10. Gutfreund, Y., Flash, T., Yarom, Y., Fiorito,G., Segev, I., and Hochner, B. (1996).Organization of octopus arm movements:a model system for studying the control offlexible arms. J. Neurosci. 16, 7297–7307.

11. Sumbre, G., Gutfreund, Y., Fiorito, G.,Flash, T., and Hochner, B. (2001). Controlof octopus arm extension by a peripheralmotor program. Science 293, 1845–1848.

12. Sumbre, G., Fiorito, G., Flash, T., andHochner, B. (2005). Motor control of flexibleoctopus arms. Nature 433, 595–596.

13. Sumbre, G., Fiorito, G., Flash, T., andHochner, B. (2006). Octopuses usea human-like strategy to control precisepoint-to-point movements. Curr. Biol. 16,767–772.

14. Georgopoulos, A., and Ashe, J. (2000).One motor cortex, two different views.Nat. Neurosci. 3, 963.

15. Hannan, M., and Walker, I. (2001). Analysisand experiments with an elephant’s trunkrobot. Adv. Robotics 15, 847–858.

16. Hannan, M., and Walker, I. (2003).Kinematics and the implementation of anelephant’s trunk manipulator and othercontinuum style robots. J. Robotic Syst.20, 45–63.

17. Mochiyama, H., Shimemura, E., andKobayashi, H. (1999). Control ofmanipulators with hyper degrees offreedom: shape tracking using only jointangle information. Int. J. Syst. Sci. 30,77–85.

18. Smith, K., and Kier, W. (1989). Trunks,tongues, and tentacles: moving withskeletons of muscle. Am. Sci. 77,29–35.

19. Chiel, H., Crago, P., Mansour, J., andHathi, K. (1992). Biomechanics of

a muscular hydrostat: a model of lappingby a reptilian tongue. Biol. Cybern. 67,403–415.

20. Morris, L., and Hooper, S. (1998). Muscleresponse to changing neuronal input inthe lobster (Panulirus interruptus)stomatogastric system: slow muscleproperties can transform rhythmic inputinto tonic output. J. Neurosci. 18,3433–3442.

Department of Biological Sciences,Ohio University, Athens,Ohio 45701, USA.E-mail: hooper@ohiou.edu

DOI: 10.1016/j.cub.2006.03.045

DispatchR285

Evolution: The PlasticTranscriptome

Studies across a wide range of species point to a surprising degree ofplasticity in the transcriptional states that organisms can adopt,suggesting that organisms often respond to environmental challengesthrough wholesale reprogramming of their gene expression.

Greg Gibson

How plastic is the transcriptome?This is probably not a question thatmany of us have thought aboutmuch; yet it is fundamental to anunderstanding of how organismsadjust their physiology andbehavior to cope with the diversechallenges posed by theenvironment. The literature on thetopic is as yet quite slim, but oneprofound insight is beginning toemerge, namely that organismscan globally switch transcriptionalstates. Individuals in either statedisplay considerably moredivergent expression profiles thanthose seen across the geographicrange of the species within a givenstate. I will briefly describe fourexamples of this phenomenon,before discussing the evolutionaryand biomedical implications oftranscriptional plasticity.

In this issue of Current Biology,Lagardier and colleagues [1]describe transcriptionaldifferentiation between sedentaryand migratory salmonid fish inWestern Europe. The authorssampled livers of 90 juvenile browntrout from six localities andconducted microarray analysis ona set of 900 cDNAs. Their striking

result (Figure 1) is that the overalltranscriptional phenotypes clusterby life history strategy, despite thegenotypic data from the samesamples clearly discriminating thefish by locality. Fish destined tomigrate out of their streams thusdisplay a particular pattern of livergene expression, which indicatesthat it is not just their behavior thathas been preprogrammed, and thispattern is pretty much the samewhether the fish stem fromMediterranean or Atlantic lineages,which diverged half a million yearsago. By employing a novelapplication of a Mantel statistic,they further estimate that lifehistory (45%) explains three timesmore of the transcriptionalvariation than genetic ancestry(15%). The remainder is ascribed torandom differences amongindividuals, but by restricting theiranalysis to 268 of the genes theyare able to generate a molecularsignature that predicts whethera fish will be sedentary ormigratory.

This is, of course, not to say thattranscriptional variation betweenindividuals is not significant.A different perspective on thepopulation structure of expressionvariation emerges from

a microarray study of 192metabolic genes in the brain, liverand heart of three populations ofFundulus fish [2]. After the fish wereraised in a common laboratoryenvironment, three quarters of thetranscripts were found to differ inabundance between tissues, asmight be expected given thevarying metabolic requirements ofthe tissues, but only one third ofthese patterns were consistentacross the three populations.Furthermore, half of the transcriptsdiffered between individuals,implying that caution should beraised in assuming thatmeasurements on a singlelaboratory strain are representativeof the entire species.

Remarkable reprogramming ofglobal gene expression has alsobeen documented in relation to thebehavioral occupations of adulthoneybees [3]. A highly replicatedexperimental design was used toshow that in the honeybee brain theabundance of 39% of 5,500 geneschanges with the transition fromworking inside the hive to foragingoutside. This plasticity was shownto be independent of the agingprocess, and as with the trout,a molecular signature derived froman informative subset of the genescorrectly predicts behavior in 95%of their sample of 60 individuals.More recently, the same group[4] has shown that the transitionaloccupations — such ascomb-building, guarding andundertaking — are by contrastassociated only with verymodest transcriptional changes.This suggests that dramaticshort-term behavioral differences

Current Biology Vol 16 No 8R286

Current Biology

D3

D4

D5

F1

F2

Med

D3

D4

D5

F2

F1

Med

Genetic differentiation Transcriptional differentiation

Sedentary

Migratory

Figure 1. Transcriptional variation matches life history rather than relatedness inbrown trout.

Clustering of liver samples from juvenile brown trout populations by transcriptionalprofile correlates with sedentary (red) versus migratory (blue) life history strategy, irre-spective of genetic and geographic differentiation. ‘Med’ indicates the Mediterraneanlineage, in contrast with the Atlantic lineage represented by samples from France (F)and Denmark (D). Adapted from [1].

need not require wholesaletranscriptional change, butthat long-term behavioraldifferences may typically beassociated with substantialphysiological remodelingmediated at the transcriptionallevel.

Moving on to a mammalianexample, considerabledifferentiation between wild andfarm-raised foxes in Sweden hasbeen observed [5], in so far as2,500 of almost 30,000 humancDNAs on their microarrays weredifferentially expressed in threebrain regions. We have observeda similar differentiation betweenwhole blood samples of captiveand free-range red wolves(E.M. Kennerly and G. Gibson, inpreparation), again favoring thenotion that environmentalcircumstances can dramaticallyalter global gene expression. TheSwedish study also contrasted twosubsets of the farm-raised foxes,one selected for tameness and theother representing the base farmpopulation, and found that fewerthan 0.1% of the genes, just 40transcripts, had responded toselection. Contrast this result withthat of a study of similar selectionintensity, for copulation speed inDrosophila melanogaster [6], inwhich as many as 25% of the geneschanged in abundance, and it isclear that the effect of behavioral

selection in the foxes was verylimited.

A last example concerns droughtand heat stress response in plants[7]. Desiccation and elevatedtemperature induce global geneexpression responses inArabidopsis thaliana, as might beexpected, but these are notadditive phenomena. Exposed toboth stresses simultaneously, theplants induce an extra set of over450 genes and turn off the inductionof some of the major stressresponse mechanisms involved ineither drought or heat alone. Theseglobal plastic changes areassociated with differentiationin the metabolic pools measuredin the plants, which seem to switchenergy sources as required. Thecoordinated response points toa highly evolved mechanism forcoping with environmentalchallenges.

The combined message fromthese diverse studies is thattranscriptomes can adopt highlydifferentiated states involving largepercentages of the genesexpressed in a particular tissue. Itis no longer surprising that singletranscription factors can regulatethe activity of batteries of hundredsof target genes; after all, this is howdevelopment is orchestrated.However, what is important here isthat the magnitude and extent ofthe effects have been found to be

large in comparison to geneticdifferentiation between individualsand even populations. It will beimportant to ascertain to whatextent the profiles are qualitativelydistinct, representing phase shiftsas opposed to simply beingextremes of a continuousdistribution of transcription.

The relevance to evolutionarybiology is that these findings onceagain remind us to look beneath thesurface of the phenotype if we areto understand the forces shapinggenetic variation. Much of theclassical literature on phenotypicplasticity deals with polyphenism,which is the adoption ofqualitatively distinct traits bydifferent individuals of a species,such as winged and winglesscastes of ants or horn size inbeetles [8]. Now we have asense that polyphenism mayindeed be prevalent at themolecular level across a broadarray of organisms.

It has also become apparentrecently that transcriptionmeasured under a standard set oflaboratory conditions is typicallyunder strong stabilizing selection[9,10]. This means that cis- andtrans-acting mutations thatpromote divergence from normallevels of transcript abundance areeither quickly purged by naturalselection, or suppressed byhomeostatic mechanisms. If thereare typically two or more states thatgene expression profiles canadopt, then the question arises asto whether variants that modifytranscript abundance have anequivalent effect on both states.How often is a mutation that isdeleterious in one condition,advantageous or neutral underanother? Widespread plasticitywould imply greater potential forthe balancing of variation: just asthe sexes provide alternatephysiological environments, so tooshould conditions of stress,behavioral caste, or otherresponses to environmental shifts.

The work mentioned above is notwithout biomedical implicationseither. Disease states such asdiabetes, asthma, and depressionare often thought to arise asa threshold response thatfundamentally shifts the relevantaspects of an organism’s

DispatchR287

physiology. The role ofcorticosteroid releasing factor inregulation of amphibianmetamorphosis and mammalianeclampsia provides precedent forimplication of a conservedregulatory pathway that normallyregulates environmental response,promoting disease under stressfulcircumstances [11]. It is worthwondering whether the moleculesthat regulate thresholds in diseasemore generally coordinate globaltranscriptional switches as part ofnormal physiological function.

References1. Giger, T., Excoffier, L., Day, P.J.R.,

Champigneulle, A., Hansen, M.M., Powell,R., and Lagardier, C.R. (2006). Life historyshapes gene expression in salmonids.Curr. Biol. 16, R281–R282.

2. Whitehead, A., and Crawford, D.L. (2005).Variation in tissue-specific gene

Chromatin RemodMotors along the

Single-molecule experiments showcomplex RSC, a member of the SNFof a negatively supercoiled DNA loo

Terence Strick andAudrey Quessada-Vial

Chromatin remodeling complexessuch as RSC and SWI/SNF usethe energy of ATP hydrolysis tomodify nucleosome structure andthereby regulate DNA function(for a review see [1]). This could inprinciple involve altering theposition or stability ofnucleosomes along DNA, forinstance via physical interactionwith the histone or even chemicalmodification of its octamercomposition [2]. An oft-discussedpossibility is that chromatin-remodeling complexes translocateDNA, directly pushingnucleosomes along, or off of, theDNA, and perhaps even modifyingthe underlying higher-order DNAstructure to further alter the bindingstability of the nucleosome [3,4].This hypothesis is supported byexperimental work showing thatDNA structures, such assupercoiling and looping, can beinduced by chromatin-remodeling

expression among natural populations.Genome Biol. 6, R13.

3. Whitfield, C.W., Cziko, A.M., andRobinson, G.E. (2003). Gene expressionprofiles in the brain predict behaviorin individual honey bees. Science 302,296–299.

4. Cash, A.C., Whitfield, C.W., Ismail, N., andRobinson, G.E. (2005). Behavior and thelimits of genomic plasticity: power andreplicability in microarray analysis ofhoneybee brains. Genes Brain Behav. 4,267–271.

5. Lindberg, J., Bjornerfeldt, S., Saetre, P.,Svartberg, K., Seehuus, B., Bakken, M.,Vila, C., and Jazin, E. (2005). Selection fortameness has changed brain geneexpression in silver foxes. Curr. Biol. 15,R915–R916.

6. Mackay, T.F., Heinsohn, S.L., Lyman,R.F., Moehring, A.J., Morgan, T.J., andRollmann, S.M. (2005). Genetics andgenomics of Drosophila mating behavior.Proc. Natl. Acad. Sci. USA 102 Suppl 1,6622–6629.

7. Rizhsky, L., Liang, H., Shuman, J.,Shulaev, V., Davletova, S., andMittler, R. (2004). When defensepathways collide. The response ofArabidopsis to a combination of

eling: RSCDNA

that the chromatin-remodeling2 ATPase family, induces formationp by active translocation.

complexes [3,5]. Understandinghow chromatin remodeling canaffect gene regulation would alsorequire additional insight into thekinetics of the process: howquickly can such changes begenerated, and how stable arethey? These questions are noteasily answered using classicalbiochemical techniques, but newexperimental work using real-timesingle-molecule DNAnanomanipulation begins toaddress these issues.

Lia et al. [6] have recently useda magnetic-trap based single-molecule DNA nanomanipulationsetup to study the interactionsbetween a single RSC complex anda single, 3.6 kilobase linear DNAmolecule. Single-molecule DNAnanomanipulation has gainedwidespread interest for itsapplicability to the real-time studyof protein–DNA interactions. Byattaching one end of a linear DNAmolecule to a glass coverslip andthe other end to a small magneticbead, the DNA can be mechanically

drought and heat stress. Plant Physiol.134, 1683–1696.

8. Miura, T. (2005). Developmental regulationof caste-specific characters in social-insect polyphenism. Evol. Dev. 7,122–129.

9. Denver, D.R., Morris, K., Streelman, J.T.,Kim, S.K., Lynch, M., and Thomas, W.K.(2005). The transcriptional consequencesof mutation and natural selection inCaenorhabditis elegans. Nat. Genet. 37,544–548.

10. Rifkin, S.A., Houle, D., Kim, J., and White,K.P. (2005). A mutation accumulationassay reveals a broad capacity for rapidevolution of gene expression. Nature 438,220–223.

11. Crespi, E.J., and Denver, R.J. (2005).Ancient origins of human developmentalplasticity. Am. J. Hum. Biol. 17, 44–54.

Department of Genetics, North CarolinaState University, Raleigh,North Carolina 27695-7614, USA.E-mail: ggibson@unity.ncsu.edu

DOI: 10.1016/j.cub.2006.03.041

manipulated — stretched andtwisted — by acting on the beadwith a magnetic tweezer. Theend-to-end extension of thenanomanipulated DNA isdetermined in real-time bymeasuring the position of thebead above the surface, and isa robust metric for theconformational state of theDNA molecule.

Such single-moleculetechniques are particularly usefulfor the study of DNA translocases,enzymes which use the energy ofnucleotide hydrolysis to drivethemselves along the DNA ina directional manner. Indeed,whereas some translocases suchas RNA polymerase advancebase-by-base [7], rotate with theDNA double-helix [8] and leavebehind an easily identifiable, easilyquantifiable biochemical product,most translocases may in factcouple their motion to DNAstructure in a nontrivial fashionand not produce anything morethan ephemeral, physicalwork — movement along DNA [9].Using an appropriateexperimental geometry, enzymetranslocation can be simplydetected in real-time bymonitoring the resulting changesin the nanomanipulated DNA’send-to-end extension.

When Lia et al. [6] placed thenanomanipulated DNA in the