A tRNA splicing operon: Archease endows RtcB with dual GTP/ATP cofactor specificity and accelerates RNA ligation

A tRNA splicing operon Archease endows RtcBwith dual GTPATP cofactor specificity andaccelerates RNA ligationKevin K Desai1 Chin L Cheng1 Craig A Bingman1 George N Phillips Jr12 and

Ronald T Raines13

1Department of Biochemistry University of WisconsinndashMadison Madison WI 53706 USA 2Department ofBiochemistry and Cell Biology and Department of Chemistry Rice University Houston TX 77005 USA and3Department of Chemistry University of WisconsinndashMadison Madison WI 53706 USA

Received October 27 2013 Revised December 12 2013 Accepted December 13 2013

ABSTRACT

Archease is a 16-kDa protein that is conserved in allthree domains of life In diverse bacteria andarchaea the genes encoding Archease and thetRNA ligase RtcB are localized into an operonHere we provide a rationale for this operon organ-ization by showing that Archease and RtcB fromPyrococcus horikoshii function in tandem withArchease altering the catalytic properties of theRNA ligase RtcB catalyzes the GTP and Mn(II)-de-pendent joining of either 2030-cyclic phosphate or30-phosphate termini to 50-hydroxyl termini We findthat catalytic concentrations of Archease are suffi-cient to activate RtcB and that Archease acceler-ates both the RNA 30-P guanylylation and ligationsteps In addition we show that Archease canalter the NTP specificity of RtcB such that ATPdGTP or ITP is used efficiently Moreover RtcBvariants that have inactivating substitutions in theguanine-binding pocket can be rescued by theaddition of Archease We also present a 14 A-reso-lution crystal structure of P horikoshii Archeasethat reveals a metal-binding site consisting ofconserved carboxylates located at the protein tipSubstitution of the Archease metal-bindingresidues drastically reduced Archease-dependentactivation of RtcB Thus evolution has sought toco-express archease and rtcB by creating a tRNAsplicing operon

INTRODUCTION

Archease is a small (16 kDa) acidic protein that isconserved in eukarya bacteria and archaea (1) Thearchease gene generally localizes adjacent to genes

encoding enzymes and other proteins involved in DNAor RNA processing (12) including the tRNA ligaseRtcB The conservation of genomic context for archeasesuggests that Archease could function broadly as a modu-lator or chaperone of nucleic acid modifying proteins (1)In accord with a suspected role in assisting nucleic acidprocessing enzymes Archease has been shown to increasethe specificity of a tRNA m5C methyltransferase (2) Thecommon organization of rtcB and archease into an operonsuggests that Archease could also function to modulatethe activity of RtcB In a preliminary report Martinezhas put forth human Archease as an activator of humanRtcB (3)The RNA ligase RtcB catalyzes the GTP and

Mn(II)-dependent joining of 2030-cyclic phosphate or30-phosphate termini to 50-hydroxyl termini (4ndash12) RtcBis an essential enzyme for the ligation of tRNAs inmetazoa (4) and likely archaea (513) after intronremoval by the tRNA splicing endonuclease (1415)Ligation proceeds through three nucleotidyl transfersteps with 2030-cyclic phosphate termini being hydrolyzedto 30-P termini in a step that precedes 30-P activation withGMP (7910) (Figure 1A) In the first nucleotidyl transferstep RtcB reacts with GTP to form a covalent RtcBndashhistidinendashGMP intermediate and release PPi in thesecond step the GMP moiety is transferred to the RNA30-P in the third step the 50-OH from the opposite RNAstrand attacks the activated 30-P to form a 3050-phospho-diester bond and release GMPWe sought to discover whether Archease and

RtcB encoded within the archaeon Pyrococcus horikoshiicould function in tandem We envisioned that thehyperthermophilic archaeal proteins would be an excellentmodel system due to their ease of purification and theirinherent conformational stability Moreover the highconservation of Archease and RtcB across the threedomains of life suggests that studies on the archaealproteins would be broadly applicable (15) Here we

To whom correspondence should be addressed Tel +1 608 262 8588 Fax +1 608 890 2583 Email rtraineswiscedu

Published online 16 January 2014 Nucleic Acids Research 2014 Vol 42 No 6 3931ndash3942doi101093nargkt1375

The Author(s) 2014 Published by Oxford University PressThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby30) whichpermits unrestricted reuse distribution and reproduction in any medium provided the original work is properly cited

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

show that Archease and RtcB from P horikoshii do indeedfunction in unison and that Archease affects the catalyticproperties of RtcB Archease expands the cofactor speci-ficity of RtcB enabling the efficient use of dGTP ATP orITP We also demonstrate that Archease accelerates theRNA 30-P activation and ligation steps of catalysis byRtcB In addition our structural and mutationalanalyses of Archease identify an essential metal-bindingsite consisting of conserved carboxylate groups locatedat a tip of the protein Together these data demonstratethe existence of a tRNA splicing operon

MATERIALS AND METHODS

Archease production and purification

The gene encoding P horikoshii Archease (AccessionO59205) was synthesized by Integrated DNATechnologies using codons optimized for expression inEscherichia coli The gene was cloned between the SphIand BamHI recognition sites of vector pQE70-lacI

(Qiagen) Cloning into the SphI site required insertion ofa codon that begins with a cytosine immediately after theATG start codon A leucine codon was therefore insertedafter ATG however throughout the manuscript wemaintain the amino-acid numbering corresponding tonative P horikoshii Archease The resulting plasmid wastransformed into the BL21 strain of E coli and theArchease protein was produced by growing cells inTerrific Broth medium to an OD600 of 07 inducing geneexpression by adding IPTG to 05mM and continuinggrowth for 3 h at 37C Cells were harvested by centrifu-gation and resuspended in 8ml per gram of wet pellet inbuffer A (50mM TrisndashHCl buffer pH 86 containing50mM NaCl) Cells were lysed by passage through a celldisruptor (Constant Systems) at 20 000 psi and the lysatewas clarified by centrifugation at 20 000g for 1 h Bacterialproteins were precipitated and removed by incubating thelysate at 70C for 25min followed by centrifugation at20 000g for 20min The clarified lysate was then loadedonto a 5-ml HiTrap Q XL anion-exchange column (GEHealthcare) The column was washed with 50ml of buffer

Figure 1 The three nucleotidyl transfer steps of catalysis by RtcB a putative tRNA splicing operon and the titration of Archease into RNA ligationreactions with RtcB (A) The three nucleotidyl transfer steps of catalysis by RtcB are (1) RtcB guanylylation (2) RNA 30-P guanylylation and (3)phosphodiester bond formation (B) The operon organization of rtcB and archease in diverse bacteria (Pelobacter propionicus Synechococcus spJA-3-3Ab and Syntrophus aciditrophicus) and archaea (Halobacterium sp NRC-1 Methanosaeta thermophila P horikoshii and Thermococcuskodakarensis) (C) RtcB-catalyzed RNA ligation reactions titrated with increasing concentrations of Archease as specified Reaction mixturescontained 50mM BisndashTris buffer (pH 70) NaCl (300mM) MnCl2 (025mM) GTP (010mM) P horikoshii RtcB (5 mM) 50 RNA fragment(10 mM) and 30 RNA fragment (10 mM) (RNA substrates are shown at top) Reaction mixtures were incubated at 70C for 30min andquenched with an equal volume of RNA gel-loading buffer The reaction products were resolved by electrophoresis through an 18 wv ureandashpolyacrylamide gel and visualized by fluorescence scanning of the FAM label (D) Ligation product (nM) plotted versus Archease concentration(nM) Values are the meanplusmnSE for three separate experiments

3932 Nucleic Acids Research 2014 Vol 42 No 6

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

A and Archease was eluted with a NaCl gradient of bufferA (50mMndash10M) over 20 column volumes Fractionscontaining purified Archease were dialyzed against 2 l ofbuffer (10mM HEPESndashNaOH pH 75 containing200mM NaCl) overnight at 4C and the protein wasflash-frozen in liquid nitrogen and stored at 80CConcentrations of RtcB and Archease were calculatedfrom A280 values and calculated (ExPASy) extinction co-efficients of e280=62 340M1 cm1 and 19 940M1 cm1respectively

RNA ligation and guanylylation assays

Native P horikoshii RtcB was produced and purifiedas described previously (12) Ligation reaction mixturesincluded a 10-nt 50 RNA fragment and a 10-nt 30

RNA fragment The 50 RNA fragment had a6-carboxyfluorescein (FAM) label on its 50 end and wasphosphorylated on its 30 end The 30 RNA fragment hadhydroxyl groups on each end The sequence of the50 fragment was FAM-50-AAAUAACAAA-30-P and thesequence of the 30 RNA fragment was 50-AAAUAACAAA-30 Ligation reactions were performed at 70C in 25 or50 ml solutions consisting of 50mM BisndashTris buffer(pH 70) containing NaCl (300mM) MnCl2 (025mM)NTP (010mM) P horikoshii RtcB (5 mM) 50 RNAfragment (10 mM) and 30 RNA fragment (10 mM) RNA30-P guanylylation reactions were performed in identicalreaction mixtures except the RNA ligation substrates werereplaced with a 50 RNA fragment (10 mM) that had a 20-Freplacing the 20-OH that is vicinal to the terminal 30-PReactions were quenched by the addition of an equalvolume of RNA gel-loading buffer (5 TBE containing7M urea 20 vv glycerol and 15mgml blue dextran)Reaction products were separated on an 18 wv ureandashpolyacrylamide gel and the RNA was visualized by fluor-escence scanning with a Typhoon FLA9000 imager(GE Healthcare) Product quantification was performedusing ImageQuant TL (GE Healthcare)

[14C]GTP binding assays

Binding assays were performed in 250 ml solutionsconsisting of 50mM BisndashTris buffer (pH 70) containingNaCl (300mM) MnCl2 (20mM) and [8-14C]GTP(10mM) (Moravek Biochemicals) RtcB was assayed ata concentration of 30 mM and Archease was assayed at100 mM After incubation at 70C free GTP was removedby applying the reaction mixture to three 5-ml HiTrapdesalting columns (GE Healthcare) connected in seriesThe desalting columns were equilibrated with elutionbuffer (50mM HEPESndashNaOH buffer pH 75 containing200mM NaCl) and protein was eluted in 05-ml fractionsAbsorbance readings at 260 and 280 nm were obtained foreach fraction The protein fractions have high A280

readings whereas the fractions with free GTP have highA260 readings The concentration of [8-14C]GTP in theprotein fractions was determined by liquid scintillationcounting Each 05-ml fraction was mixed with 35ml ofUltima Gold MV liquid scintillation cocktail (PerkinElmer) in a 4-ml vial and radioactivity was quantifiedwith a MicroBeta TriLux liquid scintillation counter

(Perkin Elmer) The concentration of GTP in eachfraction was determined by comparing the counts perminute in these samples to that in standards of knownconcentration Although we are assaying for GTPbinding to RtcB we are assuming that the radioactivityremaining bound to RtcB is bound covalently in the formof RtcBndashpG as has been shown previously (1112)

Archease crystallization data collection andstructure determination

Selenomethionine labeled (Se-Met) Archease wasproduced in the E coli methionine auxotroph strainB834 in auto-induction medium (16) Archease waspurified as described above concentrated to 29mM(48mgml) by ultrafiltration using a spin concentrator(5000 MWCO Amicon) The protein was flash-frozen inliquid nitrogen and stored at 80C Archease crystalsdiffracted to a higher resolution when MgCl2 (115mM)was added to the protein solution before crystallizationArchease was crystallized using the hanging drop vapordiffusion method Crystals were grown by mixing 1 ml ofprotein solution with 1 ml of reservoir solution consistingof sodium acetate (010M pH 45) (+)-2-methyl-24-pentanediol (40 vv) and CaCl2 (10mM) Trays wereincubated at 20C and crystals appeared within 1 weekCrystals were harvested cryoprotected in MiTeGen lowviscosity cryo oil and flash-frozen in liquid nitrogen X-raydiffraction data were collected at the General Medicineand Cancer Institutes Collaborative Access Team(GMCA-CAT) at Argonne National Laboratory Datasets were collected at the selenium peak and edge wave-lengths from a single crystal Data sets were indexed andscaled using HKL2000 (17) The structure was solvedusing Phenix Autosol (18) and completed using alternatingrounds of manual model building using COOT (19) andrefinement with phenixrefine (18) Structure quality wasassessed with MolProbity (20) and figures were generatedusing PyMOL (21) Omit maps were calculated withPhenix

Sedimentation equilibrium analysis

Sedimentation equilibrium studies were performed with aBeckman Optima XL-A analytical ultracentrifuge in theBiophysics Instrumentation Facility at the University ofWisconsinndashMadison The identical Se-Met Archease prep-aration used for crystallography was analyzed at concen-trations of 84 167 and 25mM in HEPESndashNaOH(10mM pH 75) containing NaCl (200mM) A runthat included MnCl2 (025mM) was also performedEquilibrium data were collected at multiple speedsat 20C

RESULTS

Archease activates RtcB

Genomic context analysis shows that the archease gene ismost commonly located directly adjacent to the rtcB geneand the two genes are localized into an operon in diversebacteria and archaea (2223) (Figure 1B) Moreover the

Nucleic Acids Research 2014 Vol 42 No 6 3933

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

gene encoding a tRNA m5C methyltransferase is some-times found in the same operon as rtcB and archeaseproviding an additional clue that archease is involved intRNA maturation (2) To investigate a possible functionalinteraction between Archease and RtcB fromP horikoshii the archease gene was synthesized usingcodons optimized for expression in E coli SolubleArchease was produced at high levelsmdashwe isolated80mg of protein routinely from a 1 -l culture (seelsquoMaterials and Methodsrsquo section) To assay for RtcB-catalyzed RNA ligation we used two 10-nt RNA sub-strates that can be joined to produce a 20-nt productThe 50 RNA half contains a 50 fluorescent label and a30-P whereas the 30 RNA half has hydroxyl groups ateach terminus (Figure 1C) When Archease was titratedinto RNA ligation reaction mixtures containing 5 mMRtcB and 10 mM of each RNA half the extent of RtcBactivation leveled off at an Archease concentration of100 nM (Figure 1C and D) An Archease concentrationof 100 nM enabled 5 mM RtcB to produce 316 nM ofligation product whereas only 92 nM of product wasformed in the absence of Archease (Figure 1D) Thus100 nM Archease enabled 5 mM RtcB to produce 224 nMmore ligation product The ability of a sub-stoichiometricratio of ArcheaseRtcB to enable the formation of add-itional ligation product that exceeds the concentration ofArchease suggests that Archease acts catalyticallyReactions to determine the pH-dependence of RtcB acti-vation by Archease showed that maximal activationoccurs at pH 70 (Supplementary Figure S1) Controlreactions that did not include RtcB demonstrated thatArchease does not have RNA 30-P guanylylation orligation activity (data not shown)

Effect of Archease on RNA ligation kinetics

Next we examined the effect of Archease on the overallrate of the three-step RtcB ligation pathway by monitor-ing the rate of ligation over time under single-turnoverconditions Reaction mixtures containing 5 mM RtcBalone and including 100 nM Archease were incubated at70C and aliquots were removed and quenched at varioustime intervals (Figure 2A) Plots of the concentrationof ligation product formed over time were fitted to asingle-exponential to obtain apparent rate constants of(0011plusmn0001) min1 for RtcB alone and (0033plusmn0001) min1 with the inclusion of Archease (Figure 2B)Thus 100 nM Archease accelerates the overall three-stepRtcB ligation pathway by 3-fold under our reaction con-ditions Despite having RtcB in 5-fold excess over RNAsubstrates the ligation reaction with RtcB alone went toonly 43 completion and the reaction with Archeaseincluded went to only 49 completion (Figure 2B) Theability for only 10 of RtcB molecules to catalyze aligation event suggests that RtcB has a preferred RNAsubstrate or that the system requires unknown compo-nents for increased ligation efficiency

Effect of Archease on RNA 30-P guanylylation kinetics

Then we analyzed the effect of Archease on the rateof RNA 30-P guanylylation To monitor RNA 30-P

guanylylation directly we envisioned that we could firstreact RtcB with GTP and Mn(II) to form the RtcBndashpGintermediate and also exclude an RNA 50-OH terminusfrom the reaction thus eliminating the first andthird nucleotidyl transfer steps respectively We firstdemonstrated that Archease could activate RNA ligationwhen RtcB is replaced with RtcBndashpG Using [14C]GTPwe found that RtcB is converted fully to RtcBndashpG within5min when incubated with GTP (10mM) and MnCl2(20mM) at 70C (see lsquoMaterials and Methodsrsquo section)The RtcBndashpG intermediate was then subjected to theaction of Archease in RNA ligation reactions Weobserved that substituting RtcB with the preformed

Figure 2 Effect of Archease on the single-turnover rate of RNAligation by RtcB (A) RNA ligation reactions with RtcB alone orwith the inclusion of 100 nM Archease Reaction mixtures contained50mM BisndashTris buffer (pH 70) NaCl (300mM) MnCl2 (025mM)GTP (010mM) P horikoshii RtcB (5mM) 50 RNA fragment (10 mM)and 30 RNA fragment (10 mM) Reaction mixtures were incubated at70C and aliquots were removed at the indicated times and quenchedwith an equal volume of RNA gel-loading buffer (B) Plots of ligationproduct formation over time fitted to a single-exponential equationValues are the meanplusmnSE for two separate experiments

3934 Nucleic Acids Research 2014 Vol 42 No 6

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

RtcBndashpG intermediate had little effect on Archease-dependent activation (Supplementary Figure S2)

The activated RNAndashppG intermediate could becaptured stably if RNA with a 20-F30-P terminus wasused as the guanylylation substrate thus facilitatingkinetic analyses RNA 30-P guanylylation reactionmixtures containing 10mM RNA and 5 mM RtcB aloneand including 100 nM Archease were incubated at 70Cand aliquots were removed and quenched at various timeintervals (Figure 3A) The concentration of RNAndashppGproduct formed over time was plotted and fitted to asingle-exponential to obtain apparent rate constantsof (0018plusmn0001) min1 for RtcB alone and(0063plusmn0003) min1 in the presence of 100 nMArchease (Figure 3B) Thus Archease at 100 nM acceler-ates the RNA 30-P guanylylation step by 35-fold underour reaction conditions Significantly Archease enabledthe RNA 30-P guanylylation reaction to reach near com-pletion (gt90) whereas the reaction in the absence ofArchease went to only 54 completion In the absenceof Archease only 10 of RtcB molecules are competentfor catalyzing RNA 30-P guanylylation a finding similarto our observed ligation efficiency The rate constantsobtained for RNA 30-P guanylylation are faster thanthose obtained for the overall ligation pathway consistentwith RNAndashppG being a kinetically competent intermedi-ate Considering that formation of the RtcBndashpGintermediate is fast comparison of the apparent rate con-stants demonstrates that the ligation step is rate-limitingunder the tested reaction conditions Thus Archease ac-celerates both the second and the third nucleotidyl transfersteps of catalysis by RtcB

The apparent stalling of the RNA 30-P guanylylationreaction in the absence of Archease suggests that not allthe RtcB molecules are active for the second nucleotidyltransfer step or that RtcB arrests after the histidineguanylylation step The increase in reaction completionwhen Archease is included could indicate that Archeasefunctions by activating stalled RtcB molecules The effectof Archease on reaction completion should become moreapparent when the concentration of RtcB is less than theconcentration of substrate RNA In reactions with 05mMRtcB and 10 mM RNA 30-P the reaction rate was sluggishand only 30 of RtcB molecules catalyzed a singleturnover after 3 h at 70C (Figure 3C) When Archeasewas included 05 mM RtcB produced 065mM RNAndashppG within 2 h consistent with 100 of enzyme mol-ecules catalyzing at least one turnover Moreover thereaction in the absence of Archease was activatedrapidly on Archease addition (Figure 3C) This experi-ment demonstrated that each RtcB molecule is competentfor catalyzing RNA 30-P guanylylation when Archease isincluded

The formation of activated RNA 30-P at a concentra-tion that exceeds the RtcB concentration suggests that theligase can release the activated RNA into solution TheRNAndashppG released into solution would be expected toform a 2030-cyclic phosphate (RNAgtp) rapidly via intra-molecular attack by the terminal 20-OH releasing GMPThe RNAgtp could again be a competent substrate forRtcBndashpG In contrast the release of activated RNA or

Figure 3 Effects of Archease on the rate of RtcB-catalyzedguanylylation of RNA with a 20-F30-P terminus RtcB was pre-guanylylated by incubation with GTP and Mn(II) and the 30 RNAfragment was not included to prevent ligation (A) The guanylylationrate of a 20-F30-P RNA terminus by RtcB alone or with the inclusionof 100 nM Archease RNA guanylylation reaction mixtures contained50mM BisndashTris buffer (pH 70) NaCl (300mM) MnCl2 (025mM)GTP (010mM) P horikoshii RtcB (5mM) and RNA substrate(10 mM) (RNA substrate is shown at top) Reaction mixtures wereincubated at 70C and aliquots were removed at the indicated timesand quenched with an equal volume of RNA gel-loading buffer(B) Plots of RNAndashppG product formation over time fitted to asingle-exponential equation (C) Plots of RtcB-catalyzed RNAndashppGproduct formation over time in reactions that had the 20-F30-P RNAsubstrate in excess (05 mM RtcB and 10 mM RNA substrate) Archease(100 nM) was added where indicated Values in the plots are themeanplusmnSE for two separate experiments

Nucleic Acids Research 2014 Vol 42 No 6 3935

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

DNA by classical ATP-dependent ligases poses a greatchallenge to cells (2425) The released adenylylatedRNA or DNA intermediate is not a ligase substratebecause the ligase reacts quickly with ATP to formligasendashpA thus occupying the AMP-binding pocketHence essential repair pathways are needed to removethe nucleic acid 50-adenylyl group so as to regenerate acompetent ligase substrate (25)

Archease expands the NTP cofactor specificity of RtcB

While investigating a possible Archease-dependent effect onthe first nucleotidyl transfer step of catalysis by RtcB wefirst discovered that Archease enhances the weak dGTP-dependent ligation activity of RtcB Then we wonderedwhether Archease could also enable RtcB to use NTPswith different nucleobases We tested ATP dATP ITPCTP and UTP at 010mM in ligation reactions withRtcB In the absence of Archease RtcB-catalyzed RNAligation proceeds efficiently with GTP and substantiallyless efficiently with dGTP and ITP (Figure 4A and B)Notably inclusion of Archease in ligation reactionmixtures enabled efficient utilization of GTP dGTP ATPor ITP (Figure 4A and B) Thus Archease enabled RtcB touse all tested purine nucleobases though not the pyrimi-dines CTP and UTP We also observed that the amount ofligated product obtained with ATP was greatest whenthe Archease concentration was increased to 800 nM(Figure 4C) Archease increased the amount of ligationproduct formed with the cofactors dGTP and ITP by 10 -and 8-fold respectively during incubation of the reactionmixtures at 70C for 30min The ability for RtcB ligationto proceed with ATP is confounding given the hydrogen-bonding constraints to the nucleobase apparent in thecrystal structure of the RtcBndashpG intermediate (1112)(Figure 4D) Importantly we did not observe binding of[14C]GTP to Archease suggesting that it is unable to bindand deliver purine NTPs directly to the RtcB active site (seelsquoMaterials and Methodsrsquo section)The NTP dependence of RtcB and RtcBmiddotArchease

under single-turnover conditions appeared to followMichaelisndashMenten behavior allowing determination ofNTP cofactor Michaelis constant values (KM is definedas the apparent KM value under single-turnover reactionconditions Figure 4EndashI) The GTP KM values for RtcBand RtcBmiddotArchease are (24plusmn02) and (15plusmn03) mMrespectively The dGTP and ITP KM values forRtcBmiddotArchease are (18plusmn01) and (19plusmn04) mM re-spectively We were unable to determine KM values fordGTP and ITP in the absence of Archease owing to theirlow reactivity Nevertheless the ATP KM value forRtcBmiddotArchease is (34plusmn11) mM demonstrating efficientuse of this cofactor The ATP KM values for human DNAligase 1 and T4 DNA ligase are 12 and 14 mM respectively(2426) The apparent rate constant for the overall ligationpathway with ATP including 800 nM Archease is(0024plusmn0001) min1 (Figure 4J)

Crystal structure of P horikoshii Archease

To gain greater insight into the mechanism of Archeaseaction we solved an x-ray crystal structure of P horikoshii

Archease Archease crystals diffracted to a resolution of14 A and contained four protein molecules per asymmet-ric unit (Supplementary Table S1) Archease appeared tocrystallize as a dimer of dimers however sedimentationequilibrium results were consistent with the existence of apredominantly monomeric species in both the presenceand absence of Mn(II) (Figure 5A) Archease subunits Aand B are essentially identical to subunits C and D re-spectively The structure of each subunit consists of twostrandndashhelixndashstrand domains each consisting of a three-strand core and a single helix in a strandndashhelixndashstrandconfiguration (27) The crystallization buffer includedCaCl2 and the omit density map showed two Ca(II) ionsper asymmetric unit with each Ca(II) ion bound to asingle Archease subunit in octahedral coordinationgeometry (Supplementary Figure S3) The metal-bindingsites are identical and consist of two strictly conservedaspartate residues the C-terminal carboxylate group andthree water molecules (Figure 5B and SupplementaryFigure S4) The metal-binding site is located at the inter-face between two subunits subunit A binds a Ca(II) ionwhile the analogous residues in subunit B are not in aposition for metal binding (Figure 5B) The N and Ctermini of each Archease subunit are proximal and theN terminus exists as an extended protrusion that forms abeta sheet with its partner subunit The Ca(II)-binding siteresides at the base of the N-terminal protrusion on theprotein exterior Rotation of the Archease structure dem-onstrates that the small protein is slender spanning only20 A on one side (Figure 5C) The electrostatic surfacepotential of Archease is dominated by regions of negativecharge (Figure 5D)

Structure-guided mutagenesis of Archease

Site-directed mutagenesis of residues in the P horikoshiiArchease metal-binding site revealed their importance forthe Archease-dependent activation of RtcB The metal-binding variants D12A D141A and I142 drasticallyreduced the Archease-dependent activation of RtcB(Figure 6A and Supplementary Figure S5) The D12AArchease variant had the most diminished RtcB-activation activity Residue His9 is a highly conservedresidue in the N-terminal tail and Lys117 is a highlyconserved residue adjacent to the metal-binding siteSubstitution of each of these two residues also severelyreduced Archease-dependent activation of RtcB thoughan E8A substitution had no detrimental effect The effectsof the Archease substitutions in the metal-binding sitewere recapitulated when ATP was used as a cofactor(Figure 6B and C) In contrast to the observation withGTP the E8A variant of Archease had a substantialeffect on ATP utilization by RtcB

Archease rescues inactive RtcB variants that havesubstitutions in the guanine-binding pocket

If Archease is perturbing the binding of the NTPnucleobase we reasoned that residues interacting withguanine as observed in the RtcBndashpG crystal structurewould become irrelevant for catalysis by RtcB in thepresence of Archease To test this hypothesis active-site

3936 Nucleic Acids Research 2014 Vol 42 No 6

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

variants of RtcB were assayed in ligation reactions withand without the inclusion of Archease The RtcB amino-acid substitutions assayed were D65A D95A N202AH203A F204A E206A H404A and K480A Each of

these substitutions rendered RtcB alone inactive Yetwhen Archease was included the F204A E206A andK480A variants were rescuedmdashthey became active cata-lysts of the ligation reaction (Figure 7A) These three

Figure 4 RNA ligation reactions demonstrating the NTP cofactor specificity of RtcB and an active-site view of the RtcBndashpG intermediate(A) Reactions testing NTP cofactor specificity with RtcB alone or RtcB with 100 nM Archease NTP cofactors were tested at 010mM andreaction mixtures were incubated at 70C for 30min (B) Graph of the ligation product obtained for each NTP cofactor Values are themeanplusmnSE for two separate experiments (C) ATP-dependent RtcB-catalyzed RNA ligation reactions titrated with increasing concentrations ofArchease as specified Reaction mixtures were incubated at 70C for 20min Values are the meanplusmnSE for three separate experiments (D) Crystalstructure of the P horikoshii RtcBndashpG intermediate (PDB entry 4it0) illustrating the residues that contact the guanine nucleobase (EndashI) MichaelisndashMenten plots of reaction rate versus NTP cofactor concentration for RtcB-catalyzed RNA ligation reactions under single-turnover conditions Whereindicated Archease was included at a concentration of 100 nM for reactions with GTP dGTP and ITP while reactions with ATP included 800 nMArchease Values are the meanplusmnSE for three separate experiments (J) Single-turnover kinetics of ATP-dependent RNA ligation catalyzed by RtcBwith the inclusion of Archease (800 nM) Values are the meanplusmnSE for two separate experiments Ligation reaction mixtures contained 50mM BisndashTris buffer (pH 70) NaCl (300mM) MnCl2 (025mM) NTP as indicated P horikoshii RtcB (5mM) 50 RNA fragment (10 mM) and 30 RNAfragment (10 mM)

Nucleic Acids Research 2014 Vol 42 No 6 3937

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

variants had substitutions of residues that interact directlywith the guanine base in the RtcBndashpG structure (1112)(Figure 4D) The ligation activity of RtcB variants withsubstitutions of residues involved directly in bindingMn(II) (D95A and H203A) interacting with the triphos-phate moiety (N202A) and forming the histidinendashGMPcovalent bond (H404A and D65A) were unable to berescued by Archease We did however observe thatArchease rescued the RNA 30-P guanylylation activityof D95A RtcB (Figure 7A) a variant perturbed inthe binding site for the second Mn(II) ion Mn2(Figure 4D) The rescue of ligation activity with GTP asa cofactor was recapitulated with ATP as a cofactor

except for undetectable activity with F204A RtcB(Figure 7B and C) When assayed with ATP as acofactor K480A RtcB catalyzed the formation of2-fold more ligation product than did the wild-typeenzyme in the presence of Archease during incubation ofthe reaction mixtures at 70C for 30min This finding sug-gested that eliminating the clash between the amino groupof Lys480 and the adenine exocyclic amine facilitates ATPutilization by lowering its KM value Yet the ATP KMvalue for K480A RtcB was nearly identical to the ATPKM value for wild-type RtcB when 800 nM Archease wasincluded (Figure 7D and E) Instead the increased ligationproduct formed by K480A RtcB can be explained by an

Figure 5 Crystal structure of P horikoshii Archease (A) A cartoon representation of the four Archease subunits per asymmetric unit with the Ca(II)ions represented as spheres (B) The Ca(II) ion-binding site at the interface of subunits A and B The Ca(II) ion-binding residues are depicted assticks and the Ca(II) ion and water molecules are shown as spheres (C) Archease subunit A depicting the Ca(II) ion-binding site at the base of theN-terminal protrusion A turn of 90 demonstrates the slenderness of the protein (D) Electrostatic surface potential of Archease subunit A with blueand red indicating regions of positive and negative charge respectively

3938 Nucleic Acids Research 2014 Vol 42 No 6

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

increased reaction rate under single-turnover RNAligation conditions (Figure 7F) The apparent overallligation rate constant for K480A RtcB with ATP was(0032plusmn0004) min1 a value 33 greater than thatobtained for wild-type RtcB with ATP under identicalreaction conditions (Figure 4J)

DISCUSSION

We have demonstrated the evolutionary rationale fororganizing rtcB and archease into an operon in diversebacteria and archaea Likewise the E coli operon organ-ization of rtcA (RNA 30-terminal phosphate cyclase A)and rtcB was initially used as a rationale to test RtcBfor the ability to ligate RNA 2030-cyclic phosphate and50-OH termini (28) an activity which was confirmedlater (6) RtcB and Archease are highly conserved in allthree domains of life and RtcB is known to be the essen-tial catalytic component of a tRNA splicing complex inhumans (4) Our functional studies of Archease and RtcBfrom P horikoshii have shown that Archease modulatesall three nucleotidyl transfer steps during catalysis byRtcB It is especially notable that Archease convertsRtcB from a ligase that displays no ATP-dependentactivity to a ligase displaying low micromolar KMvalues for GTP dGTP ITP and ATP This demonstratesthat Archease affects binding not only to the nucleobasebut also to the ribose The ability for Archease to endowRtcB with altered NTP specificity as well as recover the

activity of RtcB variants with substitutions in the guanine-binding pocket suggests that there might be a purine-binding pocket on Archease We were however unableto detect direct binding of [14C]GTP to ArcheaseAlternatively a novel composite purine-binding pocketcould form on interaction of the two proteins orArchease could enforce an alternative nucleoside bindingconformation within the RtcB active site The small sizeand negative surface charge of Archease are consistentwith its binding in the positively charged cleft of RtcBwhich is also presumed to be the binding site for substrateRNA Elucidation of the mechanism of altered NTP spe-cificity will likely require obtaining a crystal structure ofthe RtcBndashpAmiddotArchease complexAlthough we have identified an essential metal-binding

site in Archease we cannot ascertain if that site binds toits own metal ion or to one of the two Mn(II) ions in theRtcB active site We do note that a metal-binding siteconsisting of absolutely conserved carboxylates locatedat the tip of Archease is reminiscent of GreB a bacterialRNA polymerase transcription factor (29) GreB is a small(185 kDa) protein that functions to rescue a stalled RNApolymerase complex RNA polymerase arrests when thetranscript 30-end loses base-pair contact with the DNAtemplate thereby disengaging the 30-OH from the activesite GreB stimulates endonucleolytic cleavage of the tran-script 30-end allowing RNA polymerase to restart GreBhas two conserved carboxylates at the tip of a coiled-coilthat are placed into the polymerase active site and stabilize

Figure 6 Structure-guided mutagenesis of conserved Archease residues (A B) Archease variants with alanine substitutions were tested for theirability to activate RtcB in RNA ligation reactions with GTP (A) or ATP (B) as a cofactor Reaction mixtures included 100 nM Archease wherespecified and were incubated at 70C for 30min (C) Graph of the ligation product obtained for each Archease variant Values are the meanplusmnSE fortwo separate experiments Ligation reaction mixtures contained 50mM BisndashTris buffer (pH 70) NaCl (300mM) MnCl2 (025mM) GTP or ATP(010mM) P horikoshii RtcB (5mM) 50 RNA fragment (10 mM) and 30 RNA fragment (10 mM)

Nucleic Acids Research 2014 Vol 42 No 6 3939

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

Figure 7 Effect of Archease on the activity of active-site variants of RtcB in RNA ligation assays with GTP or ATP as a cofactor (A) Reactionswith GTP (010mM) as a cofactor (B) Reactions with ATP (010mM) as a cofactor Reaction mixtures included 100 nM Archease where indicatedand were incubated at 70C for 30min (C) Graph of the ligation product obtained for each RtcB variant Values are the meanplusmnSE for twoseparate experiments (D) ATP-dependent K480A RtcB-catalyzed RNA ligation reactions titrated with increasing concentrations of Archease asspecified ATP was included at 010mM and reaction mixtures were incubated at 70C for 15min Values are the meanplusmnSE for three separateexperiments (E) MichaelisndashMenten plot of reaction rate versus ATP cofactor concentration for K480A RtcB-catalyzed RNA ligation reactions undersingle-turnover conditions Values are the meanplusmnSE for three separate experiments (F) Single-turnover kinetics of ATP-dependent RNA ligationcatalyzed by K480A RtcB with the inclusion of Archease (800 nM) Values are the meanplusmnSE for two separate experiments Ligation reactionmixtures contained 50mM BisndashTris buffer (pH 70) NaCl (300mM) MnCl2 (025mM) NTP as indicated P horikoshii RtcB (5mM) 50 RNAfragment (10 mM) and 30 RNA fragment (10 mM)

3940 Nucleic Acids Research 2014 Vol 42 No 6

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

binding of the polymerase Mg2 ion which stimulatesan intrinsic endonucleolytic transcript cleavage activity(2930)

Previously we showed that RtcB uses a two-metalmechanism during reaction with GTP a mechanism thatis similar among all nucleotidyl transferases includingRNA polymerase (1231) By analogy to GreB thelsquosticky fingerrsquo carboxylates of Archease might functionto position or reposition an Mn(II) ion in the RtcBactive site The structure we present of Archease appearsto capture it in a position appropriate for lsquohanding offrsquo itsmetal ion to a partner protein (Figure 5B) A role forArchease in positioning Mn(II) in the Mn2 site is sug-gested by the observation that Archease can rescue theRNA 30-P guanylylation activity of an RtcB variant thathas an inactive D95A substitution in that site (Figure 7A)During the RtcB histidine-guanylylation reaction Mn(II)in the Mn2 site reduces the negative charge on the GTPtriphosphate moiety and orients the PPi leaving groupapically to the histidine nucleophile The subsequentRNA 30-P guanylylation reaction would require theRNA 30-P to bind where PPi was located directlyadjacent to the Mn2 site such that an in-line attack canoccur on the phosphorous atom of GMP It is unknownwhether Mn(II) leaves the Mn2 site along with PPi or staysbound to RtcB If Mn(II) is indeed displaced after histi-dine guanylylation then the remarkable rescue of RNA30-P guanylylation activity by Archease could be explainedby an effect on the Mn2 site Additionally the findingthat only catalytic concentrations of Archease arerequired for RtcB activation suggests that the twoproteins interact only transiently with RtcB recruitingArchease as necessary

An alternative hypothesis is that Archease assists in therelease of GMP or ligated RNA formed as a product ofthe RtcB ligation step (Figure 1A) Product release isthought to be the rate-limiting step for nick-sealing byT4 DNA ligase (32) Yet the rate accelerations weobserved were under single-turnover conditionsmdashthereaction rates were not dependent on GMP or ligatedRNA product release Also pertinent to consider is thatArchease could affect the release of PPi generated on for-mation of RtcBndashpG Obtaining a crystal structure of theRtcBmiddotArchease complex will likely be important forunderstanding the mechanism of RtcB activation

Activators of enzymatic activity are typically small mol-ecules (33) Posttranslational modifications such as phos-phorylation and acetylation can also enhance enzymaticactivity (34ndash36) Protein activators of enzymes arerare though the activation of a nucleic acid ligase by aprotein partner is known The bacterial proteins poly-nucleotide kinase-phosphatase (Pnkp) and hua enhancer1 (Hen1) form a complex that repairs ribotoxin-cleavedRNAs The PnkpHen1 complex has evolved such thatthe methyltransferase Hen1 is required for activating theligase activity of Pnkp This requirement ensures thatHen1 has the opportunity to methylate the 20-OH at therepair junction thus preventing future ribotoxin-cleavageat the same site (37) The activity of eukaryotic DNAligase IV involved in nonhomologous end-joining isactivated by the proteins XRCC4 and XLF (3839)

The previous demonstration that Archease also modulatesthe specificity of a tRNA m5C methyltransferase (2)strengthens our conclusion that Archease is a criticalfactor for tRNA maturation Here we have demonstratedthat Archease not only activates a tRNA ligase but alsobroadens its NTP specificity

ACCESSION NUMBERS

PDB 4n2p

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online

ACKNOWLEDGMENTS

The authors are grateful to A A Hoskins (University ofWisconsinndashMadison) for helpful discussions

FUNDING

National Institutes of Health [F32GM100681 to KKD]Protein Structure Initiative grants [U01GM098248 toGNP] and [U54GM074901 Center for EukaryoticStructural Genomics] [R01CA073808 to RTR] Federalfunds from the National Cancer Institute and theNational Institute of General Medical Sciences [Y1-CO-1020 and Y1-GM-1104 respectively toward the GeneralMedicine and Cancer Institute Collaborative Access Team(GMCA-CAT)] US Department of Energy Basic EnergySciences Office of Science [DE-AC02-06CH11357 towarduse of the Advanced Photon Source] Funding for openaccess charge NIH [R01CA073808]

Conflict of interest statement None declared

REFERENCES

1 CanavesJM (2004) Predicted role for the Archease proteinfamily based on structural and sequence analysis of TM1083 andMTH1598 two proteins structurally characterized throughstructural genomics efforts Proteins 56 19ndash27

2 AuxilienS El KhadaliF RasmussenA DouthwaiteS andGrosjeanH (2007) Archease from Pyrococcus abyssi improvessubstrate specificity and solubility of a tRNA m5Cmethyltransferase J Biol Chem 282 18711ndash18721

3 MartinezJ (2013) Keep digging and you will find an exoticpartner of the human tRNA ligase complex revealed bycombining biochemistry and phyletic distribution Abstract S006Biogenesis and turnover of small RNAs Royal SocietyEdinburgh UK January 15thndash17th

4 PopowJ EnglertM WeitzerS SchleifferA MierzwaBMechtlerK TrowitzschS WillCL LuhrmannR SollD et al(2011) HSPC117 is the essential subunit of a human tRNAsplicing ligase complex Science 331 760ndash764

5 EnglertM SheppardK AslanianA YatesJR III and SollD(2011) Archaeal 30-phosphate RNA splicing ligase characterizationidentifies the missing component in tRNA maturation Proc NatlAcad Sci USA 108 1290ndash1295

6 TanakaN and ShumanS (2011) RtcB is the RNA ligasecomponent of an Escherichia coli RNA repair operon J BiolChem 286 7727ndash7731

7 TanakaN ChakravartyAK MaughanB and ShumanS (2011)Novel mechanism of RNA repair by RtcB via sequential

Nucleic Acids Research 2014 Vol 42 No 6 3941

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from

2030-cyclic phosphodiesterase and 30-phosphate50-hydroxylligation reactions J Biol Chem 286 43134ndash43143

8 DesaiKK and RainesRT (2012) tRNA ligase catalyzes theGTP-dependent ligation of RNA with 30-phosphate and50-hydroxyl termini Biochemistry 51 1333ndash1335

9 ChakravartyAK SubbotinR ChaitBT and ShumanS (2012)RNA ligase RtcB splices 30-phosphate and 50-OH ends viacovalent RtcB-(histidinyl)-GMP and polynucleotide-(30)pp(50)Gintermediates Proc Natl Acad Sci USA 109 6072ndash6077

10 ChakravartyAK and ShumanS (2012) The sequential 2030-cyclic phosphodiesterase and 3rsquo-phosphate5rsquo-OH ligation steps ofthe RtcB RNA splicing pathway are GTP-dependent NucleicAcids Res 40 8558ndash8567

11 EnglertM XiaS OkadaC NakamuraA TanavdeV YaoMEomSH KonigsbergWH SollD and WangJ (2012)Structural and mechanistic insights into guanylylation of RNA-splicing ligase RtcB joining RNA between 30-terminal phosphateand 50-OH Proc Natl Acad Sci USA 109 15235ndash15240

12 DesaiKK BingmanCA PhillipsGN Jr and RainesRT(2013) Structures of the noncanonical RNA ligase RtcB reveal themechanism of histidine guanylylation Biochemistry 522518ndash2525

13 SarmientoF MrazekJ and WhitmanWB (2013) Genome-scaleanalysis of gene function in the hydrogenotrophic methanogenicarchaeon Methanococcus maripaludis Proc Natl Acad Sci USA110 4726ndash4731

14 PopowJ SchleifferA and MartinezJ (2012) Diversity and rolesof (t)RNA ligases Cell Mol Life Sci 18 1197ndash1209

15 AbelsonJ TrottaCR and LiH (1998) tRNA splicing J BiolChem 273 12685ndash12688

16 SreenathHK BingmanCA BuchanBW SederKDBurnsBT GeethaHV JeonWB VojtikFC AcetiDJFrederickRO et al (2005) Protocols for production ofselenomethionine-labeled proteins in 2-L polyethyleneterephthalate bottles using auto-induction medium Protein ExprPurif 40 256ndash267

17 OtwinowskiZ and MinorW (1997) Processing of X-raydiffraction data collected in oscillation mode Methods Enzymol276 307ndash326

18 AdamsPD AfoninePV BunkocziG ChenVB DavisIWEcholsN HeaddJJ HungLW KapralGJ Grosse-KunstleveRW et al (2010) PHENIX a comprehensive Python-based system for macromolecular structure solution ActaCrystallogr D Biol Crystallogr 66 213ndash221

19 EmsleyP and CowtanK (2004) Coot model-building tools formolecular graphics Acta Crystallogr D Biol Crystallogr 602126ndash2132

20 DavisIW Leaver-FayA ChenVB BlockJN KapralGJWangX MurrayLW ArendallWB III SnoeyinkJRichardsonJS et al (2007) MolProbity all-atom contacts andstructure validation for proteins and nucleic acids Nucleic AcidsRes 35 W375ndashW383

21 DelanoWL (2002) The PyMOL Molecular Graphics SystemDeLano Scientific San Carlos CA

22 CiriaR Abreu-GoodgerC MorettE and MerinoE (2004)GeConT gene context analysis Bioinformatics 20 2307ndash2308

23 MaoF DamP ChouJ OlmanV and XuY (2009) DOOR adatabase for prokaryotic operons Nucleic Acids Res 37D459ndashD463

24 TaylorMR ConradJA WahlD and OrsquoBrienPJ (2011)Kinetic mechanism of human DNA ligase I reveals magnesium-dependent changes in the rate-limiting step that compromiseligation efficiency J Biol Chem 286 23054ndash23062

25 AhelI RassU El-KhamisySF KatyalS ClementsPMMcKinnonPJ CaldecottKW and WestSC (2006) Theneurodegenerative disease protein aprataxin resolves abortiveDNA ligation intermediates Nature 443 713ndash716

26 WeissB Jacquemin-SablonA LiveTR FareedGC andRichardsonCC (1968) Enzymatic breakage and joining ofdeoxyribonucleic acid VI Further purification and properties ofpolynucleotide ligase from Escherichia coli infected withbacteriophage T4 J Biol Chem 243 4543ndash4555

27 AnantharamanV and AravindL (2004) The SHS2 module is acommon structural theme in functionally diverse protein groupslike Rpb7p FtsA GyrI and MTH1598Tm1083 superfamiliesProteins 56 795ndash807

28 GenschikP DrabikowskiK and FilipowiczW (1998)Characterization of the Escherichia coli RNA 30-terminalphosphate cyclase and its s54-regulated operon J Biol Chem273 25516ndash25526

29 OpalkaN ChlenovM ChaconP RiceWJ WriggersW andDarstSA (2003) Structure and function of the transcriptionelongation factor GreB bound to bacterial RNA polymerase Cell114 335ndash345

30 SosunovaE SosunovV KozlovM NikiforovV GoldfarbAand MustaevA (2003) Donation of catalytic residues to RNApolymerase active center by transcription factor Gre Proc NatlAcad Sci USA 100 15469ndash15474

31 SteitzTA (1998) A mechanism for all polymerases Nature 391231ndash232

32 LohmanGJ ChenL and EvansTC Jr (2011) Kineticcharacterization of single strand break ligation in duplex DNA byT4 DNA ligase J Biol Chem 286 44187ndash44196

33 GrimsbyJ SarabuR CorbettWL HaynesNE BizzarroFTCoffeyJW GuertinKR HilliardDW KesterRFMahaneyPE et al (2003) Allosteric activators of glucokinasepotential role in diabetes therapy Science 301 370ndash373

34 LinK RathVL DaiSC FletterickRJ and HwangPK(1996) A protein phosphorylation switch at the conservedallosteric site in GP Science 273 1539ndash1542

35 ZhaoS XuW JiangW YuW LinY ZhangT YaoJZhouL ZhengY LiH et al (2010) Regulation of cellularmetabolism by protein lysine acetylation Science 19 1000ndash1004

36 LinH SuX and HeB (2012) Protein lysine acylation andcysteine succination by intermediates of energy metabolism ACSChem Biol 7 947ndash960

37 ChanCM ZhouC and HuangRH (2009) Reconstitutingbacterial RNA repair and modification in vitro Science 326 247

38 GrawunderU WilmM WuX KuleszaP WilsonTEMannM and LieberMR (1997) Activity of DNA ligase IVstimulated by complex formation with XRCC4 protein inmammalian cells Nature 388 492ndash495

39 LuH PannickeU SchwarzK and LieberMR (2007) Length-dependent binding of human XLF to DNA and stimulation ofXRCC4DNA ligase IV activity J Biol Chem 28211155ndash11162

3942 Nucleic Acids Research 2014 Vol 42 No 6

by guest on May 13 2016

httpnaroxfordjournalsorgD

ownloaded from