郝權(quán)-基于XFEL的結(jié)構(gòu)生物學(xué)研究

3/15/2023

基于XFEL的結(jié)構(gòu)生物學(xué)研究范海福中國科學(xué)院物理研究所郝權(quán)（第四次XFEL香山會議）蛋白質(zhì)晶體結(jié)構(gòu)X射線衍射分析的幾個重要環(huán)節(jié)jhkl模型構(gòu)建結(jié)構(gòu)精修|F(hkl)|相位推演晶體數(shù)據(jù)采集試樣制備三維結(jié)構(gòu)X光激光會有重要影響的幾個環(huán)節(jié)X光激光如何影響現(xiàn)有的X射線衍射分析●相干性●高強(qiáng)度●短脈沖Imaging

AtomicStructureand

DynamicswithUltrafastX-rayScatteringK.J.Gaffney&H.N.Chapman(2007).Science

316,1444-1448.光源

實驗裝置

數(shù)據(jù)分析數(shù)據(jù)分析方法1952D.Sayre1968

D.J.DeRosier&A.Klug1971R.W.Gerchberg&

W.O.Saxton1980D.Sayre1982R.H.T.Bates1999

J.Miao,C.Charalambous,JKirz,&D.Sayre2001

J.Miao,K.O.Hodgson&D.Sayre蛋白質(zhì)單分子衍射相干成像（CDI)的相位推演模擬試驗

J.Miao,K.O.Hodgson&D.SayrePNAS,98,6641-6645(2001)“Anapproachtothree-dimensionalstructuresofbiomoleculesbyusingsingle-moleculediffractionimages”X-FEL:wavelength=1.5?;per-pulseflux=2×1012photons;pulselength=10fsec;Dataresolution:2.5?

現(xiàn)有的光強(qiáng)和信噪比還不足以解單分子蛋白結(jié)構(gòu)；但在2010年，LCLS完成了一個影響深遠(yuǎn)的衍射實驗HNChapmanetal.

Nature

470,73-77(2011)doi:10.1038/nature09750DiffractionintensitiesandelectrondensityofphotosystemI.X射線的串行晶體學(xué)（SFX)是研究高分辨X射線CDI的中間產(chǎn)物；但是它具有在X射線衍射分析領(lǐng)域引發(fā)巨大變革的潛力。X射線的串行晶體學(xué)有兩大特點：一、先衍射后損傷

(DiffractionbeforeDestruction)；二、用“粉晶”試樣產(chǎn)生“單晶”

衍射。Aone-dimensionalpowderdiffractionpatternseenusingconventionalmethods(left)

maypotentiallybeanalysedasthree-dimensionalsingle-crystalpatternsusingserialcrystallography(right).Zhangatal.

Volume2|Part3|May2015|Pages322–326Hao,

Volume2|Part3|May2015|Pages307–3083/15/2023串行晶體學(xué)的出現(xiàn)，為用X射線研究生物大分子的結(jié)構(gòu)提供了許多新的可能：可以用微米甚至亞微米的多晶樣品來測定蛋白質(zhì)的結(jié)構(gòu)，過去一般需要用比這大幾十倍的單晶體?？梢栽谑覝叵逻M(jìn)行衍射實驗。過去樣品一般都需要冷凍處理以增加樣品抗輻照的能力，但也增加了樣品的鑲嵌程度，甚至改變了蛋白質(zhì)分子結(jié)構(gòu)的細(xì)節(jié)。室溫條件更便于操作，也更接近于蛋白質(zhì)的天然活動環(huán)境。能夠?qū)ι矬w內(nèi)形成的蛋白質(zhì)微晶進(jìn)行分析研究，甚至不需要將這些微晶從生物細(xì)胞中取出來。還可以進(jìn)行蛋白質(zhì)結(jié)構(gòu)的動態(tài)研究。CrystalstructureofrhodopsinboundtoarrestinbyfemtosecondX-raylaser.Nature

523,561(2015).

入選兩院院士評選中國世界十大科技進(jìn)展中科院上海藥物所研究員徐華強(qiáng)帶領(lǐng)的國際團(tuán)隊(28個單位)利用世界上最強(qiáng)X射線激光LCLS，成功解析視紫紅質(zhì)與阻遏蛋白復(fù)合物的晶體結(jié)構(gòu).APSdata(7.7?)vs.LCLSdata(3.3

?).攻克了細(xì)胞信號傳導(dǎo)領(lǐng)域的重大科學(xué)難題,即GPCR如何激活另一條信號通路—阻遏蛋白信號通路.3/15/2023StructuralbasisforselectivityanddiversityinangiotensinIIreceptors

HaitaoZhang

etal.

Nature

544,327–332(20April2017)3/15/2023ImproveSFX?Challenges:partialreflections;massivenumberofcrystals(~1M)needed;difficulttoprocess.ProposedSwissFELbandwidth~3%bystrongpulsecompression(savetimeandsampleby30-foldcomparingtomonomode)3/15/2023Howtogeneratechirp?縱向動力學(xué)阻抗元件激光等離子體物理過程過壓縮、偏峰optimizationCorrugatedstructure尾波加速場作用位置直線直線出口直線出口Chirp量1-2%1-2%1-2%FEL帶寬2-4%2-4%2-4%Remarks影響正常FEL運行~10MV/m可開關(guān)10m以上距離被動穩(wěn)定1-100GV/m可開關(guān)10cm作用距離？穩(wěn)定性？HaixiaoDeng,SINAP3/15/2023LaueDiffraction:

astationarycrystalduringexposure;

polychromaticbeam;

wavelengthnormalizationrequired.3/15/2023Laue

DiffractionLargenumberofspotsonasingleimage.Non-oscillationsample(e.g.insituplates)Potentialtime-resolvedapplicationExamples:Moffat,Szebenyi&Bilderback(1984),Science,223,1423-25.Hadjuetal(1987)Nature,329,178-181.BilderbackHoffman&Thiel(1994)Science,263,201-203.Genick,…,Moffat&Getzoff(1997)Science,275,1471-75.Lauev.s.MonoLauePro:mostfullreflections;fewercrystalsneeded;easiertoscale/merge.Con:moresensitivetomosaicity;possiblespatialoverlapforlargeunitcell.MonoPro:lesssensitivetomosaicity.Con:partialreflections;morecrystalsneeded;difficulttoprocess.

3/15/20233/15/2023ExperimentalProceduresMicroMesh(Mitegen)mountsandflash-cooled,orLiquidjetorLCPorinsituplatesatroom-TDatawillberecordedfromonecrystalatatimeusinganareadetector.Foreachcrystal,onestillimagewillberecordedDatawillbeindexed,refinedandintegratedwiththeprogramLAUEGEN3/15/2023

Feasibilitystudy

@CHESS-D1

Cornaby,S.,D.M.E.Szebenyi,D.-M.Smilgies,D.J.Schuller,R.Gillilan,Q.HaoandD.H.Bilderback(2010)ActaCryst.D66,2-11.

3/15/2023ΔE/E

~30%3/15/2023NarrowbandpassOpticsKazmirovetal(2006)J.SynchrotronRad.13,204-210.MultilayerX-rayopticsatCHESS:ΔE/E:0.2%,2%,5%and10%W/B4C:mostcommonlyusedmultilayer@2%3/15/2023CHESS-G1experimentMulti-layeroptic(~1.7%bandpass:0.975-0.992?)Onesetofslits(~0.1x0.1mm2)Lysozymecrystal(a=b=79,c=38?,P43212)Roomtemperature72stillimages;=2.5;1sexposureCrystal-detector(Q4)dist.=188mm3/15/2023Diffractionimage3/15/2023DataprocessingUsingDENZO/MOSFLMtofindcrystalorientationUsingLAUEGEN*torefineorientation,improvesoftlimits(lmin,lmax&dmin)andintegrateUsingLAUENORMtonormalize,scaleandmerge.*Campbell,J.W.,Hao,Q.,Harding,M.M.,Nguti,N.D.&Wilkinson,C.(1998).LAUEGENversion6.0andINTLDM.J.Appl.Cryst.

31,496-502.3/15/20233/15/2023DatastatisticsResolutionrange:12.0–2.2?Uniquereflections:4370I/(I)=10.0Rmerge=13.4%(~6%between0.981-0.985?)Redundancy=2.0Completeness=70.7%3/15/2023StructuresolutionandrefinementStraightforwardMolecularreplacementsolutionwithAMoRe:c.c.=75%,R=31.7%RefinementwithREFMAC:R~20%.3/15/2023Electrondensitymap(residues40-45,50-55,58-60)3/15/20233/15/20233/15/20233/15/20233/15/20233/15/20233/15/2023DiscussionIdealbandwidth? smallunitcells(<100?),~10%; largeunitcells(>100?);~6%.At1.7%bandwidth,~1°.3/15/2023Beamline:aPXbeamlinewithMultilayer-opticsScientificGoal：Todetermineproteinstructureusingsub-microncrystalsExperimental：Lauediffraction;One-shot-per-crystalstructuredeterminationSpecifications：Photonenergy：7-16keVBandwidth：>3%Beams