DNAWorks全基因合成介绍.ppt
Gene Synthesis using DNAWorksDr.David HooverHelix Systems,SCB,CIT,NIHGene SynthesisSeveral methodsligation-incredibly tedious and inefficientFokI-sequence dependent(type IIs.)serial cloning-sequence dependentassembly or self-priming PCRGene Synthesis MethodsThermodynamically Balanced ConventionalThermodynamically Balanced Inside-OutProtein ExpressionProtein/Structure Independent Factors:promoters and upstream elementstranslational initiation and terminationmRNA stabilitycodon biasProtein/Structure Dependent Factors:folding and aggregationproteolysis and degradationsecretion and localizationCodon BiasSynthetic GenesBenefits:Codon use optimized for hostFlexibility in subcloningEase of complex mutagenesisProblems:Time consumingComplicatedError-proneCommercial SourcesBlue Heron Biotechnology(http:/)DNA 2.0(http:/ Script Corporation(http:/)BioNexus Inc.(http:/)Entelechon(http:/)GeneArt(http:/)Codon Devices(http:/)Commerical SourcesTypical costs:$0.79-$3.60/bpComplexities?Intellectual property?800 bp=$1000(Gene Script)Genes From Scratcholigos$0.20/nt(NIH discount)PCR reagents$2/reaction sequencing$20/600 bpelectrophoresis$3/gellabor$20/hrGFP,238 aa,714 bp,20 oligos,1134 nt,2 reactions,2 gels,4 sequences,10 hrs=$517How to design oligosreverse-translate protein into DNA,optimum codon usagebreak into fragments of equal overlap Tmoptimize:hairpins/mRNA structurerepeats/misprimingrestriction site inclusion/exclusionlengthDNAWorkshttp:/DNAWorks Output 181 TCTGGTGAAGGCGAGGGTGACGCGACCTACGGTAAACTCACTCTCAAAT agaccact TGCCATTTGAGTGAGAGTTTAAGTAGACGTGG 241 ggttccttggccgaccctggttactaccttctcttacggtgttcag TGCCCGTTTGACGGCCAAGGAACCGGCTGG tc 12 nt(Short)Tm range should not be 3C(TmRange)Dont depend entirely on scoresArbitrary,somewhat dependent on lengthTricksChoosing codonsrandom-slower optimization,less constrainedstrict-for the fussyscored-if codon score really mattersTm,Length ranges,Number of SolutionsTo find the very best solutionno more than 999TricksDesign multi-use and interchangeable oligosFlanking primers with standard overlapsIntersperse nucleotide elements between protein elementsGap-fix restriction sitesAllow for mutations later onRandom mutagenesisNucleotide sequences can be degenerateTricksThermodynamically Balanced Inside-Out ModeMulti-step PCRMore controlled,reliable methodGao X.,et al.,Nucleic Acids Res 2003Random oligo lengthsFaster,better optimizationFor the not-so-fussyProbably best for DNA-only genesTricksSet Tm higher64C-70Clonger oligos,extra purification($)Always double check!Nothing is foolproofThink carefully about what you need BEFORE starting workAlways run final sequences through alternate program(EMBOSS,GCG-Lite)Make sure oligos are what you intendedPCRMix all oligos and additivesSpecific PCR protocolsAnalytical gelIsolate desired productsAssembly ProtocolOligos 1 l625 nM each25 nM eachdNTPs 2 l 2.5 mM each0.25 mM eachH2O19 lBuffer2.5 l10X1XPfu pol.0.5 l95C2.0 1X95C0.5 6555(-0.5)0.5 20X72C0.5 72C5 1X4CholdAmplification ProtocolPCR mix 2 l?dNTPs 8 l 2.5 mM each0.2 mM each3 primer4 l10 M400 nM5 primer4 l10 M400 nMBuffer10 l10X1XH2O70 lPfu pol.2 l95C2.0 1X95C0.5 62C0.5 20X72C0.5 72C5 1X4CholdProblemsNo product(complete failure)Wrong size product(mispriming)Mutations(2 out of 3 correct,2 errors/kb)Sequencing is warranted.FixesOptimize PCR conditionsBreak gene synthesis into steps(TBIO)Errorsp=mutation rate/1000 nt/duplication(Cline et al.,Nucleic Acids Res 24(1996)TaqKOD(NovagenPfuUltra(StratageneThe probability of a gene n bp in length having no errors using a polymerase with mutation rate p:p=(1-p)nTherefore,p for a 738 bpErrorsThe number of clones needed to screen to find a correct gene with 95%confidence:N=log(0.05)/log(1-p)Thus,log(0.05)/log(1-0.728)=3 clones need to be sequenced.From Wu et al.,J Biotech 124(2006)TimeFind protein of interest,design oligos,order oligosRun PCR,integrate into sequencing vector,transformPick colony,grow overnight cultureMiniprep construct,integrate into expression vector,transformPick colony,grow overnight cultureRun expression growth trials 1 week between concept and initial trial(at best!)Can be automated and parallelized(96 well plates?)
