Sensitivity of the Yeast Mitochondrial RNA Polymerase to �1and �2 Initiating Nucleotides*

Received for publication, September 7, 2006 Published, JBC Papers in Press, September 26, 2006, DOI 10.1074/jbc.M608638200

Elizabeth A. Amiott1 and Judith A. Jaehning2

From the Department of Biochemistry and Molecular Genetics and the Program in Molecular Biology, University of Colorado atDenver and Health Sciences Center, Aurora, Colorado 80045

Despite a simple consensus sequence, there is considerablevariation of promoter strengths, transcription rates, and thekinetics of initiating nucleotide incorporation among the pro-moters found in the Saccharomyces cerevisiae mitochondrialgenome. We asked how changes in the initiating (�1 and �2)nucleotides, conformation of the promoter DNA template, andmutation of the mitochondrial RNA polymerase (mtRNAP)affect the kinetics of nucleotide (NTP)utilization.Using ahighlypurified in vitro mitochondrial transcription system, we foundthat 1) the mtRNAP requires the highest concentrations of the�1 and �2 initiating NTPs, intermediate concentrations ofNTPs at positions 5 to 11, and low concentrations of elongat-ingNTPs; 2) themtRNAP requires a higher concentration of the�2 NTP than the �1 NTP for initiation; 3) the kinetics of �2NTP utilization are altered by a point mutation in the mtRNAPsubunitMtf1; and4) a supercoiledorpre-meltedpromoterDNAtemplate restores normal �2 NTP utilization by the Mtf1mutant. Based on comparisons to the structural and biochem-ical properties of the bacterial RNAP and the closely relatedT7 RNAP, we propose that initiating nucleotides, particularlythe �2 NTP, are required at high concentrations to drivemitochondrial promoter opening or to stabilize a productiveopen complex.

The core subunit of the budding yeast mitochondrial RNApolymerase (mtRNAP),3 encoded by the nuclear geneRPO41, isa 145-kDa protein that shares significant amino acid similaritywith the T7/T3 bacteriophage family of single subunit RNAPs(1, 2). However, T7 RNAP is a single subunit enzyme capable ofspecific and regulated activity in the absence of any transcrip-tion factors (3), but mtRNAPs from all eukaryotes examined todate require one or more auxiliary factors for selective tran-scription (reviewed in Ref. 4). The accessory factor for the yeastmtRNAP, Mtf1, is a 43-kDa protein that interacts directly with

Rpo41, creating a functional “holoenzyme” with promoter-spe-cific activity (5). Following the synthesis of a short RNA chain(�12 nucleotides), Mtf1 is released, and Rpo41 continues elon-gation of the transcript (6).Recently it has been demonstrated that Rpo41 alone can ini-

tiate transcription in a promoter-specific manner in vitro in theabsence ofMtf1 (7). However, thisMtf1-independent activity isonly possible on supercoiled or pre-melted “bubble” promotertemplates in which the promoter opening stages of transcrip-tion initiation have been facilitated or bypassed. Mtf1 alters theratio of full-length to abortive transcripts on nonlinear tem-plates (7) demonstrating that Mtf1 influences both promoteropening and clearance by the mtRNAP. Despite these findings,themechanistic function ofMtf1 and its homologs in promoterselective transcription is still not well understood.The yeast mtRNAP mediates gene expression in the mito-

chondrial compartment by binding and initiating transcriptionat a collection of short, nonanucleotide (consensus ATATA-AGTA) promoters encompassing the transcriptional start site(�1) and the eight nucleotides directly upstream. Despite theapparent similarity of promoters, transcription rates of mito-chondrial genes can vary over 50-fold in vitro andup to 100-foldin vivo, which in combinationwith different stabilities results inup to 1000-fold variations in vivo in transcript abundance (8, 9).A series of careful in vitro studies looking at the relationship

between promoter sequence and mtRNAP activity have charac-terized the elements ofmitochondrial promoters essential for rec-ognition by the polymerase as well as the NTPs that contribute tothe efficiency of transcription or promoter strength (10–14). Sev-eral nucleotides within the consensus sequence are essential, butsubstitution is tolerated at other positions. Of note, the�1 initiat-ingNTPcanbechangedwithmodest affectsonpromoter strengthin vitro, but this position is conserved as an “A” in all yeast mito-chondrial promoters (13). In addition to the �8 to �1 consensuspromoter region, DNA sequences as much as 300 bp upstream,and at least 9 bp downstream of the promoter can affect observedpromoter utilization by the mtRNAP. The nucleotide at the �2position determines promoter strength (12–15). Specifically, a�2purine (A or G) results in a “strong” or highly efficient promoter,whereas promoters with a pyrimidine (C or T) in the �2 positionare classified as “weak” (14, 16). In vitro promoter strength andtranscription rate correlate well with steady-state mitochondrialRNAabundance (8, 9).Therefore, changes inmtRNAPactivity areresponsible for the regulation of mitochondrial transcript levelsin vivo.Like other RNAPs, the mtRNAP requires higher levels of its

initiating (�1 ATP) nucleotide than those used for elongation

* This work was supported by Grant MCB 0235354 from the National ScienceFoundation (to J. A. J). The costs of publication of this article were defrayedin part by the payment of page charges. This article must therefore behereby marked “advertisement” in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

1 Present address: Dept. of Biochemistry, University of Utah School of Medi-cine, 15 North Medical Dr. East, Rm. 4100, Salt Lake City, UT 84112.

2 To whom correspondence should be addressed: Dept. of Biochemistry andMolecular Genetics, University of Colorado at Denver and Health SciencesCenter, Mail Stop 8101, P. O. Box 6511, Aurora, CO 80045. Tel.: 303-818-1931; Fax: 303-724-3215; E-mail: Judith.Jaehning@UCHSC.edu.

3 The abbreviations used are: mtRNAP, mitochondrial RNA polymerase; GST,glutathione S-transferase; WT, wild type.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 281, NO. 46, pp. 34982–34988, November 17, 2006© 2006 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

34982 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 46 • NOVEMBER 17, 2006

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

(10, 11, 17). We have recently extended this observation todemonstrate a clear relationship between the in vitro kinetics ofATPutilization (KmATP) and the in vivo regulation ofmtRNAPtranscription (18). This property of themtRNAP couples mito-chondrial gene expression to mitochondrial function. How-ever, in vivo and in vitro responses to fluctuating ATP concen-tration (KmATP) vary over 10-fold between promoters (18, 19),suggesting that factors in addition to the �1 NTP influencetranscriptional activity and NTP utilization by the mtRNAP.To explore these additional factors we looked more closely

at the kinetics of NTP usage by the mtRNAP when both theDNA template and the polymerase itself are altered. We foundthat the requirement for high concentrations of initiatingNTPswas specific for an initiation step prior to or including firstphosphodiester bond formation. By altering the initiating (�1and �2) nucleotides of several promoters we observed that, asfor bacterial rRNA promoters (20, 21), both the �1 and �2nucleotides were very important for promoter utilization. Infact, like T7 RNAP (22) we found that the NTP used for the �2position of the transcript was needed at higher concentrationthan the�1NTP. Finally, we found that a pointmutation in themtRNAP subunit Mtf1 specifically raised the Km for the �2NTP on linear DNA templates but not on supercoiled or pre-melted bubbleDNA.Our results suggest that an open promoterconformation can bypass one of the NTP-dependent steps ofthe polymerase compromised by the Mtf1 mutation.

EXPERIMENTAL PROCEDURES

In Vitro Transcription Templates—COX2, 21 S rRNA, 14 SrRNA, tRNAGlu, ATP9, COX1, and tRNACys linear plasmidDNA templates were as described previously (18). COX2, 14 SrRNA, and tRNACys promoter templates with substitutions atthe �1 and �2 nucleotide positions were created by annealingcomplementary 70-bp oligonucleotides containing WT or thedescribed sequence variations with EcoRI restriction sites oneach end (Table 1). The annealed fragments were cloned intothe pRS2.1 vector (Invitrogen) and, when linearized, generatedrun-off transcripts 35 nucleotides in length. Uncut plasmidsserved as supercoiled promoter templates, and themismatched14 S rRNA bubble template was prepared from 70-bp oligonu-cleotides as described (7).In Vitro Transcription Reaction Conditions—Recombinant

Rpo41 (pJJ1399: His-Rpo41) andWT or mutantMtf1 (pJJ1286:

GST-Mtf1/pJJ1292:GST-C192FMtf1) proteinswere expressedand purified as described (23, 24).Multiple round transcriptionreactions were performed using 20 �g/ml plasmid DNA tem-plate or 1 nmol bubble template. Transcription was initiated bythe addition of 0.8 pmol of Rpo41 to a reaction containing 0.8pmol of Mtf1, 50 mM Tris (pH 7.9), 20 mM MgCl2, 1 mM dithi-othreitol, 50 �M UTP, [�-32P]UTP (1000 cpm/pmol UTP), 250�M ATP, GTP, and CTP (as described in figure legends) for afinal volume of 20 �l. Reactions were incubated at 30 °C for 7min, and transcription was terminated by adding 25 �l of stop/loading buffer (90% formamide, 50 mM EDTA, bromphenolblue, and xylene cyanol) on ice. Samples were denatured at70 °C for 5 min prior to loading on 7 M urea, 8% or 10% polyac-rylamide gels for electrophoresis. Radiolabeled run-off tran-scription products were visualized on a STORM 860 phospho-rimager (Amersham Biosciences) and quantified usingQuantity One analysis software (Bio-Rad). Statistical analyses,KmNTP calculations, and all graphs were prepared usingPRISM� software (GraphPadTM Software, Inc.). The dinucle-otide ApA (Sigma) was used at 50 �M.

RESULTS

The mtRNAP Requires High Concentrations of Both the �1and �2 NTPs—Using purified recombinant mtRNAP subunitsand varying concentrations of a single NTP, we determined theconcentration of ATP, GTP, CTP, or UTP substrate at whichhalf-maximal enzymatic activity was achieved (Km) on severalmitochondrial promoters in vitro. The run-off transcripts gen-erated for each ATP or GTP concentration on the COX2 (�1�2 AG) promoter template are shown in Fig. 1A. The signalfrom each transcript was quantified and normalized to themaximal signal within each titration (% max cpm), and the Kmfor each NTP was determined by nonlinear regression analysisusing theMichaelis-Menten equation (Fig. 1A, lower panel andFig. 2). As reported previously (18), the Km for the �1 ATP foreach promoter is always higher than the Km for the elongatingNTP CTP (�5 �M, data not shown).

To confirm that the high Km values observed are in factbecause of a requirement for initiation, we repeated KmATPexperiments on the COX1 (�1 �2 AA) promoter template inthe presence or absence of the dinucleotide ApA to eliminatethe energy barrier of first bond formation and stabilize the ini-tiation complex (10, 11). Fig. 1B shows that addition of ApA

TABLE 1Oligonucleotides for constructing in vitro transcription templates with altered �1 and �2 nucleotidesComplementary A and B oligonucleotides were annealed to create a double-stranded DNA fragment. Promoter sequences are in bold and the �1 and �2 nucleotides inparentheses. Oligonucleotides for the altered promoters are the same as WT except for the indicated changes in the �1 and �2 nucleotides.

Promoter �1 �2 5� to 3� sequenceCOX2 (WT) AG (A) 5�-CCGGAATTCAAATTATAAATAAATTTTAATTAAAAGT(AG)TATTAATATATTATAAATAGATAAAAGAATTCC-3�

(B) 5�-GGAAATCTTTTATCTATTTATAATATATTAATA(CT)ACTTTTAATTAAAATTTATTTATAATTTGAATTCCGG-3�(�2A) AA(�1G) GG(�1G �2A) GA14 S rRNA (WT) AA (A) 5�-CCGGAATTCATTAATAATTTATTTATTATTATATAAGT(AA)TAAATAATAGTTTTATAATAAGAATTCC-3�

(B) 5�-GGAATTCTTATTATATAAAACTATTATTTA(TT)ACTTATATAATAATAAATAAATTATTAATGAATTCCGG-3�(�1G) GA(�2G) AGtRNACys (WT) AU (A) 5�-CCGGAATTCATAGAATAAAGATATAAATAATTATAAGT(AT)ATAAAGTAATAAAGGAGATGTTGAATTCC-3�

(B) 5�-GGAATTCAACATCTCCTTTATTACTTTAT(AT)ACTTATAATTATTTATATCTTTATTCTATGAATTCCGG-3�(�2A) AA(�2G) AG

�1 and �2 Nucleotide Control of mtRNAP

NOVEMBER 17, 2006 • VOLUME 281 • NUMBER 46 JOURNAL OF BIOLOGICAL CHEMISTRY 34983

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

lowers the KmATP on the COX1 promoter nearly 10-fold from59 to 6.5 �M, indicating that transcription initiation is muchless dependent on ATP if the bond between the first two nucle-otides in the RNA chain has already been formed. Therefore,the elevated ATP requirement of the mtRNAP on this pro-moter is specific to initiation events prior to or including bondformation between the �1 and �2 NTPs.

Unexpectedly, we observed thaton the COX2 promoter the Km forthe �2 nucleotide, GTP (57 �M), ishigher than the Km for the �1 ATP(12 �M) (Figs. 1A and 2). This sug-gested that the mtRNAP requiresboth the �1 and �2 NTPs at higherconcentrations for initiation, andthat the�2NTPmight be needed athigher concentrations than the �1NTP.When we lookedmore closelyat the sequence of each promoter,we observed that both 21 S rRNAand COX2, the only transcripts car-rying a GTP in the �2 position,required more GTP than ATP (Fig.2). These are also the mitochondrialpromoters with the lowest KmATP(Fig. 2). All other promoters testedhave ATP in both the �1 and �2position except for the weaktRNACys promoter. Synthesis of thistranscript requires the pyrimidineUTP at the �2 position, andalthough transcription from thisweak promoter requires very highlevels of ATP, the KmUTP (190 �M)is even greater than the KmATP(144 �M). Thus, initiation by the

mtRNAP requires high concentrations of both the �1 and �2NTP, but the �2 NTP appears to be required at the highestconcentration.

�1 �2 Promoter Sequence Determines NTP ConcentrationDependence—To confirm that the Km of the mtRNAP for the�2 NTP is higher than that for the �1 NTP, we created andanalyzed promoter templates with substitutions in the �1 and�2 positions. Mutations in the 14 S rRNA (a strong, consensuspromoter), COX2 (a strong promoter that varies from the con-sensus), and tRNACys (a weak, non-consensus promoter) pro-moters were generated to create various combinations of A, G,andT at�1 and�2 as shown in Fig. 3. TheKmATPandKmGTPwere determined for each template as described above.For the 14 S rRNA templates we found that the KmATP for

the �2 A (AA and GA, 41 and 47 �M) templates was nearly10-fold higher than for the �2 G (AG, 6.5 �M) template (Fig.3A). Conversely, theKmGTPwas up to 10-fold higher on the�2G (AG, 28 �M) template than the �2 A (AA and GA, 1.5 and 4�M) promoters (Fig. 3B). Similarly, in the COX2 promoter con-text we observed much higher KmATP for �2 A (AA and GA,63 and 75�M) constructs than for�2G (AGandGG, 5 and 26.6�M) constructs (Fig. 3A).KmGTPwas highest for the COX2GGpromoter (�500�M, Fig. 3B). Not unexpectedly, theKmATP orKmGTP was lowest (�5 �M) when there was no ATP (GG,COX2) or GTP (AA, 14 S rRNA), respectively, required forinitiation. Also consistent with this pattern, we saw that theKmfor anNTP found only in the�1 positionwas intermediate. Forexample, on the 14 S rRNA template, �1 G (GA) raised theKmGTP slightly over that of an AA start (from 1.5 to 4 �M, Fig.

FIGURE 1. Measurement of KmNTP for the �1 and �2 positions of the COX2 promoter and use of adinucleotide primer to bypass first bond formation on the COX1 promoter. Standard in vitro transcriptionreactions were performed using COX2 (panel A) or COX1 (panel B) promoter templates. The concentration of asingle NTP (ATP or GTP) varied from 0 to 1000 �M, and the others were held constant at 250 �M. (�) ApAreactions (panel B) were supplemented with 50 �M ApA dinucleotide corresponding to �1 and �2 NTP of thetranscript. The amount of each run-off transcript (upper panels) was normalized to the maximal transcript signalwithin an experiment (% max cpm) and plotted as a function of NTP concentration (�M) (lower panels). Datapoints from replicate experiments were averaged and subjected to nonlinear regression analysis using theMichaelis-Menten equation to calculate the average KmNTP.

FIGURE 2. The Km for the �2 NTP is higher than the Km for the �1 NTP.Promoters and their average KmATP and KmGTP (KmUTP included for the �1�2 AU tRNACys promoter) are graphed and arranged in groups by �1 �2 NTPsequence as indicated at the bottom of the figure. For KmATP and KmGTP, barsrepresent the average, and error bars are the standard deviation from at leasttriplicate experiments. For KmUTP, the bar represents the average, and theerror bars represent the range of duplicate experiments.

�1 and �2 Nucleotide Control of mtRNAP

34984 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 46 • NOVEMBER 17, 2006

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

3B) but to a value much lower than when G was at �2 (AG, 28�M). The samepatternwas seen forATPon theCOX2 template(Fig. 3A), where the KmATP for the AG start was much higherthan the very lowKmATP for the GG template (27 versus 5�M).Finally, the KmUTP of the weak tRNACys promoter (AU) wasvery high (190 �M, Fig. 2), but switching the �1 position fromATP to GTP resulted in a switch from a high Km for ATP (163�M) to a high Km for GTP (98 �M) (Fig. 3). These experimentsconfirm and extend our observation that the mtRNAP requireshigher concentrations of both the �1 and �2 NTPs than thelevels needed for elongating positions, with the �2 positionneeding the highest NTP concentration.Intermediate NTP Concentrations Are Required Prior to Pro-

moter Escape—InFig. 2 it is apparent that evenwithin the groupof promoters that initiate with AA, there is a range of KmGTP.The range is from �5 �M to over 20 �M, similar to that seen forthe elongating NTP CTP. Examination of the DNA sequencedownstreamof the promoters revealed that the higherKmGTPscorresponded to the presence of a G in the transcript shortlyafter initiation. This relationship is seen more clearly in Fig. 4where we have plotted average KmGTP relative to the positionof the first GTP in the transcript. When the first GTP occursbetween 5 and 11 nucleotides of the transcript, the KmGTP iselevated relative to the �5 �M values observed for G at �12 orfurther. We note that Mtf1 release and the transition of the

mtRNAP to an elongating form occurs within this same region(6). This supports the idea that the requirement of the polymer-ase for NTPs changes as it takes on different conformations,with the most dramatic transition occurring between initiationand elongation as Mtf1 is released.A Mutation in the mtRNAP Subunit Mtf1 Results in a Selec-

tive Increase in Km for the�2NTP,WhichCan be Suppressed byPre-opening the Promoter—Because the association betweenthe mitochondrial accessory factor, Mtf1, and the core poly-merase, Rpo41, has been compared with interactions with fac-tors that regulate NTP affinity of other RNAPs (5, 7), we askedhowMtf1mightmodulate the kinetics of NTP utilization orKmof the coremtRNAP.Wehave previously reported the effects ofan Mtf1 point mutation (C192F) on in vitro transcription effi-ciency (24) and in vivomitochondrial transcriptional regulationof theATP9 transcript (18). To extend these findings and deter-mine the effects of the C192F mutation on �2 NTP utilization,we performed in vitro transcription reactions to determineKmNTPs on other promoter templates using purified recombi-nant WT and C192F mutant Mtf1. As reported for the ATP9(AA) promoter template (18), the mutant Mtf1 dramaticallyraised theKmATP of the 14 S rRNA (AA) promoter (from 29 to349 �M, Fig. 5A). However, the mutant Mtf1 had essentially noeffect on theKmATP of the COX2 (AG) promoter (16 versus 20�M, Fig. 5B). To test for the effect of the C192Fmutation on the�2 NTP, we measured the KmGTP for the COX2 (AG) pro-moter. We observed that the C192F mutant raised the Km forthe �2 GTP over 10-fold from 54 to 677 �M (Fig. 5C) similar tothe over 10-fold change in KmATP observed on the 14 S rRNA(AA) promoter (Fig. 5A). Therefore, the C192F point mutationhas the general effect of lowering the polymerase affinity for, orthe ability to utilize, the �2 NTP of a transcript.The observation that Mtf1 is required for transcription in

vitro, but only on linearDNA templates, suggests that one of thefunctions of Mtf1 is to assist the polymerase in the promoter-opening step of initiation (7). Because initiating (�1 �2) NTPconcentration appears to regulate other RNAPs by affecting theformation and/or stability of the open promoter complex (20,

FIGURE 3. Substitution of the �1 �2 promoter positions alters the KmNTPconfirming the requirement for the highest level of the �2 NTP. In Vitrotranscription reactions with varied concentrations of ATP and GTP were usedto determine the KmATP (panel A) and KmGTP (panel B) for 14 S rRNA (AA),COX2 (AG), and tRNACys (AU) promoter templates with the indicated substi-tutions at the �1 and �2 NTP positions of the transcripts. Bars indicate theaverage Km value, and error bars represent the standard deviation from atleast triplicate experiments.

FIGURE 4. The KmGTP for positions prior to promoter escape are higherthan for those in elongation. The average KmGTP determined from in vitrotranscription reactions is plotted relative to position of the first G in the tran-script for the indicated mitochondrial promoter templates. Error bars indicatethe standard deviation from at least triplicate experiments.

�1 and �2 Nucleotide Control of mtRNAP

NOVEMBER 17, 2006 • VOLUME 281 • NUMBER 46 JOURNAL OF BIOLOGICAL CHEMISTRY 34985

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

22, 25), we askedwhether theNTP requirement of themtRNAPmight be altered on “open” or pre-melted promoter templates.Supercoiled templates facilitate promotermelting, andpartiallysingle-stranded or bubble templates with mismatched bases inthe non-template strand can be used tomimic amelted or openpromoter conformation (26). We used both supercoiled (Fig.5D) and bubble (Fig. 5E) 14 S rRNA templates to measure invitro KmATP with both WT and the mutant C192F Mtf1. Wefound that when mtRNAP-dependent promoter melting wasreduced (supercoiled template) or eliminated (bubble tem-plate), the KmATP in reactions containing WT Mtf1 was rela-tively unaffected (KmATP of 29, 25, and 46 �M on linear, super-coiled, and bubble 14 S rRNA templates, respectively).However, with the C192F mutant Mtf1, the high KmATP seenon the linear template (KmATP of �350 �M, Fig. 5A) wasreduced on the supercoiled (Fig. 5D) and bubble (Fig. 5E) tem-plates nearly 10-fold to values near that forWTMtf1 on a linearor supercoiled template (33–37 versus 25–29 �M). Therefore,the C192Fmutation appears to impair�2NTP utilization dur-ing promoter opening, which can be suppressed by providing apartially or fully open promoter.

DISCUSSION

This work demonstrates that the mtRNAP requires highconcentrations of the first twoNTPs, in particular the�2NTP,

used to initiate RNA synthesis. A similar requirement for highlevels of initiatingNTPs is shared by other RNAPs including theclosely related single polypeptide T7 RNAP (22, 27) and themulti-subunit bacterial RNAP (20, 21, 28). The necessity forhigh levels of these initiating nucleotides serves as the basis forsimple but powerful regulation of these enzymes by changingnucleotide pools and the associationwith accessory factors thatmodulate the affinity of the RNAPs for NTPs.The bacteriophage T7 RNAP has a strong preference for ini-

tiating with GTP (27, 29–31), and during the infectious cycle,expression of late class II and III genes is inhibited, at least inpart, by T7 lysozyme-induced destabilization of the initiationcomplex and changes in the KmGTP of the pre-initiation com-plex (32–34).Aswehave found for themtRNAP, the�2NTPofphage transcripts is required at high levels and is a determinantof transcription efficiency (22). Patel and colleagues have pre-sented a model whereby binding of the �2 NTP in the activesite of the polymerase drives the ternary complex into a fullyopen state; a transition not seen upon binding of just the �1NTP (22). They also hypothesize that this open conformationfacilitates entry and positioning of the�1NTP in the active siteto create a productive initiation complex. Hence, the efficiencyof transcription initiation by T7 RNAP relies on the affinity ofthe pre-initiation complex for the �2 NTP, which then influ-

FIGURE 5. The C192F Mtf1 mutation increases the Km of the �2 NTP on linear templates but not on partially or fully open promoters. In vitro transcrip-tion reactions were performed on the 14 S rRNA (AA) and COX2 (AG) promoters using purified WT (E) or C192F (�) mutant Mtf1. Data points on theMichaelis-Menten plots are the average of at least two experiments, and the average KmATP and KmGTP values are presented on the graphs. A, average KmATPon the 14 S rRNA (AA) promoter. B, average KmATP of the COX2 (AG) promoter. C, average KmGTP of the COX2 (AG) promoter. D, average KmATP on a supercoiled14 S rRNA promoter template. Uncut 14 S rRNA plasmid DNA was used as a template, and reactions were carried out except CTP was omitted from the reactionsto produce a 107 nucleotide transcript. E, average KmATP on a pre-melted bubble 14 S rRNA promoter produced as described under “Experimental Procedures.”

�1 and �2 Nucleotide Control of mtRNAP

34986 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 46 • NOVEMBER 17, 2006

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

ences �1 NTP incorporation. Based on the strong similarity ofthe mtRNAP to the T7 RNAP and our finding that the�2 NTPis required at higher levels than the �1 NTP, we think it verylikely that a similar mechanism for initiation pertains for themtRNAP.Both the �1 and �2 nucleotides of bacterial rRNA promot-

ers are also required at high concentrations to stabilize the ini-tiation complex (35) and possibly drive the isomerization of thepolymerase into an initiation-competent conformation (21,36). These initiation-specific NTP requirements are responsi-ble for changes in ribosomal RNA transcriptional patterns thatare sensitive to fluctuations in intracellular ATP and GTP con-centrations (35). The nucleotide sensitivity of the bacterialRNAP is modulated by the recently discovered DksA protein,which binds the polymerase and, like T7 lysozyme, reduces thestability of promoter-polymerase complexes and raises theNTP requirement for transcription of the rrnB promoters (37).Our discovery that amutation in themtRNAP accessory fac-

tor Mtf1 affects the affinity for the �2 NTP is interestingbecause the association of this factor with the core RNAP hasbeen comparedwith bothT7 lysozyme-T7RNAP (7) and sigmafactor-RNAP (5) interactions. In the absence of structuralinformation about the mtRNAP in combination with its acces-sory factor, it is difficult to speculate about how Mtf1 mightmodulate nucleotide binding byRpo41. The fact that the crystalstructure of Mtf1 resembles neither sigma factor nor T7lysozyme, but instead is most like the family of RNA methyl-transferases (38) does not facilitate this speculation. However,the knowledge thatMtf1 interacts with Rpo41 in regions sharedwith T7 RNAP that include residues important for interactionswith the open promoter (39) indicates that Mtf1 interactionmay modify Rpo41 structure to help stabilize the open pro-moter. Clearly both high concentrations of the �1 and �2nucleotides as well as Mtf1 play an important role in theformation and/or stabilization of this critical intermediate ininitiation.Our elucidation of the importance of the �2 NTP for initia-

tion also offers an explanation for why promoters with pyrimi-dines in the �2 position are functional but weak (14). Based onthe model proposed for the T7 RNAP in which the �2 NTPenters the active site first and promotes�1NTP binding (22), ifthere is a very low affinity between the polymerase and a �2pyrimidine, this would, by default, raise the Km for the �1 ATPresulting in a very weak promoter. This view is consistent withthe very high KmATP and even higher KmUTP we observed forthe weak tRNACys promoter (�1 �2 AU). In addition to thehigh concentrations of the �1 and �2 NTPs required for initi-ation, we also observed that nucleotides from �5 to �11 areneeded at levels higher than those required for elongation. Thismay reflect a requirement for additional stabilization of theRNAP�DNA�RNA complex prior to release ofMtf1 and entry ofthe core RNAP into processive elongation.Conservation of ATP at the �1 position of yeast mitochon-

drial promoters suggests that either themtRNAP has an intrin-sic preference for initiating with ATP or that retaining ATP asthe �1 nucleotide is important for regulation. Because theclosely related T7 RNAP initiates with GTP (27, 31) andbecause themtRNAPcan initiatewith otherNTPs (13), the first

possibility seems less likely. Regulation of the mtRNAP by ATPlevels in vivo (18) has probably created selective pressure toretainA in both the�1 and�2 positions ofmostmitochondrialpromoters. An A at �1 and �2 creates promoters exquisitelysensitive to changing ATP pools within the mitochondrion.Promoters with nucleotides other than A at the �2 positionrequire lower levels of ATP for maximal initiation, and thelevels of their transcripts change less dramatically inresponse to changes in respiration (18). Therefore thesedeceptively simple changes in the �2 position of the pro-moter actually have significant consequences for patterns ofgene regulation. It will be interesting to determine whetherother mitochondrial genomes also demonstrate such sophis-tication in the organization and information content of theirrelatively small promoters.

Acknowledgments—We thank Sei-Heon Jang, Michio Matsunaga,and Mark Karlok for strains, mutants, purification protocols, andhelp with expression and isolation of the mtRNAP, and ChristopherKorch and the University of Colorado at Denver and Health SciencesCenter (UCDHSC), Cancer Center Sequencing Core for sequencingthe promoter constructs. We also thank Sei-Heon Jang and Joan Betzfor their comments on the manuscript.

REFERENCES1. Masters, B. S., Stohl, L. L., and Clayton, D. A. (1987) Cell 51, 89–992. Kelly, J. L., Greenleaf, A. L., and Lehman, I. R. (1986) J. Biol. Chem. 261,

10348–103513. McAllister, W. T. (1993) Cell. Mol. Biol. Res. 39, 385–3914. Shadel, G. S. (2004) Trends Genet. 20, 513–5195. Jang, S.-H., and Jaehning, J. A. (1991) J. Biol. Chem. 266, 22671–226776. Mangus, D. A., Jang, S.-H., and Jaehning, J. A. (1994) J. Biol. Chem. 269,

26568–265747. Matsunaga, M., and Jaehning, J. A. (2004) J. Biol. Chem. 279,

44239–442428. Mueller, D. M., and Getz, G. S. (1986) J. Biol. Chem. 261, 11756–117649. Mueller, D. M., and Getz, G. S. (1986) J. Biol. Chem. 261, 11816–1182210. Biswas, T. K. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 9338–934211. Biswas, T. K., and Getz, G. S. (1990) J. Biol. Chem. 265, 19053–1905912. Biswas, T. K., Edwards, J. C., Rabinowitz, M., and Getz, G. S. (1985) Proc.

Natl. Acad. Sci. U. S. A. 82, 1954–195813. Biswas, T. K., Ticho, B., and Getz, G. S. (1987) J. Biol. Chem. 262,

13690–1369614. Biswas,T.K., andGetz,G. S. (1986)Proc.Natl. Acad. Sci.U. S. A.83,270–27415. Biswas, T. K. (1999) Biochemistry 38, 9693–970316. Wettstein-Edwards, J., Ticho, B. S., Martin, N. C., Najarian, D., and Getz,

G. S. (1986) J. Biol. Chem. 261, 2905–291117. Winkley, C. S., Keller, M. J., and Jaehning, J. A. (1985) J. Biol. Chem. 260,

14214–1422318. Amiott, E. A., and Jaehning, J. A. (2006)Mol. Cell 22, 329–33819. Ulery, T. L., Jang, S.-H., and Jaehning, J. A. (1994) Mol. Cell. Biol. 14,

1160–117020. Gaal, T., Bartlett, M. S., Ross, W., Turnbough, C. L., and Gourse, R. L.

(1997) Science 278, 2092–209721. Lew, C. M., and Gralla, J. D. (2004) J. Biol. Chem. 279, 19481–1948522. Stano, N. M., Levin, M. K., and Patel, S. S. (2002) J. Biol. Chem. 277,

37292–3730023. Matsunaga, M., Jang, S.-H., and Jaehning, J. A. (2004) Protein Expression

Purif. 35, 126–13024. Karlok, M. A., Jang, S.-H., and Jaehning, J. A. (2002) J. Biol. Chem. 277,

28143–2814925. Schneider, D. A., Gaal, T., and Gourse, R. L. (2002) Proc. Natl. Acad. Sci.

U. S. A. 99, 8602–8607

�1 and �2 Nucleotide Control of mtRNAP

NOVEMBER 17, 2006 • VOLUME 281 • NUMBER 46 JOURNAL OF BIOLOGICAL CHEMISTRY 34987

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

26. deHaseth, P. L., and Helmann, J. D. (1995)Mol. Microbiol. 16, 817–82427. Ikeda, R. A., and Richardson, C. C. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,

3614–361828. Nierman, W. C., and Chamberlin, M. J. (1979) J. Biol. Chem. 254,

7921–792629. Kuzmine, I., and Martin, C. T. (2001) J. Mol. Biol. 305, 559–56630. Jia, Y., and Patel, S. S. (1997) J. Biol. Chem. 272, 30147–3015331. Nierman, W. C., and Chamberlin, M. J. (1980) J. Biol. Chem. 255,

1819–182332. Huang, J., Villemain, J., Padilla, R., and Sousa, R. (1999) J. Mol. Biol. 293,

457–475

33. Villemain, J., Guajardo, R., and Sousa, R. (1997) J. Mol. Biol. 273, 958–97734. Stano, N. M., and Patel, S. S. (2004) J. Biol. Chem. 279, 16136–1614335. Schneider, D. A., Ross,W., andGourse, R. L. (2003)Curr. Opin.Microbiol.

6, 151–15636. Lew, C. M., and Gralla, J. D. (2004) Biochemistry 43, 12660–1266637. Paul, B. J., Barker, M. M., Ross, W., Schneider, D. A., Webb, C., Foster,

J. W., and Gourse, R. L. (2004) Cell 118, 311–32238. Schubot, F. D., Chen, C. J., Rose, J. P., Dailey, T. A., Dailey, H. A., and

Wang, B. C. (2001) Protein Sci. 10, 1980–198839. Cliften, P. F., Park, J. Y., Davis, B. P., Jang, S.-H., and Jaehning, J. A. (1997)

Genes Dev. 11, 2897–2909

�1 and �2 Nucleotide Control of mtRNAP

34988 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281 • NUMBER 46 • NOVEMBER 17, 2006

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from

Elizabeth A. Amiott and Judith A. JaehningNucleotides

Sensitivity of the Yeast Mitochondrial RNA Polymerase to +1 and +2 Initiating

doi: 10.1074/jbc.M608638200 originally published online September 26, 20062006, 281:34982-34988.J. Biol. Chem. 

  10.1074/jbc.M608638200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/281/46/34982.full.html#ref-list-1

This article cites 39 references, 25 of which can be accessed free at

by guest on June 30, 2020http://w

ww

.jbc.org/D

ownloaded from