 (in ␤11) that imparts stability to the floor of the tunnel.