a technicolour approach to the conectome-brainbow

5/11/2018 A Technicolour Approach to the Conectome-Brainbow - slidepdf.com

http://slidepdf.com/reader/full/a-technicolour-approach-to-the-conectome-brainbow

Undrstanding ho th prcis intrcon-nctions of nurons account for brainfunctions has bn a proccupation of nuroscintists for or a cntury. Toaccompish this aim, nuroscintists ind arious typs of data. First, i nd a physica map of th nuronsthat compris a circuit and th sits at

hich thy form synapss ith ach othr.Scond, i nd to undrstand hoctrica signas fo through th circuithn a stimuus is prcid, a dcisionis mad or an action is takn. Third, if thcircuitry undrgos modifications, ind information on ho circuits and thirsignas chang or tim. Finay, and argu-aby most dmanding, i nd aysto transform th ra data into usfu formsso that can xtract maning from thphysica maps. Nuroscintists ar hard atork doping th toos that ar nddto accompish ach of ths; hr discuss

sra tchnoogica adancs that mightproid soutions to th first probm:obtaining a compt physica map of thnrous systm.

Circuit-mapping strategies

Or th past cntury, nuroscintistsha usd thr main sts of anatomicaapproachs to study nura connctiity:sing-c imprgnation, opticay basdtract-tracing and ctron microscopy (FIG. 1). In this sction, brify discussths mthods and thir imitations.

Single-cell staining by dye impregnation. This as th first and most infuntiaapproach. Sing-c imprgnation is rootdin th xtraordinary ork of SantiagoRamón y Caa (1852–1934), ho usdGogi’s ‘back raction’ — a sir-stainingtchniqu that abs nurons ith a dark prcipitat. This mthod stains th ntirty

of ust a f indiidua nr cs anda thir procsss (axons and dndrits),thrby making idnt shaps that oudothris b indiscrnib in th dnsmshork of th nuropi1. Using th Gogistain, Caa as ab to not ony idntify many diffrnt nurona typs in th brain,but aso dscrib thir pattrn of intrcon-nctions and discor princips of nuronacircuit organization (FIG. 1a).

Dspit thir por, th Gogi mthodand othr sing-c abing tchniqus— for xamp, intracuar inction of dys— ha significant imitations. First, bcaus

th chmicas ith hich th nurons arimprgnatd must b appid to ratiy sma bocks of tissu, thy cannot b usdto trac ong-distanc connctions. Scond,bcaus of diffraction, nurona procsssthat com cosr to ach othr than approxi-maty 0.25 µm ar not rsoab in th ightmicroscop. It is thrfor difficut to tracsndr axons and dndrits ithout thrbing any ambiguitis. Third, if a or nmany of th mnts in a circuit ar staind,thy bcom impossib to distinguish fromon anothr. Thrfor, circuit dscriptions

ar amost aays basd on a statisticaanaysis of sma amounts of data from mu-tip indiiduas, rathr than on fu anayssof any actua circuit. Indd, ths probmsar so sr that, ith th xcption of Caa, f scintists ha bn ab to dragnra concusions from such matria thatha stood th tst of tim2.

Diffusion or transport of labelling agents

placed at discrete positions. Th probmof isuaizing ong-distanc connctiity can b addrssd, at ast in part, by circuit-

tracing tchniqus, hich ino markingtracts that proct from on brain rgionto anothr. Tracts ar gnray abd by mchanicay introducing a spcific abingagnt — for instanc, through a surgica pro-cdur. Th first tract-tracing mthods to bappid, hor, took adantag of th factthat axons dgnrat fooing inury anda bhind dbris that prsists for days toks bfor it is card. This dbris coudb staind ith sir sats. Thus, fooingoca damag to a nurona popuation ortract, th sir stain markd sits to hich

nurons had proctd3

.By th 1970s, dgnration mthods hadargy bn rpacd by a suit of tchniqusin hich abs r pacd into circum-scribd rgions of th cntra or priphranrous systms and thn diffusd or actiy transportd aong axons to othr sits(FIG. 1b). A id rang of abs prod usfufor this appication: radioacti amino acidsor sugars, protins such as th chora toxin Bsubunit, nzyms such as horsradish proxi-das, and fuorscnt organic mocus suchas DiI, DiO and Fuorogod4. Most rcnty,gnticay ncodd tracrs, incuding

fuorscnt protins (s bo), ha bnxprssd undr th contro of subtyp- orrgion-spcific rguatory mnts or from

ira ctors5–9. Ths mthods ha sraadantags or mthods that us chmicaabs, incuding th possibiity of indibmarking and, for th transgnic approachs,aoidanc of th nd for surgry. An addi-tiona xciting dopmnt is th abiity tointroduc gnticay ncodd trans-synaptictracrs that can cross from a nuron to itssynaptic partnrs8–12, thus raing (ithsom caats; s reF. 9) connctiity.

A technicolour approachto the connectome

Jeff W. Lichtman, Jean Livet and Joshua R. Sanes

Abtact | A cntal aim f nucinc i t map nual cicuit, in d t lan

hw thy accunt f mntal activiti and bhaviu and hw altatin in thm

lad t nulical and pychiatic did. Hwv, th mthd that a

cuntly availabl f viualizin cicuit hav v limitatin that mak it

xtmly difficult t xtact pci wiin diaam fm hitlical ima.H w viw cnt advanc in thi aa, aln with m f th pptuniti

that th advanc pnt and th btacl that main.

NATURe RevIewS | neuroscience vOlUMe 9 | jUNe 2008 | 417

Progress



a b c

Dspit th ncouraging progrss thatthr has bn aong ths ins, difficutisrmain. First, tracing from a st of somatato a trmina fid, or ic rsa, i in most

cass a oca intrnurons unabd orimpossib to rso. Scond, bcaus thsmthods typicay ab groups of nuronsrathr than indiidua cs, tract-tracingmthods can sdom b usd to trac thhighy compx branching pattrns of ththousands of intrdigitatd axons anddndrits in ach cubic µm of brain.

Electron microscopy. Sria sction ctronmicroscopy has bn th mthod of choicfor orcoming th imitd rsoutionof ight microscopy and th insufficincis of tract-tracing. In pionring ork, lintha,

loprsti and Macagno usd this mthod tomap connctions in th optic ob of th smacrustacan Daphnia13, and whit et al. usdit to proid a nary compt map of con-nctiity in th roundorm, Caenorhabditis elegans14. Sria ctron microscopy is soaborious, hor, that its appication hasbn infrqunt. Fortunaty, this situation isno changing thanks to rcnt dopmntsin th automation of ctron microscopy 15–17.On promising approach is to imag th thinsurfac ayr (tns of nanomtrs) of a bock of pastic-mbddd and hay-mta-staind

brain tissu by ooking at backscattrd c-trons ith a scanning ctron microscop15.Aftr taking th imag, th top sction of thbock is rmod ith a diamond knif and

th bock fac is rimagd. This procdurcan no b itratd thousands of timsto gi prfcty aignd stacks of digitactron micrographs from hich nuronaprocsss can b tracd (FIG. 1c). Othrapproachs us n microtom dsignsinoing a rotary ath and a tap transportsystm to automat th cutting and coc-tion of sria thin sctions. Aong ith nmthods in computationa sgmntationthat impro tracing and thr-dimnsionarconstruction (s bo), ths innoationsar iky to rstor ctron microscopy toa position of prominnc in circuit anaysis.

Nonthss, arg-sca rconstruction,spciay or ong distancs, rmains a dis-tant hop. For xamp, rconstructing pathsfrom th rtina to th thaamus or fromth thaamus to th cortx oud rquirhundrds of thousands or n miions of sctions to b acquird and anaysd, andn thn thr oud b no obious ay to

rify th accuracy of th tracing.

The multicolour solution

Introducing a markr into th rconstructdnuron aos its idntity to b rifid in

ach sction. Hor, th nd for sparsabing hn using monochromatic absrmaind an intractab imitation untircnty. Th obious soution is to us

mutip, distinct abs so that narby pro-csss can b distinguishd from ach othr.This approach is xmpifid by th us of ‘rainbo’ cabs in ctronic dics: th

arious coours of th irs hp pop kptrack of thm. Sra mthods dsigndaong this princip ha bn appid toth task of nurona circuit isuaization.On, cad Dioistics, inos shootingtissus ith bioistic mta bads that habn coatd ith arious combinationsof diffrnt coourd ipophiic dys, suchas DiI, DiO, and DiD, ith th rsut thatimprgnatd cs ar abd ith diffrnt

coours18. Hor, Dioistic abing is bstdon ex vivo on sics, so it is poory suitdfor ong-rang rconstruction.

Anothr approach is basd on th grnfuorscnt protin (GFP) roution19,hich has d to a tchnica rnaissanc inimaging. Or th past sra yars, doznsof spctra and photophysica fuorscntprotin ariants (XFPs) ha bn disco-rd or nginrd20, and many of ths habn introducd as intrinsic nurona absin transgnic mic. In som of th micthat produc ths XFPs, th xprssion is

Fiu 1 | ct-mappgtatg. exampl f th th mainappach that hav bn ud t tudy th cnnctivity f nun.

a | sinl-cll tainin by dy impnatin. Th bt-knwn f th tch-

niqu i th gli tainin mthd, which wa ud by ramón y Cajal t

dcib th pincipl f nunal cicuit anizatin39.b | Th intductin

f diffuibl tanptabl labllin ant t dict aa. In m ca

th labl c a ynap t ach a cnd-d tat. A claical xam-

pl i th idntificatin f cula dminanc clumn in th viual ctx

fllwin th injctin f adiactiv plin int n y40. | sial lctnmiccpy can b ud t cntuct nun and thi pc with th

bt attainabl lutin. shwn h i a cntuctin fm th adult at

bal ctx that wa mntd fm a th-dimninal ima tack

btaind by ial blck-fac imain16. Pata pducd, with pmiin,

fm reF. 39 (1914) Hd d santia ramón y Cajal. Pat b p-

ducd, with pmiin, fm reF. 40 (1977) ryal scity f Lndn.

Pat pducd, with pmiin, fm reF. 16 (2006) elvi scinc.

Progress

418 | jUNe 2008 | vOlUMe 9 www.at.m/vw/



d

ca

b

Yes No Yes

loxP loxP lox2272 lox2272

Construct Construct

Promoter Promoter

OutcomesOutcomes

Cre recombinaseTransientCre recombinaseactivity

RFP RFPOFP YFP GFP YFP

GFP YFP

YFP GFP

CFP CFP

RFP CFP

CFP

RFP YFP CFP

YFP CFP

CFP

OFP GFP

RFP

GFP

CFP

YFP

RFP

YFP

CFP

or

or

1

1 2 3

2

3

loxN loxPlox2272i

i

ii iii

iii

RFP

iv v

loxP loxP loxP loxP

confind to sma, quasi-random substs of cs, crating th quiant of a ita Gogistain21. By crossing to such ins of micxprssing spctray distinct XFPs, on canobtain ‘bicoour’ or n ‘tricoour’ animasin hich th intractions btn to nu-rons can b isuaizd21–24. Anothr mthod

for xprssing to XFPs in a mosaic man-nr is th ingnious mosaic anaysis ithdoub markrs (MADM) stratgy 25. In thismthod, th XFP gns ar artificiay spitby a synthtic intron that contains a loxP sit(Box 1). Rciprocay chimric gns madby pairing th xons that ncod th to

XFPs ar insrtd at idntica positions ina pair of homoogous chromosoms. Thsmnts rmain non-functiona unti Cre rcombinas (Box 1) triggrs intrchromo-soma rcombination to rconstitut actiXFP-ncoding gns. Subsqunt mitosssgrgat th as such that diffrnt cs

inhrit diffrnt products of rcombinationand thus xprss distinct coours.Dspit thir utiity, bi- and tricoour

mic ha far too f abs to disambiguaton nura procss from th thousandsof axons and dndrits that it i cosy approach as it courss through th nroussystm. Rcnty, dopd a combi-natoria coour mthod, cad Brainbo26,that addrsss this probm by markingindiidua nurons in on of >100 cooursrathr than on of 2–3 coours. Th mthodis basd on combinatoria and stochasticxprssion of mutip XFPs from a sing

transgn. In princip, combinatoriaxprssion of thr diffrnt XFPs coudcoour nurons in on of tn hus (FIG. 2).Hor, thr XFPs can produc a muchargr numbr of hus if thy ar xprssdin diffrnt amounts in diffrnt nurons.A good anaogy is a Tv monitor, hichcombins diffrnt intnsitis of thr chan-ns (rd, grn and bu) to gnrat amostth ntir coour spctrum that th human

isua systm can prci (FIG. 2b). Ourthought as to us such coour combina-tions to distinguish nurons in a circuit by

gnrating a random mixtur of thr fuo-rscnt protins in ach nuron. This sto-chastic coour xprssion as accompishdin to stps. First, dsignd gnticconstructs that coud b rcombind insra mutuay xcusi ays to gi any on of thr or four coours (s Box 1 for adtaid dscription of th mthod). Scond, insrtd mutip copis of th transgninto th gnom. Indpndnt xprssionfrom ach copy aos a rang of coours torsut from combinatoria xprssion of thXFPs (FIG. 2c). As a rsut, indiidua nuronsxprss on of many distinguishab hus

(~100), nabing th spctra diffrntiationof indiidua nurons and thir procsss(FIG. 3).

The role of Brainbow in connectomics

Brainbo transgns ha gnratd strikingimags and xcitd considrab intrst, butho can thy actuay b xpctd to con-tribut to th mapping of nurona circuits? Thir potntia is iustratd by our rcntxprinc ith prhaps th simpst con-nctom: th innration of skta muscsby motor axons. A f yars ago, bfor

Bx 1 | Brainbow strategies

Brainbow transgenes drive the combinatorial expression of several fluorescent proteins (XFPs) in

neurons, resulting in the colour-tagging of individual cells. Here we outline the methodology that

we used to generate these constructs.

The strategy makes use of the Cre/lox recombination system.Crerecombinase specifically

recognizes and catalyses recombination between a pair of 34-nucleotide sequences called loxP.

This reaction leads to the excision of a DNA segment that is flanked by two loxP sites of the

same orientation, and to the inversion of a DNA segment that is flanked by loxP sites of opposite

orientation (see figure, part ).

One strategy, Brainbow-1, makes use of mutant lox sites that can recombine among themselves

but which bear mutations that render them incompatible with the canonical loxP site7; see figure,

part a). Alternating canonical (wild-type) loxP sites with variant sites creates mutually exclusive

excision possibilities (see figure, part b). Implementation of this strategy with three loxvariants

creates four possible outcomes following excision by Crerecombinase: three different products ofrecombination plus the unrecombined initial state. In all cases, only the XFP directly following the

promoter is expressed. A second strategy, Brainbow-2, usesCre-recombinase-mediated DNA

inversion. The DNA segment can invert repeatedly while Crerecombinase is present but will

stabilize in a random orientation when Crerecombinase stops acting, offering two possibilities of

expression. Moreover, when two such invertible segments are arranged in tandem, the number

of possibilities is increased: additional inversion and excision possibilities create a total of four

expression possibilities (see figure, part d).

Both of these strategies are in essence stochastic: the Crerecombinase enzyme is given a choice

and the outcome is unpredictable. This rolling of the ‘molecular dice’ allows each cell to

independently make a random choice from several XFPs. In the transgenes, a choice is initially given

between several recombination possibilities, each leading to a distinct final configuration (that is,

the different sides of the molecular dice) corresponding to the expression of a particular gene.

For the generation of the first-generation Brainbow mice26, transgenes were placed under the

control of Thy1 genomic elements to obtain strong neuronal expression; these animals were

mated to other transgenic mice that expressed Cre recombinanse or a ligand-activated variant,

CreER. CFP, cyan flucnt ptin; gFP, n flucnt ptin; oFP, an flucnt

ptin; rFP, d flucnt ptin; YFP, yllw flucnt ptin. Figure modified, with

permission, from reF. 26 (2007) Macmillan Publishers Ltd.

Progress




Red Green Blue

RFP YFP CFP

2 3

RFP YFP CFPP RFP YFP CFPP RFP YFP CFPP

CFP

CFP Light blue

Cyan

Green

Yellow-green

Orange

Red

Magenta

Purple

Grey

CFP

YFP

YFP

YFP

RFP

RFP

RFP

RFP

CFP

CFP

YFP

YFP

YFP

RFP

RFP

RFP

CFP

CFP

CFP

YFP

YFP

YFP

RFP

RFP

RFP

CFP

CFP

YFP

Blue

Resulting colour

a

c

b

1

Brainbo ins r aaiab, bgan by

using a mous in in hich a motor axonsar abd th sam coour21 to ra thntir connctiona map in a sing musc.w orkd on a sing sma musc that asinnratd by ony 15 or so motor axons(j. lu, prsona communication). A com-pt map rquirs that a axons b abd,and this is on of th imitations of th no-xisting ‘first gnration’ Brainbo mic (fora dtaid discussion of ths imitations sbo). Our thought as that, ith ony 15axons innrating ~200 musc fibrs, sucha mapping oud not b difficut. w r

rong. In ordr to gt sufficint rsoution

to track a of th axons and at th sam timimag th ho innration fid, hadto assmb tns of thousands of high mag-nification confoca imags into hundrds of stacks and thn stitch th stacks togthr tognrat a montag of th ho musc. Usof a motorizd computr-automatd stagmad it possib to coct th imags, hichtogthr comprisd hundrds of gigabytsof data. Thn, using computr-assistd trac-ing, trackd th ntir branching tr of ach motor axon and its synaptic contacts(a motor unit) to rconstruct th ntir

‘connctom’ of th musc. This ntrprisrquird sra months of ork for achindiidua musc, highighting th sca of th tchnica hurds that oud ha to bsurmountd to accompish th sam task forany rgion of th CNS. By contrast, as ha bgun to rconstruct nuromuscuar

connctoms ith Brainbo ins, ha found that th ho procss can baccompishd in an ordr of magnitud sstim. Indd, in som cass it is possib toidntify th dstination of an axon ithouthaing to trac its path, as its uniqu coourrmains rcognizab aong its ngth.

Basd on ths initia ncouragingrsuts, ar hopfu that Brainbotchnoogy can aso b appid to modnrous systms ith imitd numbrs of axons in ordr to furthr dduc principsof nrous systm connctiity. Brainbotransgns ar arady bing usd for mu-

ticoour abing of th zbrafish snsory nrous systm (Y.A. Pan and A. Schir,prsona communication), and attmpts arundray to gnrat Brainbo Drosophila.

A much mor ambitious goa is to usBrainbo mthods in th mammaian CNS.Hr th changs incud th shr dnsity of th iring, th ry fin caibr of much ofth nuropi, th normous dirsity of nurona typs and th intrsting fact thatmany axons tra ry ong distancs toinnrat nurons throughout th brain.Ths obstacs mak it uniky that com-

pt connctoms of arg brain rgions ib obtainab in th nxt f yars. workingtoards that nd sms a auab goa, ho-r, and n partia connctoms — forxamp, of th rtina innr pxiform ayror of a cortica coumn — oud proidconsidrab insight into th dsignprincips of th brain.

Moror, such partia connctomsmight b usfu in studying mous mods of psychiatric disordrs. Thr is a groing sus-picion that dfcts in th pattrn, numbr orproportions of connctions might undribhaioura disordrs ith a dopmnta

componnt, such as schizophrnia andautism. Th primary causs of such disor-drs might incud arious gntic ariantsor nironmnta insuts that might affct

arious mocuar cascads or dopmn-ta nts, but th fina common pathay coud b a quantitati or quaitati dfctin circuitry. That is, som disordrs of thissort might b ‘connctopathis’. BcausBrainbo abing faciitats th suryingof quantitati and quaitati aspcts of circuitry in dirs brain rgions, it mightnab this hypothsis to b tstd.

Fiu 2 |cmbatalxpfthdttfltptagatalag

ptmfl.a | sval pctally ditinct flucnt ptin (XFP) a nw availabl,

includin n that mit in d (rFP), n (YFP) and blu (CFP) fqunci. b | Th cmbinatial

xpin f d, n and blu XFP at vaiu lvl i ufficint t ncd a clu pac anal-

u t th n that i natd by an rgB vid mnit. | An xampl hwin hw tn ditinct

clu can b natd by xpin a timic cmbinatin f th diffnt XFP.In Bainbw

mic, thi utcm wuld ult if th cpi f a tichmatic tann (illutatd at th tp f

th panl; Box 1 f dtail) ach cmbind indpndntly (Box 1). Tianl pnt loxit

( Box 1 f dtail). CFP, cyan flucnt ptin; P, pmt; rFP, d flucnt ptin; YFP,

yllw flucnt ptin.

Progress




a b c

It i aso b auab to us Brainbotransgns to arn ho connctionapattrns chang or th ifspan of ananima. For xamp, th nurobioogicaundrpinnings of hathy aging rmain amystry. Soing this mystry is not ony important in its on right, but might asob rquird if ar to undrstand thpathoogica atrations to hich th agingnrous systm is spciay unrab.likis, circuit modifications that undr-i th critica priod in ary postnata

if ar incompty undrstood, as arth compnsatory changs that mak thyoung nrous systm so rsiintfooing inury.

In studying ths issus, th microscopicanaysis that ha usd so far i b us-fu. Gratr insights might ntuay comfrom tim-aps imaging of circuit atra-tions as thy occur. XFPs mak ida absfor this purpos, and tim-aps mthodsha no bn appid to many parts of thpriphra and cntra nrous systms of on- and to-coour XFP mic27–30. Furthrimpromnts in spctra sparation, acqui-

sition rat and dtctor snsitiity shoudmak it possib to xpand th mthod toBrainbo mic.

So far, ha ony discussd con-nctions among th minority cuarpopuation of th brain: nurons. Brainbomthods can aso b usd to map connc-tions among th maority popuation, gia,or btn nurons and gia. Indd, somBrainbo ins arady ab som gia sub-sts, incuding astrocyts, Schann cs andBrgman gia; it shoud b straightforard tognrat othrs.

Finay, Brainbo can b usd to highightconnctions of anothr sort: th inag ra-tions among c popuations. If rcombina-tion is inducd in nura prognitors ith aigand-actiatd Cre rcombinas (for xam-p, Tamoxifn-dpndnt CreR) (Box 1), aoffspring of a particuar bast c shoud babd ith th sam coour. This approachmight ao th anaysis of th intractions of many cona sts of cs (ach ith thir oncoour) in th sam pic of tissu.

Technological challengesFor mor compx circuits, th diffraction-imitd rsoution of th ight microscopis a srious imitation. Fortunaty, srasoutions to this probm ar arady mrg-ing. On is to cut xtrmy thin sctions of abd tissu31,32. By cutting sctions thatar sra-fod thinnr than th opticasction thicknss, th suprimposition of mutip nurits in th sam foca thicknssis aoidd. In addition, th thin sampsscattr ight much ss, and this impros thcontrast and carity of th abd matria.

To mak such thin sctions, th nura

tissu must b mbddd in a hard rsin.This, in turn, rquirs th XFPs to rtainthir fuorscnc fooing th dhydra-tion and mbdding stps. To som xtntthis might b possib; hor, Brainbotransgns coud atrnatiy b rdsigndto incud pitop tags that oud b rc-ognizd by antibodis33. This oud aotraditiona staining mthods to b usd aftrth tissu has bn cut and oud thrforao th abing by th ndognous fuor-scnt protin to b augmntd or rpacd.Th smingy simp soution of using

antibodis to dtct th XFPs thmssis unfortunaty not yt fasib, bcaussra of th XFPs (for xamp, GFP, cyanfuorscnt protin and yo fuorscntprotin) diffr by ony a f amino acidsand ar, in ffct, antignicay idntica. If immunohistochmica dtction is usd, oncan aso ocaiz sits of synaptic contacts onor btn markd cs, using antibodis topr- or postsynaptic componnts. Mthodssuch as array tomography, in hich thinsctions of brain matria ar assayd ith

an assmbag of antibody probs, proid angant protoco for dtcting many antigns(or pitop tags) in sing thin sctions31.

A mor radica soution to th rsoutionprobm might b proidd by th n andrapidy groing fid of nanoscopy, hichorcoms th rsoution imitations of ight diffraction34–38. Ths n imagingmodaitis proid rsoutions of tns of nmhich, aon or ith th addition of spctrainformation, might b sufficint to trac thfinst nuropi.

In para, it i b important to dsignn Brainbo transgns to incras th

rang of nurona typs and dopmntastags to hich th mthod can b appid.Thr ar sra maor imitations of th first-gnration Brainbo ins. Thrguatory mnts that ar usd (fromth Thy1 gn) dirct high s of xprs-sion in many proction nurons, but mostintrnurons ar poory markd21. Aso,in most ins, transgn xprssion is o inmbryos and during th first postnata k,so ths ins ar unsuitab for dop-mnta studis. Furthrmor, som cooursar dim (spciay rd shads), primariy

Fiu 3 |MltlallabllgBabwtagm.

a | A mt nv innvatin a mucl.b | An axn tact in th baintm.

| Th hippcampal dntat yu. In th Bainbw mic fm which th

ima w takn, up t ~160 clu w bvd a a ult f th

c-intatin f val tandm cpi f th tann int th mu

nm and th indpndnt cmbinatin f ach by Crecmbina

( FIG. 2c). Th ima w btaind by th uppitin f paat d,

n and blu channl. Th ima in pat a i cuty f ryan Daft.

Progress




bcaus th protins ar insufficinty photo-stab. Additionay, sra bright XFPsar difficut to distinguish from ach othr(for xamp, yo rsus grn or orang

rsus rd) bcaus thir spctra ar insuf-ficinty distinct. Moror, athough thnumbr of coours is prsnty arg (~100),

it is nonthss insufficint for som pur-poss. Thr ar f concptua difficutisin dsigning ‘scond-gnration’ Brainbotransgns that circumnt ths imitationsbut, as ith th first gnration, arg num-brs of atrnatis might nd to b assaydbfor succss is achid.

It aso rmains to b sn hthrmuticoour abing can gnrat compxconnctoms onc ths tchnica probmsha bn sod. If it cannot, or n if itcan, th rapid adancs in high-throughputctron microscopy might man thistchniqu mrgs as a iab atrnati16.

Finay, gin that a of ths mthods aroing in para, it might b that thoptima soution to connctomics i b ahybrid of tchniqus that incuds not ony Brainbo but aso ctron microscopy,array tomography and adancd pathay-tracing mthods.

Conclusion

w ca iring diagrams connctoms tohighight th anaogy to gnoms. Obtainingth squnc of th human gnom asdifficut and xpnsi. Th qust for con-

nctomic information i probaby b nmor difficut. lik gnomics, connctomicsrquirs th dopmnt of n tchniqus,th comparison of atrnatis and aysto mak mundan and automatic hatis initiay difficut and abour-intnsi.Unik gnomics, hor, connctomicsmight rquir th para appication of mutip approachs for abing, imaging,rconstructing and anaysing nurons.Fortunaty, th fid is no moing at anastonishing pac, and th approachs ha outind hr rprsnt ony a substof thos that ar bing dopd indpnd-

nty and nrgticay to orcom thchangs that i ahad.

Jeff W. Lichtman and Joshua R. Sanes are at the

Department of Molecular and Cellular Biology and

Center for Brain Science, Harvard University,

Cambridge, Massachusetts 02138, USA.

Jean Livet is at INSERM, UMR S592, Institut de la

Vision, F-75012, Paris, France and at UPMC University

of Paris 06, UMR S592, F-75005, Paris, France.

Correspondence to J.W.L.

e-mail: [email protected]

di:10.1038/nn2391

Publishd nlin 30 Apil 2008

1. Sotelo, C. Viewing the brain through the master hand

of Ramon y Cajal. Nature Rev. Neurosci. 4, 71–77

(2003).2. Callaway, E. M. & Sanes, J. R. New technologies. Curr.

Opin. Neurobiol. 16, 540–542 (2006).

3. Nauta, W. J. Some early travails of tracing axonal

pathways in the brain. J. Neurosci. 13, 1337–1345

(1993).

4. Callahan, C. A., Yoshikawa, S. & Thomas, J. B. Tracing

axons. Curr. Opin. Neurobiol. 8, 582–586 (1998).5. Feinstein, P., Bozza, T., Rodriguez, I., Vassalli, A. &

Mombaerts, P. Axon guidance of mouse olfactory

sensory neurons by odorant receptors and the β2

adrenergic receptor. Cell 117, 833–846 (2004).

6. Bareyre, F. M., Kerschensteiner, M., Misgeld, T. &

Sanes, J. R. Transgenic labeling of the corticospinal

tract for monitoring axonal responses to spinal cord

injury. Nature Med. 11, 1355–1360 (2005).

7. Dymecki, S. M. & Kim, J. C. Molecular

neuroanatomy’s “Three Gs”: a primer. Neuron 54,

17–34 (2007).

8. Song, C. K., Enquist, L. W. & Bartness, T. J. New

developments in tracing neural circuits with

herpesviruses. Virus Res. 111, 235–249 (2005).

9. Luo, L., Callaway, E. M. & Svoboda, K. Genetic

dissection of neural circuits. Neuron 57, 634–660

(2008).

10. Horowitz, L. F., Montmayeur, J. P., Echelard, Y. &

Buck, L. B. A genetic approach to trace neural circuits.

Proc. Natl Acad. Sci. USA 96, 3194–3199 (1999).11. Yoshihara, Y. et al. A genetic approach to visualization

of multisynaptic neural pathways using plant lectin

transgene. Neuron 22, 33–41 (1999).12. Wickersham, I. R. et al. Monosynaptic restriction of

transsynaptic tracing from single, genetically targeted

neurons. Neuron 53, 639–647 (2007).

13. Macagno, E. R., Levinthal, C. & Sobel, I. Three-dimensional computer reconstruction of neurons and

neuronal assemblies. Annu. Rev. Biophys. Bioeng. 8,

323–351 (1979).

14. White, J. G., Southgate, E., Thomson, J. N. &

Brenner, S. The structure of the nervous system of the

nematode Caenorhabditis elegans. Philos. Trans. R.

Soc. Lond. B Biol. Sci. 314, 1–340 (1986).

15. Denk, W. & Horstmann, H. Serial block-face scanning

electron microscopy to reconstruct three-dimensional

tissue nanostructure. PLoS Biol. 2, e329 (2004).

16. Briggman, K. L. & Denk, W. Towards neural circuit

reconstruction with volume electron microscopy

techniques. Curr. Opin. Neurobiol. 16, 562–570

(2006).

17. Hayworth, K., Kasthuri, N., Schalek, R. & Lichtman,

J. W. Automating the collection of ultrathin serial

sections for large volume TEM reconstructions.

Microsc. Microanal . 12, 2 (2006).18. Gan, W. B., Grutzendler, J., Wong, W. T., Wong, R. O. &

Lichtman, J. W. Multicolor “DiOlistic” labeling of thenervous system using lipophilic dye combinations.

Neuron 27, 219–225 (2000).

19. Tsien, R. Y. The green fluorescent protein. Annu. Rev.

Biochem. 67, 509–544 (1998).

20. Shaner, N. C., Steinbach, P. A. & Tsien, R. Y. A guide to

choosing fluorescent proteins. Nature Methods 2,

905–909 (2005).21. Feng, G. et al. Imaging neuronal subsets in transgenic

mice expressing multiple spectral variants of GFP.

Neuron 28, 41–51 (2000).22. Walsh, M. K. & Lichtman, J. W. In vivo time-lapse

imaging of synaptic takeover associated with naturally

occurring synapse elimination. Neuron 37, 67–73

(2003).23. Kasthuri, N. & Lichtman, J. W. The role of neuronal

identity in synaptic competition. Nature 424,

426–430 (2003).24. Lichtman, J. W. & Sanes, J. R. Watching the

neuromuscular junction. J. Neurocytol . 32, 767–775

(2003).

25. Zong, H., Espinosa, J. S., Su, H. H., Muzumdar, M. D.

& Luo, L. Mosaic analysis with double markers in mice.

Cell 121, 479–492 (2005).

26. Livet, J. et al. Transgenic strategies for combinatorial

expression of fluorescent proteins in the nervous

system. Nature 450, 56–62 (2007).27. Gan, W. B., Kwon, E., Feng, G., Sanes, J. R. &

Lichtman, J. W. Synaptic dynamism measured over

minutes to months: age-dependent decline in an

autonomic ganglion. Nature Neurosci. 6, 956–960

(2003).28. Grutzendler, J., Kasthuri, N. & Gan, W. B. Long-term

dendritic spine stability in the adult cortex. Nature 420, 812–816 (2002).

29. Nguyen, Q. T., Sanes, J. R. & Lichtman, J. W.

Pre-existing pathways promote precise projection

patterns. Nature Neurosci. 5, 861–867 (2002).

30. Trachtenberg, J. T. et al. Long-term in vivo imaging

of experience-dependent synaptic plasticity in adult

cortex. Nature 420, 788–794 (2002).

31. Micheva, K. D. & Smith, S. J. Array tomography: a

new tool for imaging the molecular architecture and

ultrastructure of neural circuits. Neuron 55, 25–36

(2007).

32. Peters, M. F. et al. Differential membrane localization

and intermolecular associations of α-dystrobrevin

isoforms in skeletal muscle. J. Cell Biol. 142,

1269–1278 (1998).33. Fritze, C. E. & Anderson, T. R. Epitope tagging: general

method for tracking recombinant proteins. Methods

Enzymol. 327, 3–16 (2000).

34. Bates, M., Huang, B., Dempsey, G. T. & Zhuang, X.

Multicolor super-resolution imaging with photo-

switchable fluorescent probes. Science 317,

1749–1753 (2007).35. Heintzmann, R. & Ficz, G. Breaking the resolution limit

in light microscopy. Methods Cell Biol. 81, 561–580

(2007).

36. Hell, S. W. Far-field optical nanoscopy. Science 316,

1153–1158 (2007).37. Huang, B., Wang, W., Bates, M. & Zhuang, X. Three-

dimensional super-resolution imaging by stochastic

optical reconstruction microscopy. Science 319,

810–813 (2008).38. Shroff, H. et al. Dual-color superresolution imaging of

genetically expressed probes within individual

adhesion complexes. Proc. Natl Acad. Sci. USA 104,

20308–20313 (2007).

39. Ramón y Cajal, S. Estudios Sobre la Degeneración y

Regneración del Sistema Nervioso (Moya, Madrid,

1913–1914); reprinted and edited with additional

translations by DeFelipe, J. & Jones, E. G. Cajal’s

Degeneration and Regeneration of the Nervous

System (Oxford Univ. Press, New York, 1991).

40. Hubel, D. H., Wiesel, T. N. & LeVay, S. Plasticity of oculardominance columns in monkey striate cortex. Philos.

Trans. R. Soc. Lond. B Biol. Sci. 278, 377–409 (1977).

Progress


mailto:[email protected]



a technicolour approach to the conectome-brainbow

Documents