Current Biology Vol 21 No 17R650

interference resulted in some cohesinloading onto meiotic chromosomes.This partial loading of cohesin resultedin similar defects in DNA repair thatwere now competent to activate theDNA damage checkpoint.

The involvement of cohesin inthe DNA damage response during aspecialized cell division in which thesister chromatid is not the preferredpartner in repair raises the question ofwhat role the cohesin complex playsin DNA damage repair and checkpointactivation. The straight-forwardconcept that the complex holds sisterchromatids in close proximity as atemplate for repair is not relevant in thissituation. The additional observationthat a fraction of cohesin on meioticchromosomes, while not enoughto support proper inter-homologrecombination, can supportcheckpoint activation, presentsan alternative hypothesis. Cohesinmay contribute to chromosomearchitecture in a way that promotescheckpoint activation and DNA repairindependent of sister chromatidcohesion [12]. Indeed, this possibilityhas been suggested by experimentsin mitotic vertebrate cells, in whichdepletion of cohesin subunitsabrogated the DNA damagecheckpoint in G2. However, depletionof an accessory factor required forestablishment of cohesion did notalter checkpoint activation,suggesting a role independent of sisterchromatid cohesion in checkpointactivation [13]. Thus, in both mitosisand meiosis, the cohesin complexmay act as a molecular platform onchromosomes that promotes DNAdamage checkpoint activation andDNA repair [12].

Additional questions are raisedby the studies performed byMartinez-Perez and his colleagues.Since cohesin is required for the DNAdamage response in meiosis, is themechanism of its regulation the sameas in mitosis? Are the same residuesin the same subunits of the cohesincomplex phosphorylated bycheckpoint kinases in responseto persistent recombinationintermediates? Experiments frombudding yeast suggest that this maybe an oversimplification. Koshlandand colleagues showed that Scc1,the mitotic kleisin, supportsDSB-dependent cohesion and DNArepair in G2. However, if the meiotickleisin, Rec8, is expressed during themitotic cell cycle, it cannot generatecohesion in G2 and DSB repair isdisrupted. They attribute this differenceto a single amino acid residue in Scc1that is phosphorylated by a DNAdamage checkpoint kinase and is notconserved in Rec8 [14]. However, theSMC members of the cohesin complexare also targets of checkpoint kinasesduring the DNA damage response invertebrate cells. Since it is becomingapparent that multiple organisms havemore than one meiotic cohesincomplex defined by different kleisinsubunits, it is possible that the DNAdamage checkpoint may target thecommon members of these meioticcomplexes, Smc1 and Smc3 [4–6].

References1. Sancar, A., Lindsey-Boltz, L.A.,

Unsal-Kacmaz, K., and Linn, S. (2004).Molecular mechanisms of mammalian DNArepair and the DNA damage checkpoints.Annu. Rev. Biochem. 73, 39–85.

2. Watrin, E., and Peters, J.M. (2006). Cohesin andDNA damage repair. Exp. Cell Res. 312,2687–2693.

3. Unal, E., Arbel-Eden, A., Sattler, U., Shroff, R.,Lichten, M., Haber, J.E., and Koshland, D.(2004). DNA damage response pathwayuses histone modification to assemble adouble-strand break-specific cohesin domain.Mol. Cell 16, 991–1002.

4. Kitagawa, R., Bakkenist, C.J., McKinnon, P.J.,and Kastan, M.B. (2004). Phosphorylation ofSMC1 is a critical downstream event in theATM-NBS1-BRCA1 pathway. Genes Dev. 18,1423–1438.

5. Yazdi, P.T., Wang, Y., Zhao, S., Patel, N.,Lee, E.Y., and Qin, J. (2002). SMC1 isa downstream effector in the ATM/NBS1branch of the human S-phase checkpoint.Genes Dev. 16, 571–582.

6. Kim, S.T., Xu, B., and Kastan, M.B. (2002).Involvement of the cohesin protein, Smc1, inAtm-dependent and independent responses toDNA damage. Genes Dev. 16, 560–570.

7. Bhalla, N., and Dernburg, A.F. (2008). Preludeto a division. Annu. Rev. Cell Dev. Biol. 24,397–424.

8. Macqueen, A.J., and Hochwagen, A. (2011).Checkpoint mechanisms: the puppet mastersof meiotic prophase. Trends Cell Biol. 21,393–400.

9. Revenkova, E., Eijpe, M., Heyting, C., Gross, B.,and Jessberger, R. (2001). Novelmeiosis-specific isoform of mammalianSMC1. Mol. Cell Biol. 21, 6984–6998.

10. Wood, A.J., Severson, A.F., and Meyer, B.J.(2010). Condensin and cohesin complexity:the expanding repertoire of functions.Nat. Rev. Genet. 11, 391–404.

11. Lightfoot, J., Testoori, S., Barroso, C., andMartinez-Perez, E. (2011). Loading of meioticcohesin by SCC-2 is required for earlyprocessing of DSBs and for the DNAdamage checkpoint. Curr. Biol. 21, 1421–1430.

12. Jessberger, R. (2009). Cohesin’s dual role in theDNA damage response: repair and checkpointactivation. EMBO J. 28, 2491–2493.

13. Watrin, E., and Peters, J.M. (2009). Thecohesin complex is required for the DNAdamage-induced G2/M checkpoint inmammalian cells. EMBO J. 28, 2625–2635.

14. Heidinger-Pauli, J.M., Unal, E., Guacci, V., andKoshland, D. (2008). The kleisin subunit ofcohesin dictates damage-induced cohesion.Mol. Cell 31, 47–56.

Department of Molecular,Cell and Developmental Biology,University of California, Santa Cruz,Santa Cruz, CA 95064, USA.E-mail: nbhalla@ucsc.edu

DOI: 10.1016/j.cub.2011.07.039

Sensory Neurophysiology:Motion Vision during Motor Action

A recent study identifies mechanisms of state-dependent modulation ofvisual processing, using a comprehensive approach of electrophysiologyin the behaving animal, pharmacology and computational modelling.

Kit D. Longden and Holger G. Krapp

Imagine you are running for your life:the faster you run, the faster the worldrushes past. Your survival cruciallydepends on properly analysing the

image flow across your eyes which isessential to coordinate your escape.How does your visual processingadjust to the situation? Recentrecordings of visual neurons in awake,behaving animals have shown that

locomotion can alter the gain andvelocity tuning of central visual neurons[1–7]. A new study [8] now extendsthese exciting findings by identifyingbiophysical and computationalmechanisms which suggest how thelocomotor state changes the gain, setpoint and velocity tuning of neuronsin the fly motion vision pathway.The fly’s flight and your pursuit

share at least three issues that yourrespective visual systems must copewith. First, the range of neuralresponses can be matched to thedynamic range of the stimuli [9]. This

A

LP HP

×

Eyefacet 1

Eyefacet 2

10−1 100 101

Temporal frequency (Hz)

1

0

Res

pons

e (a

.u.)

10-1 100 101

DecreaseLP

Increase LP

Increase HP

DecreaseHP

B C

H1 cell output

HP LP

×

Current Biology

Figure 1. The velocity tuning of the Reichardt detector model [19].

(A) Diagram of the Reichardt model used by Jung et al. [8]. The model has high pass (HP), lowpass (LP), multiplication (3), and subtraction stages (–). (B) An increase in the time constant ofthe high pass filter, tHP, increases the response of the model to low velocities. The Reichardtmodel is tuned to the temporal frequency of the stimulus, so the velocity is shown as temporalfrequency. All curves are plotted using the equations and parameters of Jung et al. [8], normal-ised to the non-flight responses in black [8]: tHP = 170 ms, tLP = 69 ms. Red: tHP = 290 ms;grey: tHP = 100. (C) An increase in the time constant of the low pass filter, tLP, increasesthe response of the model to high velocities. Black: non-flight responses as in (B): tHP =170 ms, tLP = 69 ms. Red: tLP = 30 ms; grey: tLP = 140 ms.

DispatchR651

means setting the gain and set pointappropriately to avoid saturating thesystem, and to make full use of theavailable signal bandwidth. Second,the speed of the visual processingmustbe fast enough to allow for smooth andefficient motor actions. Finally, andperhapsmost importantly, energymustnot be wasted on unnecessary neuralsignalling. Neural activity can consumea substantial proportion of the restingenergy budget, at around an estimated20% in humans [10]. During yourescape, energy is at a premium. On theone hand, the high gain states that mayoptimise the signalling bandwidth andspeed will use up valuable energyresources [10], but on the other hand,you cannot afford the fatalconsequences of an uncoordinatedescape.

Given these constraints, howshould the visual system adjust itsperformance while the animal ismoving? Pioneering work in the locustdemonstrated that flyingmodulates theresponsiveness of a visual interneuronthat helps the animal to avoidcollisions — a real problem forswarming locusts — and to escapeits predators [5,11]. Now, technicaladvances in recording neural activityhave renewed interest in the questionof how neural processing andlocomotor state are linked together inother behaving animals [7,12,13]. Inparticular, rapid progress has beenmade in the well-characterised visualsystem of the fly.

Recent studies have shown howlocomotion modulates the propertiesof a population of identified neurons inthe fly’s visual pathway, the lobulaplate tangential cells. These neuronsare named after the part of the fly’sbrain they are found in, the lobula plate,and they play a fundamental role inthe analysis of optic flow [14]. So far, ithas been established that locomotionalters the gain of their directionalresponses [1,2], the set point of thecells [1,2], and the velocity tuning inwalking flies [3]. Meanwhile, studiesin rodents have shown that walkingand running increase the gain of theprincipal excitatory cells of the primaryvisual cortex [6,7], the origin of whichappears to be dependent on gainchanges in the upstream visualpathways providing input to the cells.

Of course, it has long beenunderstood that visual processing istightly coupled to motor activity [15].What is exciting about these more

recent discoveries is that the visualsystem appears to up-regulate itsactivity during movement in astate-dependent manner, as thechanges seem not to be tightly coupledto a specific movement. Secondly, thediscovery of state-dependent visualgain-modulation in rodents and flies atthe same time indicates that generalprinciples of visual processing mayapply across different phyla.

Jung et al. [8] needed to performsome very technically challengingexperiments to establish their keyresult, that flight alters the velocitytuning of a lobula plate tangential cell,the blowfly H1 cell. An integral part ofthe fly’s flight mechanism is theoscillation of the thorax at the wingbeat frequency of 150 Hz, generatingmechanical vibrations throughoutthe whole preparation. The scientistsfrom the Max-Planck-Institute ofNeurobiology chose a long, thinrecording electrode which allowedthe head and electrode to move inunison sufficiently to maintain stablerecordings, an elegant solution to adelicate problem.

With the setup perfected, they foundthat flying broadens the velocity tuningof the H1 cell, and shifts the peak tohigher velocities. This result makesintuitive sense, as the animal willexperience a broader range and ahigher mean velocity compared tositting still. What are the mechanismsresponsible? The pioneering studiesin the locust had established that

flight was associated with the releaseof octopamine, the invertebratehomologue of norepinephrine.Octopamine alters many aspectsof the animal’s physiology duringfight-or-flight situations, including themetabolism, muscle tone and sensoryprocessing. Later studies could showit was the action of octopamine thatmodulated the sensitivity of the locustvisual interneurons used to avoidcollisions and triggering escapejumps [5,16]. Octopamine agonistscan also increase the gain state andalter the velocity tuning of lobulaplate tangential cells [17,18], so couldoctopamine agonists reproduce thechange in velocity tuning inducedby flight?Jung et al. [8] found that applying the

octopamine agonist chlordimeformwas able to broaden the velocity tuningand increase the sensitivity to highervelocities in away that was qualitativelysimilar to the effects of flight. They thenmodelled the response of the cellusing the Reichardt detector [19] — awell-established model that underliesthe detection of directional motion inmany insects and accounts for severalresponse properties of the lobula platetangential cells. Their implementationof the model involves comparing thebrightness at two locations in the eye,after the signals have been high-passfiltered down one channel andlow-pass filtered along the other(Figure 1A). Changing the timeconstant of the low pass filter tunes the

Current Biology Vol 21 No 17R652

model to the higher image velocities:the smaller the time constant, thehigher the velocity the model is tunedto (Figure 1B,C). Both flight andoctopamine agonist substantiallyreduced its value, consistent withthe observed shifts in the H1 cell’svelocity tuning [8].

An exciting parallel development isthat the circuitry believed to generatethe motion inputs corresponding to theReichardt detector are, for the firsttime, becoming accessible to detailedstudies, thanks to developments ingenetics in the fruitfly, Drosophila [20].Future studies will now be able to buildon Jung et al.’s results to identify themechanisms involved in detail. Alreadywe know from intracellular recordingsthat locomotion alters the propertiesof the lobula plate tangential cellsthemselves, as well as the propertiesof their motion inputs [1,2]. It is certainthat more signalling pathways thanthose using octopamine are involved,but how and where remains a mystery.Tying down the functional motivationfor state-dependent vision in anymodel organism remains a bigchallenge, but based on currentprogress, work on the fly looks likelyto succeed.

References1. Maimon, G., Straw, A.D., and Dickinson, M.H.

(2010). Active flight increases the gain of visualmotion processing in Drosophila. Nat.Neurosci. 13, 393–399.

2. Rosner, R., Egelhaaf, M., and Warzecha, A.K.(2010). Behavioural state affectsmotion-sensitive neurones in the fly visualsystem. J. Exp. Biol. 213, 331–338.

3. Chiappe, E.M., Seelig, J.D., Reiser, M.B., andJayaraman, V. (2010). Walking modulatesspeed sensitivity in Drosophila motion vision.Curr. Biol. 20, 1470–1475.

4. Haag, J., Wertz, A., and Borst, A. (2010). Centralgating of fly optomotor response. Proc. Natl.Acad. Sci. USA. 107, 20104–20109.

5. Rind, F.C., Santer, R.D., andWright, G.A. (2008).Arousal facilitates collision avoidance mediatedby a looming sensitive visual neuron in a flyinglocust. J. Neurophysiol. 100, 670–680.

6. Niell, C.M., and Stryker, M.P. (2010). Modulationof visual responses bybehavioral state inmousevisual cortex. Neuron 65, 472–479.

7. Szuts, T.A., Fadeyev, V., Kachiguine, S., Sher, A.,Grivich, M.V., Agrochao, M., Hottowy, P.,Dabrowski, W., Lubenov, E.V., Siapas, A.G.,et al. (2011). Nat. Neurosci. 14, 263–269.

8. Jung, S.N., Borst, A., and Haag, J. (2011).Flight activity alters velocity tuning of flymotion-sensitive neurons. J. Neurosci. 31,9231–9237.

9. Laughlin, S.B. (1981). A simple codingprocedure enhances a neuron’s informationcapacity. Z. Naturforsch. C 36, 910–912.

10. Laughlin, S.B. (2001). Energy as a constrainton the coding and processing of sensoryinformation. Curr. Opin. Neurobiol. 11, 475–480.

11. Rowell, C.H.F. (1971). Variable responsivenessof a visual interneurone in the free-movinglocust and its relation to behaviour and arousal.J. Exp. Biol. 55, 727–747.

12. Seelig, J.D., Chiappe, M.E., Lott, G.K., Dutta, A.,Osborne, J.E., Reiser, M.B., and Jayaraman, V.(2010). Two-photon calcium imaging from

head-fixed Drosophila during optomotorwalking behavior. Nat. Methods 7, 535–540.

13. Naumann, E.A., Kampff, A.R., Prober, D.A.,Schier, A.F., and Engert, F. (2010).Monitoring neural activity with bioluminescenceduring natural behavior. Nat. Neurosci. 13,513–520.

14. Krapp, H.G., and Wicklein, M. (2008). Centralprocessing of visual information in insects. InThe Senses: A Comprehensive Reference,A.I. Basbaum, A. Kaneko, G.M. Shepherd, andG. Westheimer, eds. (San Diego: AcademicPress), pp. 131–204.

15. Land, M.F. (2006). Eye movements and thecontrol of actions in everyday life. Prog. Retin.Eye Res. 25, 296–324.

16. Bacon, J.P., Thompson, K.S., and Stern, M.(1995). Identified octopaminergic neuronsprovide an arousal mechanism in the locustbrain. J. Neurophysiol. 74, 2739–2743.

17. Longden, K.D., and Krapp, H.G. (2009).State-dependent performance of optic-flowprocessing interneurons. J. Neurophysiol. 102,3606–3618.

18. Longden, K.D., and Krapp, H.G. (2010).Octopaminergic modulation of temporalfrequency coding in an identified opticflow-processing interneuron. Front. Syst.Neurosci. 4, 153.

19. Hassenstein, B., and Reichardt, W. (1953). DerSchluss von Reiz-Reaktions-Funktionen aufSystem-Strukturen. Z.Naturforsch. B. 8, 518–524.

20. Joesch, M., Schnell, B., Raghu, S.V., Reiff, D.F.,and Borst, A. (2010). ON and OFF pathways inDrosophila motion vision. Nature 468, 300–304.

Department of Bioengineering, ImperialCollege London, South Kensington Campus,London, SW7 2AZ, UK.E-mail: h.g.krapp@imperial.ac.uk

DOI: 10.1016/j.cub.2011.07.016

The Circe Principle: Are PollinatorsWaylaid by Attractive Habitats?

How do pollinators move across fragmented landscapes? Attractive habitatshave been viewed as facilitating pollinator movement; however, they mayactually be distracting the pollinators.

Ignasi Bartomeusand Rachael Winfree

In order to understandvector-mediated ecological processes,we need to know how vector speciesmove across landscapes. This isespecially challenging when the vectorspecies is an insect. Nevertheless, itis critical to understand the movementpatterns of key insect functionalgroups — such as pollinators, whichfacilitate the reproduction of most ofthe world’s plant species [1]. Severalapproaches have been used thus farto measure pollinator movement, butknowledge of how pollinators connectplants at the landscape scale remainselusive. The fundamental problem is

that large-scale approaches, whichcan inform us about how pollinatorsmove among habitats, generallydon’t provide information on whichindividual plants are pollinated;whereas smaller-scale approachesthat can measure pollinator movementamong plants aren’t feasible at thelandscape scale. For example,capture–recapture methods can tellus how pollinators move amonghabitat types [2], but not about whichplants are pollinated. Conversely,fluorescent dye techniques canidentify the individual plants visitedby pollinators [3], but such methodsare generally not feasible forlandscape-scale questions (but see [4]for an exception). Encouragingly,

recent technological innovations havemade direct tracking of pollinatorspossible. However, direct trackingis still limited to species largeenough to carry transmitters [5], or tospecies that move within reasonablyopen areas [6]. Perhaps the mostpromising technique estimatespollinator movements indirectly byusing genetic methods on obligatoryanimal-pollinated plants [7]. Thesekinds of data are becoming easierand cheaper to obtain, and promiseto greatly enhance the understandingof pollinator movement patterns atthe landscape scale.In a recent issue of Current Biology,

Lander et al. [8] show that by mappingall of the individual trees in onepopulation of a forest tree species anddoing a paternity analysis, they cantrack pollination events between trees.The novel finding of the paper has to dowith how habitat types in the largerlandscape affect pollinatormovements. The researchers use thedata on pollination events, derived fromthe paternity analysis, to characterize