Large-scale expression screening by automated whole-mount in situ(1).pdf

Technical reportLarge-scale expression screening by automated whole-mountin situ hybridizationRebecca Quiringa,Beate Wittbrodta,Thorsten Henricha,Mirana Ramialisona,Carola Burgtorfb,Hans Lehrachb,Joachim Wittbrodta,*aDevelopmental Biology Programme,European Molecular Biology Laboratory,Meyerhofstrasse 1,D-69012 Heidelberg,GermanybMax-Planck-Institut fu r Molekulare Genetik,Ihnestrasse 73,D-14195 Berlin,GermanyReceived 22 December 2003;received in revised form 12 March 2004;accepted 31 March 2004AbstractGene expression profiling is an important component offunctional genomics.We present a time and cost efficient high-throughput whole-mount in situ technique to perform a large-scale gene expression analysis in medaka fish(Oryzias latipes)embryos.Medaka is a modelsystem ideally suited for the study of molecular genetics of vertebrate development.Random cDNA clones from an arrayed stage 20 medakaplasmid library were analyzed by whole-mount in situ hybridization on embryos of three representative stages of medaka development.cDNA inserts were colony PCR amplified in a 384-format.The PCR products were used to generate over 2000 antisense RNA digoxigeninprobes in a high-throughput process.Whole-mount in situ hybridization was carried out in a robot and a broad range of expression patternswas observed.Partial cDNA sequences and expression patterns were documented with BLAST results,cluster analysis,images anddescriptions,respectively;collectively this information was entered into a web-based database,MEPD(http:/www.embl-heidelberg.de/mepd/),that is publicly accessible.q 2004 Elsevier Ireland Ltd.All rights reserved.Keywords:Oryzias latipes;Partial cDNA sequencing;Expressed Sequence Tags;Screening;Automated;Whole-mount in situ hybridization;Gene expression;High-throughput;Database;Expression profile;Embryogenesis;Organogenesis1.IntroductionThe understanding of the molecular mechanisms under-lying a given developmental process has profited from geneexpression studies by in situ analysis.This approach hasbeen instrumental in the isolation of a large number of novelgenes that are expressed in a specific spatial and temporalmanner.It is a complementary approach to systematicmutagenesis screens since genes with no obvious phenotypeor functionally redundant genes will be largely inaccessibleto mutant screens.In situ analysis is an importantcontribution to micro-array analysis as well,since thetemporal and quantitative data set is complemented by thespatial expression profile.Furthermore,novel region-,tissue-,organ-and cell-specific markers are identified,facilitating a variety ofstudies in wild-type embryos and are especially valuable asmolecular markers for the analysis of mutant embryos frommutagenesis screens.Genes with interesting expressionpatterns are readily sequenced and cloned from the cDNAlibrary.In addition,the combination of cDNA sequence andexpression pattern provides the scientific community with adetailed analysis of orthologs of genes identified in otherorganisms.Large-scale in situ hybridization screens have beencarried out for a number of organisms(Gawantka et al.,1998;Kawashima et al.,2000;Kudoh et al.,2001;Lynchet al.,1995;Neidhardt et al.,2000;Thisse et al.,2000).Insome cases,these screens were designed to focus onparticularaspects(e.g.secretedproteins)(Christiansenetal.,2001;Crosier et al.,2001;Nguyen et al.,2001;Thut et al.,2001;Yoda et al.,2003).0925-4773/$-see front matter q 2004 Elsevier Ireland Ltd.All rights reserved.doi:10.1016/j.mod.2004.03.031Mechanisms of Development 121(2004) author.Tel.:49-6221-387-576;fax:49-6221-387-166.E-mail address:jochen.wittbrodtembl-heidelberg.de(J.Wittbrodt).The high fecundity and the external development of thecompletely transparent embryos make small fresh waterteleosts particularly suited for large-scale in situ analysis(Kudoh et al.,2001;Thisse et al.,2000).Since,even at laterstages of development,the embryos are still highlytransparent,a unique opportunity is given to look at genesexpressed in internal organs during organogenesis.In medaka(Wittbrodt et al.,2002),in situ screens havebeen reported(Henrich and Wittbrodt,2000;Nguyen et al.,2001)and large-scale whole-mount in situ screens are inprogress(reviewed in Shima et al.,2003 and the presentwork).Here,we describe an efficient strategy for screening alarge number of genes for their expression pattern by whole-mount in situ hybridization in medaka embryos.UsingcDNA libraries arrayed in microtiter plates,high-through-put processing of RNA probes and automated whole-mountin situ hybridization,this technique is fast,reliable and costefficient.2.ResultsThe cDNA library was prepared by directional cloning ofcDNA inserts from stage 20 medaka embryos into a plasmidvector and bacterial transformation.Single clones werearrayed by a robot into 384-well microtiter plates.The initial complexity of the library was two millionindependent clones.The cDNA inserts were flanked bypromoters for SP6 and T7 polymerase allowing antisenseprobe preparation and cDNA sequencing,respectively;the PCR primers were designed to include both promoters inthe PCR product.4224 inserts were PCR amplified bycolony PCR and confirmed by gel electrophoresis.Twopercent of the analyzed clones contained no inserts;theaverage insert length was between 1 and 2 kb.Over 2000 digoxigenin-labeled RNA antisense probeswere transcribed directly from the non-purified PCRproducts and purified using a commercially available 96-format kit.For the whole-mount in situ analysis the embryonicstages 18,24 and 32(Iwamatsu,1994)were chosen torepresent key events during vertebrate development namely:neurulation,somitogenesis and organogenesis,thus cover-ing early as well as late embryonic development.The in situ hybridization of 96 probes was carried outsimultaneously in a robot(InsituPro,Intavis AG),and allthreestagesofmedakaembryoswerehybridizedtogetherforeach probe.To date all of the probes have been hybridized.A broad range of staining patterns was observed rangingfrom examples where the staining is observed exclusively inone organ or a few cells,to highly complex patterns inwhich several but discrete tissues express the associatedgene(Fig.1).Fig.1.Examples for the expression patterns observed.(A)and(B)Lateral view,anterior is to the left.(C)Frontal view.(D)Dorsal view of the embryo.(A)and(C)Show original pictures taken from the MEPD.(B)and(D)Show pictures of manual whole-mount in situ,these embryos were cleared in glycerol andremoved from yolk.(A)and(B)Region specific expression of clone 631-135-03-D in the epidermis of the pectoral fin blade as well as in ventrolateralepidermis(stage 32).Insert in(B)shows dissectedpectoralfin;note the stainingin the epidermisof the blade.(C)and(D)Expressionof clone 631-136-05-D inthe heart as well as in a few cells at the rostralmost poleof the optic tectum in the area where optic fibers of the retinal ganglioncells enter the tectum(stage 32).EY,eye;H,heart;LS,lens;OT,optic tectum;PF,pectoral fin;VLE,ventrolateral epidermis;YO,yolk.R.Quiring et al./Mechanisms of Development 121(2004)971976972Systematic 50and 30end sequencing of all cDNA clonesis in progress and the sequences and expression patternsare accessible on a web-based database,MEPD(MedakaExpression Data Base),at http:/www.embl-heidelberg.de/mepd/(Henrich et al.,2003).The expressed sequence tags(ESTs)are clustered upon entry into the database andblasted against public databases.So far,922 clones havebeen sequenced and clustered to represent 694 genes.Based on amino acid sequence comparison,140 of thesegenes have no significant homology hit in the publicdatabases.Currently,the expression patterns from the in situscreen are being documented by representative pictures ateachembryonicstageaswellasbydescriptions,according to a comprehensive ontology of anatomicalterms;collectively this data is entered into the database.The anatomical terms following gene ontology have beensubmitted to:http:/ by Henrich andfurtherdetailsontheontologywillbepublishedelsewhere.The database contains to date 395 sequencedgenes with fully annotated expression patterns;thesegenes were amenable for calculations and gave thefollowing results:42.5%of the genes were expressedspecifically in one or a few embryonic structures,in atleast one of the embryonic stages;18%were ubiqui-tously expressed and 2%gave no staining.The remaining37.5%were differentially expressed;the genes in thiscategorywereexpressedeitherubiquitouslybutatdifferent intensities or they were expressed in severalstructures and tissues(Fig.2).This project is stillongoing and a final statistical overview will be presentedupon completion.3.Conclusions and perspectiveIn situ hybridization screening is a robust and straight-forward method to examine the expression profile of a largeset of genes.The resolution of the pattern analysis can reachcellular levels.Furthermore,this approach is independent ofthe function of the genes analyzed,therefore,functionalredundancy or severe cellular or temporal restrictions whichmay cause them to be missed in screens for lethal mutantphenotypes,can be circumvented.With the presented protocol for high-throughput geneidentification,it is possible for one person alone to analyze50 genes per week including probe generation,in situhybridization,imaging and pattern annotation in thedatabase.Thisincombinationwiththesequenceinformationof the analyzed cDNAs should allow rapid identification ofdevelopmental regulators.The protocol can easily beadapted to other species(Arendt,personal communication).In this study,922 sequences from random picked cDNAclones resulted in 694 gene clusters;of these 591 weresingletons.This reflects the high complexity of the arrayedlibrary.We therefore conclude that using a highly complexFig.2.Examples for differential and specific expression patterns.(A),(C)and(D)Dorsal view.(B)Lateral view,anterior is to the left.(A)(C)Show originalpictures taken from the MEPD.(D)and insert in(B)show pictures of manual whole-mount in situ from the tail region of embryos.(C)and(D)Embryos werecleared in glycerol and removed from yolk;pictures were taken from a Zeiss Axiophot using Normarski optics.(A)Differential expression of clone 631-134-09-I along body axis(stage 18).(B)Expression of clone 631-136-07-H in the hypochord(arrow)and floorplate(arrowhead)(stage 24).(C)and(D)Specificstaining of clone 631-134-02-E in the tailbud(arrow);arrowhead indicates staining in the latest pair of forming somites(stage 24).R.Quiring et al./Mechanisms of Development 121(2004)971976973cDNA library that has not been pre-screened,normalized orsubtracted,results in a low redundancy in the screen,if thenumber of clones analyzed is in the order of 20005000.With the presented protocol for the ongoing whole-mount in situ screen,we have identified novel genes andpossible markers to study medaka embryonic development.Within the context of the medaka genome initiative(Shima et al.,2003),we plan to systematically analyzeclones from a medaka unigene library(Himmelbauer,personal communication).In a collaborative project,weare aiming to complete the expression analysis by whole-mount in situ hybridization of the medaka trancriptomewithin the next 3 years.Based on their expression,cDNAswill then be directly available for gain-of-function exper-iments by injection and over-expression of the cDNA cloneor using synthetic mRNA.Upon conclusion of data sampling,comparison ofexpression patterns will define groups of genes with similarcomplex expression domains,since these genes are ofteninvolved in a common process(Niehrs and Pollet,1999).Such synexpression groups(Gawantka et al.,1998)have aprognostic value on the function of novel components ofknown pathways as well as of novel genes with no sequencesimilarity(Niehrs and Pollet,1999;Onichtchouk et al.,1999;Tsang et al.,2000).Whole genome sequencing of medaka is in progress.Once genes are identified,this type of combined approachwill clearly be an important contribution to functionalgenomics,which in turn will contribute to an understandingof overall gene function in medaka as one vertebrate modelsystem.4.Experimental procedures4.1.EmbryosEmbryos were staged according to Iwamatsu(1994).Forstage 32,albino(Heino)embryos were used(Loosli et al.,2000).Embryos were fixed over night in 4%paraformalde-hyde/2PTW at 4 8C and subsequently dechorionatedmanually,washed in PTW according to published pro-cedures(Loosli et al.,1998)and stored in methanol at220 8C until usage.4.2.Probe synthesisAn oriented stage 20 cDNA library in the plasmid vectorpSPORT1(Life Technologies)was used.Individual cloneswere arrayed with a robot into 384-well microtiter platesand grown in liquid microculture.Fresh replicas from the plated library were made bytransferring bacteria from the 384-well microtiter plate(Nalge Nunc International)with a 384-pin replicator(NalgeNunc International)into a fresh 384-well plate containing70 ml LB/ampicillin/10%glycerol per well.Replicas wereincubated for 16 h without shaking at 37 8C.For PCR amplification of the cDNA inserts,smallaliquots of the microculture were transferred from thereplica plates into 384-well PCR-microtest plates(La-Bio-Med)using the 384-pin replicator.Plates were coveredwith a silicon cover.Two-step colony PCR was performedin a 10 ml volume(vector based primers:M13fwd primer:50GCTATTACGCCAGCTGGCGAAAGGGGGATGTG-30and 3/86 primer:50-CCGGTCCGGAATTCCCGGGT30).PCR conditions were as follows:an initial denaturationstep at 96 8C for 3 min was followed by 30 cycles of twosteps:96 8C for 10 s and 70 8C for 3 min,and a finalextension at 72 8C for 8 min.PCR reactions wereanalyzed by gel electrophoresis in batches of 384.ThePCRproductsincludeapromoterforSP6polymerase allowing the direct transcription of antisenseRNA probes.Digoxigenin-labeled antisense RNA probeswere synthesized from PCR-amplified templates usingSP6 RNA polymerase(Roche).Reactions were carriedout in 96-well plates with a 10 ml reaction volumecontaining 0.5 ml of the PCR reaction,1 mM of ATP,CTP and GTP each,0.65 mM UTP/0.35 mM digoxi-geninUTPmix(Roche),15unitsofRNAguard(AmershamPharmacia),0.4unitsofthermostableinorganicpyrophosphatase(NewEnglandBioLabs),2 mM MgCl2,20 units SP6 polymerase and 1 ml of10 SP6 transcription buffer(Roche).Incubation was at37 8C for 5 h;DNase I treatment was omitted.To reducethe costs of the purification of the riboprobes,theQuiaquick PCR purification kit 96(QIAGEN)was usedwith the QIAvac 96 vacuum device according tomanufacturers instructions;the important exception wasthat in the first step buffer RLT from the RNAesyKIT(QIAGEN)was used.To the eluate(60 ml)150 mlof Hybridization Buffer was added.4.3.In situ hybridizationEmbryos were rehydrated and processed according tostandard procedure(Loosli et al.,1998).Proteinase Kdigestion was according to stages between 7 min and1.5 h.In situ hybridization was performed with 96 probessimultaneously in a robot(In situ Pro,Intavis AG).Inthe robot,the incubation volu