Proc. Nail. Acad. Sci. USAVol. 86, pp. 6503-6507, September 1989Biochemistry

Isolation and expression of the Pneumocystis carinii thymidylatesynthase geneURSULA EDMAN*, JEFFREY C. EDMANt, BETTINA LUNDGRENt, AND DANIEL V. SANTIt§*Intercampus Program for Molecular Parasitology and Departments of tLaboratory Medicine, §Biochemistry and Biophysics, and §Pharmaceutical Chemistry,University of California, San Francisco, CA 94143; and tCritical Care Medicine Department, National Institutes of Health, Bethesda, MD 20892

Communicated by Thomas C. Bruice, May 22, 1989

ABSTRACT The thymidylate synthase (TS) gene fromPneumocystis carin has been isolated from complementaryand genomic DNA libraries and expressed in Escherichia coli.The coding sequence of TS is 891 nucleotides, encoding a297-amino acid protein ofMr 34,269. The deduced amino acidsequence is similar to TS from other organisms and is mostclosely related to the enzyme from Saccharomyces cerevisiaewith 65% identity. TS is found on a 330-kilobase-pair chro-mosome in P. carnii. While TS and dihydrofolate reductasereside on a single polypeptide chain in all protozoa studied todate, TS is not linked to dihydrofolate reductase in P. carinii.The TS gene shows the presence of four small interveningsequences, some of which interrupt the coding sequence inhighly ordered structural regions of the protein. Heterologousexpression of P. carinii TS in E. coli was accomplished bycloning the coding sequence into plasmid vectors under controlof the lac and tac promoters. These constructs direct thesynthesis of catalytically active enzyme to the extent of 2% oftotal soluble protein.

Pneumocystis carinii, an organism originally thought to be aprotozoan but reclassified as a fungus (1), causes disease onlyin the immunocompromised host (2, 3). In developed nations,P. carinji pneumonia (PCP) is the most common lethalinfection of patients with acquired immunodeficiency syn-drome (AIDS). Although PCP responds to treatment withtrimethoprim-sulfamethoxazole or pentamidine, these drugsfrequently cause allergic and toxic side effects when used inAIDS patients. There is, therefore, a need for other effectivetherapies for the treatment of PCP.The development of therapeutic agents for PCP has been

hampered by the lack of an adequate in vitro culture systemof P. carinii. The only model system available for study ofPCP is the steroid-induced immunosuppressed rat (4). In theabsence of continuous in vitro cultures, there is currently nosource of P. carinii that can provide sufficient material forbiochemical studies. In fact, no protein from this organismhas been purified. However, DNA libraries have been pre-pared from rat P. carinii (1) allowing the isolation andheterologous expression of genes encoding proteins that arepotential targets for drug development.We have an ongoing interest in studies of the structure,

function, and inhibition of thymidylate synthase (TS). Thisenzyme catalyzes the reductive methylation of dUMP todTMP with the concomitant conversion of 5,10-methylene-tetrahydrofolate (CH2H4folate) to 7,8-dihydrofolic acid.Since TS is the sole de novo source for dTMP synthesis, it isan attractive target for the development of chemotherapeuticagents. Furthermore, the demonstrated effectiveness of an-tifolate therapy in the treatment ofP. carinii infections showsthat thymine depletion is cytotoxic to these organisms. Basedon the large body of information available for TS, inhibitors

may be developed that may specifically affect P. carinii. Inthe present work, we describe the isolation and expression ofthe TS gene from P. carinii.¶

MATERIALS AND METHODSReagents. The following oligonucleotides were synthesized

in the Biomolecular Resource Center at the University ofCalifornia, San Francisco, CA: ONi, 5'-GGAATTCCRTN-TAYGGNTTYCARTGG-3'; ON2, 5'-GGGATCCCGCNA-TRTTRAANGGNAC-3'; ON4, 5'-GGAGCAGAATATAT-TGATTGC-3'; ON5, 5'-GCAATCAATATATTCTGCTCC-3'; ON6, 5'-GGCATTACCTCCGTGCCATATG-3'; ON7, 5'-CATATGGCACGGAGGTAATGCC-3'; ON8, 5'-GGGCG-GTCTTCCCCATGG-3'; ON9, 5'-AGCTTGAATTCAGGA-GGTAATTAACCATGGTAAACGCAGAAGAACAACAA-TATCTCAACTTAGTACAATATATTATTAAT-3'; ON10,5'-CATGATTAATAATATATTGTACTAAGTTGAGATA-TTGTTGTTCTTCTGCGTTTACCATGGTTAATTA-CCTCCTGAATTCA-3'; ON11,5'-CATGGAATCCTGCAG-ATTTAGAAAAA-3' (where R is a purine, N is any nucle-otide, and Y is a pyrimidine). AZAP was obtained fromStratagene, GeneClean from BIO101, the Random PrimedDNA Labeling Kit from Boehringer Mannheim, and theSequenase Kit from United States Biochemicals. All otherenzymes were obtained from New England Biolabs, Saccha-romyces cerevisiae chromosome blocks were from Clonetek,nitrocellulose membranes were from Schleicher & Schull.EN3HANCE was purchased from DuPont, [a-32P]dCTP,deoxyadenosine 5'-[a- 5S]thio]triphosphate, and nylon fi-ters from Amersham, and 5-fluoro-2'-deoxy[6-3H]uridine 5'-monophosphate ([6_3H]FdUMP) was from Moravek Bio-chemicals (Brea, CA).Polymerase Chain Reaction (PCR). The PCR was per-

formed using a Cetus/Perkin-Elmer DNA thermocycler.Reaction mixtures (50 Al) contained 25 pmol of each of thetwo oligonucleotide primer mixtures ON 1 and ON 2, 500 ngof P. carinji genomic DNA, all four dNTPs, each at 400 ,M,60mM KCl, 25 mM Tris HCl (pH 8), 10mM MgCl2, and 0.1%bovine serum albumin. The reaction mixture was overlaidwith a drop ofparaffin oil and denatured at 940C for 5 min, andamplification was initiated by addition of2.5 units of Thermusaquaticus DNA polymerase. PCR parameters were 36 ther-mal cycles consisting of a 30-sec denaturation at 940C fol-lowed by a 25-sec annealing period at 370C and a 1-minextension period at 700C. After analysis of the PCR producton a 5% polyacrylamide gel, the amplification product waspurified on a 2% low-melting-point agarose gel. The DNAwas digested in agarose with EcoRI and BamHI, purified by

Abbreviations: TS, thymidylate synthase; TAFE, transverse alternat-ing-field electrophoresis; CH2H4folate, 5,10-methylenetetrahydrofo-late; [6-3H]FdUMP, 5-fluoro-2'-deoxy[6-3H]uridine 5'-monophos-phate; IVS, intervening sequence; PCP, Pneumocystis carinii pneu-monia; PCR, polymerase chain reaction.IThe sequence reported in this paper has been deposited in theGenBank data base (accession nos. M25415 and M25416).

6503

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

July

26,

202

0

Proc. Natl. Acad. Sci. USA 86 (1989)

phenol extraction, and ligated into M13mpl8 and -mpl9.Several templates from each ligation were sequenced.

Transverse Alternating-Field Electrophoresis (TAFE). Low-melting-point agarose blocks from three isolates of P. cariniiwere prepared as described (5). Blocks from each isolate andfrom S. cerevisiae were loaded onto a 0.8% agarose gel andchromosomes were separated by TAFE on a Gene-Linesystem (Beckman) with the following program parameters:stage 1, 30 min at 170 mA with a pulse time of 4 sec; stage 2,15 hr at 150 mA with a pulse time of 25 sec; stage 3, 7 hr at170 mA with a pulse time of 35 sec. The gel was transferredto a Hybond-N nylon filter and subjected to Southern blotanalysis (6).

Probe Preparation and Hybridization Conditions. Probefragments were purified by low-melting-point agarose gelelectrophoresis and the GeneClean method. The fragmentswere labeled with [a-32P]dCTP using the random-primingmethod. Hybridizations were performed in 50%o (vol/vol)formamide/Sx SSC/10 mM NaH2PO4/0.05% Na4P2O7/5XDenhardt's solution/0.1% SDS at 420C by addition of probeat 106 cpm/ml (lx SSC = 0.15 M NaCl/0.015 M sodiumcitrate, pH 7.0; lx Denhardt's = 0.02% polyvinylpyrroli-done/0.02% Ficoll/0.02% bovine serum albumin). The filterswere washed in 0.1x SSC/0.1% SDS/0.05% Na4P207 at42-50°C and subjected to autoradiography using KodakXAR-5 x-ray film.

Preparation of cDNA Library. The cDNA library wasprepared in AZAP as described (7, 8).

Screening of Complementary and Genomic DNA Libraries.Approximately, 120,000 phage from each of the genomicDNA and cDNA libraries in AZAP were plated to BB4 cellson NZCYM agar (6) and transferred to nitrocellulose filters.Hybridization and wash conditions were as described above.All isolates were plaque-purified and plasmids were rescuedfrom AZAP as described (9). Plasmid DNA from five genomicclones and one cDNA clone was mapped with EcoRI,BamHI, HindIII, Pst I, and Nco I, transferred to nitrocellu-lose, and probed with the cloned PCR product. EcoRI,EcoRI-Pst I, or BamHI-HindIII fragments from genomicclones pUETS-2 and pUETS-6 and cDNA clone pUETS-1were subcloned into M13mpl8 and -mpl9 for sequenceanalysis with the Sequenase system using the universalprimer as well as four oligonucleotide primers (ON4-7) basedon the PCR product sequence. One additional primer, ON8,was required to obtain the sequence of the 5' flanking region.With the exception of the 5' flanking region, all sequencesshown in Fig. 3 were determined on both strands for thecDNA and both genomic DNA clones.

Expression of P. carinfi TS in Escherichia coli. To assessexpression of functional TS by complementation, plasmidconstructs were introduced into E. coli K12 X2913 (Athy-A572) (R. Thompson, University of Glasgow, U.K.) andplated on minimal agar lacking thymidine. Aliquots (2 ml) ofminimal medium lacking thymidine were inoculated withsingle colonies, and cultures were grown to stationary phase.These served as inocula for 20-ml LB cultures containingampicillin (50 ,ug/ml) and 1 mM isopropyl (3-D-thiogalacto-pyranoside. Cells were harvested by centrifugation 12 hrlater. Pellets were suspended in 2 ml of 100 mM Tris-HCl,pH7.4/1 mM EDTA/10 mM 2-mercaptoethanol at 4°C, and cellswere disrupted by sonication. Cell debris was removed bycentrifugation at 10,000 x g for 15 min at 4°C. Proteinconcentrations were determined by using the method ofBradford (10). Total cell lysates and soluble extracts fromcomplementing (Thy') transformants were analyzed bySDS/PAGE and TS activity was monitored spectrophoto-metrically using conditions described (11). TS concentrationwas determined by measuring the formation of the covalentcomplex of [63H]FdUMP, CH2H4folate, and enzyme (12).Soluble extracts containing protein (3 mg/ml) were incubated

for 15 min at 250C with 0.1 mM CH2H4folate and 0.1 gM[6-3H]FdUMP (16 Ci/mmol; 1 Ci = 37 GBq) in 50 mMTES/25 mM MgCl2/6.5 mM formaldehyde/l mM EDTA/75mM 2-mercaptoethanol (where TES is N-tris(hydroxymeth-yl)methyl-2-aminoethanesulfonic acid). The protein-bound[6-3H]FdUMP was isolated on nitrocellulose filters (13) orvisualized by subjecting 5 ,ug of total soluble protein toSDS/PAGE on 12% polyacrylamide gels followed by fluo-rography and autoradiography for 12 hr.

RESULTSIsolation of the P. cannii TS Gene. Degenerate oligonucle-

otide primers based on conserved regions of the TS aminoacid sequence were used for PCR amplification of P. carinjiTS. Restriction sites were added to the 5' end of eacholigonucleotide mixture to facilitate cloning of the PCRproducts. The 5' oligonucleotide mixture (ON1, 5'-RTN TAYGGN TTY CAR TGG-3') was based on residues 145-150(lle/Val-Tyr-Gly-Phe-Gln-Trp) relative to the Lactobacilluscasei sequence; the tyrosine, glycine, and glutamine residuesare invariant in all known TS sequences; Phe-148 is found inall eukaryote TS sequences, and Val/lle-145 and Trp-150 arefound in all except TS from phage 43T. At the 5' end of thisprimer, we placed the sequence 5'-GGGAATTCC, whichcontains an EcoRl site. The 3' oligonucleotide mixture (ON2,5'-GC NAT RTT RAA NGG NAC-3') was designed fromamino acid residues 226 to 231 (Val-Pro-Phe-Asn-lle-Ala) ofthe L. casei (14) sequence. Proline, phenylalanine, andasparagine residues are conserved in all known TS se-quences, and Val-226, Ile-230, and Ala-231 are found in alleukaryotic sequences. At the 5' end of this antisense strandprimer, we placed the sequence GGGATCCC, which con-tains a BamHI site.A single 403-base-pair (bp) product was amplified from

genomic P. carinji DNA (Fig. 1, lane 3). Amplification ofgenomic S. cerevisiae DNA (Fig. 1, lane 2) using the sameprimers yielded a major product of 311 bp corresponding withthe predicted fragment size for TS from this source (15). Noamplification products were obtained for a negative controlplasmid (Fig. 1, lane 4) or normal rat lung DNA (data notshown). The size of the P. carinii PCR product was 100 bplarger than that expected from known TS sequences, whichwas later shown to be due to the presence of two intervening

1 2 3 4

-1353- 872

- 603

-310

- 234

- 194

- 118

FIG. 1. PCR using oligonucleotides ON1 and ON2. Lanes: 1, HaeIII-digested OX molecular size markers (in bp); 2, S. cerevisiaegenomic DNA (311-bp PCR product is indicated by >); 3, rat P.carinji genomic DNA (403-bp PCR product is indicated by >); 4,negative control plasmid.

6504 Biochemistry: Edman et al.

Dow

nloa

ded

by g

uest

on

July

26,

202

0

Biochemistry: Edman et al.

A B1 2 3 1 2 3

600 -

460 -

370 -

290 -

FIG. 2. (A) Ethidium bromide stain ofchromosomes fractionatedon 0.8% agarose gel by TAFE. Lanes: 1, S. cerevisiae; 2, rat P.carinji isolate PcIII; 3, rat P. carinji isolate PcH. Molecular sizes aregiven in kbp. (B) Autoradiograph of nylon transfer of gel shown in Aprobed with 32P-labeled pUETSpcr.

sequences. The PCR product was cleaved at the EcoRI andBamHI restriction sites contained in the oligonucleotide

Proc. Natl. Acad. Sci. USA 86 (1989) 6505

primers and cloned into M13mpl8 (pUETSpcr) for sequenceanalysis and used as a hybridization probe. DNA sequencingrevealed a'predicted amino acid sequence corresponding tothat region of TS flanked by the primers used.To confirm that the PCR product was derived from P.

carinii, pUETSpcr was used to probe a Southern blot ofTAFE-fractionated chromosomes from three rat P. carinjiisolates and from S. cerevisiae. Specific hybridization to a330-kbp chromosome of P. carinii was seen and no hybrid-ization to S. cerevisiae chromosomes was detectable (Fig. 2).The cloned 403-bpPCR probe was used to screen 4120,000

plaques each from cDNA and genomic DNA libraries. OnecDNA clone (pUETS-1) containing an insert of 1.2 kilobases(kb) and five genomic clones -(pUETS-2 to 6) containinginserts of 2.8-4.1 kb were 'isolated. Restriction enzymemapping and Southern blot analysis using pUETSpcr as aprobe established that all genomic DNA clones overlappedone another and carried the 5' portion of the TS gene on a1.8-kb BamHI-HindIII fragment and the 3' portion of thegene on a 1.3-kb HindHI fragment (data not shown). Thenucleotide sequence ofcoding and flanking regions presentedin Fig. 3 was derived from cDNA clone pUETS-1 and thegenomic DNA clones pUETS-2 and -6.

In sequencing the'coding region ofcDNA clone pUETS-1,there was a compression at position 532 that could not beresolved. Comparison to the genomic sequence suggested a

-413aaagacgataccaattcatctatgctatcatcttgctcattaaaactagaatcacaatcaattgcgtgttttcgttctt

ctaactacacagacgtgatacaagagaaactgaagagatcaatcaaaaaaacccgtcataaatattactttatttttat

tacaattttaaagttctgaatttgatcgaatatcaa*tttttgcaataattaataagttttttttttttaaaaaaaaac

ttatttttcagaaagtcctaaataatactttcttttttccccaaaactttattatttataaagcaaatgaaaaaaattg

-335

-256

-177

-99

-201 10

Met Val Asn Ala Glu Glu Gln Gln Tyr Leu Asn Leu Val Gln Tyrttacttaatactttaaaac ATG GTA AAC GCA GAA GAA CAA CAA TAT CTC AAC TTA GTA CAA TAT +45

20 30Ile Ile Asn His Gly Glu Asp Arg Pro Asp Arg Thr Gly Thr Gly Thr Leu Ser Val PheATT ATT AAC CAT GGG GAA GAC CGC CCT GAT AGG ACA GGG ACA GGG ACT TTA TCT GTT TTT +105

40 50Ala Pro Ser Pro Leu Lys Phe Ser Leu Arg Asn Lys Thr Phe Pro Leu Leu Thr Thr LysGCA CCG TCT CCG CTA AAG TTT TCT CTT CGA AAT AAA ACA TTT CCA CTC TTA ACA ACC AAG +165

60Arg Val Phe Ile Arg Gly Val Ile Glu Glu Leu Leu IVS ICGA GTT TTT ATA CGT GGA GTA ATT GAA GAA TTG CTT T gtaaaaaatatctgaaacttatagatatg +231

70 80Trp Phe Ile Arg Gly Glu Thr Asp Ser Leu Lys Leu Arg Glu Lys Asn

ctcttaatgcatatag GG TTT ATT CGT GGA GAA ACA GAT TCT TTA AAA CTT AGA GAG AAA AAT +29490 100

Ile His Ile Trp Asp Ala Asn Gly Ser Arg Glu Tyr Leu Asp Ser Ile Gly Leu Thr LysATT CAT ATA TGG GAT GCA AAT GGA TCA CGA GAG TAC CTT GAC TCA ATT GGA TTG ACT AAA +354

110 120Arg GIn Glu Gly Asp Leu Gly Pro Ile Tyr Gly Phe Gin Trp Arg His Phe Gly Ala GluCGC CAG GAA GGT GAT CTT GGT CCC ATT TAT GGG TTT CAA TGG AGG CAT TTT GGA GCA GAA +414

130 140Tyr Ile Asp Cys Lys Thr Asn Tyr Ile Gly Gln Gly Val Asp Gln Leu Ala Asn Ile IleTAT ATT GAT TGC AAA ACC AAT TAT ATT GGA CAA GGT GTT GAT CAA TTA GCC AAT ATT ATC +474

150 160Gln Lys Ile Arg Thr Ser Pro Tyr Asp Arg Arg Leu Ile Leu Ser Ala Trp Asn Pro AlaCAA AAA ATA CGA ACA TCA CCA TAT GAT CGT CGA CTC ATA CTT TCA GCA TGG AAT CCT GCA +534

170IVS II Asp Leu Glu Lys Met Ala Leu Pro

G gcatactcttttatcaattcaatatcattgcttattactatttag AT TTA GAA AAA ATG GCA TTA CCT +603

Pro Cys His Met Phe Cys IVS IIICCG TGC CAT ATG TTT TG gtaatattcatatttaacctatatatgaaaccataaagcttacaaccacatatag +675

180 190Gln Phe Tyr Val His Ile Pro Ser Asn Asn His Arg Pro Glu Leu Ser Cys Gln Leu

T CAG TTT TAT GTT CAT ATA CCA TCG AAT AAC CAC CGA CCT GAA CTC TCA TGT CAA TTA +733200 210

Tyr Gln Arg Ser Cys Asp Het Gly Leu Gly Val Pro Phe Asn Ile Ala Ser Tyr Ala LeuTAC CAA CGT TCT TGT GAC ATG GGG CTA GGT GTT CCC TTT AAT ATT GCA TCG TAT GCT CTA +793

220 230Leu Thr Cys Met Ile Ala His Val Cys Asp Leu Asp Pro Gly Asp Phe Ile His Val MetCTA ACA TGC ATG ATT GCA CAT GTT TGT OAT CTT GAT CCA GGT GAT TTT ATC CAT GTC ATG +853

240IVS IV Gly Asp Cys His Ile Tyr Lys

G gtaaacactatttttttttttaaaaaaagcaaattttaattatctgtag GT GAT TGT CAT ATT TAT AAG +923250 260

Asp His Ile Glu Ala Leu Gln Gln Gln Leu Thr Arg Ser Pro Arg Pro Phe Pro Thr LeuGAT CAT ATT GAA GCA CTA CAA CAA CAA CTT ACA CGC TCC CCA CGA CCT TTT CCC ACT CTT +983

270 280FI.3 NuloiesqecofPcaiiTSgnSer Leu Asn Arg Ser Ile Thr Asp Ile Glu Asp Phe Thr Leu Asp Asp Phe Asn Ile Gln FIG. 3. Nucleotide sequence of P. carinii TS geneTCT TTA AAT CGC TCA ATC ACT GAT ATT GAA GAT TTT ACA TTA GAT GAT TTT AAT ATT CAA +1043 with 5' and 3' flanking regions and inferred amino acid

290 'oni290 ~~~~~~~~~~~~sequence.Flanking regions and IVSs are shoniAsn Tyr His Pro Tyr Glu Thr Ile Lys Met Lys Met Ser Ile OPAAC TAT CAT CCA TAT GAA ACT ATA AAA ATG AAA ATG TCC ATT TGA aaaatatatttacattata +1107 lowercase; the coding region is in uppercase. Numbering

of nucleotide sequence begins at the start codon. The 5'actttttctgtacactctactagattcacttttaaaaatatcatgttaaattaataacaacttatagaatatataactg +1186 and 3' ends of cDNA clone pUETS-1 are indicated by

atctatacataatt*aagcatttgaacaattggcacattataacg +1230 asterisks.

Dow

nloa

ded

by g

uest

on

July

26,

202

0

Proc. Natl. Acad. Sci. USA 86 (1989)

deletion of 4 bp in this region of the cDNA. To clarify thesequence of the mature cDNA, the mixed PCR primers wereused to amplify oligo(dT)-primed P. carinii cDNA. A 311-bpproduct corresponding to the predicted size ofmature mRNAwas obtained. Sequence analysis of the cloned PCR productclearly showed the presence ofGCAG at nucleotide positions532-535, which confirmed the deletion in pUETS-1, preciselyat the exon-intron junction. This sequence was also inagreement with that ofboth genomic clones, which shows theGAT codon for Asp-164 split after the first base by interven-ing sequence (IVS) II. The putative deletion in pUETS-1 wascorrected by site-directed mutagenesis (see below).The cDNA sequence contained an open reading frame of

1086 bp (Fig. 3). No other in-frame ATGs in the nucleotidesequence were present 5' to the predicted initiation codon.The first in-frame ATG codon downstream from the pre-dicted initiation codon occurs well into the highly conservedcoding sequence (amino acid 168). Also, the predicted initi-ation codon is preceded by an adenosine at position -3 andfollowed by a guanosine at position +4, an environmentfavorable for translation initiation (16). The cDNA clone alsocontained 143 bp of 5' flanking sequence and 114 bp of 3'flanking sequence preceding a short poly(A)tail (Fig. 3).

Translation of the open reading frame yields a polypeptideof 297 amino acids with a molecular weight of 34,269. Thededuced amino acid sequence shows conservation of all theamino acid residues that are invariant in known TSs. Overallamino acid identities range from 40% with T4 phage TS (17),to 60% with human TS (18), to 65% with S. cerevisiae TS (15).The coding sequence has an A+T content of64% and 78% Aor T in the third position of codons.IVSs in the TS Gene. The nucleotide sequence of genomic

DNA clones pUETS-2 and pUETS-6 was identical over thecoding and flanking regions presented in Fig. 3. Four IVSs atnucleotides 203-247 (IVS I, 45 bp), 536-580 (IVS II, 45 bp),621-675 (IVS III, 55 bp), and 855-903 (IVS IV, 49 bp) wereidentified by comparison to the cDNA clone pUETS-1. Fromthe available data, the hexanucleotide G Y AWW M, whereW is A or T and M is A or C, serves as a 5' splice junctionconsensus sequence for P. carinii IVSs. With exception ofthe first three nucleotides, the 5' splicejunction sequences donot correspond to, and are less well conserved than, thoseidentified in other fungi (19). No conserved branch sitesequences similar to the TACTAAC consensus sequence ofS. cerevisiae IVSs or the more variable C T R A Y consensussequence of Schizosaccharomyces pombe IVSs are identifi-able in these IVSs (20, 21). The conserved TAG sequencecharacteristic ofmost 3' splice junctions, however, is presentin all four P. carinii TS IVSs.

Expression of P. carini TS. Three plasmids were con-structed for the expression of P. carindi TS in E. coli. TheEcoRI fragment of pUETS-1 was ligated into pUC9 to yieldpUETS-1.1. Two complementary oligonucleotides (ON9 andON10) were prepared as an adapter that contains 5' HindIIIand EcoRI restriction sites followed by a Shine-Dalgarnosequence, stop codons in all three reading frames, the startcodon, the nucleotides encoding the first 18 amino acids ofP.carindi TS and a 3' Nco I overhang. Two silent sequencemodifications were made in the coding sequence: (t) An NcoI site was introduced at the start codon to allow transfer ofthegene into other expression vectors, and (ii) the first base ofthe natural Nco I site at position +54 was changed fromcytidine to thymidine to remove this restriction site. Neitherchange alters the amino acid sequence. These adapters werephosphorylated, annealed, and ligated into HindIII and NcoI sites of pUETS-1.1 from which the HindIII-Nco I fragmentcontaining the 5' flanking and coding region had been re-moved to give pUETS-1.2. The HindIII-BamHI fragment ofpUETS-1.2 was subcloned into M13mpl8. The putativedeletion at position 535 was corrected by oligonucleotide-

directed mutagenesis using ON9 and the entire coding se-quence was confirmed by nucleotide sequence analysis. TheEcoRI fragment was then isolated from this subclone andligated into pUC9 (pUETS-1.6) and pUC18 (pUETS-1.7) forexpression of the P. carinii TS under control of the lacpromoter. It was also inserted into M13mpl9 (pUETS-1.5) toattain desired restriction sites from the vector. An NcoI-HindIII fragment was isolated from pUETS-1.5 and sub-cloned into pKK233-2 to yield pUETS-1.8 for expressionunder the control of the tac promoter (22).

Plasmids pUETS-1.6, -1.7, and -1.8 complemented Thy-E. coli X2913 on minimal medium that lacked thymidine.Soluble extracts from cells harboring either plasmid showedTS activity (0.03 unit-min-1m-1), which was completelyinhibited by FdUMP. Using [6 H]FdUMP as an active-sitetitrant in a filter-binding assay (13), soluble extracts showed485 pmol of bound nucleotide per mg of protein in thepresence of CH2H4folate. Assuming 100% filtration effi-ciency, and a molecular weight of 34,000, this indicates that1.7% of the protein is TS and the specific activity of the pureenzyme is about 1.7 units mg-1 min-1, which is reasonablyclose to that of L. casei TS (3 units-mg 1'min-') (23, 24).SDS/PAGE analysis of soluble extracts revealed the pres-ence of a Mr 32,000 protein that represents -2% of the totalsoluble protein; the band is absent in extracts from cellsharboring vector alone. When an extract was treated with[6_3H]FdUMP and CH2H4folate prior to SDS/PAGE, thelabel comigrated with the M, 32,000 protein, verifying that itis TS (Fig. 4). The discrepancy between the predicted subunitmolecular weight of P. carinii TS and the apparent molecularweight by SDS/PAGE is also seen in S. cerevisiae TS (25).

DISCUSSIONOur initial attempts to isolate the TS gene from P. carinii byscreening the genomic library using a variety of heterologoushybridization probes failed. As probes, we tried the TS genesfrom Leishmania major (26, 27), Plasmodium falciparum(28), and Herpesvirus saimiri (29), as well as two degenerateoligonucleotides based on conserved sequences of TS; thehybridization conditions were also varied to such low strin-gency that numerous false positives were uncovered. Toobtain a suitable hybridization probe, we decided to use256-fold degenerate oligonucleotides as PCR primers to am-plify a fragment of the TS gene from P. carinii genomic DNA.The strategy used was to choose two highly conservedregions of eukaryotic TSs and to prepare a mixture ofsynthetic oligonucleotide primers that contains all possiblesequences of these regions. By using this approach, we wereable to amplify a 403-bp sequence of P. carinii TS. The aminoacid sequence deduced from this sequence showed signifi-

1 2 3 4 5

97.466.2

427-_

31.0 _

21.5 -

14.4 -

FIG. 4. SDS/PAGE of 5 ,ug of soluble protein extract from x2913containing pUETS-1.7 (lanes 1 and 3), pUETS-1.8 (lanes 2 and 4),and x2913 without plasmid (lane 5) incubated with [6-3H]FdUMP inthe presence of CH2H4folate. Lanes 1 and 2 represent autoradio-graph of the fluorographed gel shown in lanes 3, 4, and 5; molecularweights (x 1o-3) are given.

6506 Biochemistry: Edman et al.

Dow

nloa

ded

by g

uest

on

July

26,

202

0

Proc. Nati. Acad. Sci. USA 86 (1989) 6507

cant similarity to residues 145-231 of the L. casei TS. ByTAFE and Southern hybridization, we showed that the PCRproduct hybridized to a single P. carinii chromosome of -330kbp, which argues that the probe originated from P. carinii.The sequence of TS genomic DNA clones showed four

small IVSs 45-55 bp long. Since, to our knowledge, these arethe first IVSs reported for P. carinji, several features areworthy of note. (i) The 5' splice junctions are not highlyconserved, although a consensus sequence of GYA WWYwas apparent. (it) Except for the GYA as the first three bases(GTA in most eukaryotes), there is no sequence conservationof the 5' splice junction with junctions reported for IVSs inother eukaryotes; in particular, there is no consensus with 5'junctions of S. cerevisiae, Schizosaccharomyces pombe, orNeurospora crassa, which are presumed to be related to P.carinji (1, 19). (iii) A branch site sequence is not discernibleby comparison to other IVSs. (iv) The conserved 3' splicejunctionYAG was present in all four IVSs ofP. cariniiTS (forreview, see ref. 30). (v) Finally, in contrast to the mouse TSgene (31), the location ofIVSs for the P. carinji TS gene is notrestricted to positions external to functional domains of theprotein. Although IVS II interrupts TS on the surface ofexposed a-helix I and coincides exactly with that of the fourthIVS of the mouse TS gene, IVS I, III, and IV disruptsequences that encode secondary structural elements impor-tant for maintaining core structure or directly contributing tothe active site of the enzyme (32).

E. coli expression vectors containing the P. carinii TScoding sequence complemented Thy- E. coli, and crudeextracts harboring such plasmids contained high TS activity.SDS/PAGE and active-site titration with FdUMP andCH2H4folate showed that expressed TS represents some 2%of the total soluble protein (Fig. 4).

P. carinji has long been classified as a protozoan but smallsubunit rRNA gene sequence analysis has indicated that it isa fungus (1). In protozoa, TS and dihydrofolate reductaseexist as a bifunctional protein on the same polypeptide chain,whereas in other organisms, perhaps with exception of plants(33), the enzymes are clearly distinct and monofunctional.Limited nucleotide sequence analysis of the flanking se-quences ofP. carinii TS did not reveal proximal open readingframes that could encode dihydrofolate reductase. Further,TAFE demonstrated that TS and dihydrofolate reductasereside on different P. carinji chromosomes (35). Thus, ourresults support the contention that P. carinii is not a proto-zoan.

P. carinii TS contains all of the conserved amino acidresidues contributing to important structural elements iden-tified in the x-ray structure of L. casei TS and is, therefore,likely to have a similar core three-dimensional structure (32).The amino acid sequence shows both of the invariant eu-karyote-specific insertions; a 12-amino acid loop at position90 (between a-D and E helices) and the 8-amino acid loopafter residue 157 (between a-G and H helices) of the L. caseiTS. The single-amino acid insertion after residue 34 has thusfar been observed only in the other fungal TSs characterized(serine in P. carinii, proline in S. cerevisiae and Candidaalbicans) (15, 34). This insertion is found in the first 3-strandof TS, which forms an important interface for subunit con-tact. Another single-amino acid insertion after residue 43 isalso present in fungal and kinetoplastid TSs. The insertion islocated on a loop that would connect the first /-strand i to thea-helix B. A 6-amino acid insertion after residue 210 formspart of an exposed surface loop between -3-strands iii and iv.Insertions at this position of TS have thus far only beewseenas a 9-amino acid insertion in S. cerevisiae TS, a 17-aminoacid insertion in C. albicans TS, and a 2-amino acid insertionin kinetoplastid TSs. Additional pertinent differences be-tween human and P. carinji TS in nonconserved residues

contributing to the active site may provide the basis formolecular design of selective inhibitors of P. carinii TS.

In summary, we have isolated, characterized, and ex-pressed the P. carinji TS gene in E. coli. Abundant catalyt-ically active P. carinii TS is now available to pursue studieson the structure, function, and inhibition of this enzyme.

We thank Dr. S. Climie for helpful comments on the manuscript.This work was supported by U.S. Public Health Service GrantsA119358 and CA19358 from the National Cancer Institute and by fundsprovided by the State of California and allocated on the recommen-dation of the Universitywide Task Force on AIDS (to J.C.E.). U.E.was supported by the John T. and Catherine D. MacArthur Founda-tion Consortium on the Biology of Parasitic Diseases. B.L. wassupported by the Danish Medical Research Council.

1. Edman, J., Kovacs, J., Masur, H., Santi, D., Elwood, H. & Sogin,M. (1988) Nature (London) 334, 519-522.

2. Mills, J. (1986) Rev. Infect. Dis. 8, 1001-1011.3. Kovacs, J., Hiemenz, J., Macher, A., Stover, D., Murray, H.,

Shelhammer, J., Lane, H., Urmacher, C., Honig, C., Longo, D.,Parker, M., Natanson, C., Parillo, J., Fauci, A., Pizzo, P. & Masur,H. (1984) Ann. Intern. Med. 100, 663-671.

4. Frenkel, J., Good, J. & Schults, J. (1966) Lab. Invest. 15, 1559-1577.5. Kovacs, J., Halpern, J., Swan, J., Moss, J., Parillo, J. & Masur, H.

(1988) J. Immunol. 140, 2023-2031.6. Maniatis, T., Fritsch, E. & Sambrook, J. (1981) Molecular Cloning:

A Laboratory Manual (Cold Spring Harbor Lab., Cold SpringHarbor, NY).

7. Gubler, U. & Hoffman, B. (1983) Gene 25, 263-269.8. Morgan, D., Edman, J., Standring, D., Fried, V., Smith, M., Roth,

R. & Rutter, W. (1987) Nature (London) 329, 301-307.9. Short, J., Fernandez, J., Sorge, J. & Huse, W. (1988) Nucleic Acids

Res. 16, 7583-7600.10. Bradford, M. (1976) Anal. Biochem. 72, 248-254.11. Pogolotti, A., Danenberg, P. & Santi, D. (1986) J. Med. Chem. 29,

478-482.12. Santi, D., McHenry, C. & Sommer, H. (1974) Biochemistry 13,

471-480.13. Santi, D., McHenry, C. & Perriard, E. (1974) Biochemistry 13,

467-470.14. Maley, G., Bellisario, R., Guarino, D. & Maley, F. (1979) J. Biol.

Chem. 254, 1301-1304.15. Taylor, G., Lagosky, P., Storms, R. & Haynes, R. (1987) J. Biol.

Chem. 262, 5298-5307.16. Kozak, M. (1983) Microbiol. Rev. 47, 1-45.17. Chu, F., Maley, G., Maley, F. & Belfort, M. (1984) Proc. NatI.

Acad. Sci. USA 81, 3049-3050.18. Takeishi, K., Kaneda, S., Ayusawa, D., Shimizu, K., Gotoh, 0. &

Seno, T. (1985) Nucleic Acids Res. 13, 2035-2043.19. Parker, R. & Patterson, B. (1987) in Molecular Biology of RNA:

New Perspectives, eds. Inouye, M. & Dudock, B. (Academic, NewYork), pp. 133-148.

20. Langford, C. & Gallwitz, D. (1983) Cell 33, 519-527.21. Mertins, P. & Gallwitz, D. (1987) EMBO J. 6, 1757-1763.22. De Boer, H., Comstock, L. & Vasser, M. (1983) Proc. Natl. Acad.

Sci. USA 80, 21-25.23. Pinter, K., Davisson, V. & Santi, D. (1988) DNA 7, 235-241.24. Pogolotti, A., Weill, C. & Santi, D. (1979) Biochemistry 13, 2794-

2798.25. Bisson, L. & Thorner, J. (1981) J. Biol. Chem. 256, 12456-12462.26. Beverley, S., Ellenberger, T. & Cordingley, J. (1986) Proc. Natl.

Acad. Sci. USA 83, 2584-2588.27. Grumont, R., Washtien, W., Caput, D. & Santi, D. (1986) Proc.

Natl. Acad. Sci. USA 83, 5387-5391.28. Bzik, D., Li, W., Horii, T. & Inselburg, J. (1987) Proc. Natl. Acad.

Sci. USA 84, 8360-8364.29. Honess, R., Bodemer, W., Cameron, K., Niller, H.-H., Flecken-

stein, B. & Randall, R. (1986) Proc. Natl. Acad. Sci. USA 83,3604-3608.

30. Padgett, R., Grabowski, P., Konarska, M., Seiler, S. & Sharp, P.(1986) Annu. Rev. Biochem. 55, 1119-1150.

31. Deng, T., Li, D., Jenh, C.-H. & Johnson, L. (1986) J. Biol. Chem.261, 16000-16005.

32. Hardy, L., Finer-Moore, J., Montford, W., Jones, M., Santi, D. &Stroud, R. (1987) Science 235, 448-455.

33. Toth, I., Lazar, G. & Goodman, H. (1987) EMBO J. 6, 1853-1858.34. Singer, S., Richards, C., Ferone, R., Benedict, D. & Ray, P. (1989)

J. Bacteriol. 171, 1372-1378.35. Edman, J. C., Edman, U., Cao, M., Lundgren, B., Kovacs, J. A.

& Santi, D. V. (1989) Proc. Natl. Acad. Sci. USA, in press.

Biochemistry: Edman et al.

Dow

nloa

ded

by g

uest

on

July

26,

202

0