Limited proteolysis of Escherichia coli cytidine 5¢-triphosphatesynthase. Identification of residues required for CTP formationand GTP-dependent activation of glutamine hydrolysis

Dave Simard, Kerry A. Hewitt, Faylene Lunn, Akshai Iyengar and Stephen L. Bearne

Department of Biochemistry and Molecular Biology, Dalhousie University, Halifax, Nova Scotia, Canada

Cytidine 5¢-triphosphate synthase catalyses the ATP-dependent formation of CTP from UTP using eitherammonia or L-glutamine as the source of nitrogen. Whenglutamine is the substrate, GTP is required as an allostericeffector to promote catalysis. Limited trypsin-catalysedproteolysis, Edman degradation, and site-directed muta-genesis were used to identify peptide bonds C-terminal tothree basic residues (Lys187, Arg429, and Lys432) ofEscherichia coliCTP synthase that were highly susceptible toproteolysis. Lys187 is located at the CTP/UTP-binding sitewithin the synthase domain, and cleavage at this sitedestroyed all synthase activity. Nucleotides protected theenzyme against proteolysis at Lys187 (CTP > ATP >UTP > GTP). The K187A mutant was resistant to pro-teolysis at this site, could not catalyse CTP formation, andexhibited lowglutaminase activity thatwas enhanced slightlyby GTP. K187A was able to form tetramers in the presenceofUTP andATP.Arg429 andLys432 appear to reside in an

exposed loop in the glutamine amide transfer (GAT)domain. Trypsin-catalyzed proteolysis occurred at Arg429and Lys432 with a ratio of 2.6 : 1, and nucleotides did notprotect these sites from cleavage. The R429A and R429A/K432Amutants exhibited reduced rates of trypsin-catalyzedproteolysis in the GAT domain and wild-type ability tocatalyseNH3-dependentCTP formation.For thesemutants,the values of kcat/Km and kcat for glutamine-dependent CTPformation were reduced � 20-fold and � 10-fold, respect-ively, relative to wild-type enzyme; however, the value ofKmfor glutamine was not significantly altered. Activation of theglutaminase activity ofR429AbyGTPwas reduced6-foldatsaturating concentrations of GTP and the GTP bindingaffinitywas reduced 10-fold. This suggests thatArg429 playsa role in both GTP-dependent activation and GTP binding.

Keywords: activation; amidotransferase; CTP synthase;glutaminase; proteolysis; site-directed mutagenesis.

CTP synthase [CTPS; EC 6.3.4.2; UTP:ammonia ligase(ADP-forming)] catalyses the ATP-dependent formation ofCTP from UTP using either L-glutamine or NH3 as thenitrogen source (Scheme 1) [1,2]. This glutamine amido-transferase is a single polypeptide chain containing 545amino acids and consisting of two domains. The C-terminalglutamine amide transfer (GAT) domain catalyses thehydrolysis of glutamine, and the nascent NH3 derived fromglutamine hydrolysis is transferred to the N-terminalsynthase domain where the amination of UTP is catalysed[3,4]. CTPS belongs to the Triad family of glutamineamidotransferases [5,6] which utilizes a Cys-His-Glu triad tocatalyse glutamine hydrolysis and also includes anthranilatesynthase, carbamoyl phosphate synthase, formylglycin-amidine synthase, GMP synthase, imidazole glycerol phos-phate synthase, and aminodeoxychorismate synthase.

CTPS catalyses the final step in the de novo synthesis ofcytosine nucleotides. Because CTP has a central role in thebiosynthesis of nucleic acids [7] and membrane phospho-lipids [8], CTPS is a recognized target for the developmentof antineoplastic agents [7,9], antiviral agents [9,10], andantiprotozoal agents [11–13]. Recently, CTP synthaseinhibition has been shown to potentiate the cytotoxic effectsof the anticancer drug 1-b-D-arabinofuranosylcytosine [14]and anti-HIV therapies [15].CTPS from E. coli is the most thoroughly characterized

CTPSwith respect to its physical and kinetic properties, andis regulated in a complex fashion [1]. GTP is required as apositive allosteric effector to increase the efficiency (kcat/Km)of glutamine-dependent CTP synthesis 45-fold but has anegligible effect on the reaction when NH3 is the substrate[16,17]. In addition, the enzyme is inhibited by the productCTP [18], exhibits negative cooperativity for glutamine [19],and displays positive cooperativity for ATP and UTP

Scheme 1. CTP-forming reactions catalysed by CTPS.

Correspondence to S. L. Bearne, Department of Biochemistry and

Molecular Biology, Dalhousie University, Halifax, Nova Scotia,

Canada B3H 1X5. Tel.: +1 902 494 1974, Fax: + 1 902 494 1355,

E-mail: sbearne@is.dal.ca

Abbreviations: CTPS, CTP synthase; GAT, glutamine amide transfer;

GF-HPLC, gel-filtration-HPLC; PVDF, poly(vinylidene difluoride).

Enzymes: CTP synthase (EC 6.4.3.2).

(Received 28 February 2003, revised 17 March 2003,

accepted 21 March 2003)

Eur. J. Biochem. 270, 2195–2206 (2003) � FEBS 2003 doi:10.1046/j.1432-1033.2003.03588.x

[18–20]. ATP and UTP act synergistically to promotetetramerization of the enzyme to its active form [20].The structure of CTPS has not yet been determined and

hence little is known about the enzyme’s tertiary structure.However, analysis of crystal structures of the Triadamidotransferases GMP synthase and carbamoyl phos-phate synthase reveal that the structures of the GATdomains are probably closely related among all Triadenzymes [21,22]. Site-directed mutagenesis studies andsequence comparisons have revealed structural and cata-lytic roles of several amino acid residues within the GATdomain of CTPS, including residues of the catalytic triad(Cys379, His515, and Glu517) [3], residues comprising theoxyanion hole (Gly351, Gly377, Gly381, and possiblyadjacent hydrophobic residues) [23], and residues betweenAla346 and Tyr355 that appear to play an importantstructural role [4]. Recently, Willemoes reported thatThr431 and Arg433 in the GAT domain of Lactococcuslactis CTPS play a role in GTP-dependent activation ofglutamine hydrolysis [24].Our knowledge about the synthase domain is much more

limited. Analyses of mutant CTP synthases fromChlamydiatrachomatis [25], hamster [26], and yeast [27] have revealedthat mutations which render cells resistant to both thecytotoxic effects of cyclopentenylcytosine and feedbackinhibition by CTP occur between residues 116 and 229(E. coli numbering), with many of the mutations clusteringbetween residues 146 through 158. Hence, this region of thesynthase domain is believed to form part of the CTP-binding site. Competitive inhibition experiments havesuggested that for E. coli CTPS, this site may also be theUTP-binding site [18]. The locations of the ATP- and GTP-binding sites have not yet been identified. Recent studiesfrom our laboratory have revealed that residues Asp107 andLeu109 in the synthase domain of E. coli CTPS facilitateefficient coupling of glutamine hydrolysis to CTP synthesis[28].To learn more about the structure of CTPS, we inves-

tigated controlled proteolysis of the enzyme. Using limitedtrypsin-catalysed proteolysis and site-directed mutagenesis,we have identified peptide bonds C-terminal to three basicresidues of E. coli CTPS that are highly susceptible toproteolysis. One residue, Lys187, is located at the CTP/UTP-binding site within the synthase domain and isessential for catalysis but not for enzyme tetramerization.The other two residues, Arg429 and Lys432 appear to residein an exposed loop that is important for both GTP bindingand GTP-dependent allosteric activation of glutaminehydrolysis.

Materials and methods

Materials

HisÆBind resin and thrombin cleavage capture kits werefrom Novagen; broad range protein markers were fromNew England Biolabs; Pfu Turbo DNA polymerase wasfrom Stratagene Inc.; nucleotides, a-chymotrypsin frombovine pancreas (54 UÆmg)1), Pronase from Streptomycesgriseus (4.7 UÆmg)1), protease V8 from Staphylococcusaureus (1000 UÆmg)1), thermolysin from Bacillus thermo-proteolyticus rokko (55 UÆmg)1), and TPCK-treated trypsin

from bovine pancreas (10 900 UÆmg)1), and all otherchemicals were from Sigma-Aldrich Canada Ltd. Oligo-nucleotide primers for DNA sequencing and site-directedmutagenesis were commercially synthesized by ID Labor-atories (London, ON, Canada). QIAprep spin plasmidminiprep kit (Qiagen Inc.) was used for the preparation ofplasmids for mutagenesis and transformation. DNAsequencing was conducted at the Dalhousie University–NRC Institute for Marine Biosciences Joint Laboratory(Halifax, NS, Canada) and the Robarts Research Institute(London, ON, Canada), while the N-terminal amino acidsequencing was carried out at the Eastern Quebec Proteo-mics Core Facility (Ste-Foy, QC, Canada). Predictions ofsecondary structure were conducted using the programs 3-DPSSM [29], GOR4 [30], HNN [31], J-PRED [32], PREDATOR [33],PSIPRED [34], and SSPRO [35]. Sequence alignments wereconducted using CLUSTALW [36].

Enzyme expression and purification

Wild-type and mutant forms of recombinant E. coli CTPSwere expressed in and purified from E. coli strainBL21(DE3) cells transformed with either mutant or wild-type plasmid pET15b-CTPS1 as described previously [16].This construct encodes the CTPS gene product with anN-terminal hexahistidine tag. Thrombin-catalysed cleavageof the histidine tag from soluble enzymes (new N-terminus,GSHMLEM1…) was carried out as described previously[16]. The resulting enzyme was dialysed into Hepes buffer(70 mM, pH 8.0) containing EDTA (0.5 mM) and MgCl2(10 mM). The results of purification and cleavage pro-cedures were routinely monitored using SDS/PAGE. Theamino acid residues in the recombinant wild-type andmutant enzymes are numbered according to the sequence ofthe wild-type E. coli enzyme starting with M1 as positionone.

Mutagenesis

The plasmid pET15b-CTPS1 [16] was used as the templatefor site-directed mutagenesis. Site-directed mutagenesis wasconducted using the Quikchange Site-DirectedMutagenesisKit (Stratagene Inc.) and following the manufacturer’sprotocol. The synthetic deoxyoligonucleotide forward (F)and reverse (R) primers used to construct the mutants were:5¢-GCGTCTGGTGAAGTCGCAACCAAACCGACTCAG-3¢ (F, K187A), 5¢-GCTGAGTCGGTTTGGTTGCGACTTCACCAGACGC-3¢ (R, K187A), 5¢-CGGCAACGTTGAAGTTGCTAGCGAGAAGAGCG-3¢ (F,R429A), 5¢-CGCTCTTCTCGCTAGCAACTTCAACGTTGCCG-3¢ (R, R429A), 5¢-GCAACGTTGAAGTTGCTAGCGAGGCGAGCGATCTCG-3¢ (F, R429A/K432A), 5¢-CGAGATCGCTCGCCTCGCTAGCAACTTCAACGTTGC-3¢ (R, R429A/K432A), where the posi-tions of the mismatches are underlined. Potential mutantplasmids were isolated and used to transform competentDH5a cells. These cells were used for plasmid maintenanceand for all sequencing reactions. The entire mutant geneswere sequenced to verify that no other alterations of thenucleotide sequence had been introduced. CompetentE. coli strain BL21(DE3) cells were used as the host fortarget gene expression.

2196 D. Simard et al. (Eur. J. Biochem. 270) � FEBS 2003

Enzyme assay and protein determinations

CTPS activity was determined at 37 �C using a continuousspectrophotometric assay by following the rate of increasein absorbance at 291 nm resulting from the conversion ofUTP to CTP (De ¼ 1338ÆM)1Æcm)1) [18]. The standardassay mixture consisted of Hepes buffer (70 mM, pH 8.0)containing EDTA (0.5 mM) and MgCl2 (10 mM), CTPS,and saturating concentrations of UTP (1 mM) and ATP(1 mM) in a total volume of 1 mL. Enzyme and nucleotideswere preincubated together for 2 min at 37 �C followed byaddition of substrate (NH4Cl or glutamine) to initiate thereaction. Total NH4Cl concentrations used in the assayswere 5, 10, 20, 30, 50, 60, 80, and 100 mM, and CTPSconcentrations were � 3.0 lgÆmL)1 (wild-type),3.0 lgÆmL)1 (R429A), and 4.0 lgÆmL)1 (R429A/K432A).For glutamine assays, concentrations of glutaminewere 0.1, 0.2, 0.3, 0.5, 1.0, 2.0, 3.0, and 6.0 mM andCTPS concentrations were � 4.0 lgÆmL)1 (wild-type),1.4 mgÆmL)1 (R429A), and 0.9 mgÆmL)1 (R429A/K432A). The concentration of GTP was maintained at0.25 mM for all assays when glutamine was used as thesubstrate. In addition, the ionic strength was maintained at0.25 M in all spectrophotometric assays by the addition ofKCl. The apparent activation constant (KA) for R429ACTPS (0.4 mgÆmL)1) with respect to GTP was determinedfor glutamine-dependent CTP formation as describedpreviously [28].All kinetic parameters were determined in triplicate and

average values are reported. Initial rate kinetic data was fitto Eqn (1) by nonlinear regression analysis using theprogram ENZYMEKINETICS v1.5 (1996) from Trinity Soft-ware (Plymouth, NH). In Eqn (1), vi is the initial velocity,Vmax (¼ kcat[E]T) is the maximal velocity at saturatingsubstrate concentrations, [S] is the substrate concentration(glutamine or NH3), and Km is the Michaelis constant forthe substrate. Values of Km for NH3 were calculated usingthe concentration of NH3 present at pH 8.0 {pKa(NH4

+) ¼ 9.24 [37]}. Values of kcat were calculated forCTPS variants with the hexahistidine tag removed usingthe molecular masses (Da) of 61 029 (wild-type), 60 944(R429A), and 60 887 (R429A/K432A). The reportederrors are standard deviations. Except where notedotherwise, protein concentrations were determined usingthe Bio-Rad Protein Assay (Bio-Rad Laboratories Ltd.)with BSA standards.

mi ¼Vmax½S�Km þ ½S� ð1Þ

Glutaminase activity

Values of kcat for the hydrolysis of glutamine, at fixedsaturating concentrations of glutamine (6 mM), UTP(1 mM), and ATP (1 mM) were determined as describedpreviously [38]. Data describing the dependence of theapparent kcat values on the concentration of GTP werefitted to Eqn (2) for hyperbolic nonessential activationkinetics where KA is the apparent activation constant, ko isthe turnover number in the absence of GTP, and kact isthe turnover number at saturating concentrations of GTP[39].

kapparentcat ¼ko þ kact

GTP½ �KA

� �

1þ GTP½ �KA

� � ð2Þ

Limited proteolysis

Initially, wild-type CTPS was subjected to limited proteo-lysis by several endopeptidases including trypsin, chymo-trypsin, pronase, thermolysin and V8 protease. Proteolysiswas conducted in Hepes buffer (70 mM, pH 8.0) containingEDTA (0.5 mM) and MgCl2 (10 mM) at 37 �C using aCTPS/protease ratio (lg protein) of 60 : 1 in a total volumeof 1 mL. Limited proteolysis using pronase and thermolysinwas also conducted in potassium phosphate buffer (50 mM,pH 7.2) containing EDTA (1 mM) and MgCl2 (10 mM).EDTA was omitted from the buffers for all thermolysin-catalysed reactions. Proteolysis experiments were conductedin the absence and presence of ATP (10 mM) and UTP(10 mM). During the proteolysis reactions, aliquots (25 lL)were removed from the reaction mixture over the course of1 h, and transferred to gel loading buffer [25 lL; Tris/HCl(170 mM, pH 6.8) containing dithiothreitol (120 mM), SDS(5.4%, w/v), Bromophenol blue (0.03%, w/v) and glycerol(27.2%, w/v)] to terminate the reaction. The samples werethen boiled for 5 min and the proteolytic fragmentswere separated using SDS/PAGE (12% gels). Fragmentswere visualized by staining with Coomassie blue R-250 andsubsequent de-staining in a solution of methanol/H2O/acetic acid (45 : 45 : 10). Detailed studies were subsequentlyconducted using trypsin as trypsin-catalysed proteolysisgave different fragments depending on whether the nucleo-tides were absent or present.Limited trypsin-catalysed proteolysis of both wild-type

and mutant CTP synthases was analysed by monitoringCTPS activity (vide infra) at specific time points and bySDS/PAGE. Trypsin proteolysis reactions (1 mL totalvolume) contained either wild-type or mutant CTPS(0.20 mgÆmL)1), and were conducted for 1 h in Hepesbuffer (70 mM, pH 8.0) containing EDTA (0.5 mM) andMgCl2 (10 mM) at 37 �C. Reactions were initiated byaddition of trypsin (0.1 lgÆmL)1) and aliquots (100 lL)were removed every 10 min over 1 h and assayed foractivity using NH3 as the substrate. The cleavage fragmentsproduced from trypsin-catalysed proteolysis were analysedusing SDS/PAGE (12% and 20% gels). Reactions wereconducted as described above, with the exception that at 10,20, 30, and 60 min, aliquots (20 lL) were removed andtransferred to gel loading buffer (15 lL) to terminate thereaction. A zero time point was obtained using 3 lL ofthe enzyme stock solution (� 1.5 mgÆmL)1) used for thereaction. The ability of various ligands to protect both wild-type and mutant CTP synthases from proteolysis wasexamined using ATP (10 mM), UTP (10 mM), ATP andUTP together (10 mM each), CTP (0.1, 0.5, and 1.0 mM),GTP (2.5 mM) and L-glutamine (10 mM).

Inactivation assays

Aliquots from two separate proteolysis reactions wereassayed using both NH4Cl (100 mM) and glutamine

� FEBS 2003 Limited proteolysis of CTP synthase (Eur. J. Biochem. 270) 2197

(10 mM) as substrates, as described above. Inactivation ofCTPS activity followed first-order kinetics, and the apparentfirst-order rate constants for inactivation were calculatedfrom plots of the percent activity remaining as a logarithmicfunction of the time of incubation. The ability of variousligands to protect both wild-type and mutant CTPS fromproteolysis was examined using ATP (0.5, 1.0, 2.0, and10 mM), UTP (0.5, 1.0, 2.0, and 10 mM), ATP and UTPtogether (10 mM each), GTP (0.25 mM), and L-glutamine(10 mM) at the concentrations indicated.

N-Terminal sequence analysis

Approximately 190 lg total protein containing CTPS andthe various trypsin-catalysed cleavage fragments, producedafter a 90-min proteolysis reaction, were separated usingSDS/PAGE (12% or 20% gels) and subsequently trans-ferred to a poly(vinylidene difluoride) (PVDF) membrane[Immun-Blot 0.2 lm (Bio-Rad Laboratories Ltd) forfragments with molecular masses of 25, 28 and 53 kDa;Immobilon-Psq0.2 lm PVDF (Millipore Ltd) for thefragment with a molecular mass of 10 kDa] as describedby Wilson and Yuan [40]. Electroblotting was conductedin CAPS buffer (0.01 M, pH 11.0) containing methanol(10%). Whole enzyme and cleavage fragments were locatedon the PVDF membrane by staining with Coomassie bluefollowed by destaining in 50%methanol. Sections (50 mm2)of the PVDF membrane with adsorbed protein weresubmitted for N-terminal amino acid sequence analysis.

CD spectra

CD spectra were obtained using a JASCO J-810 spectro-polarimeter and were recorded for both the wild-type andmutant enzymes (K187A, R429A, R429A/K432A) over therange 190–260 nm in the absence of nucleotides. A markeddecrease in buffer transparency was observed below190 nm and therefore all spectra were truncated at thiswavelength. The resulting CD spectra obtained fromenzyme solutions (0.2 mgÆmL)1) in Bis-Tris propane buffer(10 mM, pH 8.0) containingMgSO4 (10 mM) were analysedfor percent a-helix and b-sheet structure using CDNN CDSpectra Deconvolution v. 2.1 developed by G. Bohm [41].Protein concentrations were determined spectrophotomet-rically at 280 nm using an extinction coefficient equal to38 030ÆM)1Æcm)1 for the wild-type, K187A, R429A, andR429A/K432A CTP synthases.

Tetramerization of CTPS

The ability of the K187A CTPS to form tetramers wasevaluated using gel-filtration-HPLC (GF-HPLC) withnative tryptophan fluorescence detection. Wild-type andmutant CTP synthases, and standard proteins were elutedunder isocratic conditions using Hepes buffer (pH 8.0,0.07 M) containing MgCl2 (10 mM) and EDTA (0.5 mM) ata flow rate of 1.0 mLÆmin)1 on a BioSep–SEC-S 3000column (7.80 · 300 mm; Phenomenex, Torrance, CA). AWaters 510 pump and 680 controller were used for solventdelivery. Injections were made using a Rheodyne 7725isample injector fittedwith a 20-lL injection loop. The elutedproteins were detected by native protein fluorescence

(excitation and emission wavelengths of 285 nm and335 nm, respectively) using a Waters 474 scanning fluores-cence detector. GF-HPLC of both wild-type and mutantenzymes was conducted in the absence and presence of ATP(1 mM) and UTP (1 mM), and the retention times werecompared with those observed for the wild-type enzyme.The column was standardized using the following proteins(0.5 mgÆmL)1): bovine thyroglobulin (669 kDa), b-amylase(200 kDa), BSA (66 kDa), and carbonic anhydrase(29 kDa). Chromatograms were analysed using PEAKSIM-

PLE software fromMandel Scientific (Guelph,ON,Canada).The retention time of bovine thyroglobulin was used toestimate the column void volume (Vo).

Results

Limited proteolysis of CTPS

CTPS from E. coli was subjected to controlled proteolysisby five endopeptidases (pronase, a-chymotrypsin, V8 pro-tease, thermolysin, and trypsin) in the absence or presence ofthe nucleotides ATP and/or UTP (data not shown). Thesepreliminary experiments revealed that only treatment ofCTPS with trypsin produced a limited number of cleavagefragments over the course of 1 h, of which the formation ofsome fragments was suppressed in the presence of ATP and/or UTP. Thus, trypsin was used in the present study toinvestigate the accessibility of regions in CTPS to proteolyticcleavage in the presence and absence of various ligands.

Limited trypsin-catalysed cleavage of wild-type CTPS

Limited trypsin-catalysed proteolysis of wild-type CTPSproduced different cleavage fragments depending on whe-ther ATP and/orUTPwere absent or present in the reactionmixture (Fig. 1). In the absence of nucleotides, trypsin-catalysed cleavage of CTPS (63 kDa) produced fourfragments with molecular masses corresponding to 10, 25,28, and 53 kDa. However, in the presence of either ATP orUTP (data not shown), ATP and UTP together, or CTP,only fragments with molecular masses of 10 and 53 kDawere produced indicating that these nucleotides protectedCTPS from cleavage at the site which produced the 25- and28-kDa fragments. Neither glutamine nor GTP protectedCTPS from limited trypsin-catalysed digestion.The sites of trypsin-catalysed cleavage yielding each of the

fragments were identified using Edman degradation toobtain the N-terminal amino acid sequence of eachfragment (Table 1), and the known nucleotide sequenceencoding the enzyme [3]. The N-terminal sequence of the25- and 53-kDa fragments were the same as that ofthe whole recombinant wild-type protein indicating that thecleavage site was located at the C-terminus of each of thesetwo fragments. The N-terminal sequence of the 28-kDafragment indicated that one of the cleavage sites wasLys187. N-terminal analysis of the 10-kDa fragmentproduced two sequences indicating that cleavage occurredat Arg429 and Lys432 with a ratio of 2.6 : 1, respectively.The cleavage pattern observed in the absence of anyprotecting ligands, and the sites identified using N-terminalanalysis are summarized in Fig. 2. The molecular mass ofeach polypeptide fragment, calculated using the known

2198 D. Simard et al. (Eur. J. Biochem. 270) � FEBS 2003

amino acid sequence [3], is in excellent agreement with thevalues deduced from SDS/PAGE calibration (Table 1).

Limited trypsin-catalysed cleavage of mutant CTPsynthases

To confirm that the cleavage sites identified usingN-terminal analysis were indeed correct, site-directed mut-agenesis was used to construct two single mutants (K187Aand R429A) and one double mutant (R429A/K432A).Limited trypsin-catalysed proteolysis of K187A CTPS inthe absence of nucleotides produced only the 10- and53-kDa fragments (Fig. 3A), a cleavage pattern which wasidentical to that observed for wild-type CTPS in thepresence of ATP and UTP (Fig. 1B). Proteolysis ofR429A in the absence of nucleotides gave fragments withmolecular masses of approximately 10, 25, 38, and 53 kDa(Fig. 3B). The 38-kDa fragment accumulated because rapidcleavage at Arg429 no longer occurred while cleavage at

Fig. 1. SDS/PAGE analysis of trypsin-catalysed cleavage of recom-

binant wild-type CTPS in the absence and presence of ligands. For each

gel: lane 1 contains molecular mass standards and lanes 2–6 contain

wild-type CTPS treated with trypsin for 0, 10, 20, 30, and 60 min,

respectively. In the absence of any ligands (A), the wild-type protein is

rapidly cleaved to yield a 10-kDa (not shown except in E) and a

53-kDa fragment, the latter which is subsequently cleaved to yield two

fragments with molecular masses of � 25 and � 28 kDa. In the pres-

ence of ATP and UTP (10 mM each) (B), and CTP (2.5 mM) (C), only

production of the 10- and 53-kDa cleavage fragments is observed. In

the presence of GTP (2.5 mM) (D) and L-glutamine (10 mM) (E),

fragments with molecular masses of 10, 25, 28, and 53 kDa are pro-

duced. All gels are 12% except for (E) which is 20%.

Table 1. N-terminal amino acid sequences of trypsin-catalysed cleavage

fragments.

Molecular massa

(kDa) N-terminal sequence

Cleavage site

identifiedb

63 (61) GSHMLEM1 … None

53 (48) GSHML None

28 (27) T188KPTQHSVKE Lys187

25 (21) GSHML None

10 (13.1) S430EKSDLGGTM (major)c Arg429

(12.8) S433DLGGTMRL (minor)c Lys432

a Apparent molecular mass for full-length recombinant wild-type

E. coli CTPS and fragments determined from SDS/PAGE calib-

ration are given. The corresponding molecular masses calculated

using the known amino acid sequence are given in parentheses. b The

amino acid listed is that which provides the carbonyl function to the

scissile peptide bond. The numbers correspond to the numbering for

wild-type E. coli CTPS. c The ratio of the major peptide to minor

peptide was 2.6 : 1 and was determined by integration of the HPLC

chromatogram peaks corresponding to the phenylthiohydantoin

derivatives of the N-terminal serines.

Fig. 2. Fragments generated by limited trypsin-

catalysed proteolysis of wild-type CTPS.

Peptide bond cleavage occurs C-terminal to

Lys187 in the synthase domain, and Arg429

and Lys432 in the GAT domain. The CTP/

UTP-binding site and residues comprising the

catalytic triad (Cys379, His515, and Glu517)

are also shown.

� FEBS 2003 Limited proteolysis of CTP synthase (Eur. J. Biochem. 270) 2199

Lys187 divided the protein into the 25- and 38-kDafragments. The 10- and 53-kDa fragments were formed inmuch lower amounts than observed with wild-type CTPSbecause of slow cleavage at Lys432. In the presence of ATPandUTP, cleavage of R429A at Lys187 was suppressed andonly the 10- and 53 kDa fragments were formed because ofslow cleavage at Lys432 (Fig. 3C). Limited proteolysis of thedouble mutant (R429A/K432A) in the absence of nucleo-tides produced only two fragments with molecular massescorresponding to 25 and 38 kDa consistent with cleavageoccurring only at Lys187 (Fig. 3D). In the presence of ATP

and UTP, proteolysis of R429A/K432A CTPS was com-pletely suppressed (Fig. 3E).

Inactivation and protection studies

Treatment of wild-type and mutant CTP synthases withtrypsin produced time-dependent loss of both NH3-depend-ent activity and glutamine-dependent activity (Fig. 4),which followed first-order kinetics up to at least 90% ofthe reaction. The observed first-order inactivation rateconstants for CTPS activity assayed using either NH3 orglutamine as the substrate are given in Table 2. In theabsence of ligands, the observed first-order rate constant fortrypsin-catalysed proteolysis of CTPS was slightly greaterwhen glutamine-dependent CTP formation was measuredthan when NH3-dependent CTP formation was measured.A reduction in the observed first-order rate constants for theinactivation of theNH3-dependent activity was observed forincreasing concentrations of ATP, UTP, and CTP consis-tent with each of these nucleotides providing protectionfrom trypsin-catalysed cleavage. Interestingly, in the

Fig. 3. SDS/PAGE analysis of trypsin-catalysed cleavage of mutant

CTP synthases in the absence and presence of ligands. (A) Lane 1,

molecular mass standards; lane 2, trypsin (at 7000 times the concen-

tration used in the proteolysis reactions); lanes 3–7 contain K187A

CTPS treated with trypsin for 0, 10, 20, 30, and 60 min, respectively.

The wild-type protein is rapidly cleaved to yield a 10- (not shown) and

a 53-kDa fragment, the latter which is not cleaved to yield the 25- and

28-kDa fragments. For each gel shown in B through E, lane 1 contains

molecular mass standards and lanes 2–6 contain mutant CTPS treated

with trypsin for 0, 10, 20, 30, and 60 min, respectively. Limited pro-

teolysis of R429A CTPS (B) in the absence of nucleotides produced

fragments with molecular masses of 10, 25, 38, and 53 kDa. However,

in the presence of ATP and UTP (10 mM each) (C), only the 10- and

53-kDa fragments are produced. Limited proteolysis of R429A/

K432A CTPS (D) in the absence of nucleotides produces fragments

with molecular masses of 25 and 38 kDa. However, in the presence of

ATP and UTP (10 mM each) (E), no cleavage fragments were pro-

duced over the course of 1 h indicating that proteolysis was greatly

suppressed. All gels are 12%.

Fig. 4. Time-dependent inactivation of wild-type CTPS by trypsin. (A)

Inactivation of CTPS-catalysed NH3-dependent CTP formation in the

absence of ligands (s) and in the presence of ATP (10 mM, n), UTP

(10 mM, h), and ATP and UTP combined (10 mM each, ,). Panel B

shows the inactivation of CTPS-catalysed glutamine-dependent CTP

formation in the absence of nucleotides (s) and in the presence of UTP

(10 mM, n), ATP (10 mM, h), and ATP and UTP combined (10 mM

each, ,). In both panels, the activity of the enzyme in the absence of

trypsin is also shown (d).

2200 D. Simard et al. (Eur. J. Biochem. 270) � FEBS 2003

presence of ATP and UTP, the rate constant for inactiva-tion of glutamine-dependent activity was greater than thatobserved for inactivation of the NH3-dependent activity.This observation is consistent with the cleavage sites in theGAT domain not being protected by these nucleotides andthe resulting 53-kDa fragment still possessing NH3-depend-ent activity. Although all CTPS ligands tested (ATP, UTP,CTP, GTP, and glutamine) protected CTPS from inactiva-tion to some degree, the most effective protection wasafforded by CTP.The observed first-order inactivation rate constants for

the NH3-dependent activity of the R429A and R429A/K432A CTP synthases were less than that observed forwild-type CTPS. Apparently, reduced cleavage within theGAT domain results in less rapid cleavage within in thesynthase domain (i.e. at Lys187) and hence a lower value forthe rate constant for the loss of NH3-dependent activity.Inactivation of the K187A enzyme could not be studiedbecause this enzyme was inactive (vide infra).

CD

The secondary structural content of wild-type CTPS andthe three mutant enzymes was analysed using CDspectroscopy. Fig. 5 shows that the secondary structurecontent of all the mutant proteins is similar to that of thewild-type enzyme, except that the a-helix content of theK187A and R429A/K432A mutants is slightly reducedwhile the content of antiparallel b-sheet structure is slightlyincreased, relative to wild-type CTPS. Although no signi-

ficant gross perturbations in secondary structure areevident in the mutant proteins, the possibility that themutations cause a localized perturbation of secondarystructure or conformational change cannot be ruled out.

Mutant enzyme kinetics

The kinetic parameters kcat andKm for CTP formation weredetermined with respect to NH3 and glutamine for each ofthe mutant enzymes except for K187A CTPS which wasinactive (Table 3). Direct examination of the conversion ofglutamine to glutamate (glutaminase activity) revealed thatK187A CTPS was able to catalyse the hydrolysis ofglutamine, but had a value of kcat that was half of thatobserved for wild-type CTPS in the absence of GTP. Whenthe concentration of GTP was increased to 1 mM, the valueof kact was increased fivefold, compared to 50-fold for wild-type CTPS (Fig. 6).R429A and R429A/K432A CTP synthases displayed

similar kinetic properties. Each mutant had close to wild-type NH3-dependent activity; however, glutamine-depend-ent CTP formation was impaired. Interestingly, Km forglutamine only increased 1.3- to 1.7-fold indicating that themutations had little effect on glutamine binding. However,kcat was reduced � 15-fold for each mutant so that thecatalytic efficiency (kcat/Km) of glutamine-dependent CTPformation was decreased 25- and 19-fold for the R429A andR429A/K432A CTP synthases, respectively.The kinetics of R429A CTPS were investigated in detail

to determine if the impaired glutamine-dependent CTPformation was caused by an inability of GTP to activateglutamine hydrolysis. In the presence of ATP and UTP(1 mM each) and saturating glutamine (6 mM), GTP(0.25 mM) caused a 2.5-fold increase in kcat for glutamine-dependent CTP formation catalysed by R429A CTPScompared to a 30-fold increase for wild-type CTPS (datanot shown). Concentrations of GTP above 0.25 mM (up to1 mM) did not enhance the observed rate of CTP formation.Direct examination of the glutaminase activity revealed

that kcat was reduced approximately 10-fold for the R429Aand R429A/K432A enzymes relative to wild-type CTPSwith the concentration of GTP equal to 0.25 mM (Table 3).More detailed analysis of the glutaminase activity of R429ACTPS (Fig. 6) revealed that GTP binding and kact werereduced approximately 10-fold and sixfold, relative to wild-type CTPS. Thus mutation of Arg429 to alanine impairsboth GTP binding and allosteric activation of glutaminehydrolysis.Comparison of the kcat values for the glutaminase activity

and glutamine-dependent CTP formation catalysed byR429A CTPS reveals that ammonia is produced fromglutamine hydrolysis at a rate that is slightly higher than therate at which CTP is formed. This observation suggests thatthere may be a partial uncoupling of the glutaminase andsynthase reactions, however, the kcat values for the corres-ponding reactions catalysed by the R429A/K432A mutantare experimentally equal.

Oligomerization of CTPS

To determine if the K187A mutant was inactive becauseit was unable to form tetramers, we investigated the

Table 2. Observed rate constants for inactivation of recombinant wild-

type and mutant CTP synthasesa. ND, Not determined.

Protecting

ligand

Concentration

(mM)

kobs (· 10)2 min)1)

NH3 as

substrate

L-glutamine

as substrate

None 0 7 (± 2) 9.0 (± 0.3)

None (R429A) 0 3.9 (± 0.3) ND

None

(R429A/K432A)

0 1.9 (± 0.4) ND

ATP 0.5 6 (± 2) ND

1.0 1.7 (± 0.5) ND

2.0 0.31 (± 0.04) ND

10.0 0.4 (± 0.1) 1.9 (± 0.2)

UTP 0.5 5.6 (± 0.9) ND

1.0 6 (± 1) ND

2.0 4 (± 1) ND

10.0 0.7 (± 0.1) 3.9 (± 0.9)

ATP and UTP 10 + 10 0.15 (± 0.07) 0.8 (± 0.2)

CTP 0.1 3 (± 1) ND

0.5 0.26 (± 0.09) ND

1.0 0.16 (± 0.03) ND

GTP 0.25 1.7 (± 0.5) ND

L-glutamine 10.0 1.5 (± 0.3) ND

Controlb 0 0.075 (± 0.009) 0.07 (± 0.04)

a All kobs values are for inactivation of wild-type CTPS except

where indicated otherwise. b Control refers to the inactivation of

CTPS that is observed during incubation of the enzyme at 37 �Cfor 1 h in the absence of added trypsin.

� FEBS 2003 Limited proteolysis of CTP synthase (Eur. J. Biochem. 270) 2201

Fig. 5. CD analysis of wild-type and mutant

CTP synthases. (A) Spectra for wild-type,

K187A, R429A, and R429A/K432A CTP

synthases are shown. Each spectrum is the

average of three scans for each CTPS variant.

(B) The relative amount of each type of sec-

ondary structure is indicated for each CTPS

variant. Error bars represent the standard

deviation of the mean for three independent

trials.

Table 3. Kinetic parameters for wild-type and mutant CTP synthases. ND, not determined.

Reaction (substrate) Kinetic parametera

CTPS variants

Wild-type K187A R429A R429A/K432A

CTP formation Km (mM) 1.70 ± 0.08 1.60 ± 0.08 1.82 ± 0.04

(NH3) kcat (s)1) 7.8 ± 0.1 NAb 7.3 ± 0.3 6.5 ± 0.8

kcat/Km (mM)1Æs)1) 4.5 ± 0.2 4.6 ± 0.3 3.6 ± 0.5

CTP formation Km (mM) 0.24 ± 0.04 0.40 ± 0.03 0.32 ± 0.08

(L-glutamine) kcat (s)1) 5.4 ± 0.8 NA 0.36 ± 0.04 0.39 ± 0.04

kcat/Km (mM)1s)1) 22.6 ± 4.5 0.9 ± 0.1 1.2 ± 0.4

L-glutamate

formation (fixed

L-glutamine with

varying [GTP])

kcat (s)1) [GTP] ¼ 0.25 mM 5.01 ± 0.18 0.13 ± 0.03 0.48 ± 0.03 0.49 ± 0.07

ko (s)1) 0.14 ± 0.03 0.07 ± 0.02 0.14 ± 0.04 ND

kact (s)1) 7.1 ± 0.3 –c 1.19 ± 0.04 ND

KA (mM) 0.032 ± 0.006 –c 0.38 ± 0.04 ND

a Assay conditions are as described in Materials and methods. [ATP] ¼ [UTP] ¼ 1 mM. b No activity was observed (i.e. < 0.5% wild-type

CTPS activity). c Values could not be determined accurately because of the low activity.

2202 D. Simard et al. (Eur. J. Biochem. 270) � FEBS 2003

ability of this mutant to form tetramers in the presence ofnucleotides using GF-HPLC. The observed molecularmasses for wild-type CTPS in the absence of nucleotidesand in the presence of ATP and UTP were 123 and251 kDa, respectively. These values are similar to thepredicted values of 122 and 245 kDa, based on the aminoacid sequence of the recombinant mutant protein, and areconsistent with wild-type CTPS existing primarily asdimers in the absence of ATP and UTP, and with ashifting of the equilibrium to favour the tetrameric speciesin the presence of ATP and UTP [42]. The K187Amutant had an apparent molecular mass of 178 kDa inthe absence of nucleotides. This value is slightly higherthan that observed for the wild-type enzyme and corres-ponds to the enzyme existing as � 30% tetramer ascalculated using Eqn (3) [20], where X is the fraction ofthe enzyme in the tetramer form and the molecularmasses of the dimer (121 944 Da) and the tetramer(243 888 Da) are those predicted based on the monomermolecular mass of 60 972 Da for recombinant K187Alacking the histidine tag. In the presence of ATP andUTP, the observed molecular weight for K187A was259 kDa. Thus it appeared that K187A CTPS wascapable of forming tetramers in the presence of saturatingconcentrations of UTP and ATP, similar to the wild-typeenzyme.

molecular mass ¼ ð243 888Þ2Xþ ð121 944Þ2ð1�XÞð243 888ÞXþ ð121 944Þð1�XÞ ð3Þ

Discussion

Limited proteolysis has been used to delineate the structuralorganization of several amidotransferases including aspa-

ragine synthase [43], carbamoyl phosphate synthase[44–50], anthranilate synthase [51], and glucosamine-6-phosphate synthase [52]. This methodology has beenparticularly useful for identifying both ligand-binding sitesand, in the case with glucosamine-6-phosphate synthase, anexposed �hinge� region that, when cleaved by a-chymotryp-sin, led to separation of the enzyme into its GAT andsynthase domains. Our interest in delineating structuralaspects of E. coli CTPS led us to examine the susceptibilityof CTPS to controlled proteolysis. In preliminary experi-ments with endopeptidases of different specificity, weidentified trypsin as the enzyme of choice. Trypsin-catalysedcleavage of wild-type CTPS generated four fragments inthe absence of ATP and UTP, but only two fragments inthe presence of these nucleotides. Determination of theN-terminal sequence of these fragments, in conjunctionwith the known nucleotide sequence of the E. coli pyrGgene [3], permitted us to identify three principal cleavagesites: Lys187 in the synthase domain, and Arg429 andLys432 in the GAT domain. A summary of the fragmen-tation pattern arising from trypsin-catalysed cleavage atthese sites is presented in Fig. 2.Lys187 resides in a region of the synthase domain that is

highly conserved among CTP synthases from differentorganisms. This region, between residues 116 and 229, hasbeen suggested to comprise the CTP/UTP-binding site[18,25–27]. Our observation that both CTP andUTP affordeffective protection to CTPS from trypsin-catalysed clea-vage at Lys187 also supports the notion that this residue islocated in the CTP/UTP-binding site. Interestingly, ATPalso provides protection against cleavage and does so betterthan UTP. Such protection could arise because: (a) ATPbinds at an adjacent site and sterically blocks access oftrypsin to Lys187; (b) ATP-induced tetramerization yields aquaternary structure in which the Lys187 site is notaccessible to trypsin; or (c) ATP induces a conformationalchange in CTPS to yield a conformation in which Lys187is no longer exposed to bulk solvent.Replacement of Lys187 by an alanine residue yielded a

protein that was resistant to limited trypsin-catalysedproteolysis in the synthase domain, supporting our conclu-sion that cleavage occurred C-terminal to this residue. TheK187A mutant could not catalyse the formation of CTP,however, it retained the ability to form tetramers in thepresence of nucleotides, and exhibited a very low level ofGTP-dependent glutaminase activity which was enhancedslightly by GTP. Interestingly, in the absence of nucleotides,K187A existed as � 30% tetramer suggesting that neutrali-zation of positive charge at residue 187 might play a role inpromoting enzyme tetramerization. Indeed, hydrophobicinteractions between dimers of E. coli CTPS have beensuggested to play a role in the formation of tetramers [53].Predictions of the secondary structure of the highlyconserved region of amino acid sequence between residues185 and 192 suggest that Lys187 constitutes part of aconserved loop. It is not clear whether nucleotides protectthis putative loop from proteolytic cleavage becausenucleotide binding directly blocks access of trypsin to thecleavage site or, because nucleotide binding causes a changein the enzyme’s conformation or quaternary structure (i.e.tetramerization) that subsequently conceals the cleavage sitefrom trypsin.

Fig. 6. Glutaminase activity for mutant CTP synthases. The values of

kapparentcat for the hydrolysis of glutamine by K187A (d) and R429A (s)

CTP synthases are shown. Inset: values of kapparentcat for the hydrolysis of

glutamine by wild-type CTPS. The curves shown are from a fit of the

data to Eqn (2) and the values of ko, kact, and KA are given in Table 3.

� FEBS 2003 Limited proteolysis of CTP synthase (Eur. J. Biochem. 270) 2203

Finally, we note that studies on the chemical modificationof E. coli CTPS with thiourea dioxide led Roberston et al.[54] to conclude that lysine residues were important forcatalysis. To our knowledge, the present study representsthe first identification of a catalytically essential lysineresidue in E. coli CTPS involved in either amido-/NH3

transfer or UTP phosphorylation.Arg429 and Lys432 reside in a region of amino acid

sequence within theGAT domain that is partially conservedonly among CTP synthases from some sources (Fig. 7). Inaccord with our expectations, nucleotides offered noprotection against cleavage at these sites but replacementof these residues by alanine (i.e. R429A andR429A/K432A)yielded mutant enzymes that were more resistant toproteolytic cleavage in the GAT domain. These mutantenzymes displayed wild-type activity with respect toNH3-dependent CTP formation and wild-type affinity forglutamine, but glutamine-dependent CTP formation wasmarkedly impaired. Interestingly, although these mutationsin the GAT domain did not impair the enzyme’s ability toutilize NH3 as a substrate (i.e. the activity associated withthe synthase domain [16]), they did cause the rate of loss ofNH3-dependent activity during limited trypsin-catalysedproteolysis to be less than would have been predicted basedon the inactivation rate constant observed for wild-type

CTPS (Table 2). This is consistent with previous reportsthat suggested interactions between the GAT and synthasedomains within the tertiary structure of the enzyme[17,28,55]. The existence of such interactions is alsosupported by our observation that mutation of Lys187 toalanine in the synthase domain severely impairs theglutaminase activity in the GAT domain.Our observations that R429A CTPS binds GTP with

reduced affinity (� 10-fold) and, at saturating concentra-tions of GTP, the apparent kcat value for glutamine-dependent CTP formation is reduced sixfold suggest thatArg429 plays a role in both binding GTP and themechanism for allosteric activation of glutamine hydrolysis.Secondary structure predictions suggest that Arg429 andLys432 are located within a region where a b-strandundergoes a transition into a loop structure. The ability oftrypsin to catalyse cleavage adjacent to these residuessuggests that this loop is exposed to bulk solvent. Despitethe fact that this region is not highly conserved betweenorganisms, it does appear to be required forE. coliCTPS tocatalyse glutamine turnover. Arg429 and Lys432 lie close toa conserved sequence motif [GG(TS)(ML)RLG] within theGAT domain (shaded residues 436–442 in Fig. 7) that wasrecently identified by Willemoes [24]. Using site-directedmutagenesis experiments on CTPS from L. lactis,

Fig. 7. Sequence comparison of a portion of the C-terminus (GAT domain) of CTP synthases.For the protein sequences shown, invariant residues (*),

conservative substitutions (:), and semiconservative substitutions (.) are indicated. The two residues (Arg429 and Lys432) identified as cleavage sites

during limited trypsin proteolysis andmutated in the present study are indicated (›). These residues reside in a region of the primary structure that isnot conserved among different organisms. The conserved sequence motif (GG[TS][ML]RLG) identified by Willemoes [24] is shaded. The proteins

included in the alignment are as follows (accession numbers in parentheses): Girardia intestinalis (AAB41453.1), Synechococcus (Q54775), Spiro-

plasma citri (P52200), Synechocystis (P74208), Bacillus subtilis (P13242), Mycobacterium leprae (S72961), Mycobacterium bovis (AAB48045.1),

Methanococcus jannaschii (Q58574), Chlamydia trachomatis (Q59321), Haemophilus influenzae (P44341), Neisseria meningitidis (CAB84970.1),

Nitrosomonas europaea (AAC33441.1),Azospirillum brasilense (P28595),Campylobacter jejuni (CAB72520.1),Heliobacter pylori (O25116), Borrelia

burgdorferi (O51522), Cricetulus griseus (P50547), Mus musculus (P70698), Homo sapiens (NP_001896.1), Arabidopsis thaliana (AAC78703.1),

Saccharomyces cerevisiae H (URA-8, P38627), Saccharomyces cerevisiae G (URA-7, P28274), Plasmodium falciparum (AAC36385.1), Lactococcus

lactis (CAA09021.2), and Escherichia coli (AAA69290.1). Numbering shown is for the E. coli sequence.

2204 D. Simard et al. (Eur. J. Biochem. 270) � FEBS 2003

Willemoes demonstrated that Thr431 and Arg433 (Thr438and Arg440 in E. coliCTPS) within this motif play a role inGTP-dependent activation of glutamine hydrolysis butconcluded that these residues were not involved in GTPbinding [24]. However, our observation that Arg429 isimportant for GTP binding is consistent with Willemoes’observation that R433A L. lactis CTPS exhibited a 10–17-fold increase in GTP-binding affinity [24] and support thenotion that the conserved sequence motif and adjacentresidues may also be important for GTP binding.

Acknowledgements

The authors thank the Canadian Institutes of Health Research for an

operating grant (S. L. B.), the Nova Scotia Health Research Founda-

tion and Cancer Care Nova Scotia for graduate fellowships (A. I.), and

Cancer Care Nova Scotia for research training grants (K. H. & D. S).

The authors also express thanks to Prof. M. Dobson and J. Chew for

technical advice and assistance with electroblotting.

References

1. Koshland, D.E. Jr & Levitzki, A. (1974) CTP Synthetase and

related enzymes. In The Enzymes (Boyer, P.D., ed.), pp. 539–559.

Academic Press, New York.

2. Long, C. & Koshland, D.E. Jr (1978) Cytidine triphosphate syn-

thetase.Methods Enzymol. 51, 79–83.

3. Weng, M., Makaroff, C.A. & Zalkin, H. (1986) Nucleotide

sequence of Escherichia coli pyrG encoding CTP synthetase.

J. Biol. Chem. 261, 5568–5574.

4. Weng, M.L. & Zalkin, H. (1987) Structural role for a conserved

region in the CTP synthetase glutamine amide transfer domain.

J. Bacteriol. 169, 3023–3028.

5. Zalkin, H. (1993) The amidotransferases. Adv. Enzymol. Relat.

Areas Mol. Biol. 66, 203–309.

6. Zalkin, H. & Smith, J.L. (1998) Enzymes utilizing glutamine as an

amide donor. Adv. Enzymol. Relat. Areas Mol. Biol. 72, 87–144.

7. Hatse, S., De Clercq, E. & Balzarini, J. (1999) Role of anti-

metabolites of purine and pyrimidine nucleotide metabolism in

tumor cell differentiation. Biochem. Pharmacol. 58, 539–555.

8. Kennedy, E.P. (1986) The biosynthetsis of phospholipids. In

Lipids and Membranes: Past, Present and Future (Op den Kamp,

J.A.F., Roelofsen, B. & Wirtz, K.W.A., eds), pp. 171–206. Else-

vier. Scientific Publishers, Amsterdam.

9. Kensler, T.W. & Cooney, D.A. (1989) Inhibitors of the de novo

pyrimidine pathway. In Design of Enzyme Inhibitors as Drugs

(Sandler, M. & Smith, H.J., eds), pp. 379–401. Oxford University

Press, New York.

10. De Clercq, E. (1993) Antiviral agents: characteristic activity

spectrum depending on the molecular target with which they

interact. Adv. Virus. Res. 42, 1–55.

11. Hendriks, E.F., O’Sullivan,W.J. & Stewart, T.S. (1998)Molecular

cloning and characterization of the Plasmodium falciparum cyti-

dine triphosphate synthetase gene. Biochim. Biophys. Acta 1399,

213–218.

12. Hofer, A., Steverding, D., Chabes, A., Brun, R. & Thelander, L.

(2001) Trypanosoma brucei CTP synthetase: a target for the

treatment of African sleeping sickness. Proc. Natl Acad. Sci. USA

98, 6412–6416.

13. Lim, R.L., O’Sullivan, W.J. & Stewart, T.S. (1996) Isolation,

characterization and expression of the gene encoding cytidine

triphosphate synthetase from Giardia intestinalis. Mol. Biochem.

Parasitol. 78, 249–257.

14. Verschuur, A.C., Van Gennip, A.H., Leen, R., Voute, P.A.,

Brinkman, J. & Van Kuilenburg, A.B. (2002) Cyclopentenyl

cytosine increases the phosphorylation and incorporation into

DNA of 1-b-D-arabinofuranosyl cytosine in a human T-lym-

phoblastic cell line. Int. J. Cancer 98, 616–623.

15. Gao, W.Y., Johns, D.G. &Mitsuya, H. (2000) Potentiation of the

anti-HIV activity of zalcitabine and lamivudine by a CTP synthase

inhibitor, 3-deazauridine. Nucleosides Nucleotides Nucl. Acids 19,

371–377.

16. Bearne, S.L., Hekmat, O. &Macdonnell, J.E. (2001) Inhibition of

Escherichia coli CTP synthase by glutamate gamma-semialdehyde

and the role of the allosteric effector GTP in glutamine hydrolysis.

Biochem. J. 356, 223–232.

17. Levitzki, A. & Koshland, D.E. Jr (1972) Role of an allosteric

effector. Guanosine triphosphate activation in cytosine triphos-

phate synthetase. Biochemistry 11, 241–246.

18. Long, C.W. & Pardee, A.B. (1967) Cytidine triphosphate syn-

thetase of Escherichia coli B. I. Purification and kinetics. J. Biol.

Chem. 242, 4715–4721.

19. Levitzki, A. & Koshland, D.E. Jr (1969) Negative cooperativity

in regulatory enzymes. Proc. Natl Acad. Sci. USA 62, 1121–

1128.

20. Levitzki, A. & Koshland, D.E. Jr (1972) Ligand-induced dimer-

to-tetramer transformation in cytosine triphosphate synthetase.

Biochemistry 11, 247–253.

21. Tesmer, J.J., Klem, T.J., Deras, M.L., Davisson, V.J. & Smith,

J.L. (1996) The crystal structure of GMP synthetase reveals a

novel catalytic triad and is a structural paradigm for two enzyme

families. Nat. Struct. Biol. 3, 74–86.

22. Thoden, J.B., Holden, H.M., Wesenberg, G., Raushel, F.M. &

Rayment, I. (1997) Structure of carbamoyl phosphate synthetase:

a journey of 96 A from substrate to product. Biochemistry 36,

6305–6316.

23. Chittur, S.V., Klem, T.J., Shafer, C.M. & Davisson, V.J. (2001)

Mechanism for acivicin inactivation of triad glutamine amido-

transferases. Biochemistry 40, 876–887.

24. Willemoes, M. (2003) Thr-431 and Arg-433 are part of a con-

served sequence motif of the glutamine amidotransferase domain

of CTP synthases and are involved in GTP activation of the

Lactococcus lactis enzyme. J. Biol. Chem. 278, 9407–9411.

25. Wylie, J.L., Wang, L.L., Tipples, G. & McClarty, G. (1996) A

single point mutation in CTP synthetase ofChlamydia trachomatis

confers resistance to cyclopentenyl cytosine. J. Biol. Chem. 271,

15393–15400.

26. Whelan, J., Phear, G., Yamauchi, M. & Meuth, M. (1993) Clus-

tered base substitutions in CTP synthetase conferring drug

resistance in Chinese hamster ovary cells. Nature Genet. 3, 317–

322.

27. Ostrander, D.B., O’Brien, D.J., Gorman, J.A. & Carman, G.M.

(1998) Effect of CTP synthetase regulation by CTP on phospho-

lipid synthesis in Saccharomyces cerevisiae. J. Biol. Chem. 273,

18992–19001.

28. Iyengar, A. & Bearne, S.L. (2003) Aspartate 107 and leucine 109

facilitate efficient coupling of glutamine hydrolysis to CTP

synthesis by E. coli CTP synthase. Biochem. J. 369, 497–507.

29. Kelley, L.A., MacCallum, R.M. & Sternberg, M.J. (2000)

Enhanced genome annotation using structural profiles in the

program 3D-PSSM. J. Mol. Biol. 299, 499–520.

30. Garnier, J., Gibrat, J.F. & Robson, B. (1996) GOR method for

predicting protein secondary structure from amino acid sequence.

Methods Enzymol. 266, 540–553.

31. Guermeur, Y. (1997) Combinaison de Classifieurs Statistiques:

application a la Prediction de la Structure Secondaire des

Proteines, PhD Thesis, Universite Pierre et Marie Curie, Paris,

France.

32. Cuff, J.A. & Barton, G.J. (2000) Application of enhanced multiple

sequence alignment profiles to improve protein secondary struc-

ture prediction. In Proteins. 40, 502–511.

� FEBS 2003 Limited proteolysis of CTP synthase (Eur. J. Biochem. 270) 2205

33. Frishman, D. & Argos, P. (1996) Incorporation of non–local

interactions in protein secondary structure prediction from the

amino acid sequence. Protein Eng. 9, 133–142.

34. McGuffin, L.J., Bryson, K. & Jones, D.T. (2000) The

PSIPRED protein structure prediction server. Bioinformatics 16,

404–405.

35. Baldi, P., Brunak, S., Frasconi, P., Soda, G. & Pollastri, G. (1999)

Exploiting the past and the future in protein secondary structure

prediction. Bioinformatics 15, 937–946.

36. Higgins, D., Thompson, J., Gibson, T., Thompson, J.D., Higgins,

D.G. & G. (1994) CLUSTAL W: improving the sensitivity of

progressive multiple sequence alignment through sequence

weighting, position-specific gap penalties and weight matrix

choice. Nucl. Acids Res. 22, 4673–4680.

37. Jencks, W.P. & Regenstein, J. (1968) Ionization constants

of acids and bases. In Handbook of Biochemistry (Sober,

H.A., ed.), pp. J150–189. The Chemical Rubber Co, Cleveland,

Ohio.

38. Iyengar, A. & Bearne, S.L. (2002) An assay for cytidine 5¢-triphosphate synthetase glutaminase activity using high perfor-

mance liquid chromatography. Anal. Biochem. 308, 396–400.

39. Cornish-Bowden, A. (1995) Fundamentals of Enzyme Kinetics.

Portland Press Ltd, London.

40. Wilson, K.J. & Yuan, P.M. (1989) Protein and peptide mod-

ification. In Protein Sequencing: a Practical Approach (Findlay,

J.B.C. & Geisow, M.J., eds), pp. 13. Oxford University Press,

New York.

41. Bohm, G., Muhr, R. & Jaenicke, R. (1992) Quantitative analysis

of protein far UV circular dichroism spectra by neural networks.

Protein Eng. 5, 191–195.

42. Robertson, J.G. (1995) Determination of subunit dissociation

constants in native and inactivated CTP synthetase by sedimen-

tation equilibrium. Biochemistry 34, 7533–7541.

43. Hongo, S. & Sato, T. (1983) Some molecular properties of

asparagine synthetase from rat liver. Biochim. Biophys. Acta 742,

484–489.

44. Carrey, E.A. (1986) Nucleotide ligands protect the inter-domain

regions of the multifunctional polypeptide CAD against limited

proteolysis, and also stabilize the thermolabile part-reactions of

the carbamoyl-phosphate synthase II domains within the CAD

polypeptide. Biochem. J. 236, 327–335.

45. Guadalajara, A., Grisolia, S. & Rubio, V. (1987) Limited pro-

teolysis reveals low-affinity binding of N-acetyl-L-glutamate to

rat-liver carbamoyl-phosphate synthetase (ammonia). Eur. J.

Biochem. 165, 163–169.

46. Marshall, M. & Fahien, L.A. (1988) Proteolysis as a probe of

ligand-associated conformational changes in rat carbamyl phos-

phate synthetase I. Arch. Biochem. Biophys. 262, 455–470.

47. Evans, D.R. & Balon, M.A. (1988) Controlled proteolysis of

ammonia-dependent carbamoyl-phosphate synthetase I from

Syrian hamster liver. Biochim. Biophys. Acta 953, 185–196.

48. Kim, H., Kelly, R.E. & Evans, D.R. (1992) The structural orga-

nization of the hamster multifunctional protein CAD. Controlled

proteolysis, domains, and linkers. J. Biol. Chem. 267, 7177–7184.

49. Mareya, S.M. & Raushel, F.M. (1995) Mapping the structural

domains of E. coli carbamoyl phosphate synthetase using limited

proteolysis. Bioorg. Med. Chem. 3, 525–532.

50. Guy, H.I., Bouvier, A. & Evans, D.R. (1997) The smallest car-

bamoyl-phosphate synthetase. A single catalytic subdomain cat-

alyzes all three partial reactions. J. Biol. Chem. 272, 29255–29262.

51. Walker, M.S. & DeMoss, J.A. (1986) Organization of the func-

tional domains of anthranilate synthase from Neurospora crassa.

Limited proteolysis studies. J. Biol. Chem. 261, 16073–16077.

52. Denisot, M.A., Le Goffic, F. & Badet, B. (1991) Glucosamine-6-

phosphate synthase from Escherichia coli yields two proteins upon

limited proteolysis: identification of the glutamine amidohydrolase

and 2R ketose/aldose isomerase-bearing domains based on their

biochemical properties. Arch. Biochem. Biophys. 288, 225–230.

53. Anderson, P.M. (1983) CTP synthetase from Escherichia coli: an

improved purification procedure and characterization of hysteretic

and enzyme concentration effects on kinetic properties. Biochem-

istry 22, 3285–3292.

54. Robertson, J.G., Sparvero, L.J. & Villafranca, J.J. (1992)

Inactivation and covalent modification of CTP synthetase by

thiourea dioxide. Protein Sci. 1, 1298–1307.

55. Lewis, D.A. & Villafranca, J.J. (1989) Investigation of the

mechanism of CTP synthetase using rapid quench and isotope

partitioning methods. Biochemistry 28, 8454–8459.

2206 D. Simard et al. (Eur. J. Biochem. 270) � FEBS 2003