Current Biology Vol 22 No 2R54

diverse functions. Indeed, thegeometry of fission yeast may be ideal,and unique, for precisely this typeof chimeric analysis. However,given the multiple and overlappingroles of the actin and microtubulecytoskeleton in complex cellularprocesses, such as cell polarity andcell shape, the chimeric analysispresented by Lo Presti and Martin [6]can help to simplify the differentpathways even further. By eliminatingone pathway, future research canfocus on the molecular dissectionof one pathway without compoundingeffects from another overlappingpathway.

References1. Li, R., and Gundersen, G.G. (2008). Beyond

polymer polarity: how the cytoskeleton buildsa polarized cell. Nat. Rev. Mol. Cell Biol. 9,860–873.

2. Rodriguez, O.C., Schaefer, A.W.,Mandato, C.A., Forscher, P., Bement, W.M.,and Waterman-Storer, C.M. (2003). Conservedmicrotubule-actin interactions in cellmovement and morphogenesis. Nat. Cell Biol.5, 599–609.

3. Siegrist, S.E., and Doe, C.Q. (2007).Microtubule-induced cortical cell polarity.Genes Dev. 21, 483–496.

4. Brown, S.S. (1999). Cooperation betweenmicrotubule- and actin-based motor proteins.Annu. Rev. Cell Dev. Biol. 15, 63–80.

5. Ross, J.L., Ali, M.Y., and Warshaw, D.M. (2008).Cargo transport: molecular motors navigatea complex cytoskeleton. Curr. Opin. Cell Biol.20, 41–47.

6. Lo Presti, L., and Martin, S.G. (2011). Shapingfission yeast cells by rerouting actin-basedtransport on microtubules. Curr. Biol. 21,2064–2069.

7. Chang, F., and Martin, S.G. (2009). Shapingfission yeast with microtubules. Cold SpringHarb. Perspect. Biol. 1, a001347.

8. Martin, S.G. (2009). Microtubule-dependentcell morphogenesis in the fission yeast. TrendsCell Biol. 19, 447–454.

9. Piel, M., and Tran, P.T. (2009). Cell shape andcell division in fission yeast. Curr. Biol. 19,R823–R827.

10. Bendezu, F.O., and Martin, S.G. (2011). Actincables and the exocyst form two independent

morphogenesis pathways in the fission yeast.Mol. Biol. Cell 22, 44–53.

11. Nakano, K., Toya, M., Yoneda, A., Asami, Y.,Yamashita, A., Kamasawa, N., Osumi, M., andYamamoto, M. (2011). Pob1 ensurescylindrical cell shape by coupling twodistinct rho signaling events duringsecretory vesicle targeting. Traffic 12,726–739.

12. Snaith, H.A., Thompson, J., Yates, J.R., 3rd,and Sawin, K.E. (2011). Characterization ofMug33 reveals complementary roles for actincable-dependent transport and exocystregulators in fission yeast exocytosis. J. CellSci. 124, 2187–2199.

13. Feierbach, B., Verde, F., and Chang, F. (2004).Regulation of a formin complex by themicrotubule plus end protein tea1p. J. Cell Biol.165, 697–707.

1Institut Curie, CNRS-UMR144, Paris 75005,France. 2Cell and Developmental Biology,University of Pennsylvania, Philadelphia,PA 19104, USA.*E-mail: tranp@mail.med.upenn.edu

DOI: 10.1016/j.cub.2011.12.007

Epigenetic Inheritance: What Newsfor Evolution?

Whether epigenetic variation is important in adaptive evolution has beencontentious. Two recent studies in Arabidopsis thaliana significantly add to ourunderstanding of genome-wide variation and stability of an epigenetic mark,and thus help pave the path for realistically incorporating epigenetics intoevolutionary theory.

Ben Hunter, Jesse D. Hollister,and Kirsten Bomblies

Epigenetic marks such as cytosinemethylation or histone modificationscan be very dynamic and can altergene expression in response toenvironmental and developmental cueswithout changes in DNA sequence; insome cases epigenetic changes can beheritable through meiosis [1,2]. Thishas spurred interest — and heateddebates — about whether epigeneticvariation may play a significant rolein adaptive evolution [3–6]. The needto formally consider epialleles inpopulation genetics and evolutionarytheory has been emphasized (e.g.,[6,7]); however, more empirical dataare necessary to parameterize modelsand assess the actual impacts ofepigenetic variation on adaptivephenotypes (e.g., [3,8]).

Two recent studies in Arabidopsisthaliana have quantified spontaneous

genome-wide methylation variation,and are a significant step forwardin quantifying epigenetic change[9,10]. Both studies capitalized ona very useful resource: a set ofwell-characterized mutationaccumulation lines propagatedfrom one homozygous ancestor(Figure 1) [11]. This allowsquantification of the rate andaccumulation of differences in theabsence of natural selection. Suchlines exist for numerous species, whichwill allow for extensive comparativework [12]. In the A. thaliana studiestwo individuals each from five [9] andten [10] 31st generation lines wereassayed for genome-wide cytosinemethylation patterns and comparedto lines that had been propagated foronly three generations from thecommon ancestor (Figure 1). Theselines have also been used to quantifythe base mutation rate [13] as well asphenotypic divergence [11].

Both A. thaliana studies concludedthat, in general, cytosine methylation isremarkably stable over the 64generations that separate the mostdivergent lines (Figure 1). But at someloci it does vary: in the two studies1.6% [9] and 6.4% [10] of methylatedcytosines differed in methylation stateamong lines. This gives epimutationrate estimates orders of magnitudehigher than the DNA base mutationrate. Consistent with patternspreviously reported for variation amongnatural A. thaliana strains [14], variableCG-methylation sites werepreferentially located in gene regions,while the methylation states oftransposons and repeat regions weremostly stably inherited. Variation innon-CG methylation is comparativelyrare, but showed the opposite pattern,beingmore variable in transposons andintergenic regions [9].What does spontaneous variation

imply for the potential for epigeneticchange to play an important rolein evolution? First, consider thegenome-wide variation in stability ofmethylation states. Among the variablesites identified in these A. thalianagenome scans, a large proportionchanged state in multiple independentlines, suggesting that some sites areindeed ‘hotspots’ for epigeneticchange [9,10], and rates of reversionare appreciable [10]. It has been known

Generation 31

Generation 32Current Biology

Generation 3

Figure 1. Assaying the stability of DNA methylation epialleles in Arabidopsis thaliana.

Schematic illustration of the generation of mutation accumulation lines in A. thaliana [11] usedin recent studies of spontaneous variation in cytosine methylation patterns [9,10]. Many lines(for simplicity only four are shown here) are propagated from a single homozygous ancestor(grey, top) by single-seed descent. For outcrossing species, this would be achieved by sibmating. The A. thaliana lines sampled for cytosine methylation variation were propagatedfor 3 and 31 generations. Third and 31st generation individuals are separated by 34 genera-tions and any two 31st generation individuals by 62 generations. One study [10] also sampled32nd generation individuals that were progeny of sibs of the 31st generation plants sampled,and thus separated from them by two generations.

DispatchR55

for some time that epialleles at someloci are ‘metastable’ and can changedramatically over generations [15].Such instability suggests it is unlikelythat alternative epialleles cancontribute appreciably to stableevolutionary change [4].

While instability speaks against theidea that individual epialleles wouldcontribute to long-term adaptiveevolution, it does beg the questionwhy there is variation among loci inepigenetic stability in the first place.As Richards has pointed out, onepossibility is that the unstableepialleles are really just phenotypicallyinconsequential ‘‘genomic clutter’’ thatis reset with passing generations [3].On the other hand, such variationcould also be part of a plasticenvironmental response system or,if selection can stabilize epigeneticstates, then it becomes a standingsupply of potentially heritable, adaptiveepialleles [3]. A particularly intriguingpossible explanation when consideringthe role that epigenetic variation mayplay in long-term evolution is that it isthe propensity to vary, rather than anyparticular allelic state, that is underselection. Simulations have shown thatphenotypic variation and plasticitygenerated by epigenetic instability canbe beneficial in variable environments,and thus instability may itself bea target of selection [16]. Hence themethylation hotspots identified inthese A. thaliana lines may beover-represented for loci that haveexperienced selection for epigeneticinstability. An important future questionthen is whether the trans-generationalstability of methylation at a particularlocus can respond to selection, andif so, by what mechanism.

Understanding the significance ofmethylation variation, or any otherepigenetic mark, depends on how itcorrelates with gene expression andphenotypic variation. In both of theA. thaliana studies, gene expressionwas measured for several loci withvariable methylation profiles. In abouthalf the cases, methylation status didcorrelate with gene expression levels[9,10]; at one locus a methylationchange resulted in use of an alternativepromoter, which led to a change inabundance of one isoform [9]. It isalso clear from this that some loci canhave altered methylation with noappreciable effect on transcription ofthat locus. Transcriptome analysisrevealed that 320 transcripts differed in

abundance among lines, yet only sevenoverlapped with identified differentiallymethylated regions [10]. While someof this can be explained by limits indetection power of methylation, it alsosuggests there may be more todiscover — other factors, such as DNAmutations or epimutations in upstreamregulators, or other epigenetic changessuch as histone or nucleosomemodification, may also play importantroles.

The mutation accumulation linesused here, as well as those that havebeen generated in a range of othersystems [12], can provide important

insights. In addition to variation incytosine methylation, these lines canalso be used to gather data on theaccumulation of variation in otherepigenetic marks, and how all of thesemarks interact to alter gene expressionpatterns in the absence of sequencechanges [17]. As was pointed out [10],an important additional experiment willbe to expose these lines to alternative(stable or variable) environments andask how this affects epigenetic marks.This will allow assessment of thedegree to which the spontaneouslyvariable sites observed in these studiesoverlap with the set that are

Current Biology Vol 22 No 2R56

environmentally responsive. Thougha greater time investment, a valuablefollow-up study would be to propagatelines under different selection regimesfor multiple generations and thenask whether or how epigenetic generegulation responds, and whether anyobserved differences in fitness arestable adaptations or plasticacclimation.

Having realistic numbers forparameters such as allele stability,epimutation rates and reversion rates iscritical for incorporating epigeneticsinto evolutionary theory. Studies suchas the recent A. thaliana variationaccumulation studies [9,10] providesuch vital empirical data. Movingforward, we need methods forassessing whether epigenetic marksare evolving neutrally or underselection. How do we quantifyselection on methylation patterns orother epigenetic marks? What is theneutral expectation? When we observedivergence in methylation, how canwe assess whether this happenedunder selection or via random ‘noise’or plasticity in the regulatory system?Having a formal body of evolutionarytheory that incorporates epigenetics,as well as developing a clearerquantification of the connectionbetween epigenetic variation andphenotypes will allow us to more

rigorously ask whether or howepigenetics plays an important role inadaptive evolution. This area promisesinteresting new angles in the study ofevolution.

References1. Jaenisch, R., and Bird, A. (2003). Epigenetic

regulation of gene expression: how the genomeintegrates intrinsic and environmental signals.Nature Genet. 33 (Suppl. ), 243–254.

2. Wang, X., Elling, A.A., Li, X., Li, N., Peng, Z.,He, G., Sun, H., Qi, Y., Liu, X.S., and Deng, X.W.(2009). Genome-wide and organ-specificlandscapes of epigenetic modifications andtheir relationships to mRNA and small RNAtranscriptomes in Maize. Plant Cell. 21,1053–1069.

3. Richards, E.J. (2008). Population epigenetics.Curr. Op. Genet. Devel. 18, 221–226.

4. Haig, D. (2007). Weismann rules! OK?Epigenetics and the Lamarckian temptation.Biol. Philos. 22, 415–428.

5. Jablonka, E., and Lamb, M.J. (2007). Theexpanded evolutionary synthesis - a responseto Godfrey-Smith, Haig, and West-Eberhard.Biol. Philos. 22, 453–472.

6. Geoghegan, J.L., and Spencer, H.G. (2011).Population-epigenetic models of selection.Theor. Pop. Biol., in press. DOI: 10.1016/j.tpb.2011.08.001.

7. Johannes, F., Porcher, E., Teixeira, F.K.,Saliba-Colombani, V., Simon, M., Agier, N.,Bulski, A., Albuisson, J., Heredia, F.,Audigier, P., et al. (2009). Assessing the impactof transgenerational epigenetic variation oncomplex traits. PLoS Genet. 5, e1000530.

8. Bossdorf, O., Richards, C.L., and Pigliucci, M.(2008). Epigenetics for ecologists. Ecol. Lett.11, 106–115.

9. Schmitz, R.J., Schultz, M.D., Lewsey, M.G.,O’Malley, R.C., Urich, M.A., Libiger, O.,Schork, N.J., and Ecker, J.R. (2011).Transgenerational epigenetic instability isa source of novel methylation variants. Science334, 369–373.

10. Becker, C., Hagmann, J., Muller, J., Koenig, D.,Stegle, O., Borgwardt, K., and Weigel, D. (2011).Spontaneous epigenetic variation in theArabidopsis thaliana methylome. Nature 480,245–249.

11. Shaw, R.G., Byers, D.L., and Darmo, E. (2000).Spontaneous mutational effects onreproductive traits of Arabidopsis thaliana.Genetics 155, 369–378.

12. Halligan, D.L., and Keightley, P.D. (2009).Spontaneous mutation accumulation studies inevolutionary genetics. Annu. Rev. Ecol. Evol.Syst. 40, 151–172.

13. Ossowski, S., Schneeberger, K.,Lucas-Lledo, J.I., Warthmann, N., Clark, R.M.,Shaw, R.G., Weigel, D., and Lynch, M. (2010).The rate and molecular spectrum ofspontaneous mutations in Arabidopsis thaliana.Science 327, 92–94.

14. Vaughn, M.W., Tanurdzi�c, M., Lippman, Z.,Jiang, H., Carrasquillo, R., Rabinowicz, P.D.,Dedhia, N., McCombie, W.R., Agier, N.,Bulski, A., et al. (2007). Epigenetic naturalvariation in Arabidopsis thaliana. PLoS Biol. 5,e174.

15. Rakyan, V.K., Blewitt, M.E., Druker, R.,Preis, J.I., and Whitelaw, E. (2002). Metastableepialleles in mammals. Trends Genet. 18,348–351.

16. Feinberg, A.P., and Irizarry, R.A. (2010).Stochastic epigenetic variation as a drivingforce of development, evolutionary adaptation,and disease. Proc. Natl. Acad. Sci. USA 107(Suppl. 1 ), 1757–1764.

17. Henderson, I.R. (2009). Describing epigenomicinformation in Arabidopsis. In Epigenomics,A.C. Ferguson-Smith, J.M. Greally, andR.A. Martienssen, eds. (Springer), pp. 163–175.

Department of Organismic and EvolutionaryBiology, Harvard University, 26 Oxford St,Cambridge, MA 02138, USA.E-mail: kbomblies@oeb.harvard.edu

DOI: 10.1016/j.cub.2011.11.054

Auditory Neuroscience: How toEncode Microsecond Differences

Minute differences between the time of arrival of a sound at the two ears areused by humans and animals to locate the source. New in vivo recordings haveshed light on howauditory neurons solve the problemof resolvingmicrosecondtime differences.

Christine Koppl

When a sound reaches one ear beforethe other, the resulting interaural timedifference is used by humans andanimals to locate the source. Soundseasy? The catch is that these interauraltime differences are tiny, only fractionsof milliseconds. Just how neuronsresolve these is an ongoing topic ofinvestigation. In an experimental tourde force, Funabiki and colleagues [1]have now achieved the first in vivo

intracellular recordings from neuronsthat are known to perform the interauralcomparison with exquisite precision.Surprisingly, they found that the spikingof those neurons, in the barn owl,was not driven by slow changes inmembrane potential, as is the generalrule. Instead, membrane-potentialfluctuations of hitherto unknownspeed— in the kilohertz range — wereobserved that correlated with the sharptuning for specific interaural timedifferences in single cells. These results

significantlyadvanceourunderstandingof a computation that lies at the limits ofwhat neurons are capable of.

Can Neurons Be Sufficiently Fast?The fact that humans and animalsuse interaural time differences forsound localisation has long beenknown [2,3]. Ways in which this couldbe implemented neurally were alsosuggested early. Arguably the mostinfluential model was that publishedin 1948 by Lloyd A. Jeffress [4]. Onecentral tenet of Jeffress’ model wascoincidence detection betweentemporally precise inputs from bothears — neurons that would firepreferentially if their binaural inputscoincided exactly in time.Such coincidence detection has

since been demonstrated inspecialised auditory brainstemneurons of the avian (and crocodilian)