research article error rate comparison during polymerase ...downloads.hindawi.com › archive ›...

Research ArticleError Rate Comparison during Polymerase Chain Reaction byDNA Polymerase

Peter McInerney12 Paul Adams13 and Masood Z Hadi134

1 Joint BioEnergy Institute Emeryville CA USA2 Sandia National Laboratories Livermore CA USA3 Physical Biosciences Division Lawrence Berkeley National Laboratories Berkeley CA 94720 USA4 Synthetic Biology Program Space BioSciences Division NASA AMES Research Center Mail Stop 239-15 Moffett FieldCA 94035 USA

Correspondence should be addressed to Masood Z Hadi masoodhadigmailcom

Received 22 May 2014 Accepted 21 July 2014 Published 17 August 2014

Academic Editor Alessandro Desideri

Copyright copy 2014 Peter McInerney et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

As larger-scale cloning projects become more prevalent there is an increasing need for comparisons among high fidelity DNApolymerases used for PCR amplification All polymerases marketed for PCR applications are tested for fidelity properties (ieerror rate determination) by vendors and numerous literature reports have addressed PCR enzyme fidelity Nonetheless it isoften difficult to make direct comparisons among different enzymes due to numerous methodological and analytical differencesfrom study to study We have measured the error rates for 6 DNA polymerases commonly used in PCR applications including3 polymerases typically used for cloning applications requiring high fidelity Error rate measurement values reported here wereobtained by direct sequencing of cloned PCR products The strategy employed here allows interrogation of error rate across a verylarge DNA sequence space since 94 unique DNA targets were used as templates for PCR cloning The six enzymes included inthe study Taq polymerase AccuPrime-Taq High Fidelity KOD Hot Start cloned Pfu polymerase Phusion Hot Start and Pwopolymerase we find the lowest error rates with Pfu Phusion and Pwo polymerases Error rates are comparable for these 3 enzymesand aregt10x lower than the error rate observedwithTaq polymeraseMutation spectra are reported with the 3 high fidelity enzymesdisplaying broadly similar types of mutations For these enzymes transition mutations predominate with little bias observed fortype of transition

1 Introduction

With the rapid pace of developments in systems biology-based research for example genomics proteomics andmetabolomics larger-scale biological discovery projects arebecoming more common Put differently the scope of manyprojects has changed from the study of onefew targets tothe study of hundreds thousands or more An example ofresearch that has been transformed by developments in sys-tems biology is the cloning of expressed open reading frames(ORFs) from cDNA substrates The traditional path for ORFcloning has usually started with experimental observationsdriving the identification of one or several genes of interest

to a particular pathway Cloning of target(s) then typicallyresulted in further refinements of pathway details and oftenidentification of new cloning targets With the creation andcontinual refinements of databases of genomic sequencescloning now often takes place on a much larger scaleMicroarray technology and DNA sequencing breakthroughshave led to a vast increase in the number of ORFs presentin biological databases Furthermore biological observationsno longer necessarily precede target identification whichnow is often driven in large part by bioinformatics-basedpredictions and analyses Examples of large-scale cloningefforts include structural genomics projects to systematicallydetermine protein structures [1] pathogen ORF cloning to

Hindawi Publishing CorporationMolecular Biology InternationalVolume 2014 Article ID 287430 8 pageshttpdxdoiorg1011552014287430

2 Molecular Biology International

understand disease and therapeutic mechanisms [2] andcreation of the entire human ORFeome which will furtherdevelopments in basic and applied biomedical sciences [3]

DNA polymerases used to amplify targets during PCRcloning are high fidelity enzymes with error frequenciestypically in the range of 10minus6 mutationsbp amplified [4]Minimizing PCR-generated errors is especially important forlarger-scale cloning projects because given a sufficiently largepool of target DNA sequence even high fidelity enzymeswill produce clones with mutations There are a variety ofmethods to assay the fidelity of a DNA polymerase Howevererror frequencies for PCR enzymes are almost always assayedusing one (or a few) defined DNA target that samples alimited portion of DNA sequence space Early studies usingthe relatively low-fidelity Taq DNA polymerase relied onthe sequencing of cloned PCR products (eg [5 6]) Directsequencing of clones was a practical approach at the time dueto the low fidelity of the polymerase that is most clones thatwere sequenced would contain at least one mutation

With the introduction of higher fidelity polymerases newscreening methods were developed to rapidly interrogatelarge numbers of PCRproducts for the presence ofmutationsThese assays were based on a forward mutation fidelity assaydeveloped by Kunkel and colleagues which used a gap-filling reaction with a DNA polymerase on a lacZ templatesequence followed by ligation and transformation into Ecoli Colorimetric screening based on a functional lacZgene allowed rapid identification of mutations which weresubsequently sequenced to determine the nature of the DNAalteration [7] A similar approach was used to screen PCRproducts for mutations by cloning a lacZ fragment amplifiedby PCR as opposed to simple gap filling byDNApolymerasesThis method sometimes using a different reporter genehas been used to screen a variety of high fidelity PCRenzymes and to optimize PCR reaction conditions to mini-mize mutations [4 8] Finally methods that rely on assayingPCR mutations based on differing chemical properties (iemelting temperature) of reaction products with mismatchesrelative to perfect duplexes have been developed and appliedto a variety of enzyme systems [9 10] While reported fidelityvalues differ among research groups and assaymethods thereis a general consensus that a relatively low-fidelity enzymesuch as Taq has a fidelity value in the 10minus5 range and higherfidelity enzymes have values that are in the 10minus6 range (usuallyreported as mutations per bp per template doubling)

A tradeoff involved in using screeningmethods like thosedescribed above is that generally only one DNA sequence isinterrogated during the assay Additionally limitations builtinto the assays further restrict the possible mutations thatcan be detected For example the assay based on screeninglacZ gene amplification products uses a single 19 kb targetof which only 349 bases will produce a color change whenmutated [11] Likewise assaying mutations based on differ-ential duplex melting profiles is restricted to unique targetsequences that are short enough typically in the 100ndash300 bprange and have thermalmelting profiles that allow resolutionof single mismatches [9 10]

Because polymerase errors are known to be stronglydependent on DNA sequence context (reviewed in [12])ideally one would use a large set of DNA sequences whenmeasuring enzyme fidelity This becomes especially relevantin the context of large-scale cloning projects which involvehundreds or thousands of targets and thus contain an almostinfinite DNA sequence space To this end we have designedand executed a study that measures enzyme fidelity by directsequencing of cloned PCR products Falling costs for DNAsequencing have made this method of fidelity determinationpractical even for enzymes that make few mistakes Ourgoals are to compare fidelity values derived from direct clonesequencing to those derived from screening-based methodsas well as to evaluate these results in the context of choosingan enzyme for a high-throughput cloning project

2 Results and Discussion

To determine error rates and observe mutational spectrafor a variety of DNA polymerases used in PCR cloningwe directly sequenced clones produced from 94 differentplasmid templates These plasmids each with a unique targetDNA sequence are a subset of a larger group of glycosyl-transferase clones that we have prepared from Arabidopsisthaliana cDNA (manuscript in preparation)The 94 plasmidshave inserts with size ranging from 360 bp to 31 kb (median14 kb) and GC content ranging from 35 to 52 (median44) A summary of the 6 DNA polymerases used in thisstudy is presented in Table 1 We included Taq polymerasein our study because of the extensive body of literaturethat exists on the fidelity properties of this enzyme Theother enzymes included are all typically classified as ldquohighfidelityrdquo and therefore are potential candidates for large-scalecloning projects And while comparison of fidelity values isdifficult due to differences in assay and quantitation methodsamong different studies a general ranking of the enzymesstudied here (lowest fidelity to highest) appears to be Taq ltAccuPrime-Taq lt KOD asymp Pfu asymp Pwo lt Phusion

Our cloning pipeline uses recombinational insertion ofpurified PCR products into a plasmid vector using theGateway cloning system a method widely used for high-throughput cloning studies (reviewed in [17]) Since ourinput plasmid DNA templates were prepared using theGateway system the target genes of interest are all flankedby att recombination sequences This allowed the use ofcommon primers for all PCR reactions thus eliminatingthe need for target-specific optimizations Purified plasmidDNA was used as template for PCR and in all cases vendor-recommended buffers were used We used small amountsof plasmid template (25 pgrxn) in order to maximize thenumber of doublings in the PCR reaction and the size ofinsert relative to total plasmid size was taken into accountto determine the amount of target fragment present in thetemplate The PCR protocol used 30 cycles of amplificationwith an extension time of 2 minutescycle for targets le2 kb(82 of 94 targets) and 4 minutescycle for targets gt2 kb (12 of94 targets) Figure 1 shows gel images for a representative setof PCR reactions for each enzyme In all cases a single majorproduct band migrating at the expected size was observed

Molecular Biology International 3

Table 1 Published fidelity (error rate) values for DNA polymerases used in this study Due to the numerous methodological and analyticaldifferences among studies values are often reported as ranges Furthermore the references listed are meant to provide representative but notnecessarily exhaustive documentation for error rate values All values are given using Taq as the reference (1x)

Enzyme Published error rate(errorsbpduplication) Fidelity relative to Taq References

Taq 1ndash20 times 10minus5 1x [4 7 9]AccuPrime-Taq HF NA 9x better Vendor websiteKOD NA 4x better 50x better [13 14]Pfu 1-2 times 10minus6 6ndash10x better [4 15]

Phusion Hot Start 4 times 10minus7 (HF buffer)

95 times 10minus7 (GC buffer)

gt50x better (HF buffer)24x better (GC buffer) [16] Vendor website

Taq

AccuPrime-Taq

KOD Hot Start

Pfu (cloned)

Phusion Hot Start

Pwo

Figure 1 Representative agarose gel electrophoresis images ofproducts of PCR amplification of 24 unique DNA targets usingsix different enzymes Each lane contains 125 of the entire PCRreaction Expected product sizes range from 14 to 17 kb in size

Amplification efficiency was measured by quantitation ofPCR product using a dsDNA-specific dye and calculatingthe fold-amplification based on a known quantity of inputDNA template The fold-amplification is used to determinethe number of template doublings that occurred during PCRAs reported in Table 2 amplification efficiency values werefairly uniform for all samples within a plate We observesimilar amplification efficiencies between different enzymeswith the exception that we routinely observed fewer templatedoublings in reactions with Pfu polymerase We have keptthermocycling protocols constant for all enzymes and thus

it is possible that some parameters were not optimal foramplification by Pfu

Following amplification PCR products were purified byprecipitation with PEGMgCl

2

which is known to selectivelyfractionate DNA on the basis of size [18] to remove shortproducts lt300 bp in size This precipitation step can beperformed in 96-well plate format which is a requirementwhen the number of samples becomes largeWe have adoptedthis protocol for routine use and have observed a higherefficiency for insertion of correct-size DNA into the vectorcompared to purification using kit-based PCR purificationswhich typically have size cutoffs of sim100 bp (data not shown)In the case where off-target PCR products of gt300 bp arepresent gel extraction is used to isolate the desired productPurified PCR products were incubated with vector DNA andBP Clonase II and transformed into competent cells Threecolonies per plate were picked and grown up in 96-well platesand cultures were screened for correct-size insert by colonyPCR Insertion efficiency values for BP Clonase II expressedas the average number of clones having an insert atnear theexpected size (out of 3 colonies screened per transformation)were typically 80ndash90 (data not shown) For each target oneor more clones for each target containing a correct-size insert(if obtained) were cultured and used for DNA sequencing

For method validation purposes we used Taq DNApolymerase a Family A DNA polymerase and the enzymeused in the earliest PCR experiments [6] As an earlyworkhorse in PCR technology Taq polymerase has beenstudied extensively for purposes of fidelity determinationTaq DNA polymerase lacks a 31015840 rarr 51015840 exonuclease activityand thus is unable to correct misincorporated nucleotidesthat occur during DNA synthesis Various assays have beenused to assay Taq fidelity and depending on the methodused error rate values (expressed as mutations per base pairper template duplication) for Taq polymerase range from sim1times10minus5 (eg [4 19]) to 2times10minus4 (eg [7 20]) Furthermore the

mutational spectrum of Taq polymerase has been character-ized with A∙T rarr G∙C transitions predominating due to thepropensity for the enzyme tomisincorporate incoming dCTPwith a template thymine nucleotide [6 9 21]

By direct sequencing of clones from two independentPCR experiments with Taq polymerase we observed 99unique mutations out of gt100 kbp of target DNA sequenceThe type and number of individual mutations are listed


Table 2 Error rate values for six PCR enzymes included in this study are presented Each enzymewas used in two independent PCR reactionsAverage doublingsPCR reaction (119889) is the average of doubling values for each of the 94 PCR reactions in one plate where doublings arecalculated from the formula 2119889 = (ng DNA after PCRng DNA input) Error rate (f ) is calculated as f = 119899119878 (target size times 119889) where 119899 is thenumber of mutations observed for all clones that were sequenced and the (target size times 119889) for each target that was cloned

Enzyme ExptAvg

doublingsPCRreaction

Number ofclones

sequencedTotal bp sequenced

Number ofmutationsobserved

Error rate

Taq 1 205 plusmn 12 65 88 times 104 54 30 times 10minus5

2 167 plusmn 07 37 47 times 104 45 56 times 10minus5

AccuPrime-Taq 1 170 plusmn 12 75 10 times 105 18 10 times 10minus5

2 169 plusmn 06 ND ND ND ND

KOD 1 208 plusmn 15 70 10 times 105 16 76 times 10minus6


Pfu (cloned) 1 165 plusmn 11 151 20 times 105 9 28 times 10minus6


Phusion 1 210 plusmn 19 175 24 times 105 13 26 times 10minus6


Pwo 1 225 plusmn 12 170 24 times 105 13 24 times 10minus6

2 176 plusmn 06 ND ND ND NDND not determined

Table 3 Mutational spectra of six different PCR enzymes are presented Results with Taq polymerase are mutations observed in sequencingclones from two independent PCR reactions For indel mutations the type of insertion or deletion is indicated

Taq AccuPrime-Taq HF KOD Pfu Phusion PwoTransitions

A∙TrarrG∙C 67 12 9 3 6 6G∙CrarrA∙T 28 5 5 5 5 4

TransversionA∙TrarrT∙A 1 1 1 1A∙TrarrC∙G 2G∙CrarrT∙A 3

Indels 1(T del)

1(T del)

2(A ins TCT del in

(TCT)5 run)Insertion = ins deletion = del

in Table 3 Given the amplification efficiency of each PCRreaction the error rate (average of 2 experiments) for Taqpolymerase is 43 times 10minus5 plusmn 18 mutationsbp per templateduplication This value is in excellent agreement with otherpublished values for this enzyme and the relatively highvariance suggests that calculated error values differing byup to 2-fold are probably not significant relative to theexperimental noise The majority of the mutations (67 of 99)are A∙T rarr G∙C transitions which could result from eitherincoming dCTP mispairing with template A or incomingdGTP mispairing with template T Transitions of the G∙CrarrA∙T type resulting from either incoming TTP mispairingwith template G or incoming dATPmispairing with templateC are the second most prevalent mutation (28 of 99)There were 3 transversion mutations with 1 A∙TrarrT∙A and2 A∙TrarrC∙G changes Overall the spectrum of the basesubstitutionmutations agreeswell with previous observationson Taq polymerase reported in the literature [7] There was

only one insertion or deletion (indel) mutation observed inour data set a single T deletion in a T

3

template sequenceTaqpolymerase has been reported to produce indel mutationswith a significant frequency as much as approximately 25of total mutations with all occurring in homopolymericruns [7] Since our target pool contains 1481 instances ofhomopolymer runs of at least 4 bp we suspect that other dif-ferences between the earlier assay conditions and those usedhere explain the discrepancy Specifically the earlier exper-iments were performed with elevated magnesium (10mMversus 15mM used here) and elevated dNTP levels (1mMversus 02mM used here) Both elevated magnesium anddNTP levels were subsequently shown to elevate frameshift(indel) mutations preferentially relative to base substitutionmutations [21]

An important control for these experiments is neces-sitated by the method used to generate template forDNA sequencing For larger-scale cloning projects DNA


sequencing using cell culture is advantageous because of thesaving in time and resources relative to purifying plasmidDNA However sequencing using cell culture requires aPCR or another amplification step and this step could inprinciple be a source of ldquoadditional mutationrdquo To address thisdirectly we sequencedminiprepDNAprepared from a subsetof clones produced with Taq polymerase Every one of thefourteenmutations detected in the subset using cell culture asthe source for sequencing template was also observed whensequencing from plasmid DNA template (data not shown)We conclude that our method has a false positive rate of lt7(114) and is acceptable for assaying PCR-induced mutationsFurthermore based on our results with Taq polymerase weconclude that our method for fidelity determination givesresults in excellent agreement with other studies and is thusan accurate measure of polymerase accuracy

Our results indicate that 3 of the enzymes includedin the study Pfu polymerase Phusion Hot Start and Pwopolymerase have error rates that are significantly lowerthan the others This is consistent with previous findingsdemonstrating very high fidelity PCR amplification for theseenzymes Interestingly error frequency values for these threeenzymes are extremely similar to each other approximately2-3 times10minus6 mutationsbptemplate doubling The slight errorfrequency value differences are probably not significant giventhat the small number of mutations is produced by thesehigh fidelity polymerases in addition to the experimentalvariability discussed above for the results with Taq Giventhe costs of cloning and sequencing and finite researchbudgets mutation detection by DNA sequencing of clonesgenerates a relatively small data set of mutations when theenzyme fidelity is high This is a drawback to our assayand despite the fact that DNA sequencing costs continueto drop screening bacteria is still a far more economicalmethod of interrogating a large number of clones For allmutant clones produced by Pfu Phusion Hot Start and Pwopolymerases samples were resequenced to rule out sampleprocessing or DNA sequencing as a source of error In allcases the original mutation was present confirming the PCRreaction as the most likely source of the mutation From thestandpoint of use in a large-scale cloning project any one ofthese enzymes would be acceptable judged on the criteria ofminimizing error rate Other factors need to be consideredof course such as amplification efficiency mutation spectraperformance with high GC content templates and cost toname a few As far as mutation spectra the 3 high fidelitypolymerases all produced predominantly (gt75) transitionmutations with no significant template bias With Phusionenzyme we observed 15 (213) indel mutations whichare problematic for cloning applications where translationreading frame should be maintained Both indel mutationsoccurred in repeat regions with one being an A insertioninto an A

3

template sequence and the other being a (TCT)deletion within a (TCT)

5

template sequence This resultwas unexpected in light of the high processivity of Phusionpolymerase relative to other commonly used PCR enzymes(vendor website) Because multiple studies have found thatincreased polymerase processivity reduces the frequency of

slippage mutations that result in indel mutations [22 23] weexpected Phusion to produce the fewest of this class of errorsIt should be noted however that this conclusion is based ona small sample size and a larger number of mutations shouldbe analyzed for confirmation

It was interesting to us that none of the enzymes testedhere was found to have an error rate below sim2 times 10minus6Other studies in the literature have reported sub-10minus6 errorfrequencies for PCR enzymes 65 times 10minus7 [10] for Pfupolymerase assayed by differential duplex Tm measurementand 42 times 10minus7 for Phusion using HF buffer assayed witha method called BEAMING [16] For the study of Phusionfidelity the PCRused a different buffer than the one employedhere which according to the vendor does result in a 2-3-fold lower error rate In addition that study uses theBEAMING method an extremely sensitive flow cytometricprotocol that screens large numbers of beads that containPCR products for the presence of nucleotide variationsHowever only one specificmutation aG∙C rarr A∙Tmutationat a single position was interrogated in that study Thuswhile the assay is extremely sensitive for detection of definedmutations results obtained with the BEAMING method formutation frequency at a single position may not necessarilyreflect the fidelity properties of an enzyme for much largersequence spaces For the study on Pfu error rate severalfundamental methodological differences are present in theearlier study the PCR was performed under ldquoalmost anaero-bicrdquo conditions with significantly shorter cycling times thetarget size was limited to 93 bp and mutation detectionrelied on a physiochemical method separation and isolationof PCR products containing mismatches by capillary elec-trophoresis [10] Andwhile thismethod has been successfullyused in the detection of rare mutations in mitochondrialDNA samples from normal and cancer tissues [24] therequirement for a mutation to result in a molecule with analtered melting profile may bias the number of mutationsthat can be detected A major discrepancy between ourresults and those from this earlier report on Pfu fidelitywhich may be connected to the differing mutation detectionmethodologies can be seen in the mutation spectra resultsin Table 3 We observed sim90 (8 of 9) transition mutationswith a slight bias for G∙C rarr A∙T alterations In con-trast the study using capillary electrophoresis for detectionresulted in predominantly (35) transversion mutations witha single A∙T rarr G∙C transition and a single 1 bp deletionmutation Transversion mutations require the polymerase tosynthesize either a purine∙purine or a pyrimidine∙pyrimidinemismatch both of which are significantly disfavored relativeto the different purine∙pyrimidine mismatches in Family Bpolymerases including Pfu polymerase [25 26] Because thetypes of mutations we observe are consistent with previouslyreported mutational spectra for other Family B polymeraseswe believe our method has detected polymerase errors in abias-free fashion

The other two enzymes included in our study KOD poly-merase andAccuPrime-TaqHigh Fidelity have fidelity valuesintermediate between Taq polymerase and the higher fidelityenzymes The error rate observed for KOD polymerase was


only sim4-fold lower than that of Taq polymerase and sim25-fold higher than for Pfu polymerase The initial report onfidelity of KOD polymerase a Family Bpola-like polymerasefrom Thermococcus kodakaraensis KOD1 reported an errorrate very slightly lower than Pfu polymerase and sim4-foldlower than forTaq polymerase [13]That study used a forwardmutation assay (not PCR) expressed fidelity simply as theratio of white colonies to blue with no accounting for PCRamplification efficiency and used experimental conditions(Mg2+ concentration) that differ significantly from typicalPCR conditions A subsequent study measuring fidelityunder PCR conditions using a different reporter gene but stilla simple ratio of mutant to wild-type colonies reported errorrates sim50x lower than those with Taq and marginally lowerthan those for Pfu polymerase [14] In neither of those studieswas there a report of themolecular changes leading tomutantcolonies The large difference between these two resultswhich are from the same research group serves to high-light the difficulties in making comparisons between studieswhere there are significant methodological differences Inthe present study we find that the mutation spectrum forKODpolymerase is similar to the other B-family polymerases(Pfu Pwo and Phusion) assayed here As shown in Table 3transitions predominate (14 of 16mutations) with a slight bias(64) for A∙T rarr G∙C mutations

For the PCR performed with AccuPrime-Taq HighFidelity system we observed a 3-fold improvement infidelity relative to Taq polymerase According to the vendorAccuPrime-Taq High Fidelity is an enzyme blend that con-tains Taq polymerase a processivity-enhancing protein anda higher fidelity proofreading polymerase from Pyrococcusspecies GB-D The lower error rate seen with AccuPrime-Taqmost likely arises from theGB-D polymerase editingmistakesintroduced by Taq polymerase as opposed to enhancedprocessivity since increased processivity has been shown tohave no significant effect on base substitution errors [2227] The mutation spectrum of the blend is almost identicalto that seen with Taq polymerase alone with transitionspredominant and a significant bias for A∙T rarr G∙C changes(71 for AccuPrime-Taq versus 73 for Taq) However itshould be noted that a study on the mutation spectra ofGB-D DNA polymerase (commercially available as DeepVent) found A∙T rarr G∙C transitions to be the predominantmutation [28] Detailed analysis on the contribution of eachenzyme to the overall mutation spectrum is also precluded bythe proprietary enzyme formulation used by the vendor

In summary we have used direct DNA sequencing ofcloned PCR products to assay polymerase fidelity and evalu-ate other aspects of enzyme suitability for large-scale cloningprojects Based on minimizing PCR errors Pfu polymerasePwo polymerase and Phusion all produce acceptably lowlevels of mutations Phusion was observed to produce moreindel mutations than Pfu or Pwo polymerases althoughthe total number of mutations was limited This type ofmutation is particularly problematic forORF cloning projectsand should be taken into account in the process of enzymeselection Aside from fidelity considerations amplificationefficiency values were significantly higher for Phusion and

Pwo compared to Pfu although further optimization ofthe PCR reaction for Pfu would likely improve efficiencyvalues Likewise for cloning projects where targets are eithervery long or very highly GC-rich fidelity may be of lesserimportance relative to the ability to amplify ldquodifficultrdquo targetDNA And finally since the application space for PCRtechnology is huge with cloning representing only a smallfraction enzymes other than those studied here need to becompared and evaluated based on project-specific needs andchallenges

3 Materials and Methods

31 PCR Reactions All enzymes and reaction buffers werefrom commercial sources Fermentas (Taq polymerase)InvitrogenLife Technologies (AccuPrime-Taq) EMDChemicalsNovagen (KOD Hot Start) Agilent (cloned Pfupolymerase) Finnzymes (Phusion Hot Start) and Roche(Pwo polymerase) PCR reactions were carried out in a finalvolume of 50120583Lusing buffer conditions and enzyme amountsrecommended by the vendor For reactions with Phusionthe GC buffer was used In all cases reactions included02mM each dNTP (Fermentas) and 02mM each primer(IDT) with the sequences (51015840 to 31015840) GGGGACAAGTTTG-TACAAAAAAGCAGGCTTCACC for the forward primerand GGGGACCACTTTGTACAAGAAAGCTGGGTC forthe reverse primer Template for PCR reactions was miniprepplasmid DNA with each plasmid template containing aunique target sequence of known sequence and size rangingfrom 03 to 3 kb The target insert was cloned in between theatt sites of a pDONR vector allowing the use of a commonprimer set for all plasmids Each PCR reaction contained0025 ng plasmid DNA quantitated using the PicoGreenDNA quantitation reagent (InvitrogenLife Technologies)and thus the amount of input target (i) was calculated as119894 = 0025 ng times (size of target divide (size of target + size ofplasmid)) The thermocycling protocol for all reactions withtarget length le2 kb was 5 minutes 95∘C then 30 cycles of15 seconds 95∘C rarr 30 seconds 55∘C rarr 2 minutes 72∘Cand finally 7 minutes at 72∘C For the targets gt2 kb in sizethe 2-minute extension step was extended to 4 minutes Foranalysis of PCR products by gel 2 120583L of each PCR reactionwas run on a 2 agarose eGel (InvitrogenLife Technologies)run according to vendor recommendations

32 Quantitation of PCR Reactions Efficiency of PCR ampli-fication was determined bymeasuring the amount of productusing a modified PicoGreen dsDNA quantitation assay Thismethod was facilitated by optimizing the PCR reaction toproduce a single product band (Figure 1) Using a Biomek FX-P (Beckman) automated liquid handing system 5 120583L of eachPCR reaction was diluted 50-fold in TE buffer (pH 8) into anew 96-well plate From this plate 5 120583L from each well wasmixed with 195 120583L of PicoGreen solution a 500-fold dilutionof dye in TE (pH 8) Fluorescence measurements were takenwith a Paradigm (Beckman) plate reader Background fluo-rescence was determined from a PCR reaction that containedno template DNA Following background subtraction DNAconcentration was determined by comparing fluorescence


readings to those obtained with a standard curve using DNAof known concentration supplied with the dye Extent oftarget amplification (e) is calculated as e = (ng DNA afterPCR)divide (ng of target DNA input) and the number of templatedoublings during PCR (d) can be calculated as 119890 = 2119889

33 Cloning of PCR Products PCR reactions were purifiedin 96-well plate format by the addition of PEG 8000 andMgCl

2

to final concentrations of 10 and 10mM respectivelydirectly to each well of the PCR plate using a multichannelpipettor The plate was spun at 4000 rpm for 60 minutes atroom temperature and the supernatant was discarded Pelletswere washed two times with cold isopropanol air-dried andresuspended in 25 120583L TE (pH 8) This protocol resulted inexcellent yields (50ndash75) of PCR products with no productslt300 bp as judged by gel electrophoresis Purified PCR prod-ucts were cloned into a pDONR223 vector (a generous gift ofDrs Dominic Esposito and JimHartley NCI FrederickMD)using BP Clonase II (InvitrogenLife Technologies) Clonasereactions were assembled using a multichannel pipettor in96-well PCR plates in a 5 120583L volume and contained 75 ngpDONR223 1 120583L purified PCR product (typically 50ndash150 ngDNA) and 1 120583L BP Clonase II Sealed plates were incubatedat least 16 hours at 25∘C and 1 120583L of each reaction was imme-diately (no proteinase K treatment) used to transform either25 120583L or 50120583L of competent TOP10 cells (InvitrogenLifeTechnologies) Following heat shock and recovery followingaddition of 250120583L of SOCmedia 100120583L of cells was plated onLB plates containing 50mgmL spectinomycin Equivalentnumbers of colonies were observed in transformations using25 120583L or 50 120583L of frozen competent cells and control BPreactions lacking BP Clonase II or PCR product resulted inno transformants

34 Screening of Transformants Three colonies from eachtransformation plate were picked and cultured in 96-wellplates (Costar 3788) sealed with gas-permeable membranewith each colony incubated in 150mL of LB media with50mgmL spectinomycin and 10 glycerol After overnightincubation at 37∘C (no shaking) 1120583L of each culture wasused to screen by colony PCR for the presence of insert withexpected size Colony PCR reactions (25mL) used the sameprimers used for cloning at a final concentration of 01mMeach with 30 amplification cycles as described above withGoTaq polymerase (Promega) Reactions were analyzed byagarose gel electrophoresis and the presence of a band at ornear the expected size was scored as a ldquohitrdquo The number ofhits (0ndash3) for each target was determined and an averagenumber of hits per target for each plate were determined andused as a measure of Clonase reaction efficiency

35 Clone Sequencing In cases where Clonase efficiencyvalues were gt66 average of at least 2 hits out of 3 coloniesscreened the entire liquid culture plate was replicated with a96-pin replicator onto an agar platewith the same dimensionsas a 96-well plate The plate was immediately submitted toan outside vendor (Quintarabio Berkeley CA) and aftergrowth overnight sequencing was performed on amplified

DNA from each clone If Clonase efficiency values werelt66 (Taq and Pfu polymerase reactions) a rearray stepwas added using a Qpix2 colony picking robot (Genetix)to maximize the number of clones with correct-size inserton one plate For comparing sequencing results using cellsversus miniprep DNA one plate of colonies picked froma Taq cloning reaction was replicated into a 96-well deepwell plate with 800mL media per well and grown overnightwith shaking at 300 rpm Cells were pelleted and DNA wasprepared using a Qiaprep 96 Turbo Miniprep Kit (Qiagen)Eluted DNA was submitted directly for sequencing

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

This work was part of the DOE Joint BioEnergy Institute(httpwwwjbeiorg) supported by the US Department ofEnergy Office of Science Office of Biological and Envi-ronmental Research through Contract DE-AC02-05CH11231between Lawrence Berkeley National Laboratory and theUSDepartment of Energy The authors would like to thank DrsDominic Esposito and Jim Hartley (NCI Frederick MD)for the gift of pDONR223 DNA Huu M Tran (JBEISandiaNational Laboratories) for assistance with the laboratoryautomation Drs Richard Shan and Sue Zhao (QuintaraBiosciences Berkeley CA) for DNA sequencing and analysisand Dr Nathan J Hillson for critical reading of the paper

References

[1] B G Fox C Goulding M G Malkowski L Stewart and ADeacon ldquoStructural genomics from genes to structures withvaluable materials and many questions in betweenrdquo NatureMethods vol 5 no 2 pp 129ndash132 2008

[2] A Rolfs W R Montor S Y Sang et al ldquoProduction andsequence validation of a complete full lengthORF collection forthe pathogenic bacterium Vibrio choleraerdquo Proceedings of theNational Academy of Sciences of the United States of Americavol 105 no 11 pp 4364ndash4369 2008

[3] G Temple P Lamesch S Milstein et al ldquoFrom genome toproteome developing expression clone resources for the humangenomerdquoHumanMolecular Genetics vol 15 no 1 pp R31ndashR432006

[4] J Cline J C Braman and H H Hogrefe ldquoPCR fidelity of PfuDNA polymerase and other thermostable DNA polymerasesrdquoNucleic Acids Research vol 24 no 18 pp 3546ndash3551 1996

[5] A M Dunning P Talmud and S E Humphries ldquoErrors in thepolymerase chain reactionrdquo Nucleic Acids Research vol 16 no21 p 10393 1988

[6] R K Saiki D H Gelfand S Stoffel et al ldquoPrimer-directedenzymatic amplification of DNA with a thermostable DNApolymeraserdquo Science vol 239 no 4839 pp 487ndash491 1988

[7] K R Tindall and T A Kunkel ldquoFidelity of DNA synthesis bytheThermus aquaticus DNA polymeraserdquo Biochemistry vol 27no 16 pp 6008ndash6013 1988


[8] JM Flaman T Frebourg VMoreau et al ldquoA rapid PCRfidelityassayrdquo Nucleic Acids Research vol 22 no 15 pp 3259ndash32601994

[9] P Keohavong and W G Thilly ldquoFidelity of DNA polymerasesin DNA amplificationrdquo Proceedings of the National Academy ofSciences of the United States of America vol 86 no 23 pp 9253ndash9257 1989

[10] P Andre A Kim K Khrapko and W G Thilly ldquoFidelity andmutational spectrum of Pfu DNA polymerase on a humanmitochondrial DNA sequencerdquo Genome Research vol 7 no 8pp 843ndash852 1997

[11] G S Provost P L Kretz R T Hamner et al ldquoTransgenicsystems for in vivo mutations analysisrdquo Mutation Research-Fundamental and Molecular Mechanisms of Mutagenesis vol288 no 1 pp 133ndash149 1993

[12] T A Kunkel andK Bebenek ldquoDNA replication fidelityrdquoAnnualReview of Biochemistry vol 69 pp 497ndash529 2000

[13] M Takagi M Nishioka H Kakihara et al ldquoCharacterizationof DNA polymerase from Pyrococcus sp strain KOD1 and itsapplication to PCRrdquo Applied and Environmental Microbiologyvol 63 no 11 pp 4504ndash4510 1997

[14] M Kitabayashi Y Nishiya M Esaka M Itakura andT Imanaka ldquoGene cloning and polymerase chain reactionwith proliferating cell nuclear antigen from Thermococcuskodakaraensis KOD1rdquo Bioscience Biotechnology and Biochem-istry vol 66 no 10 pp 2194ndash2200 2002

[15] K S Lundberg D D Shoemaker MWW Adams J M ShortJ A Sorge and E J Mathur ldquoHigh-fidelity amplification usinga thermostable DNA polymerase isolated from Pyrococcusfuriosusrdquo Gene vol 108 no 1 pp 1ndash6 1991

[16] M Li F Diehl D Dressman B Vogelstein and K W KinzlerldquoBEAMing up for detection and quantification of rare sequencevariantsrdquo Nature Methods vol 3 no 2 pp 95ndash97 2006

[17] G Marsischky and J LaBaer ldquoMany paths to many clonesa comparative look at high-throughput cloning methodsrdquoGenome Research vol 14 no 10 pp 2020ndash2028 2004

[18] J T Lis ldquoFractionation of DNA fragments by polyethyleneglycol induced precipitationrdquo Methods in Enzymology vol 65no 1 pp 347ndash353 1980

[19] J J Choi J Song H N Ki et al ldquoUnique substrate spectrumand PCR application of Nanoarchaeum equitans family B DNApolymeraserdquo Applied and Environmental Microbiology vol 74no 21 pp 6563ndash6569 2008

[20] L L Ling P Keohavong C Dias and W G Thilly ldquoOptimiza-tion of the polymerase chain reaction with regard to fidelitymodified T7 Taq and vent DNA polymerasesrdquo PCR Methodsand Applications vol 1 no 1 pp 63ndash69 1991

[21] K A Eckert and T A Kunkel ldquoHigh fidelity DNA synthesisby the Thermus aquaticus DNA polymeraserdquo Nucleic AcidsResearch vol 18 no 13 pp 3739ndash3744 1990

[22] J F Davidson R Fox D D Harris S Lyons-Abbott and LA Loeb ldquoInsertion of the T3 DNA polymerase thioredoxinbinding domain enhances the processivity and fidelity of TaqDNA polymeraserdquo Nucleic Acids Research vol 31 no 16 pp4702ndash4709 2003

[23] E Viguera D Canceill and S D Ehrlich ldquoReplication slip-page involves DNA polymerase pausing and dissociationrdquo TheEMBO Journal vol 20 no 10 pp 2587ndash2595 2001

[24] K KhrapkoH Coller P Andre et al ldquoMutational spectrometrywithout phenotypic selection human mitochondrial DNArdquoNucleic Acids Research vol 25 no 4 pp 685ndash693 1997

[25] A Niimi S Limsirichaikul S Yoshida et al ldquoPalm mutants inDNA polymerases 120572 and 120578 Alter DNA replication fidelity andtranslesion activityrdquoMolecular and Cellular Biology vol 24 no7 pp 2734ndash2746 2004

[26] E M Kennedy C Hergott S Dewhurst and B Kim ldquoThemechanistic architecture of thermostable Pyrococcus furiosusfamily B DNA polymerase motif A and its interaction with thedNTP substraterdquo Biochemistry vol 48 no 47 pp 11161ndash111682009

[27] T A Kunkel S S Patel and K A Johnson ldquoError-pronereplication of repeated DNA sequences by T7 DNA polymerasein the absence of its processivity subunitrdquo Proceedings of theNational Academy of Sciences of the United States of Americavol 91 no 15 pp 6830ndash6834 1994

[28] H Huang and P Keohavong ldquoFidelity and predominantmutations produced by deep vent wild-type and exonuclease-deficient DNApolymerases during in vitroDNA amplificationrdquoDNA and Cell Biology vol 15 no 7 pp 589ndash594 1996

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Anatomy Research International

PeptidesInternational Journal of


Hindawi Publishing Corporation httpwwwhindawicom

International Journal of

Volume 2014

Zoology


Molecular Biology International

GenomicsInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


BioinformaticsAdvances in

Marine BiologyJournal of



Signal TransductionJournal of


BioMed Research International

Evolutionary BiologyInternational Journal of



Biochemistry Research International

ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Genetics Research International


Advances in

Virolog y

Hindawi Publishing Corporationhttpwwwhindawicom

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational



Enzyme Research



Microbiology


understand disease and therapeutic mechanisms [2] andcreation of the entire human ORFeome which will furtherdevelopments in basic and applied biomedical sciences [3]

DNA polymerases used to amplify targets during PCRcloning are high fidelity enzymes with error frequenciestypically in the range of 10minus6 mutationsbp amplified [4]Minimizing PCR-generated errors is especially important forlarger-scale cloning projects because given a sufficiently largepool of target DNA sequence even high fidelity enzymeswill produce clones with mutations There are a variety ofmethods to assay the fidelity of a DNA polymerase Howevererror frequencies for PCR enzymes are almost always assayedusing one (or a few) defined DNA target that samples alimited portion of DNA sequence space Early studies usingthe relatively low-fidelity Taq DNA polymerase relied onthe sequencing of cloned PCR products (eg [5 6]) Directsequencing of clones was a practical approach at the time dueto the low fidelity of the polymerase that is most clones thatwere sequenced would contain at least one mutation

With the introduction of higher fidelity polymerases newscreening methods were developed to rapidly interrogatelarge numbers of PCRproducts for the presence ofmutationsThese assays were based on a forward mutation fidelity assaydeveloped by Kunkel and colleagues which used a gap-filling reaction with a DNA polymerase on a lacZ templatesequence followed by ligation and transformation into Ecoli Colorimetric screening based on a functional lacZgene allowed rapid identification of mutations which weresubsequently sequenced to determine the nature of the DNAalteration [7] A similar approach was used to screen PCRproducts for mutations by cloning a lacZ fragment amplifiedby PCR as opposed to simple gap filling byDNApolymerasesThis method sometimes using a different reporter genehas been used to screen a variety of high fidelity PCRenzymes and to optimize PCR reaction conditions to mini-mize mutations [4 8] Finally methods that rely on assayingPCR mutations based on differing chemical properties (iemelting temperature) of reaction products with mismatchesrelative to perfect duplexes have been developed and appliedto a variety of enzyme systems [9 10] While reported fidelityvalues differ among research groups and assaymethods thereis a general consensus that a relatively low-fidelity enzymesuch as Taq has a fidelity value in the 10minus5 range and higherfidelity enzymes have values that are in the 10minus6 range (usuallyreported as mutations per bp per template doubling)

A tradeoff involved in using screeningmethods like thosedescribed above is that generally only one DNA sequence isinterrogated during the assay Additionally limitations builtinto the assays further restrict the possible mutations thatcan be detected For example the assay based on screeninglacZ gene amplification products uses a single 19 kb targetof which only 349 bases will produce a color change whenmutated [11] Likewise assaying mutations based on differ-ential duplex melting profiles is restricted to unique targetsequences that are short enough typically in the 100ndash300 bprange and have thermalmelting profiles that allow resolutionof single mismatches [9 10]

Because polymerase errors are known to be stronglydependent on DNA sequence context (reviewed in [12])ideally one would use a large set of DNA sequences whenmeasuring enzyme fidelity This becomes especially relevantin the context of large-scale cloning projects which involvehundreds or thousands of targets and thus contain an almostinfinite DNA sequence space To this end we have designedand executed a study that measures enzyme fidelity by directsequencing of cloned PCR products Falling costs for DNAsequencing have made this method of fidelity determinationpractical even for enzymes that make few mistakes Ourgoals are to compare fidelity values derived from direct clonesequencing to those derived from screening-based methodsas well as to evaluate these results in the context of choosingan enzyme for a high-throughput cloning project

2 Results and Discussion

To determine error rates and observe mutational spectrafor a variety of DNA polymerases used in PCR cloningwe directly sequenced clones produced from 94 differentplasmid templates These plasmids each with a unique targetDNA sequence are a subset of a larger group of glycosyl-transferase clones that we have prepared from Arabidopsisthaliana cDNA (manuscript in preparation)The 94 plasmidshave inserts with size ranging from 360 bp to 31 kb (median14 kb) and GC content ranging from 35 to 52 (median44) A summary of the 6 DNA polymerases used in thisstudy is presented in Table 1 We included Taq polymerasein our study because of the extensive body of literaturethat exists on the fidelity properties of this enzyme Theother enzymes included are all typically classified as ldquohighfidelityrdquo and therefore are potential candidates for large-scalecloning projects And while comparison of fidelity values isdifficult due to differences in assay and quantitation methodsamong different studies a general ranking of the enzymesstudied here (lowest fidelity to highest) appears to be Taq ltAccuPrime-Taq lt KOD asymp Pfu asymp Pwo lt Phusion

Our cloning pipeline uses recombinational insertion ofpurified PCR products into a plasmid vector using theGateway cloning system a method widely used for high-throughput cloning studies (reviewed in [17]) Since ourinput plasmid DNA templates were prepared using theGateway system the target genes of interest are all flankedby att recombination sequences This allowed the use ofcommon primers for all PCR reactions thus eliminatingthe need for target-specific optimizations Purified plasmidDNA was used as template for PCR and in all cases vendor-recommended buffers were used We used small amountsof plasmid template (25 pgrxn) in order to maximize thenumber of doublings in the PCR reaction and the size ofinsert relative to total plasmid size was taken into accountto determine the amount of target fragment present in thetemplate The PCR protocol used 30 cycles of amplificationwith an extension time of 2 minutescycle for targets le2 kb(82 of 94 targets) and 4 minutescycle for targets gt2 kb (12 of94 targets) Figure 1 shows gel images for a representative setof PCR reactions for each enzyme In all cases a single majorproduct band migrating at the expected size was observed








Taq

AccuPrime-Taq

KOD Hot Start

Pfu (cloned)

Phusion Hot Start

Pwo





2







Enzyme ExptAvg


Number ofclones



Error rate

















Indels 1(T del)

1(T del)

2(A ins TCT del in




3






3


5
















2







Acknowledgments


References





































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology








Taq

AccuPrime-Taq

KOD Hot Start

Pfu (cloned)

Phusion Hot Start

Pwo





2







Enzyme ExptAvg


Number ofclones



Error rate

















Indels 1(T del)

1(T del)

2(A ins TCT del in




3






3


5
















2







Acknowledgments


References





































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



Enzyme ExptAvg


Number ofclones



Error rate

















Indels 1(T del)

1(T del)

2(A ins TCT del in




3






3


5
















2







Acknowledgments


References





































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology




3


5
















2







Acknowledgments


References





































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology












2







Acknowledgments


References





































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology






























Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

research article error rate comparison during polymerase ...downloads.hindawi.com › archive ›...

Documents