Cell Host & MicrobeArticlePlasmodium Sporozoite Motility Is Modulatedby the Turnover of Discrete Adhesion SitesSylvia Mu nter,1,6Benedikt Sabass,2,6Christine Selhuber-Unkel,4,6Mikhail Kudryashev,1Stephan Hegge,1Ulrike Engel,3Joachim P. Spatz,4Kai Matuschewski,1Ulrich S. Schwarz,2,5,* and Friedrich Frischknecht1,*1Parasitology, Department of Infectious Diseases, University of Heidelberg Medical School, Im Neuenheimer Feld 3242Bioquant, Im Neuenheimer Feld 2673Bioquant, Nikon Imaging Center, Im Neuenheimer Feld 267University of Heidelberg, 69120 Heidelberg, Germany4Max Planck Institute for Metals Research, Department of New Materials and Biosystems and University of Heidelberg,Department of Biophysical Chemistry, Heisenbergstr. 3, 70569 Stuttgart, Germany5Institute of Zoology, Theoretical Biophysics Group, University of Karlsruhe, Kaiserstrasse 12, 76131 Karlsruhe, Germany6These authors contributed equally to this work*Correspondence: ulrich.schwarz@bioquant.uni-heidelberg.de (U.S.S.), freddy.frischknecht@med.uni-heidelberg.de (F.F.)DOI 10.1016/j.chom.2009.11.007SUMMARYSporozoites are the highly motile stages of themalariaparasiteinjectedintothehostsskinduringamosquitobite. Inorder tonavigateinsideof thehost, sporozoites rely onactin-dependent glidingmotility. Although the major components of thegliding machinery are known, the spatiotemporaldynamics of the proteins and the underlying mecha-nismpoweringforwardlocomotionremainunclear.Here, we show that sporozoite motility is character-ized by a continuous sequence of stick-and-slipphases. Reection interference contrast and tractionforce microscopy identied the repeated turnover ofdiscrete adhesion sites as the underlying mechanismof this substrate-dependent type of motility.Transient forcescorrelatedwiththeformationandruptureofdistinctsubstratecontactsitesandweredependent on actin dynamics. Further, we showthat theessential sporozoitesurfaceproteinTRAPiscriticalfortheregulatedformationandruptureofadhesion sites but is dispensable for retrogradecapping.INTRODUCTIONCell migrationisacomplexandhighlyregulatedprocessthatplays a crucial role in a large range of physiological anddisease-related situations. Eukaryotic cell motility is dependenton the concerted action of actin and myosin and is often initiatedbyactin-drivenextensionsoftheforward(leading) membrane,followed by the establishment of newcontact sites to thesubstrate and the release of contacts at the rear end. Quantita-tive microscopy studies of migrating mammalian cells haverevealedmanydetailsabouttheroleof actinandmyosinandits relation to the continuous formation and turnover ofintegrin-mediatedadhesionsites (Vicente-Manzanares et al.,2009; Wang, 2007). In contrast, little is known about theseprocesses of protozoa, including the medically importantapicomplexan parasites that feature motile invasive stagessuch as Plasmodium or Toxoplasma (Heintzelman, 2006). Theseparasitesconstitutehighlypolarizedandsimple-shapedcellsthat show limited types of motile behavior (Frixione et al., 1996;Vanderberg, 1974). Their mainmodeof activelocomotionisactomyosin-dependent gliding motility that is important formigration across tissue barriers and host cell invasion(Heintzelman, 2006).Themaincomponentsthatconstitutethecoreoftheglidingmachineryareknown(Daher andSoldati-Favre, 2009; Fre nalandSoldati-Favre, 2009; Schu ler andMatuschewski, 2006),but the spatiotemporal dynamics of the individual proteins insidethe complex and, hence, the underlying mechanism that powersforward locomotion still remain unclear. Studies on malaria para-sites and the related apicomplexan parasite Toxoplasma gondiirevealedthat themyosinmotorproteinsareanchoredintoanalveolate-specic membrane system, the inner membranecomplex(IMC), andthat short actinlamentsmight belinkedtoplasmamembraneproteins(Baumet al., 2008; Daher andSoldati-Favre, 2009). The motility machinery of malaria parasitesis important for the merozoite and ookinete stages to invade redblood cells of the host and traverse the midgut epithelium of themosquito vector, respectively (Cowman and Crabb, 2006;Vlachouet al., 2006). Athirdinvasivestage, thePlasmodiumsporozoite, formsinoocystsat themosquitomidgut wall, isreleased into the hemolymph, and penetrates the salivary glands(Matuschewski, 2006). Sporozoitesisolatedfromthesalivaryglands of an infected mosquito move on solid substrates alongcircular paths, without changingtheir shape, at an averagespeedof12mm/s(Vanderberg,1974).Withinthecavitiesandducts of salivary glands, only fewsporozoites move at lowvelocities(Frischknecht et al., 2004). However, onceinjectedinto the skin of a host, sporozoites move on seemingly randompathsatsimilarspeedsasinvitroandeventuallyinvadebloodorlymph vessels(Amino et al., 2006). Sporozoitesthat enteredthebloodcanfurthermoremoveontheendotheliumof bloodvessels and invade hepatocytes, where they differentiateintothousandsof redbloodcell-infectingmerozoites(Frevertet al., 2005). ThemechanismsunderlyingglidingmotilityareCell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc. 551therefore essential for the progression through the malarialifecycle.In marked contrast to higher eukaryotic cells, actin laments ofPlasmodium parasites are elusive in vivo. Until now, no actin la-ments could be revealed in living or xed malaria parasites, andpuried parasite actin only polymerizes into laments of less than100 nm length in vitro (Sahoo et al., 2006; Schmitz et al., 2005;Schu leretal.,2005).AhighturnoverofshortactinlamentsinPlasmodiumparasitesmightbenecessarytoachievethehighspeedofsporozoites.TheactinlamentsofthesporozoitearethoughttobelinkedviatheglycolyticF-actin-bindingenzymealdolase to the substrate by the thrombospondin-related anony-mousprotein(TRAP) (Baum et al., 2008; Morahan et al., 2009).TRAP, like its orthologs in other apicomplexan parasites and ma-laria parasite stages, is stored in specic organelles at the apical(front) tip of the parasites and secreted during cell motility(Morahanet al., 2009). Inanalogy toactin-basedmotility ofhigher eukaryotic cells, it was predicted that sporozoites estab-lish contacts to the substratumover which the actomyosin motorpushes the parasite forward (Baum et al., 2008; Me nard, 2001).Eventually, TRAPiscleavedbyspecializedproteasesat theposterior (rear) endof theparasitesandisreleasedintothemembraneandprotein-richtrail left behindmovingparasites(Bakeretal.,2006;Dowseetal.,2008).Thus,thesequenceofsecretion,adhesionformation,andcleavageisthoughttoleadto continuous gliding.Here, weusedreectioninterferencecontrast microscopy(RICM) and traction force microscopy (TFM) to show that sporo-zoite motility is limitedby the dynamic turnover of discreteadhesionsites, whichleadstoastick-and-slipmovement ofthe cell.RESULTSMotility and Adhesion Pattern of Sporozoites In Vivoand In VitroInvivoimagingat lowframerates(0.5Hz) hasrevealedthatsporozoitesmoveirregularlywithintissues(Aminoetal.,2006;Frischknechtet al., 2004). Toresolve parasitemotilityat highertemporal resolution, we rst imaged sporozoites in salivarycanals, in the skin, and on cultured hepatocytes at 2 Hz(Figure 1A). Speedplots revealedthat sporozoites move ina stop-and-go fashion, with periods of rapid movement followedby periods of slow movement. This irregularity observed in vivoandonculturedcellscouldbecausedbybothinternal andexternal factors. We therefore imagedsporozoite glidingonglasssurfacesandfoundsimilar motionpatterns(Figure1B),suggestingthat internal processescontributetotheirregulartrajectories. No difference in the overall speed was foundbetweensporozoitesmovingonglass, oncellsor intheskin(Figure1CandseealsoExperimental Procedures).Quantica-tionof thetracksfromparasitesglidinginsideof thedifferenttissues and in vitro revealed that, during periods of rapid move-ment, sporozoites traveled up to one parasite length (Figure 1D).In order to investigate whether, as in mammalian cells,discrete adhesion sites are formed and possibly modulatemotility, weemployedreectioninterferencecontrast micros-copy (RICM) (Ra dler andSackmann, 1992; Senguptaet al.,2006) (Figure S1A available online). For the initial stage ofadhesion, RICM clearly showed that the sporozoite rst adhereswith one end, then with the second end, and nally with the cellbody(FigureS1B). Importantly, RICMrevealedanonuniformdistributionofadhesionsitesonthesurfaceoftheparasite,asdenedbydarkareasalongthecell (Figures2AandS1CS1Eand Movie S1). The absolute intensity level of the reected lightdiffered fromone sporozoite to the other, with parasitesshowinglargeradhesionzonesmovingataslowerspeed(Fig-ure 2B and Movie S1). Analysis of more than 100 motile sporo-zoitesshowedthat,usually,asmall adhesionsitewaspresentat the front and one or two adhesion sites at the rear of the spor-ozite(Figures2Aand2C).Adhesion Turnover Rate Correlates with OverallSporozoite SpeedThe occurrence of discrete sites of adhesion and their dynamicsduring motility suggest that parasite motility might functiondifferently than previously anticipated (Baumet al., 2008;Me nard, 2001). RICM revealed the turnover of discrete adhesionsites to be linked to changes in gliding speed, showing a thrustwhen the rear end of the parasite detached from the substrate.Moreover, theformationof anewadhesionsitewasfoundtoslowdownmotility(Figures2C,2D,andS2).Asslowlymovingsporozoitesshowedfewerchangesintheiradhesionpatterns,they underwent fewer cycles of speed changes per time (Figures2E and S3). Examination of the dynamics of readily distinguish-ableadhesionsitesat thefront andrear of 22independentsporozoitesshowedalinearcorrelationbetweenadhesionsiteturnover and speed (Figure 2F), conrming that speed is stronglydeterminedbyadhesion. Importantly, thefrequencyof speedpeaks (periods of fast movement; seeFigureS4) correlatedwiththefrequencyof observedadhesioncycles(Figure2F).Speed peaks could be dened and automatically detected(Figure S4). This enabled an alternative calculation of therunlengthduringaspeedpeak, whichessentiallyyieldedthesame result as the wavelet analysis: sporozoites typically travellessthantheirownlengthduringoneperiodoffastmovement(Figures1DandS4F). Wealsonotedthat sporozoitesusuallydidnotchangetheiroverall averagespeedovertime(FiguresS3DandS3E). Eventhosesporozoitesthat interruptedtheirmotilityfor sometimeusuallycontinuedglidingat thesamespeedasbefore. All examinedsporozoitesmovedinastop-and-go fashion, and even slow-moving ones relied on adhesionturnover for short periods of rapid movement.We next mapped the crescent shapes of the sporozoites ontolinearones,whichallowedustomeasuretemporal changesintheintensitydistributionofthereectedlightalongthecentralaxisof thecell body. Theresultingkymographsrevealedthattheadhesionsitesattheveryfrontandrearmovedalongwiththe parasite during locomotion. In contrast, the larger adhesionsitesinthecentral part showmorecomplexdynamics, witha mixed pattern of staying xed with the parasite and undergoingretrograde movement (Figure 3). Our data suggest that parasitemotility can be described in a rst approximation as analternating sequence of periods of run and rm adhesion. Mostimportantly, resting time dominates the average velocity,suggestinga functional relevance of the observedstick-slipfrequencies (Supplemental Text).Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite Gliding552 Cell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc.TractionForce Appliedby theParasites ontheSubstrateOur observations of the adhesive dynamics suggest that sporo-zoites generate variable forces at discrete adhesion sites. To testthis prediction, we used traction force microscopy (TFM) on softelasticsubstrateswithembeddedmarker beads(DemboandWang, 1999; Sabasset al., 2008) (Figures4A, S5A, andS5BandMovieS2). Imageprocessingandcomputational recon-structionofthecellulartractionpatternofsporozoitesshowedthat,duringcontinuousgliding,thecentral partoftheparasiteexerted forces perpendicular to its direction of movement(Figures4Aand4BandMoviesS3andS4). Strikingly, forcespointingalongthedirectionof movement werepreferentiallylocalizedat bothends, wherethelong-livedadhesionswereobservedinRICM(Figure2A).Statistical analysisofsixglidingsporozoites shows that forces in the center are larger than forcesproduced at the front or rear (Figure 4C). These data suggest thatonehastodistinguishbetweenperpendicularforcesalongthecell bodyandlongitudinal forcesat theends(Figure4D, left).Forcesperpendiculartothecell axiscouldbepartiallyduetononspecic adhesionbetweenthe elasticsubstrate andthesporozoite, as predicted by contact mechanics of elastic bodies(Johnson, 1985) (Figure 4D, right).Intriguingly, analysis of ten independent parasites showed thatstrongforcesattherearwereinvariablylinkedtotimeframesduringwhichsporozoitesbecameimmobilizedwiththeir rearbut continued to pull forward with their front, therefore stretchingthe parasite (Figures 4E and 4F). These apparent stalling forcespointed into the direction of movement. At the front of the para-site, similar forces are transmitted independently of the forwardspeed of the parasite (Figure 4F). In contrast, the forcesFigure 1. Motility Pattern of Sporozoites In Vivo and In Vitro(A) Speed plot and uorescence images of sporozoites tracked inside of the salivary duct, inside of the skin, and on cultured hepatocytes. Three consecutiveuorescence images and the maximum projection of the uorescence intensity from an entire time-lapse sequence (green) are shown; for salivary duct and livercells, the transmission image was overlayed. Numbers indicate time in seconds. Scale bars, 10 mm.(B) Speed plots and uorescence images of sporozoites tracked on glass cover slides. Scale bar, 10 mm.(C) Plot of instantaneous speeds fromparasites gliding in the different environments. Boxes contain 50%of data distributed around its median (red bar). Whiskersshow the range of data. Sporozoites in the skin, which migrate within one focal plane, are labeled with Skin 2D. This distinction is necessary to avoid trackingerrors (see Experimental Procedures). Numbers above the bars indicate the number of tracked sporozoites. For each sporozoite, the instantaneous speed wasdetermined at 2 Hz over 50 to 300 frames.(D) Global wavelet power spectrum (see Experimental Procedures) showing a peak (arrows) at the typical length that sporozoites move during one speed peak(runlength). Intuitively, this method can be pictured as tting half of a period of a sine into each speed peak and thus nding the average width of the peaks (smallinset). Abscissa: half of the center period of the wavelets. Ordinate: global power spectrum normalized to maximum.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite GlidingCell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc. 553transmittedbythesameparasiteat thecenterandrearendswerelargerduringslowermovement(sticking), supportingthehypothesisthat theadhesionsitesat therearendslowdownmovement (Figure 4F).Correlation of Force Measurements with AdhesionSite TurnoverThe incompatibility of a combined RICMandTFManalysisprecludesdirect comparisonsof adhesionsitedynamicswithlocal forces.However,themovementoftheparasitesonglassandtheexiblesubstrateishighlysimilar,andsomeparasitesdo repeatedly get immobilized at the rear end during the contin-uouscircularmotilityonbothsubstrates. Therefore, wecouldreadilyndstickingparasitesusingRICM(MovieS5).Similarlytoparasitesmovingonglassor elasticgels, theseparasiteschangedtheir curvaturefrom0.24mm1to0.17mm1duringstretching, while their length stayed constant. Analysis ofstretchingparasitesbyRICMrevealedthat, duringstretching,the parasites built a newadhesion site at their center (Figure 4G).Duringsticking,thefrontoftheparasitemovedfasterthantherear, andonlyafter releaseof therear, itsspeedsurpassedFigure 2. Turnover of Discrete AdhesionSites Results in Irregular Sporozoite Motility(A) Time-lapseRICMimagesof oneslow(Spz1)and one fast (Spz2) gliding sporozoite. Mostadhesionsitesstayxed(arrowheads)inrespecttothemovingsporozoite.Numbersindicatetimein seconds. Scale bar, 10 mm. See also Movie S1.(B) Analysisof thesporozoitesin(A). Theslow-moving Spz1 (speed plot; top) has largeradhesions (darker average RICM intensities;bottom).(C) A sporozoite accelerates upon disassembly ofadhesion sites at the front and rear. Red line indi-cates tip of the sporozoite in the rst frame.(D) Speed plot from the frames shown in (C). Highspeed correlates with lighter areas in RICM,reecting less adhesion.(E)Theinstantaneousspeedplottedagainsttimeof thetwoparasitesshownin(A) (Spz1inred;Spz2 in blue). Arrowheads indicate speed peaks.(F) Thefrequencyof adhesioncyclesat front orrearisproportional totheaveragespeedaswellas tothefrequency of speedpeaks of glidingparasites (R2= 0.84 and 0.77, respectively).Twenty-two sporozoites were examined for100200 frames each.briey the speed of the front (Figure 4H).Interestingly, the main traction forceduring stretching originatednot at theadhesionsiteat thefront end, but wasdistributedtotheadhesionsitetowardthecenter of thesporozoite, indicatingaatteningof theparasite(Figures4Eand 4F). Every stretching event wentalong with a subsequent thrust in speed,suggesting that the presence of the rear-wardadhesionisthemain restrictionformovement (Figures 4H and S5C andMovieS5). Measuringthefront-to-rear distanceof aparasiteandanalyzingthespeedpeakstogether withtheintensityofthe rear-end adhesion, as well as the stretching of the parasite,showedthat thesethreeparameters arehighly coordinated.Decrease in adhesion, which is revealed as an increase in pixelintensity, strongly correlateswith an increase in speed (Figures5Aand5B).Quantitativeanalysisconrmedthatweakeningofthe adhesion at the rear end positively correlates with the speed,and in accordance to this, the stretching shows a negative corre-lationwiththespeedoftheparasites(Figure5C).Accordingly,the traveled path of the rear end is shorter after a sticking periodof the parasite (Figure 5D). Therefore, the TFMdata together withRICMshowedthatsporozoitemotilityresultsfromactiveforcegeneration transmitted through distinct adhesion sites.Inuence of Actin-Interfering Drugs on Parasite MotilityThe irregular motion patterns most likely result fromthe interplayoftheforce-generating(actomyosin)and-transmitting(surfaceadhesion receptors) machinery. We therefore investigatedsporozoites gliding on glass under a range of concentrations ofjasplakinolide(Jas), whichinhibitsactindepolymerization, andcytochalasin D (CytoD), which inhibits actin polymerizationCell Host & MicrobeAdhesion Dynamics during Malaria Parasite Gliding554 Cell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc.(Cooper,1987;Cramer,1999).Averageandmaximumspeedswere highest at low concentrations of Jas. Both speedsdecreasedwithincreasingconcentrationsfor JasandCytoD(Figures6AandS6A). Parasitesglidingonelasticsubstratesunder50nMCytoDthatmovedwithhalfoftheaveragespeedof controls showed traction forces of about half of control forces(Figures6Band6C). Curiously, parasitesmovingat thesameaveragespeedascontrol sporozoitesbut under 100nMJasshowedforcesof onlyabout aquarter of thosefromcontrols(Figures6Band6C). Imagingsporozoitesmovingat 25%ofcontrol speed in 100 nM CytoD with RICM showed an increasedadhesion turnover rate. In contrast, sporozoites moving at 50%of control speedunder 200nMJasshowedalower rateofadhesionturnover,withthefrontandrearendsstayinglongerincontactwiththesubstrate(Figures6DandS6BandMoviesS6 and S7). This indicates that increased F-actin weakensadhesion strengths and turnover. As expected, at high concen-trations of bothinhibitors, sporozoites stoppedmoving, andturnover of adhesionsites droppeddramatically (Figure6E).However, whenasuddenhydrodynamicowwasappliedtosporozoites, they remained better attached under high concen-trations of CytoD than in the absence of drugs, and they readilydetached under high concentrations of Jas (Figures 6F and S7),conrming that more F-actin leads to weaker adhesion.The Role of TRAP on the Adhesion Turnover duringGliding MotilitySporozoite adhesion is mediated through different surfaceproteins, includingTRAP. Moreover, it hasbeenshownthatTRAPformsalinkbetweentheextracellularsubstrateandtheactin cytoskeleton (Morahan et al., 2009). trap() sporozoites stillforminside of the mosquito midgut and circulate in thehemolymphbutcannolongerenterthesalivaryglands(Sultanet al., 1997). In order to investigate the role of TRAP in sporozoitelocomotion,weimagedtrap()andwild-type(WT)sporozoitesisolatedfromthe mosquitohemolymph. As expected, mostattachedwild-typeparasiteswereglidingincircles, andthemajority of attached trap() sporozoites were not moving.However, trap() parasites were still able to build adhesion sites(Figure 7A). In addition, both trap() and WT parasites from thehemolymph showed a new type of motility that we termed patchgliding(Figures7BandS8). Asimilar typeof motility, termedpendulumgliding,hasbeendescribedearlierforparasitesex-pressingonlyatruncatedormutatedcytoplasmictail ofTRAP(Kappeetal.,1999).Accordingtothisreport,mutantparasitesappeartoglidefullyattachedforonebodylength,followedbyanarrest for 12sandbackwardmovement alongthesameaxis. Inmarkedcontrast, patchglidingsporozoitesfromthehemolymphofWTortrap() parasitesdonotstop,canmoveatspeedsexceedingthoseofcircularglidingsporozoites,anddescribe noncircular trajectories. Using RICMand fast DICimaging,wefoundthatpatchglidingsporozoitescontinuouslymoveoverasinglespotinaback-and-forthmanneratsimilarspeeds in both directions (Movie S8). We also observed trap()parasitesthat wereattachedtothesubstrateat oneendtotranslocate particles in both anterograde and retrogradedirection with indistinguishable speeds (Figure 7C). Both particlemovements and patch gliding were abolished with CytoD (MovieS9). trap() sporozoites could undergo patch gliding forextendedperiodsoftime(several minutes)withoutcompletingadhesion. However, once they formed a second adhesion,trap()sporozoiteswerenolongerabletorupturethiscontactsite(MovieS10). Comparingthebehavior of WTwithtrap()parasitesconrmedthattrap()parasiteswerenotcapableofexhibitingnormal glidingmotility(Sultanet al., 1997), andthepercentage of oating and waving parasites was higherFigure 3. Generation of Kymographs from Gliding Sporozoites(A) The central axis of the parasite was dened by tting an ellipse.(B) The intensity prole of the parasite along the ellipse (radial average: 6 pixels) was extracted.(C) The intensities along the ellipse were visualized by expanding the intensity prole to a rectangle.(D) The kymograph was generated by stringing together a series of rectangular proles obtained from (C) for subsequent images (t1,t2,.).(E) Kymograph of a fast gliding sporozoite (average speed: 1.8 mm/s) for 70 s. The front and rear adhesions (red and green arrowheads, respectively) stay xedwithrespecttothemovingparasite.Inthecenterofthecell body,amorecomplexscenarioisobserved,withmixedpatternsofstableadhesionsandtheirretrograde motion. Scale bars, 2 s (white) and 2.5 mm (yellow).(F and G) (F, left) Kymograph of the slow-moving sporozoite (Spz 1 fromFigure 2A) showing only some irregularities in the central adhesions. Again, front and rearadhesions (red and green arrowheads, respectively) stay xed with respect to the moving parasite. (F, right) Kymograph of the fast-moving sporozoite (Spz 2 fromFigure 2A). The light-green arrowhead indicates an adhesive region moving rearward at a higher speed than the sporozoite forward movement. Scale bars, 2 s(white)and2.5mm(yellow).(G)Cartoonshowingthatthefasterrearwardmovementcanbeexplainedbyazipper-likeadhesionseeminglymovingathigherinverse speed than the parasite as the respective parts of the sporozoite approach the substrate.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite GlidingCell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc. 555Figure 4. Traction Forces Correlate with Adhesion Dynamics during Sporozoite Gliding(A) (Left) Sporozoiteonaexiblepolyacrylamidegel containinguorescentredandfar-red(pseudocoloredinblue) markerbeads. Thearrowsshowbeaddisplacements. (Right) Traction vectors (red) are reconstructed from displacement data. See also Movies S2S4.(B) Pseudocolored traction force maps (Pa, Pascal). Numbers indicate time in seconds. Bar in force scale indicates noise (2s). See also Movie S3.(C) Comparing forces at the front end, the central region, and the rear end of six gliding sporozoites (each fromabout 50 frames). Nonoverlapping notches indicatethat distributions differ at a signicance level below 5%, as also determined by a Kruskal-Wallis test.(D) Cartoon showing the directionality of traction (force vectors) in gliding sporozoites. Central forces might result mainly from contact adhesion.(E) Strong forces are detected at the rear and at the center during sporozoite stretching. The white line indicates the front (red arrowhead) to rear (green arrow-head) distance (Pa, Pascal).(F) Bootstrap analysis of the force distribution of ten parasites during slipping and stretching. During sticking, large forces appear at the rear. Bar corresponds to95% condence of Bootstrap analysis.(G) Adhesion dynamics and speed changes during sporozoite stretching as visualized with RICM. A sporozoite getting stuck twice at the rear (green arrowheadsand green line) stretches before releasing the rear and regaining its crescent shape. The white line indicates the front-to-rear distance. See also Movie S5.(H) The graph shows that stretching increases until release of the rear leads to a speed peak.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite Gliding556 Cell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc.compared to WT parasites (Figure 7D). These observations showthat TRAPis important, but not essential, for initial parasite adhe-sion.TheyalsoshowthatTRAPisimportantindetachmentofsporozoiteadhesionsites,suggestingthatTRAPplays amajorrole in coordinating continuous gliding motility through the turn-over of contact sites. Most importantly, TRAPappearstobedispensableforslipping. IncontrasttotheroleofTRAP,actinseemstohavearoleinbothadhesionturnover andslipping(Figure 7E).DISCUSSIONStick-Slip Motility of SporozoitesHere, we showthat Plasmodiumsporozoites move in a stick-slipmanner that is regulated by the formation and turnover of distinctadhesion sites. Our in vitro approach enabled us to use surface-sensitivemicroscopytechniquestoinvestigatetheroleof theadhesion dynamics for gliding motility of malaria parasites.RICMrevealeddistinctadhesionsitesduringinitial sporozoiteadhesion and motility. Surprisingly, the sporozoite did nottranslocateovertheinitial adhesionsiteatthefront,aspostu-latedpreviously(Baumet al., 2008; Me nard, 2001). Instead,sporozoitescouldmoveatslowspeedwithoutchangingfrontandrear adhesionpatternsinrespect tothemovingparasite(Figure 2A). Slow-moving sporozoites could speed up by disen-gaging the front or rear adhesion sites. Strikingly, the frequencyof these events correlated with the overall average speed of theparasites (Figures 2E and 2F). It thus appears that, in addition toclassicgliding, i.e., backwardcappingof adhesionproteins,sporozoitesalsouseadhesionturnover at thefront andrearendstoregulatetheirspeed.Usingtractionforcemicroscopy,we found a large stalling force at the rear adhesion sites,Figure 5. Correlation of Speed with Attachment and Stretching of the Parasites(A) RICM image of a sporozoite. The two red circles indicate the spots on which the parasites were manually tracked at the front and rear ends. The green lineindicates the distance calculated fromthe tracking coordinates of the front and rear. For the intensity analysis, the pixels located in a radius of 7 pixels around thex/y coordinates of the tracked points were summed.(B) Top graph shows the speed of the sporozoite rear end (black curve) and the intensity measured for the rear adhesion spot (red curve). The bottomgraph showsstretching of a different parasite (green curve) in addition to the speed and the intensity of the spot at the rear. Note the almost perfect correlation of stretching withthe increase in speed and distance from the substrate.(C)Quantitativeanalysisandcorrelationoftheeventshighlightedin(B)fromsevenparasites(about200timeframeseach).(Top)Positivecorrelationoftheintensity of the rear-end adhesion (dened in A) with the average speed. (Middle) Negative correlation of the stretching with the average speed of the parasites.(Bottom) The maximumstretch of a parasite is regularly followed by a speed peak. Correlation between stretch and speed displays a clear minimumfor a lag of 0.5s, showing that the speed peak occurs during the shortening rather than after it. Standard deviations of correlations were calculated with the Bootstrap method.(D) For a parasite stuck with the rear end, the traveled path of the rear end (the two adhesion dots at the rear end of the parasites are tracked in red and green,respectively) after rupture is shorter than the traveled path of the front end (light blue dots).Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite GlidingCell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc. 557especially right before rupture of this adhesion. The adhesion siteon the front exerted much less force. Large forces also appearedinthecenteroftheparasite,buttheirregularmotionpatterninthisregionmakesitdifcult tocorrelateforcegenerationwiththeappearanceof adhesionsites(Figure4). Furthermore, theforces at the center of the parasite were directed perpendicularlyto the direction of movement. It is likely that the central forces aremainly nonproductive adhesion forces that are important inkeepingtheparasiteincontact withthesubstrate. Therapidchangesin theforceand speedpatternsafter ruptureof adhe-sion sites suggests that elastic energy is built up during stickingand then released during slipping, similar to stick-slipphenomena in the sliding friction of nonbiological material.TRAP and Actin Dynamics Modulate Adhesion TurnoverApart fromthesemechanical insights, wealsoidentiednewrolesfor actin and TRAP duringgliding. Investigationof mutantsporozoites lacking TRAP as well as WT sporozoites in the pres-ence of actin dynamic-inhibiting molecules revealed a number ofunexpected ndings. Most importantly, TRAP is not essential forthe translocation of the parasite over an adhesion site becausemutant parasites still perform a fast and actin-dependant, albeitnondirectional, slippingmotilitythat wetermedpatchgliding(Figure7). Instead, TRAPappearstoplayanessential roleininitial parasite adhesion and a major role in coordinating regulargliding motility through deadhesion of contact sites. As there aretwo TRAP-like proteins present in Plasmodium sporozoites, TLPandS6(alsonamedTREP) (Combeet al., 2009; Heisset al.,2008; Moreiraet al., 2008; Steinbuechel andMatuschewski,2009),aninterestingquestionishowtheseproteinscontributetobothadhesionmodulationandgliding. Thisissuecouldbeaddressedinthefuturebythegenerationof doubleknockoutparasitesanddomain-swapexperiments. It will befurther ofinterest whether these proteins are colocalized on the sporozoiteFigure 6. Actin-Disrupting Drugs Modulate Traction Force and Adhesion Turnover(A)ParasitespeedonglassatdifferentconcentrationsofCytoDandJas(n>10ineachcondition).Inbothcases,speeddecreasesatsufcientlyhighdrugconcentrations.(B) Speed of selected parasites gliding on the elastic substrate under control and drug conditions.(C) Traction forces generated by the sporozoites shown in (B) decrease for both drug treatments. Parasites under Jas show even lower forces despite unalteredspeed.(D) For each speed value, the frequency of adhesion cycles increases for 100 nMCytoD(blue diamonds) and decreases for 200 nMJas (red circles), respectively.Linear t (black line) shows the control parasites (Figure 2F). See also Movies S6 and S7.(E) Gliding sporozoites exposed to high concentrations of CytoD show extended arrest and enhanced adhesion lifetimes. Red and green arrow and line indicatefront and rear, respectively, of the parasites. Numbers indicate time in seconds.(F) Effect of a disrupting shear ow. (Left graph) Under shear, the percentage of adherent sporozoites increases and decreases if treated with CytoD and Jas,respectively.(Rightgraph) Sporozoitemotionundershearcanbeclassiedasglidingincircles,driftingwiththeow,orarrest.Undershear,thedominanttransition for CytoD is gliding arrest, whereas for Jas, considerably more transitions to drift occur, indicating that adhesion strength is weakened.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite Gliding558 Cell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc.Figure 7. trap() Mutants Are Decient in Deadhesion(A) trap() mutants can build the same adhesion sites compared to WT parasites (one at the front and two at the rear); however, they are not able to glide once thediscrete adhesions at the front and rear are formed. A single frame (top) and the maximum projection (Max) of a movie of 110 s.(B) Sporozoites fromthe hemolymph, including the trap() mutant shown here, often display patch gliding during which the cell moves forth and back over a singleadhesion site (orange line). See also Movie S8.(C) A particle (orange arrowhead) is observed to be translocated in both directions over the trap() sporozoite surface in a way reminiscent of patch gliding. Seealso Movie S9.(D) Classication and quantication of hemolymph sporozoite motility patterns into arrested, waving, patch gliding, gliding in circles, and oating free in solution.(E) Model illustrating that TRAP plays a role in adhesion formation and turnover (deadhesion), whereas actin plays a dual role in adhesion dynamics and slipping.TRAPisdispensableforrapidpatchgliding(slipping)andmightregulateaswitch,whichconvertsadhesiondynamicsintodirectedmotion.Arrowsindicatedirection of slipping. Molecules are marked with matching colors.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite GlidingCell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc. 559surface. Interestingly, TRAP appears not to be uniformlydistributedonthecell surface(Ganttetal.,2000;Kappeetal.,1999).Howthisdistributionrelatestoourndings,however,isunclear. Furthermore, it will beinterestingtoinvestigatehowthe TRAP-cleaving rhomboid protease (Baker et al., 2006;Dowse et al., 2008) modulates adhesion turnover. Actindynamics appeared to play a dual role during sporozoite motility.SurprisinglywefoundthatincreasedlevelsofF-actin,inducedbytheadditionofincreasingconcentrationsofJas,ledtolessadhesiondynamics, weaker adhesionforces, and, eventually,detachment of thesporozoite(Figures6andS7). Thisndingsuggests that elevatedF-actin leads to less traction at thesamespeed. Astheforcesdecreasewithincreasingspeedintheabsenceof drugs(Figure4F), it appearsthat fast-movingsporozoitescontainmoreF-actin. Similarly, it wasshownforT. gondii tachyzoitesthat theseparasitesmovefaster undersmall concentrationsof Jas(Wetzel et al., 2003). Incontrast,the actin lament-disrupting drug CytoD led to increased adhe-sion dynamics. Curiously, despite reduced traction forces in thepresenceof CytoD, moresporozoites stay attachedat highconcentrationsof thedrugunder owconditions(Figure6F).Therefore, traction force seems to peak at an intermediateconcentration of F-actin. We thus postulate that actin lamentsarenot just requiredfor retrogrademovement of theplasmamembraneproteins,butalsoforestablishmentandturnoverofdiscreteadhesionsitesat thefront andrear of theparasites(Figures6Dand7E). Adirectdemonstrationofactinlamentsusingelectronmicroscopywouldclearlybe helpfulbut hasnotyet beenachievedfor intact sporozoites (M.K., unpublisheddata; Lepper et al., 2010).Inconclusion, wesuggestthatactinlamentstogetherwithTRAPlikelyregulatetheturnoverofdiscreteparasiteadhesionsites. MoreF-actincouldleadtoweaker adhesionandlessadhesion dynamics. During motility, TRAP possibly coordinatestheformationof contact sites andthedissociationof thesecontact sites from the substrate. The observed interplay of forcegeneration and adhesion is reminiscent of the way that mamma-lian cells link the actomyosin systemwith integrins to achieve celladhesion. It also raises the possibility that a parasite myosin, aswell as aset of signalingproteins, participateinsporozoiteadhesion dynamics.EXPERIMENTAL PROCEDURESPreparation and Imaging of SporozoitesPlasmodium berghei (strain NK65) sporozoites expressing the greenuorescent protein(GFP) andtrap() parasites (Sultanet al., 1997) wereproducedinAnopheles stephensi mosquitoes andsporozoites harvestedessentially as described in Frischknecht et al. (2004) and Hegge et al. (2009).Imaging inside of the salivary duct was performed as described(Frischknecht et al., 2004) with 3%bovine serumalbumin. For imaging on cells,Huh7 cells (cultivated in RPMI supplemented with 10% fetal calf serum) weretransferred on a glass bottom Petri dish 1 day prior to the experiment, and thesporozoiteswereaddedinRPMI 3%BSA. TheparasiteswerelmedontheHuh7cellsaswellasinbetweenthecellsontheglasssurface,at37

C.The latter parasites served as controls for this experiment (Figure 1). All imageacquisitions were performed on an inverted Axiovert 200M Zeiss microscopeusing a GFPlter set. Images were collected with a Zeiss AxiocamHRmat 2 Hzusing Axiovision 4.6 software and a 253LCI Plan-Neouar objective (NA 0.8).In vivo imaging was essentially performed as described in Amino et al. (2006)usingthePerkinElmer UltraViewspinningdiscconfocal unit onaninvertedNikon TE2000-Emicroscope through a 203objective (PlanFluor multi-immersion NA 0.75) but at a 2 Hz imaging rate.For drugtreatment, isolatedsporozoiteswereplacedina96-well glassbottom plate (Corning, Germany), and the drugs (CyoD and Jas) were addedintheindicatedconcentrations.Imagingwasperformedat216HzalsoonthePerkinElmer spinningdiscconfocal witha203(NikonPlanFluor multi-immersionNA0.75) or1003(NikonPlanApoVCNA1.4) objective. Speedplots represent either average speed over several tens of seconds or instan-taneousspeedwhenatleastoneimagewasrecordedpersecondandthespeedfromonetothenext framewasplotted. For limitationsof accuratespeed measurements, see Hegge et al. (2009).For owmeasurements, sporozoites were placedinanuncoatedowchamber (Ibidi, Germany) andimagedonaninvertedAxiovert 200MusingaGFPlterset at roomtemperature. Imageswerecollectedat 1Hzusinga 103Apoplan objective (NA 0.25). For the perfusion of drugs, a 100 mM stocksolutionin PBSwasusedandmixedinsidetheowchamberduringcontin-uousimageacquisition. Unidirectional owwasappliedbyadding2ml ofmedium to one buffer container of the ow chamber.RICM is a surface-sensitive optical technique that has been frequently usedfor studying the adhesion of vesicles and cells (Sengupta et al., 2006)(Figure S1A). RICM was set up on an inverted Axiovert 200 Zeiss microscopewith a Antiex Plan-Neouar 633objective (NA 1.25) using the 546.1 nmline ofa mercury lamp (HBO 103, Osram, Germany). Images of sporozoites in glassbottomdishes were recorded at 2 Hz with a CCDcamera (ORCA-ER,Hamamatsu, Japan) using simplePCI software (Hamamatsu, Japan).All image series were eventually imported to ImageJ for analysis, and gureswere generated using the Adobe Creative Suite software package. Pixel sizeswere preserved during image processing.Wavelet-Based Runlength MeasurementTheglobal waveletpowerspectrumquanties, analogoustoausual powerspectrum, the scale dependence of a functions behavior. However, thewavelet transformuses local basis functions, thus providing informationabout frequencyandtime. Here, weusedaMorlet (s=5) wavefunction.Thetrackedsporozoitemovementwasparameterizedasv(x) [speed(path)],andthewidthof thespeedpeaksisthenthedesiredrunlength. Awaveletanalysis was conducted on the concatenated data of all sporozoites. Multiplepermutations of the sequences served to suppress any possible effect due totheconcatenationofendpointsofindividual sequences.Theglobal waveletpowerspectrumistheaveragex,inwhichsisthescaleofthe wavelet, x is the sporozoite path, andFis the wavelet transformeddata. A scale corresponding to the average width of the speed peaks contrib-utes a peak to the wavelet power spectrum. One advantage of using waveletanalysisinsteadof regular Fourier analysisliesinthefact that thewaveletpowerspectrumisusuallysmootherthanisthefrequencyspectrum. Moreimportantly, we also found that the irregular duration of slow periods betweenthe speed peaks does not affect the wavelet analysis as strongly as it does fortheFourieranalysis.Parasite TrackingSporozoites wereeithermanuallytrackedat the apicalend(RICM)usingthemanual tracking plug-in of ImageJ, semiautomatically tracked using theMTrack2plug-infor ImageJ, oranalyzedwith a recently developed softwarebasedonMTrack2that ispart of Fiji (http://pacic.mpi-cbg.de/wiki/index.php/Main_Page) (Hegge et al., 2009). Adhesion cycles were counted manuallyon RICMacquisitions and normalized to the number of cycles occurring during100s. Oneadhesioncyclewasdenedasfollows: parasitefront or reartouching the surface (i.e., a dark spot) until either of both is released (dark spotdisappeared) and reattached (dark spot reappeared).For trackingof invivomovement, zprojectionswereperformedprior totracking. Thisresultsinthelossofthespeedvectorinzdirectionandthusanunderestimationof averagespeed. Italsoprovidesapotential pitfall forspeeddetectionofmovementinthedirectionofthelightpath(zdirection).We therefore identied parasites that moved only within a single optical planeand tracked and analyzed those. This showed an increase of speedcomparedtothecombineddatafromall sporozoites movingintheskin(Figures1A1C).Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite Gliding560 Cell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc.Speed Peak and Runlength DetectionSpeed peak detection was performed on time series frommanually (RICM) andautomatically tracked parasites imaged at 14 Hz. Aspeed peak was dened ifthe following conditions were fullled: (1) the instantaneous speed was largerthanthe medianvalueoftheentiresequenceofthe trackedparasite; (2)theinstantaneousspeedwashighestcomparedtothethreetimepointsbeforeandafter;(3)theinstantaneousspeedsurpasses150% ofthe minimalvaluefromthreeorseventimepointsbeforeandaftertheactualposition(thetwodifferentwindowsizeswereusedtodetectbothverysharpandratherwidepeaks); and (4) two adjacent peaks could not be closer than 2 s. For compar-isonbetweentracks, thefrequencyof peakoccurrencewascalculatedforevery track as number of speed peaks detected divided by the total durationof the track.Inordertotest thestabilityof thepeakdeterminationroutine, weaddedextraconditionsas:aspeedpeakshould beoveraglobalthresholdkcanabsolute value determining the minimum value of the speed peak.Therunlengthofaparasiteduringaspeedpeakwascalculatedasasumoftheinstantaneousspeedsaroundaspeedpeak(threevaluesbeforeandafter thespeedpeak) andthetwoadjacent valuesoneachside(radiusofvetimepointsaroundonespeedpeak). Thisvaluewasmultipliedbythelapsed time to get the runlength in mm.Traction Force MicroscopyElastic gel substrates were prepared as described before (Wang and Pelham,1998). We mixed two different marker beads: either 0.1 mm yellow-green and0.2 mm red uorescent beads or 0.2 mm red and 0.2 mm dark-red uorescentbeads(Invitrogen, Germany). TherelativeconcentrationsofacrylamideandbisacrylamidewerechosenaccordingtoYeunget al. (2005) topermit theusageofpublishedvaluesfortheelasticmoduli.Theparasiteswereplaceddirectlyontothegel andwerecoveredwitha22322mmglasscoverslip.ImageacquisitionwasperformedontheNikon-PerkinElmer spinningdiscconfocal witha1003objective(NikonPlanApoVCNA1.4). Imageswerecollected with an EM-CCDcamera (Orca ERHamamatsu, Japan) at 1 Hz usingthe UltraView software (PerkinElmer).Traction forces were reconstructed as described previously (Sabass et al.,2008). Displacements of red and green (red and far-red, respectively) uores-centmarkerbeadsembeddedinthepolyacrylamidegel reecttheeffectoftraction applied by the sporozoite to the surface of the gel. We thus comparedimages of sporozoites adhering at different sites in the eld of view to nd therelativesubstratedisplacements. APoissonratioaround0.5justiedtheassumption that the effect of vertical pulling of the sporozoite wasdecoupledfrom thelateral traction;hence,we interpretedallbeaddisplace-ments as a sole result of lateral traction. We used either a boundary elementmethodorFouriertransformtractioncytometrytodeterminethetractionatthenodes of our prescribedmesh. Inthecaseof theFourier Transformmethod, we implemented an iterative procedure similar to the one describedin Butleret al.(2002)to includethe informationaboutthe cellcontour.Here,wefoundit necessarytousezero-patterningtechniquestoavoidtractionartifacts at the outer rim of the cell.Statistical AnalysisSignicant differencesinthetractionforceor speeddataweredisplayedvisually with the help of the notched whisker plots using groups of data throughve-number summaries: the sample minimum, the lower and upper quartiles,the median, and the sample maximum. Maxima and minima are dened as theextreme data points within a range of 2 times the distance between upper andlower quartile, centered on the average of the quartiles. Outliers, marked as redpoints, are those data points that are beyond the above range. Nonoverlappingnotchesareusedtoindicatesignicant differencesbetweenmedians. Thenotchdisplaysthe95%condenceinterval for themedianbasedontheassumptionof anormal distribution. Signicant differenceswerevalidatedwith a nonparametric Kruskal-Wallis test.SUPPLEMENTAL DATASupplemental Data include Supplemental Text, eight gures, and ten moviesand can be found with this article online at http://www.cell.com/cell-host-microbe/supplemental/S1931-3128(09)00384-9.ACKNOWLEDGMENTSWethankR. RichterforhelpwithRICM;A. Besser, M. Cyrklaff, M. Ganter,G. Gerisch, S. Lepper, andM. Meissner for discussionsand/or readingofthemanuscript;D.Scheppanformosquitoinfection;andtheNikonImagingCenterattheUniversityofHeidelbergforaccesstotheirmicroscopes.Thiswork was supported by grants fromthe German Federal Ministry for EducationandResearch(BMBF-BioFuture) toF.F.; theGermanResearchFoundation(DFG, SFB 544) to F.F.; and the Frontier Program of the University of Heidel-berg to U.S.S.and F.F. This work wassupported,in part, bythe Institute forComputational Modeling; Siberian Branch of the Russian Academy of Science,Krasnoyarsk; the Center for Modelling andSimulation (BIOMS); andtheClusterof Excellence CellNetworksand the Medical School at the UniversityofHeidelberg.TheMaxPlanckSocietyandtheKarlsruheInstituteofTech-nology (KIT) and its Concept for the Future are acknowledged. K.M. and F.F.are members of the European Network of Excellence BioMalPar.Received: August 11, 2009Revised: October 21, 2009Accepted: November 19, 2009Published: December 16, 2009REFERENCESAmino, R., Thiberge, S., Martin, B., Celli, S., Shorte, S., Frischknecht, F., andMe nard, R. (2006). Quantitativeimagingof Plasmodiumtransmissionfrommosquito to mammal. Nat. Med. 12, 220224.Baker, R.P., Wijetilaka, R., and Urban, S. (2006). Two Plasmodium rhomboidproteasespreferentiallycleavedifferent adhesinsimplicatedinall invasivestages of malaria. PLoS Pathog. 2, e113.Baum, J., Gilberger, T.W., Frischknecht, F., and Meissner, M. (2008). Host-cellinvasionbymalariaparasites: insightsfromPlasmodiumandToxoplasma.Trends Parasitol. 24, 557563.Butler, J.P., Tolic -Nrrelykke, I.M., Fabry, B., and Fredberg, J.J. (2002).Traction elds, moments, and strain energy that cells exert on theirsurroundings. Am. J. Physiol. Cell Physiol. 282, C595C605.Combe, A., Moreira, C., Ackerman, S., Thiberge, S., Templeton, T.J., andMe nard, R. (2009). TREP, a novel protein necessary for gliding motility of themalaria sporozoite. Int. J. Parasitol. 39, 489496.Cooper, J.A. (1987). Effects of cytochalasin and phalloidin on actin. J. Cell Biol.105, 14731478.Cowman, A.F., and Crabb, B.S. (2006). Invasion of red blood cells by malariaparasites. Cell 124, 755766.Cramer, L.P. (1999). Role of actin-lament disassembly in lamellipodiumprotrusion in motile cells revealed using the drug jasplakinolide. Curr. Biol. 9,10951105.Daher, W., and Soldati-Favre, D. (2009). Mechanisms controlling glideosomefunction in apicomplexans. Curr. Opin. Microbiol. 12, 408414.Dembo, M., and Wang, Y.L. (1999). Stresses at the cell-to-substrate interfaceduring locomotion of broblasts. Biophys. J. 76, 23072316.Dowse, T.J., Koussis, K., Blackman, M.J., and Soldati-Favre, D. (2008). Rolesof proteasesduringinvasionandegressbyPlasmodiumandToxoplasma.Subcell. Biochem. 47, 121139.Fre nal, K., and Soldati-Favre, D. (2009). Role of the parasite and hostcytoskeleton in apicomplexa parasitism. Cell Host Microbe 5, 602611.Frevert, U., Engelmann, S., Zougbe de , S., Stange, J., Ng, B., Matuschewski,K., Liebes, L., and Yee, H. (2005). Intravital observation of Plasmodiumbergheisporozoite infection of the liver. PLoS Biol. 3, e192.Frischknecht, F., Baldacci, P., Martin, B., Zimmer, C., Thiberge, S., Olivo-Marin, J.C., Shorte, S.L., and Me nard, R. (2004). Imaging movement of malariaparasitesduringtransmissionbyAnophelesmosquitoes. Cell. Microbiol. 6,687694.Frixione, E., Mondrago n, R., and Meza, I. (1996). Kinematic analysis ofToxoplasma gondii motility. Cell Motil. Cytoskeleton 34, 152163.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite GlidingCell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc. 561Gantt, S., Persson, C., Rose, K., Birkett, A.J., Abagyan, R., andNussenzweig, V.(2000). Antibodiesagainst thrombospondin-relatedanonymous proteindonot inhibit Plasmodiumsporozoite infectivity in vivo. Infect. Immun. 68,36673673.Hegge, S., Kudryashev, M., Smith, A., and Frischknecht, F. (2009). Automatedclassicationof Plasmodiumsporozoitemovement patternsrevealsashifttowardsproductivemotilityduringsalivaryglandinfection. Biotechnol.J.4,903913.Heintzelman, M.B. (2006). Cellular and molecular mechanics of glidinglocomotion in eukaryotes. Int. Rev. Cytol. 251, 79129.Heiss, K., Nie, H., Kumar, S., Daly, T.M., Bergman, L.W., and Matuschewski, K.(2008).Functional characterizationofaredundantPlasmodiumTRAPfamilyinvasin, TRAP-like protein, by aldolase binding and a geneticcomplementation test. Eukaryot. Cell 7, 10621070.Johnson, K.L. (1985). Contact mechanics (Cambridge: Cambridge UniversityPress).Kappe, S., Bruderer, T., Gantt, S., Fujioka, H., Nussenzweig, V., and Me nard,R. (1999). Conservationof aglidingmotilityandcell invasionmachineryinApicomplexan parasites. J. Cell Biol. 147, 937944.Lepper,S.,Merkel,M.,Sartori,A.,Cyrklaff,M.,andFrischknecht,F.(2010).Rapid quantication of the effects of blotting for correlation of light and cryo-lightmicroscopy.J.Microsc.PublishedonlineOctober22,2009.10.1111/j.1365-2818.2009.03327.x.Matuschewski, K. (2006). Gettinginfectious: formation andmaturationofPlasmodium sporozoites in the Anopheles vector. Cell. Microbiol. 8,15471556.Me nard, R. (2001). Gliding motility and cell invasion by Apicomplexa: insightsfrom the Plasmodium sporozoite. Cell. Microbiol. 3, 6373.Morahan, B.J., Wang, L., andCoppel, R.L. (2009). NoTRAP, noinvasion.Trends Parasitol. 25, 7784.Moreira, C.K., Templeton, T.J., Lavazec, C., Hayward, R.E., Hobbs, C.V.,Kroeze, H., Janse, C.J., Waters, A.P., Sinnis, P., andCoppi, A. (2008). ThePlasmodium TRAP/MIC2 family member, TRAP-Like Protein (TLP), is involvedin tissue traversal by sporozoites. Cell. Microbiol. 10, 15051516.Ra dler, J., and Sackmann, E. (1992). On the measurement of weak repulsiveandfrictional colloidalforcesbyreectioninterferencecontrastmicroscopy.Langmuir 8, 848853.Sabass, B., Gardel, M.L., Waterman, C.M., andSchwarz, U.S. (2008). Highresolution traction force microscopy based on experimental andcomputational advances. Biophys. J. 94, 207220.Sahoo,N.,Beatty,W.,Heuser,J.,Sept,D.,andSibley, L.D.(2006).Unusualkinetic and structural properties control rapid assembly and turnover of actinin the parasite Toxoplasma gondii. Mol. Biol. Cell 17, 895906.Schmitz, S., Grainger, M., Howell, S., Calder, L.J., Gaeb, M., Pinder, J.C.,Holder,A.A.,andVeigel,C.(2005).Malariaparasiteactinlamentsareveryshort. J. Mol. Biol. 349, 113125.Schu ler, H., and Matuschewski, K. (2006). Regulation of apicomplexanmicrolamentdynamicsbyaminimalsetofactin-bindingproteins. Trafc 7,14331439.Schu ler, H., Mueller, A.K., and Matuschewski, K. (2005). Unusual properties ofPlasmodiumfalciparumactin: newinsightsintomicrolament dynamicsofapicomplexan parasites. FEBS Lett. 579, 655660.Sengupta,K.,Aranda-Espinoza,H.,Smith,L.,Janmey,P.,andHammer,D.(2006). Spreadingof neutrophils: fromactivationtomigration. Biophys. J.91, 46384648.Steinbuechel, M., andMatuschewski, K. (2009). Rolefor thePlasmodiumsporozoite-specic transmembrane protein S6 in parasite motility and efcientmalaria transmission. Cell. Microbiol. 11, 279288.Sultan, A.A., Thathy, V., Frevert, U., Robson, K.J., Crisanti, A., Nussenzweig,V., Nussenzweig, R.S., and Me nard, R. (1997). TRAPis necessary for glidingmotility and infectivity of Plasmodium sporozoites. Cell 90, 511522.Vanderberg, J.P. (1974). Studies on the motility of Plasmodium sporozoites. J.Protozool. 21, 527537.Vicente-Manzanares, M., Choi, C.K., and Horwitz, A.R. (2009). Integrins in cellmigrationthe actin connection. J. Cell Sci. 122, 199206.Vlachou, D., Schlegelmilch, T., Runn, E., Mendes, A., and Kafatos, F.C. (2006).The developmental migration of Plasmodiumin mosquitoes. Curr. Opin.Genet. Dev. 16, 384391.Wang, Y.L. (2007). Flux at focal adhesions: slippage clutch, mechanical gauge,or signal depot. Sci. STKE 2007, pe10.Wang, Y.L., andPelham, R.J., Jr. (1998). Preparationof aexible, porouspolyacrylamidesubstrateformechanical studiesof culturedcells. MethodsEnzymol. 298, 489496.Wetzel, D.M., Ha kansson, S., Hu, K., Roos, D., and Sibley, L.D. (2003). Actinlamentpolymerizationregulates glidingmotility byapicomplexan parasites.Mol. Biol. Cell 14, 396406.Yeung, T., Georges, P.C., Flanagan, L.A., Marg, B., Ortiz, M., Funaki, M., Zahir,N., Ming, W., Weaver, V., andJanmey, P.A. (2005). Effects of substratestiffness on cell morphology, cytoskeletal structure, and adhesion. Cell Motil.Cytoskeleton 60, 2434.Cell Host & MicrobeAdhesion Dynamics during Malaria Parasite Gliding562 Cell Host & Microbe 6, 551562, December 17, 2009 2009 Elsevier Inc.