new fluorescent tools for watching nanometer-scale...

ANRV311-BB36-17 ARI 3 April 2007 15:17

New Fluorescent Tools forWatching Nanometer-ScaleConformational Changesof Single MoleculesErdal Toprak1 and Paul R. Selvin1,21Center for Biophysics and Computational Biology and 2Physics Department,University of Illinois, Urbana-Champaign, Urbana, Illinois 61801;email: [email protected], [email protected]

Annu. Rev. Biophys. Biomol. Struct. 2007.36:349–69

First published online as a Review in Advance onFebruary 13, 2007

The Annual Review of Biophysics and BiomolecularStructure is online at biophys.annualreviews.org

This article’s doi:10.1146/annurev.biophys.36.040306.132700

Copyright c© 2007 by Annual Reviews.All rights reserved

1056-8700/07/0609-0349$20.00

Key Words

fluorescence microscopy, motor proteins, measuring step sizes,motility mechanism

AbstractSingle-molecule biophysics has been serving biology for more thantwo decades. Fluorescence microscopy is one of the most commonlyused tools to identify molecules of interest and to visualize biologi-cal events. Here we describe some of the most commonly used flu-orescence imaging tools to measure nanoscale movements and therotational dynamics of biomolecules.

349

Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

Contents

MEASURING DISTANCES ANDSTUDYING ROTATIONALDYNAMICS OF PROTEINSUSING SINGLE-MOLECULEFLUORESENCE IMAGINGTECHNIQUES . . . . . . . . . . . . . . . . . 350Single-Particle Tracking:

Resolution and Orientation . . . 350Fluorescence Imaging with One

Nanometer Accuracy . . . . . . . . . . 351Possible Methods of Decreasing

σμx,y . . . . . . . . . . . . . . . . . . . . . . . . . . . 351FIONA Applied to Processive

Cytoskeleton Motor Proteins . . 353In Vivo FIONA . . . . . . . . . . . . . . . . . . 354Pros and Cons of FIONA . . . . . . . . . 355Two-Color FIONA Imaging or

SHREC . . . . . . . . . . . . . . . . . . . . . . 355

Measuring Small Distances UsingSHRImP. . . . . . . . . . . . . . . . . . . . . . 357

STUDYING ROTATIONALDYNAMICS WITHSINGLE-MOLECULEFLUORESCENCETECHNIQUES . . . . . . . . . . . . . . . . . 357Taking Advantage of the

Polarization of Light . . . . . . . . . . 357Combining FIONA and

Polarization Studies . . . . . . . . . . . 360Removing Angular Degeneracies

Using Defocused Orientationsand Position Imaging . . . . . . . . . . 360

Studying Lever Arm Dynamics ofMyosin V with DOPI. . . . . . . . . . 363

CONCLUSIONS AND FUTUREDIRECTIONS . . . . . . . . . . . . . . . . . . 365

MEASURING DISTANCES ANDSTUDYING ROTATIONALDYNAMICS OF PROTEINSUSING SINGLE-MOLECULEFLUORESENCE IMAGINGTECHNIQUES

Single-Particle Tracking: Resolutionand Orientation

It has been known for more than a century thatthe far-field diffraction limit of light is approx-imately λ/2NA, where λ is the wavelength andNA is the numerical aperture of the collect-ing lens. For a microscope, the maximum NAis

ANRV311-BB36-17 ARI 3 April 2007 15:17

and there is a need for tools that measurethe rotational movements of biomolecules.The existing tools for measuring rotationaldynamics are usually based on the polarizedcharacteristic of light. A fluorophore typicallyacts like a transition dipole moment, i.e.,it has an excitation dipole and an emissiondipole. Dickson and coworkers (2–4) showedthat the shape of a fluorophore’s image col-lected with a defocused lens depends on thedye’s orientation. Goldman and colleagues(15) used total internal reflection (TIR)fluorescence microscopy with orthogonalpolarizations to excite a single molecule inorder to determine the dye’s orientationwhen bound to myosin V molecules. Finally,we have utilized the work of both Dicksonand colleagues and Goldman and colleaguesto determine orientations and centroids usinga technique called defocused orientation andpositioning imaging (DOPI) (42), as well aspolarization microscopy (39). The work issummarized here.

Fluorescence Imaging with OneNanometer Accuracy

Fluorescence imaging is widely used to studybiological mechanisms. With classical fluo-rescence imaging methods, point-like fluores-cent emitters (e.g., small organic dyes) can-not be localized to an accuracy better than∼250 nm because of the diffraction limit.However, better localization accuracy is nec-essary in order to study the mechanisms ofbiological molecules. For example, processivemolecular motors such as myosin, kinesin,and dynein take discrete steps, which occurin the nanometer range (49–51). The studyby Cheezum et al. (9) showed that fittinga two-dimensional (2D) Gaussian functionto the emission pattern of a single fluores-cent probe to obtain the centroid of a flu-orescent probe is the best approach amongother possible algorithms. Thompson et al.(40) later quantified all the parameters that af-fect the standard error of the mean (SEM) ofthe resulting fit and, using a charge-coupled

TIR: total internalreflection

DOPI: defocusedorientation andposition imaging

CCD: charge-coupled device

FIONA:fluorescence imagingwith one nanometeraccuracy

device (CCD), demonstrated experimentallythat fluorescent beads can be localized to∼2 nm. The main contributions to the SEMof a least-squares fit of the 2D Gaussian func-tion arise from five sources: pixelation of theCCD camera, fluorescent background, read-out noise from the CCD camera, number ofphotons, and photon noise. This approachwas extensively developed by Yildiz et al. (49),and it was successfully applied to measure thestep sizes of fluorescently labeled myosin Vmolecules. Yildiz et al. named the method flu-orescence imaging with one nanometer accu-racy (FIONA) (49). FIONA has been widelyused by several groups to study the transloca-tion of molecular motors or to measure smalldistances.

The emission intensity distribution [I(x,y)]of a fluorophore on the CCD imaging planeis fit to the following 2D Gaussian functionin order to get the lateral position of the flu-orophore in the specimen plane:

I (x, y) = Ibackground + A.

exp{−1/2

[(x − x0

s x

)2+

(y − y0

s y

)2]}, 1.

where x0 and y0 are the coordinates of thecenter, and sx and sy are the standard devi-ations of the distributions in each direction(Figure 1a). The SEM (σμx,y ) of the 2D Gaus-sian function is

σμx,y

=√

s 2x,yNphotons

+ a2

12.Nphotons+ 8.π.s

4x,y .b2

a2.Nphotons, 2.

where Nphotons is the number of the collectedphotons during one exposure period, a isthe effective pixel size (a = CCD pixel sizeper magnification), and b is the background(40).

Possible Methods of Decreasing σμx,yIt is possible to reduce the SEM to as lowas 1 nm (or less) by increasing the numberof collected photons (>10,000), decreasing

www.annualreviews.org • Fluorescence Imaging of Single Molecules 351

Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

Figure 1(a) The observed CCD image (inset) and the resulting point spread function of a fluorescent bead; theGaussian fit is shown with wires. The standard error of the mean for the Gaussian fit is ∼1.2 nm. (b) Afluorescent bead stuck to the coverslip was moved by 20-nm steps using a sample-holding piezo stage(Nano-Bio2�, Mad City Labs Inc.), and each frame was acquired for 0.5 s. (c) A cartoon illustration oftwo consecutive steps of a fluorescently labeled myosin V molecule with hand-over-hand fashion. Leverarms are blue and red. The motility direction is leftward. Only one of six IQ domains is labeled with abifunctional rhodamine dye. The emission dipole of the dye is depicted by an orange, double-arrowedline. The center-of-mass displacement of myosin V is ∼37 nm at each step. Therefore the measured stepsizes are expected to alternate by 37 − 2x and 37 + 2x nm, where x is the projected distance between thedye and the center of mass of the myosin V. The lever arm is in a trailing state after short (37 − 2x) stepsand in the leading state after long (37 + 2x) steps. The lever arm might be slightly curved in the leadingstate. (d ) A sample displacement trajectory for a myosin V molecule. The average step size is ∼76 nm,with a standard deviation of 9.8 nm. The short step size is hidden likely because of the dye location. Thedye is likely to be close to the actin binding domain.

the fluorescence background and the size ofthe effective CCD pixels, or both. Objective-type TIR microscopy significantly reducesthe background, and the high NA (1.40–1.65) oil-immersion objective lenses signifi-

cantly increase the number of collected pho-tons. Prism-type TIR is another method toincrease the signal-to-noise ratio (SNR), de-spite its reduced photon collection efficiency.(We generally prefer objective-type TIR

352 Toprak · Selvin

Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

because of the increasing collection effi-ciency.) Epifluorescence microscopy is alsoamenable to FIONA experiments, however,at the expense of a much higher backgroundlevel compared with TIR imaging (49). A wayin which to understand the necessary pho-ton number and background level is to ob-tain a SNR of ∼30 for the peak value of thePSF (SNR = Ipeak/(Ipeak + b2)0.5). The moststraightforward way to achieve a higher SNRis to use longer exposure times during imag-ing. Another way of decreasing σμx,y is to de-crease the pixel size. However, decreasing σμx,yby reducing both the effective pixel size andthe background is not possible. The intensitydecreases as a function of the magnification (Iα 1/M2, where M is the magnification), andthis leads to an intrinsic background increase.

In summary, small pixels (200 nm) are not useful becausethe PSF converges to a 2D delta functionrather than to a 2D Gaussian function. Agood way of selecting the right pixel size isto predict the full width at half maximum(FWHM ≈ 250 nm) of the emission distri-bution. An effective pixel size ranges from 30to 50% of the FWHM, which corresponds toa value between ∼80 and ∼150 nm for thevisible light. We typically choose either 86or 100 nm for most of the single-moleculeexperiments.

Suppressing blinking events and increas-ing photostability of the dyes through theuse of an enzymatic oxygen scavenging sys-tem and reducing agents are also essentialfor FIONA experiments (49). The exact re-ducing agents, including the manufacturersof the glucose oxidase and catalase, are nec-essary (49). Trolox is a promising reducingagent (31). Piezo stages can be useful for test-ing both the optical adequacy of the imagingsystem (SNR, stage drifts, etc.) and the fittingalgorithm used. Figure 1b shows a trace ofartificial 20-nm steps by using a piezoelectricstage.

BR: bifunctionalrhodamine

FIONA Applied to ProcessiveCytoskeleton Motor Proteins

A number of different processive cytoskele-ton proteins exist, including actin-dependentmotors, such as myosins, in particular, myosinV and myosin VI, as well as microtubule-dependent motors, such as kinesins anddyneins. These molecular motors have beenstudied using single-molecule techniques forthe past two decades, most often with opticaltraps, and their step sizes have been success-fully measured (6, 24, 33, 34, 43). Accordingto these experiments, myosin V has a ∼37-nmstep size and myosin VI has a ∼30-nmstep size. On the other hand, microtubule-dependent molecular motors, such as con-ventional kinesin, cytoplasmic dynein, anddimeric EG5, have ∼8-nm step sizes. Thereis still not much known about the step-ping mechanism of dynein. These experi-ments were performed using optical tweez-ers, in which a polystyrene microsphere wasattached to the stalk of the motor proteinand the center-of-mass displacements andthe stall forces of these proteins were ob-tained. The center-of-mass movements, how-ever, do not give any information about thedynamics of the individual motor domains ofthese proteins. Therefore it was not possi-ble to distinguish between the two predictedwalking models: inchworm versus hand-over-hand. Each of the proposed models makestestable predictions as to the motions of theprotein during ATP-dependent translocation(49).

To address this mechanistic question, onelight chain of a myosin V molecule was la-beled with a bifunctional rhodamine (BR) ora cyanine 3 (Cy3) dye and an in vitro motil-ity assay was performed at ∼300 nM ATP(49). Because only one light chain of myosin Vwas labeled (Figure 1c), ∼37-nm steps wereexpected if the inchworm model was valid.However, alternating 37 − 2x (short) and 37+ 2x nm (long) steps were expected if thehand-over-hand model was valid, where x isthe average distance between the center of


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

the myosin V and the dye projected on thespecimen plane (Figure 1c). Yildiz et al. (49)observed alternating long and short steps (74− 0 nm, 52 − 23 nm, 44 − 30 nm), suggest-ing that myosin V was indeed walking hand-over-hand (Figure 1d ). Sakamoto et al. (35)determined that the length of the leg deter-mined the step size and that the 74 − 0 nmsteps were actually ∼64 − 10 nm steps (39,42). Furthermore, Yildiz et al. (51) showedthat conventional kinesin moves hand-over-hand, and Ökten et al. (27) and Yildiz et al.(50) showed that myosin VI also moves via ahand-over-hand mechanism. Recently, Reck-Peterson et al. (32) showed strong evidencethat cytoplasmic dynein also has a hand-over-hand stepping mechanism.

Figure 2(a) The fluorescent image of an S2 Drosophila cell treated withcytochalasin D, which formed microtubule-containing long processes.Several peroxisomes are moving along the microtubules. The region ofinterest containing one of the peroxisomes is shown within the square.(b) The 3D plot of the fluorescent image of the peroxisome within thesquare. Imaging was done using an objective-type TIR setup.

In Vivo FIONA

Most motility experiments measuring the stepsizes have been performed in vitro, becausethe experimental conditions can be controlledto reduce the effects of unknown parametersthat exist within the cell such as high fluo-rescent background, nonuniform drag coeffi-cients, complicated actin or microtubule ge-ometries, and other interacting enzymes. Thedrawback to such in vitro experiments, how-ever, is long sample preparation times. In vivoexperiments on the other hand have the ad-vantage of simpler sample preparation andthe aforementioned difficulties can be han-dled through the use of clever experimentaldesigns.

For example, Kural et al. (22) recentlyperformed an in vivo experiment to studythe motility mechanism of conventional ki-nesin and cytoplasmic dynein in DrosophilaS2 cells, which expressed eGFP with aperoxisome-targeting sequence. Labeled per-oxisomes consist of a large number of eGFPmolecules depending on their size. Further-more, the fluorescent background of the cellwas reduced by using an objective-type TIRsetup and the cells were chilled to ∼10◦C toslow down protein motion. To isolate peroxi-some movement by kinesin and dynein move-ment and to eliminate the effect of actin,the authors treated the cells with cytocha-lasin D. Removing the actin content not onlyyielded pure microtubule-dependent motility,but also created linear tracks on which the or-ganelles could move, since S2 cells tend togrow thin linear cell structures (called pro-cesses) filled with microtubules (Figure 2a).The motility of the peroxisomes was recordedin the TIR volume on these linear tracks.

Kural et al. (22) could also determinewhether the peroxisome was being moved bya kinesin (moving away from the nucleus) orby dynein (moving toward the nucleus). (Inearlier experiments, while looking at peroxi-somes within the cell body, without processes,it was not possible to determine definitelywhether the peroxisomes were driven by


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

kinesin or dynein.) Kural et al. successfullymeasured the step sizes of kinesin (∼8 nm)and dynein (∼8 nm) with a temporal reso-lution of 1.1 ms and a position accuracy of∼1.5 nm (Figure 2b).

Similar in vivo experiments were done byNan et al. (26) and Toba et al. (41), who usedquantum dots as probes to study cytoplasmicdynein and kinesin motility inside the cell.We should mention that the dynein experi-ments performed by different groups are notin absolute agreement on dynein step sizes andstall forces (22, 24, 26, 32, 41). The reason forthese different observed step sizes and the stallforces is not yet apparent.

Pros and Cons of FIONA

FIONA has the great advantage of using wide-field illumination (usually a ∼50 × ∼50 μmarea), whereby many molecules can be ob-served simultaneously. The covalent attach-ment of organic fluorophores does not ad-versely affect motor function (but this mustalways be tested) and furthermore does notimpose significant additional load. The cost ofan objective-type TIR setup is usually muchcheaper than the cost of other optical de-signs (i.e., optical tweezers). A standard flu-orescence microscope can be converted to aTIR microscope by using a laser line for theillumination and a high NA (>1.4) oil objec-tive. Commercial TIR microscopes are alsoavailable.

Furthermore, FIONA does not need anyspecial calibrations; rather, the only parame-ter needed is the effective pixel size, which canbe easily obtained by knowing the real CCDpixel size and the magnification of the imag-ing optics. Currently, the time resolution ofFIONA can be reduced to ∼20 ms for single-fluorophore imaging (data not shown) and tosubmilliseconds for an object that containsseveral fluorophores or quantum dots (22, 26).

Brightness, especially for in vivo workand on single-molecule fluorophores, is asignificant limitation. Photostability of thefluorophores is another major challenge for

FRET: fluorescenceresonance energytransfer

SHREC:single-moleculehigh-resolutioncolocalization

FIONA experiments. Optical probes such asquantum dots and gold nanoparticles, whichlast longer than conventional fluorophores,may provide attractive alternatives to the cur-rent suite of fluorescent dyes and fluorescentproteins (23, 25, 26, 40). Quantum dots areespecially attractive candidates and they havebeen used by several groups for in vivo and invitro studies (11). However, current obstaclesto the use of quantum dots, such as low la-beling efficiency, blinking, large effective size,nonuniform size and shape, and nonuniformexcitation and emission characteristics, all re-quire significant improvements (11, 26, 45).

Unlike optical or magnetic tweezers,FIONA does not apply force onto molecules.Therefore, combining force spectroscopytechniques such as optical and magnetictweezers with FIONA will constitute a sig-nificant development. We are currently com-bining optical traps with FIONA.

Two-Color FIONA Imaging orSHREC

Conventional microscopy cannot resolve twomolecules closer than ∼250 nm, the far-filediffraction limit, also known as the Rayleighlimit. Fluorescence resonance energy transfer(FRET) can effectively measure distances be-tween ∼2 and ∼10 nm. However, many con-formational changes occur in the range of 10to 250 nm, which is not accessible by con-ventional light microscopy or FRET. There-fore an alternative approach, such as differ-ential labeling of particular sites of proteins,is required. In many cases, biological macro-molecules provide sufficient reactivity suchthat two-color labeling can be achieved witha variety of organic and inorganic dyes.

A particularly powerful technique is two-color FIONA, also called single-moleculehigh-resolution colocalization (SHREC), de-veloped by the Spudich group (10). Thismethod relies on having two fluorophores thathave sparse emission peaks, which are thenimaged on the same CCD camera. Both dyesare excited by two separate laser lines and


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

the resulting emissions are split with the ap-propriate dichroic mirrors and emission fil-ters. After obtaining the images of both flu-orophores from the CCD, one can apply a

Figure 3(a) A sample motility trace for a myosin V molecule labeled with abifunctional rhodamine (BR, blue circle) and a cyanine 5 (Cy5, red circle)dye. (b) The estimated dye–myosin V geometry. The Cy5 dye ( purplecircles and red fit line) shows alternating ∼74 − 0 nm steps; therefore it islikely located very close to the actin binding domain. However, BR dye(blue circles and blue fit line) shows alternating ∼52 − 23 nm steps;therefore it is possibly located on another IQ domain that is farther fromthe actin binding domain. Both dyes are on the same lever arm.

mapping function to colocalize the dye po-sitions. Usually two-color experiments havethe difficulty of finding a sharp focal planefor both colors simultaneously because of thechromatic aberrations of the microscope ob-jective. However, it is possible to use a menis-cus lens to compensate for the focal planedifference due to the chromatic aberrations.(We use a commercial dual-view tube man-ufactured by Optical Insights, LLC for two-color imaging.)

Churchman et al. (10) colocalized Cy3 andCy5 dyes, which were attached to two endsof a short DNA, and measured distances assmall as 10 nm between the two. Furthermore,SHREC was used to measure the step size ofmyosin V, which was labeled with a Cy3 onone lever arm and a Cy5 on the other (10). Fi-nally, A. Yildiz, Y.E. Goldman, and P.R. Selvin(unpublished results) used an approach simi-lar to that used by Churchman et al. (10) tolabel myosin V with Cy3 and Cy5 and to mea-sure the step sizes when myosin V was walking(Figure 3).

Two-color FIONA has also been per-formed with two different quantum dots, oneon each head of the myosin V (45). One quan-tum dot had an emission peak at 565 nm, andthe other had an emission peak at 655 nm.Quantum dots are handy for two-color ex-periments because different quantum dots canbe excited simultaneously with the same ex-citation source. Warshaw et al. (45) mea-sured the distance between the motor do-mains to be ∼36 nm and a step size of ∼72 nmwhen the motor domain goes from the trailingstate to leading state. One problem was thatonly a small percentage (∼3%) of myosin Vwalked while having two quantum dots (D.M.Warshaw, personal communication).

Although these recent experiments didnot provide any new information regardingthe myosin V motility mechanism, they didprovide a new, general method for studyingmolecular motors. This technology enablesresearchers to observe the dynamics of dif-ferent sites of a molecule at the same time.FRET is limited to 2 to 8 nm, while SHREC is


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

capable of measuring 8 nm and longer. Onedownside to the technique is that, so far, rel-atively few available dye pairs have the req-uisite photostability. Cy3 (or rhodamine) andCy5 are nearly optimal existing organic dyepairs for SHREC experiments. We suspectthat with the addition of Trolox, an effectivephotostabilizer for Cy3 and Cy5, SHREC willbe suitable for more applications (31).

Measuring Small Distances UsingSHRImP

Single-molecule high-resolution imagingwith photobleaching (SHRImP) is a tech-nique introduced by Gordon et al. (18) tomeasure small distances (>10 nm) betweenidentical immobile fluorophores. SHRImP isa method of obtaining the individual centroidfor each of the two closely spaced fluo-rophores, which are otherwise not separatelyresolvable.

Briefly, a spot on the CCD camerathat photobleaches in two steps is iden-tified (Figure 4a). Once the first dye hasbleached, the resulting image—the secondfluorophore—is fit to a single 2D Gaussianfunction. To obtain the position of the firstdye, the image after photobleaching of thesecond dye is subtracted from the image be-fore photobleaching of the first dye. This im-age is then also fit to a single 2D Gaussianfunction (Figure 4b). The difference in cen-troids of the two dyes is then the distance be-tween the two fluorophores. Gordon et al. (18)successfully measured the distance betweentwo Cy3 dyes that are attached to short seg-ments of DNA (>10 nm).

SHRImP is also used to measure the inter-head distances of myosin VI molecules boundto actin filaments. Using an eGFP-myosinVI fusion protein, Balci et al. (1) determinedthat the average interhead distance is ∼29 nmwhen myosin VI was bound to actin. As men-tioned above, SHRImP is useful for immo-bilized molecules, unlike SHREC. However,SHRImP is easier to apply. For example, onedoes not need to find a reliable, different col-

SHRImP:single-moleculehigh-resolutionimaging withphotobleaching

ored dye pair, and efficient labeling is not asdifficult as multicolor labeling. Inherent op-tical problems such as chromatic aberrationsdo not occur with SHRImP.

STUDYING ROTATIONALDYNAMICS WITHSINGLE-MOLECULEFLUORESCENCE TECHNIQUES

Single-molecule biophysics comprises morethan simply measuring the step sizes of molec-ular motors. For instance, many proteins suchas F1-F0-ATPase (48), nonprocessive molecu-lar motors, and ion channels do not move pro-cessively, but rotate. Here we describe someof the techniques that can be used with wide-field illumination and for studying the ori-entation (as well as the translation) of singlebiological molecules.

Taking Advantage of the Polarizationof Light

The 3D orientations of the fluorophores thathave linearly polarized excitation and emis-sion dipoles can be detected by taking ad-vantage of the polarized characteristic of (in-cident and/or emitted) light. Therefore onecan study rotational dynamics of biologicalmolecules by attaching a fluorophore so thatit is rigid and does not undergo rotations in-dependent of the target molecule’s rotations.

Polarization modulation was applied tosingle fluorophores by Ha et al. (19). Thedyes immobilized on the surface were excitedwith polarized light and the polarization wascontinuously modulated (Figure 5a). The ex-citation yield varied with respect to the an-gle between the excitation dipole of the fluo-rophore and the polarization of the excitationlaser. The resulting intensity profile of thefluorophore can be fit to a quadratic cosinefunction, and the in-plane angle (the anglearound the optical axis) of the fluorophore canbe extracted (Figure 5b). However, polariza-tion modulation is not amenable to dynamicsystems unless the modulation and exposure


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

Figure 4(a) A dimeric myosin VI molecule is labeled with GFP molecules on the motor domain and excited with a488-nm laser line. Myosin VI was attached to the actin filament during the imaging time. The intensityprofile shows the two-step photobleaching events of the GFP molecules. (b) The point spread function ofthe two GFP molecules before and after the first photobleaching event (left to right). The difference ofthese two functions gives the approximate point spread function of the first photobleached GFPmolecule. Therefore it can be used to find the centroid of the first photobleached GFP. The remainingpoint spread function after the first photobleaching event is used to find the centroid of the secondphotobleached GFP molecule.

speed are high enough compared with the ro-tational dynamics of the target molecule.

An alternative way of finding the in-planeangles is to fix the excitation dipole and recordthe intensity fluctuations caused by rotations

(38). However, it is usually difficult to re-late the fluctuations and the rotations usingonly one excitation laser beam. Therefore it isessential to use two carefully calibrated beamsthat have orthogonal polarizations. Both


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

approaches have generic angular and dipo-lar degeneracy problems and do not give anyout-of-plane angle information. A similar ap-proach was used by Kinosita and coworkers(21) to obtain out-of-plane angles in order tomeasure the axial rotations of the actin fila-ments caused by myosin molecules. The ideais based on splitting the emission coming froma dye into vertical and horizontal polariza-tions and calculating the out-of-plane angles.However, both of the approaches need furtherinformation to get the full 3D orientationsfor single molecules, which was performed byForkey et al. (13–15).

Forkey et al. (15) cleverly combined the ap-proaches of both Ha et al. (19) and Kinositaet al. (21) to get the full 3D orientation infor-mation. They used an advanced prism-typeTIR setup to excite the BR molecules. TheirTIR setup included two orthogonal incidentbeams. Each of the beam’s polarization wasswitching between horizontal and vertical po-larizations, such that they had four differ-ent excitation beams (first beam vertical po-larization, first beam horizontal polarization,second beam vertical polarization, and sec-ond beam horizontal polarization). This ap-proach has the advantage of creating evanes-cent fields using TIR that are in x, y, and zdirections, depending on the excitation laser’sincident angle and the polarization of the inci-dent excitation laser. They also split the emis-sion light with respect to the polarization, andeach of the split emission lines was recordedwith avalanche photo diodes (APDs). So foreach cycle of excitation they had four exci-tation lines and two emission lines, result-ing in eight different combinations. In thisway they obtained the 3D orientation of asingle fluorophore. This method, also calledsingle-molecule fluorescence polarization mi-croscopy (SMFP), detects only one moleculeat a time because APDs are used for detection.But fast CCD cameras may be used instead,thus allowing more molecules to be observedsimultaneously.

In addition to this technical improvement,Forkey et al. (15) designed a novel experi-

3200

2800

Inte

nsi

ty (

a.u

.)

Excitation angle (°)

2400

20000 180 360 540 720

Excitationpolarization

Excitationdipole

θ

a

b

Figure 5(a) The simplified geometry of the polarization modulation method. Afrozen DiI molecule was excited by a linearly polarized 532-nm excitationlaser. The excitation laser’s polarization was continuously modulated by ahalf-wave plate mounted on a motorized rotary stage. (b) The resultingemission intensity of a DiI molecule that is excited as described above. Theintensity profile can be fit to a quadratic cosine function (I∼A +B.cos2(�excitation − �dipole)). Intensity is maximum when �excitation =�dipole or �excitation = π + �dipole because of dipolar degeneracy.

APD: avalanchephoto diode

SMFP:single-moleculefluorescencepolarizationmicroscopy

ment. They labeled myosin V molecules withBR dyes and observed the 3D rotational dy-namics of myosin V molecules when theywere walking along actin filaments. The au-thors also used an actin-based coordinatesystem for unambiguous analysis. They re-ported that the probe attached to the leverarm had two well-defined angles with respectto the actin filament. On average this angu-lar change was ∼40◦, and Forkey et al. alsoreported that almost half of the myosin Vmolecules were moving but not undergoing


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

CaM: calmodulin

any rotational changes. However, this angularchange of ∼40◦ and the existence of nontiltingmyosin molecules were not consistent withthe electron microscopy images (8, 44). Thereason for this inconsistency, as explained byForkey et al., was the fourfold angular degen-eracy of the SMFP and the inherent twofolddegeneracy of the dipoles. The angular de-generacy of SMFP can be reduced twofold bymaking extra polarization measurements (30).

Combining FIONA and PolarizationStudies

SMFP is a powerful technique used for in-vestigating rotational dynamics. However, be-cause the emitted photons are detected byAPDs, this method cannot report on trans-lational movements as do FIONA and op-tical tweezers. On the other hand, FIONAhas no capability to measure specific dye ori-entations. Achieving simultaneous localiza-tion and rotational dynamics data thereforerequires combining SMFP with a positionmeasurement technique. Syed et al. (39) suc-cessfully combined polarization studies withFIONA to study the translational and rota-tional dynamics of myosin V. Myosin V sam-ples labeled on the light chain with a BRmolecule were excited by switching betweentwo orthogonal excitation lasers with thesame polarization, which provided the abil-ity to track the intensity fluctuations causedby reorientations of the dye and to measureprecisely the motility of myosin V moleculesat the same time. Unlike SMFP this methoddoes not aim to get full 3D orientation infor-mation, rather it tracks qualitative intensitychanges caused by lever arm reorientations(Figure 6). The BR molecule in this study wasattached to the myosin by two bonds to min-imize the thermal rotations of the dye. Theexposure time for the experiments was either500 or 150 ms and the positional accuracy was∼2 nm.

Previous in vitro myosin V experimentsproposed a model in which the lever arm ofmyosin V was bent instead of straight. Snyder

et al. (37) measured the step sizes of myosin Vby labeling one of the motor domains with aneGFP molecule, and they observed alternat-ing 74 − 0 nm steps, as expected. Surprisinglythe step sizes measured by Yildiz et al. (49) forthe myosin V labeled on one of its six calmod-ulins (CaMs) were 74 − 0 nm, 52 − 23 nm,and 44 − 30 nm depending on the dye lo-cation on the lever arm. The observation of74 − 0 nm steps even when the dye was onthe lever arm suggests a bent lever arm model(37). By combining FIONA with polarizationmeasurements, Syed et al. (39) found that the74 − 0 nm steps originally observed by Yildizet al. were in fact 64 − 10 nm steps after us-ing the qualitative information from intensityfluctuations due to lever arm reorientations.

Another interesting observation was theoccurrence of ATP-dependent lever arm re-orientations without any lateral movements.Combining FIONA with polarization meth-ods is an excellent example of how monitoringmultiple observables simultaneously can yieldadditional information not provided by eithertechnique on its own.

Removing Angular DegeneraciesUsing Defocused Orientationsand Position Imaging

Conventional microscopy techniques, whichobserve the specimen in the sharp focal plane,generally cannot report on the 3D orienta-tions of the fluorophores. However, if thesample is deliberately moved away from thesharp focus by a specific distance (dependingon the optical parameters), 3D orientationsof the fluorophores can be found without anyangular degeneracy (2, 3, 7, 42).

The defocused CCD image of a single flu-orescent molecule is a combination of lobesand fringes depending on the orientation ofthe fluorophore (Figure 7a). This methodwas first introduced by Dickson and cowork-ers (2, 3) and the complete description of theelectrodynamics behind DOPI has been de-tailed elsewhere (2, 3, 7). In brief, the im-age of the single fluorophore is taken in


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

Time (s)

Po

siti

on

(10

0 n

m)

30 40 50 60 70– 2

0

2

4

6

8 2.0

1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.2

0.4

8.8

7.6

11.7

13.2

6.413.7

65.1

Inten

sity (a.u.)

68

65.4

53.3

61.3

82.8

Figure 6A myosin V molecule that has a bifunctional rhodamine on the lever arm was excited by alternatingbetween two orthogonally polarized laser lines. The intensity fluctuations due to dye reorientations areobserved ( green and pink diamonds). The displacement trajectory ( pink and green stars) was analyzed usingthe intensity information as the indication of the existence of a step. Therefore even statisticallyinsignificant events are better resolved. This trace is shown in alternating ∼64 − 10 nm steps.

a plane where the different light rays havenot converged into a single focus. The re-sulting emission pattern of the single fluo-rophore on the CCD depends on � (theaxial angle between dipole axis and opticalaxis), � (the azimuthal angle around opticalaxis), and δz (the defocusing amount fromthe focal plane) (Figure 7a). Using a pattern-match analysis technique, Patra & Enderlein(29) calculated �, �, and the lateral position ofthe fluorophore. DOPI can detect the 3D ori-entations with ∼10–15◦ uncertainty and lat-eral position with ∼15 nm accuracy when theexposure time is 0.6 to 0.8 s (for a rhodaminemolecule, ∼3 mWatt excitation power withobjective-type TIR illumination).

It is difficult to perform control experi-ments to test the angular accuracy of DOPI;therefore computer simulations are used totest the analysis algorithm. Perfect theoreticaldefocused images were computer generatedand Poissonian noise was added to each pixelof the image. The total number of photons,background level, pixel size, and other opticalparameters were all chosen according to ourexperimental setups and results, which weregenerally obtained (42). The fit results neverdeviated more than 5◦ when using the algo-rithm described by Patra & Enderlein (29).These simulations also showed that the sensi-tivity of DOPI for in-plane angles is higherwhen the out-of-plane angle of the dipole


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

is larger (i.e., more parallel to the specimenplane) (Figure 7b,c). DOPI has no degen-eracy in measuring the orientation of thedipole axis within any preselected hemisphere,but the bidirectional symmetry of theoptical dipoles still presents an ambiguitybetween (�, �) and (180◦ − �, � + 180◦)for any dipole.

An objective-type TIR setup was usedto excite molecules, and the excitation laserwas circularly polarized to excite all the dyeswith maximum efficiency regardless of dyeorientations. Choosing the right objectiveand the right effective pixel size is essen-tial to observe defocused images. We used aninfinity-corrected 100X 1.45NA oil objective(Olympus) and a 1.6X magnification unit tocollect more photons and have a proper effec-tive pixel size (total magnification was 160X,resulting in a 100-nm effective pixel size).Having pixel sizes smaller than 100 nm is pos-sible, but increasing magnification decreasesthe intensity of the fluorophore dramatically.We started our experiments by looking atDiI molecules that were either dissolved in apolymer (methyl methacrylate or poly vinylalcohol) or methanol. The dyes (∼10 nM)were spin coated (3000 rpm for 30 s) ona precleaned coverslip (0.17-mm thickness)and then dried in an oven at 100◦C for 1min. For this setup and the emission wave-length (575 nm for BR), the optimal defo-cusing amount was ∼500 nm. It is essentialthat the bond between the dye and the tar-get molecule is rigid. We did not see anyuseful defocused patterns for Cy3 dyes thatwere attached to DNA with a single bond.The defocusing of the sample was done us-ing a sample-holding piezoelectric stage (MadCity Labs Inc.). (It is not essential to startthese experiments with a piezo stage; onecan easily do the preliminary experiments bymanually adjusting the focusing knob of themicroscope.) By these experiments we op-timized our imaging and excitation opticsto ensure that we did not have any severeaberrations that could adversely affect ourexperiments.

Figure 7(a) The spherical coordinate system used in ourcalculations. � is the out-of-plane angle and � isthe in-plane angle. (b) We tested the precision ofour fitting algorithm by creating theoreticaldefocused images; the optical parameters, signal,and noise level were chosen according to ourexperimental setup. For a fixed in-plane angle wegenerated images with 3◦ � increments, and wefitted those images to theoretical patterns. Thedeviations of the � from �fit are plotted against �values. These deviations are usually less than 2◦ (asshown). (c) The same procedure in panel b isrepeated for in-plane angles �. For fixed � values(15◦, 30◦, 45◦, 60◦, 75◦, 90◦) we calculated averagedeviations of � from �fit as a function of � values.The sensitivity of our algorithm for in-plane angleis enhanced when the dipole is more parallel to thesurface.


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

-2

0 30 60 90

-1

0

1

2

< φ φ – φ fit> (°)

b

c

0 30 60 900

5

10

θ (°)

θ θ – θfit (°)

φ = 0°

θ (°)Figure 7(Continued )

Studying Lever Arm Dynamicsof Myosin V with DOPI

Evidence from previous studies (15, 44) in-dicates that the lever arm of myosin V hastwo well-defined states (trailing and lead-ing) during ATP-dependent translocation andtherefore was a perfect choice to demon-strate the capabilities of DOPI for biologi-cal studies. We labeled one of the six CaMsof myosin V with a BR. The orientation ofthe dye’s dipole with respect to the leverarm is ∼40◦. However, the azimuthal angleof the dipole around the lever arm is notknown. Also there was evidence that each

CaM had a different orientation dependingon its location on the lever arm. All thesevariables made using the lab coordinate frameunsuitable for an unambiguous analysis. In-stead we used an actin-based coordinate sys-tem (Figure 8a) to map everything on theactin frame (15). We defined two angles, α(the azimuthal angle around the actin) andβ (the axial angle with respect to the actinfilament).

We captured defocused images using ex-posure times between 0.6 and 0.8 s in thepresence of ∼300 nM ATP. The exposuretime, ATP concentration, and excitation laser


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

power (∼3 mWatts measured before theobjective) were optimized to capture four tofive frames for each dwelling period of myosinV. Every myosin molecule that moved un-derwent the expected lever arm swing. Thisfinding contradicts the results reported byForkey et al. (15). We observed 65 molecules(which walked on a linear track, sparse enoughfor analysis, and moved processively for manysteps), all of which showed well-defined leverarm rotations (Figure 8b). The average βchanges were ∼71◦, which were consistentwith electron microscopy results (44). Weobserved ∼27◦ azimuthal back-and-forth ro-tations (α changes) around the actin foralmost 90% of the molecules. These rota-tions were correlated, or anticorrelated, with

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 8(a) The actin-based coordinate system adaptedfrom Forkey et al. (15). This coordinate system isused for a less ambiguous analysis. α is theazimuthal angle between the dipole axis around theactin, whereas β is the axial angle with respect tothe actin filament. (Also see Figure 1c and Figure3a.) (b) The histogram of observed β angles whilemyosin V molecules were walking along the actin.The mean angles are 57◦ and 128◦ for the trailingand leading lever arms, respectively. (c) A sampletrace of a myosin V molecule that was imaged byswitching between focused and defocused imaging.The exposure time is 0.66 s per frame. We havecaptured repeated cycles of five consecutivedefocused images and three consecutive focusedimages. The sample is moved away from the bestfocus by 500 nm. Blue circles are the raw positiondata analyzed by FIONA, and black lines are theaveraged position within each dwelling period.Red diamonds are raw β values analyzed by DOPIand red lines are the dwell-averaged β values.Green triangles are raw α values and green linesare the dwell-averaged α values. The patternsabove the graph are representative defocusedimages for each dwell time and the correspondingtheoretically calculated patterns. (d ) A sampletrace of a myosin V molecule that was imaged bypure defocused imaging (DOPI). The exposuretime is 0.75 s and the sample is moved away fromthe focus by 500 nm.

lever arm tiltings with respect to the actin(β changes).

However, the position accuracy of DOPIwas usually not enough to measure step sizesat the same time. We measured the step sizesof myosin V by using a simple trick to takeadvantage of FIONA. By using a synchro-nized CCD and a piezo stage, we switchedbetween focused and defocused imaging toget 3D orientations and accurate position data(Figure 8c,d ). These measurements (for 32molecules) confirmed that all lever arm tilt-ings are synchronized with stepping events atour time resolution. Accurate step size infor-mation also allowed us to distinguish trailingstates from leading states of the lever arm,because the lever arm is in the leading state


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

Figure 8(Continued )

after long steps (37 + 2x nm) and in the trail-ing state after short steps (37 − 2x nm). Weconcluded that the average dipole angle withrespect to the actin (β) was ∼57◦ for the trail-ing lever arm and ∼128◦ for the leading leverarm (Figure 8b) for the 97 molecules and1151 tilting events.

CONCLUSIONS AND FUTUREDIRECTIONS

The capabilities of modern imaging tools haveshed light on biology since the invention ofmicroscopes. Fluorescence spectroscopy andconventional microscopy still need much de-velopment, both in vitro and particularly in

vivo. The inherent limitations of microscopy,such as the diffraction limit and the photosta-bility problems of organic dyes and autofluo-rescent proteins, are still valid. Recent studiesby several groups are pushing the diffractionlimit to bring the optical resolution downto the nanometer range, although there isstill room for improvement to extend the ap-plications of the proposed methods (5, 20).These new imaging tools and future imag-ing tools will likely bring their own questionsresulting from their high-resolution observa-tions. Hopefully, these studies and the exist-ing microscopy tools will be applicable forstudying live organisms and used for medicalpurposes.


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

SUMMARY POINTS

1. Fluorophores can be localized within 1 to 2 nm accuracy using FIONA.

2. The translocation dynamics of biological molecules can be studied accurately by usingfluorescence microscopy.

3. Single-molecule fluorescent techniques such as SHREC and SHRImP can be used tomeasure distances that cannot be measured with FRET.

4. Single-molecule fluorescent techniques such as DOPI and fluorescence polarizationmicroscopy can be used to study the rotational dynamics of biological molecules.

ACKNOWLEDGMENTS

This work was supported by NIH grants AR44420 and GM068625. We thank Dr. TylerLougheed for his careful proofreading and suggestions on the text. We also thank ComertKural, Hamza Balci, Sheyum Syed, and Ahmet Yildiz for generously sharing their original datafor this review paper.

LITERATURE CITED

1. Balci H, Ha T, Sweeney HL, Selvin PR. 2005. Interhead distance measurements in myosinVI via SHRImP support a simplified hand-over-hand model. Biophys. J. 89:413–17

2. Bartko AP, Dickson RM. 1999. Imaging three-dimensional single molecule orientations.J. Phys. Chem. B. 103:11237–41

3. Bartko AP, Dickson RM. 1999. Three-dimensional orientations of polymer-bound singlemolecules. J. Phys. Chem. B. 103:3053–56

4. Bartko AP, Xu K, Dickson RM. 2002. Three-dimensional single molecule rotational dif-fusion in glassy state polymer films. Phys. Rev. Lett. 89:026101

5. Betzig E, Patterson GH, Sougrat R, Lindwasser OW, Olenych S, et al. 2006. Imagingintracellular fluorescent proteins at nanometer resolution. Science 313:1642–45

6. Block SM, Goldstein LS, Schnapp BJ. 1990. Bead movement by single kinesin moleculesstudied with optical tweezers. Nature 348:384–85

7. Böhmer M, Enderlein J. 2003. Orientation imaging of single molecules by wide-fieldepifluorescence microscopy. J. Opt. Soc. Am. B 20:554–59

8. Burgess S, Walker M, Wang F, Sellers JR, White HD, et al. 2002. The prepower strokeconformation of myosin V. J. Cell Biol. 159:983–91

9. Cheezum MK, Walker WF, Guilford WH. 2001. Quantitative comparison of algorithmsfor tracking single fluorescent particles. Biophys. J. 81:2378–8810. First

demonstration ofSHREC, ortwo-color FIONA.

10. Churchman LS, Okten Z, Rock RS, Dawson JF, Spudich JA. 2005. Single moleculehigh-resolution colocalization of Cy3 and Cy5 attached to macromolecules mea-sures intramolecular distances through time. Proc. Natl. Acad. Sci. USA 102:1419–23

11. First use ofquantum dots forin vivo studies.

11. Dahan M, Levi S, Luccardini C, Rostaing P, Riveau B, Triller A. 2003. Diffusiondynamics of glycine receptors revealed by single-quantum dot tracking. Science302:442–45


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

12. Dyba M, Hell SW. 2002. Focal spots of size lambda/23 open up far-field fluorescencemicroscopy at 33 nm axial resolution. Phys. Rev. Lett. 88:163901

13. Forkey JN, Quinlan ME, Goldman YE. 2000. Protein structural dynamics by single-molecule fluorescence polarization. Prog. Biophys. Mol. Biol. 74:1–35

14. Forkey JN, Quinlan ME, Goldman YE. 2005. Measurement of single macromoleculeorientation by total internal reflection fluorescence polarization microscopy. Biophys. J.89:1261–71

15. Firstexperimentmeasuringrotational dynamicsat thesingle-moleculelevel.

15. Forkey JN, Quinlan ME, Shaw MA, Corrie JE, Goldman YE. 2003. Three-dimensional structural dynamics of myosin V by single-molecule fluorescence po-larization. Nature 422:399–404

16. Geerts H, De Brabander M, Nuydens R, Geuens S, Moeremans M, et al. 1987. Nanovidtracking: a new automatic method for the study of mobility in living cells based on colloidalgold and video microscopy. Biophys. J. 52:775–82

17. Gelles J, Schnapp BJ, Sheetz MP. 1988. Tracking kinesin-driven movements withnanometre-scale precision. Nature 331:450–53

18. Describesmeasuring smalldistances usingSHRImP.

18. Gordon MP, Ha T, Selvin PR. 2004. Single-molecule high-resolution imaging withphotobleaching. Proc. Natl. Acad. Sci. USA 101:6462–65

19. Ha T, Enderle T, Chemla DS, Selvin PR, Weiss S. 1996. Single molecule dynamics studiedby polarization modulation. Phys. Rev. Lett. 77:3979–82

20. Hofmann M, Eggeling C, Jakobs S, Hell SW. 2005. Breaking the diffraction barrier in flu-orescence microscopy at low light intensities by using reversibly photoswitchable proteins.Proc. Natl. Acad. Sci. USA 102:17565–69

21. Kinosita K Jr, Itoh H, Ishiwata S, Hirano K, Nishizaka T, Hayakawa T. 1991. Dual-viewmicroscopy with a single camera: real-time imaging of molecular orientations and calcium.J. Cell Biol. 115:67–73

22. First example ofin vivo FIONA withhigh positionalaccuracy andtemporalresolution.

22. Kural C, Kim H, Syed S, Goshima G, Gelfand VI, Selvin PR. 2005. Kinesin anddynein move a peroxisome in vivo: a tug-of-war or coordinated movement? Science308:1469–72

23. Larson DR, Zipfel WR, Williams RM, Clark SW, Bruchez MP, et al. 2003. Water-soluble quantum dots for multiphoton fluorescence imaging in vivo. Science 300:1434–36

24. Mallik R, Carter BC, Lex SA, King SJ, Gross SP. 2004. Cytoplasmic dynein functions asa gear in response to load. Nature 427:649–52

25. Michalet X, Pinaud FF, Bentolila LA, Tsay JM, Doose S, et al. 2005. Quantum dots forlive cells, in vivo imaging, and diagnostics. Science 307:538–44

26. Nan X, Sims PA, Chen P, Xie XS. 2005. Observation of individual microtubule mo-tor steps in living cells with endocytosed quantum dots. J. Phys. Chem. B 109:24220–24

27. Ökten Z, Churchman LS, Rock RS, Spudich JA. 2004. Myosin VI walks hand-over-handalong actin. Nat. Struct. Mol. Biol. 11(9):884–87

28. Park H, Ramamurthy B, Travaglia M, Safer D, Chen LQ, et al. 2006. Full-length myosinVI dimerizes and moves processively along actin filaments upon monomer clustering. Mol.Cell 21:331–36

29. Patra D, Gregor I, Enderlein J. 2004. Image analysis of defocused single-moleculeimages for three-dimensional molecule orientation studies. Phys. Chem. A 108:6836–41

30. Prummer M, Sick B, Hecht B, Wild UP. 2003. Three-dimensional optical polarizationtomography of single molecules. J. Chem. Phys. 118:9824–29


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

31. Rasnik I, McKinney SA, Ha T. 2006. Nonblinking and long-lasting single-molecule flu-orescence imaging. Nat. Methods 3:891–93

32. Reck-Peterson SL, Yildiz A, Carter AP, Gennerich A, Zhang N, Vale RD. 2006.Single-molecule analysis of dynein processivity and stepping behavior. Cell 126:335–48

33. Rief M, Rock RS, Mehta AD, Mooseker MS, Cheney RE, Spudich JA. 2000. Myosin-Vstepping kinetics: a molecular model for processivity. Proc. Natl. Acad. Sci. USA 97:9482–86

34. Rock RS, Rice SE, Wells AL, Purcell TJ, Spudich JA, Sweeney HL. 2001. Myosin VI is aprocessive motor with a large step size. Proc. Natl. Acad. Sci. USA 98:13655–59

35. Sakamoto T, Yildiz A, Selvin PR, Sellers JR. 2005. Step-size is determined by neck lengthin myosin V. Biochemistry 44:16203–10

36. Sheets ED, Lee GM, Simson R, Jacobson K. 1997. Transient confinement of aglycosylphosphatidylinositol-anchored protein in the plasma membrane. Biochemistry36:12449–58

37. Snyder GE, Sakamoto T, Hammer JA 3rd, Sellers J, Selvin PR. 2004. Nanometer local-ization of single green fluorescent proteins: evidence that myosin V walks hand-over-handvia telemark configuration. Biophys. J. 87:1776–83

38. Sosa H, Peterman EJ, Moerner WE, Goldstein LS. 2001. ADP-induced rocking of thekinesin motor domain revealed by single-molecule fluorescence polarization microscopy.Nat. Struct. Biol. 8:540–44

39. Syed S, Snyder GE, Franzini-Armstrong C, Selvin PR, Goldman YE. 2006. Adaptabilityof myosin V studied by simultaneous detection of position and orientation. EMBO J.25:1795–80340. First

characterization ofthe possiblereasons affectingthe localization ofthe dyes.

40. Thompson RE, Larson DR, Webb WW. 2002. Precise nanometer localization anal-ysis for individual fluorescent probes. Biophys. J. 82:2775–83

41. Toba S, Watanabe TM, Yamaguchi-Okimoto L, Toyoshima YY, Higuchi H. 2006. Over-lapping hand-over-hand mechanism of single molecular motility of cytoplasmic dynein.Proc. Natl. Acad. Sci. USA 103:5741–45

42. Defocusedimaging isdescribed and it isthe first applicationof DOPI forbiologicalmolecules.

42. Toprak E, Enderlein J, Syed S, McKinney SA, Petschek RG, et al. 2006. Defocusedorientation and position imaging (DOPI) of myosin V. Proc. Natl. Acad. Sci. USA103:6495–99

43. Valentine MT, Fordyce PM, Krzysiak TC, Gilbert SP, Block SM. 2006. Individual dimersof the mitotic kinesin motor Eg5 step processively and support substantial loads in vitro.Nat. Cell Biol. 8:470–76

44. Walker ML, Burgess SA, Sellers JR, Wang F, Hammer JA 3rd, et al. 2000. Two-headedbinding of a processive myosin to F-actin. Nature 405:804–7

45. Warshaw DM, Kennedy GG, Work SS, Krementsova EB, Beck S, Trybus KM. 2005.Differential labeling of myosin V heads with quantum dots allows direct visualization ofhand-over-hand processivity. Biophys. J. 88:L30–32

46. Willig K, Kellner R, Medda R, Hein B, Jakobs S, Hell S. 2006. Nanoscale resolution inGFP-based microscopy. Nat. Methods 3:721–23

47. Willig KI, Rizzoli SO, Westphal V, Jahn R, Hell SW. 2006. STED microscopy revealsthat synaptotagmin remains clustered after synaptic vesicle exocytosis. Nature 440:935–39

48. Yasuda R, Noji H, Kinosita K Jr, Motojima F, Yoshida M. 1997. Rotation of the gammasubunit in F1-ATPase; evidence that ATP synthase is a rotary motor enzyme. J. Bioenerg.Biomembr. 29:207–9


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

ANRV311-BB36-17 ARI 3 April 2007 15:17

49. FIONA isdescribed in detail;the walkingmechanism ofmyosin V isdiscovered.

49. Yildiz A, Forkey JN, McKinney SA, Ha T, Goldman YE, Selvin PR. 2003. Myosin Vwalks hand-over-hand: single fluorophore imaging with 1.5-nm localization. Science300:2061–65

50. Yildiz A, Park H, Safer D, Yang Z, Chen LQ, et al. 2004. Myosin VI steps via a hand-over-hand mechanism with its lever arm undergoing fluctuations when attached to actin.J. Biol. Chem. 279:37223–26

51. Yildiz A, Tomishige M, Vale RD, Selvin PR. 2004. Kinesin walks hand-over-hand. Science303:676–78


Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

Contents ARI 12 April 2006 20:14

Annual Reviewof Biophysics andBiomolecularStructure

Volume 35, 2006Contents

FrontispieceMartin Karplus � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �xii

Spinach on the Ceiling: A Theoretical Chemist’s Return to BiologyMartin Karplus � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Computer-Based Design of Novel Protein StructuresGlenn L. Butterfoss and Brian Kuhlman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �49

Lessons from Lactose PermeaseLan Guan and H. Ronald Kaback � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �67

Evolutionary Relationships and Structural Mechanisms of AAA+ProteinsJan P. Erzberger and James M. Berger � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �93

Symmetry, Form, and Shape: Guiding Principles for Robustness inMacromolecular MachinesFlorence Tama and Charles L. Brooks, III � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 115

Fusion Pores and Fusion Machines in Ca2+-Triggered ExocytosisMeyer B. Jackson and Edwin R. Chapman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 135

RNA Folding During TranscriptionTao Pan and Tobin Sosnick � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 161

Roles of Bilayer Material Properties in Function and Distribution ofMembrane ProteinsThomas J. McIntosh and Sidney A. Simon � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 177

Electron Tomography of Membrane-Bound Cellular OrganellesTerrence G. Frey, Guy A. Perkins, and Mark H. Ellisman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 199

Expanding the Genetic CodeLei Wang, Jianming Xie, and Peter G. Schultz � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 225

Radiolytic Protein Footprinting with Mass Spectrometry to Probe theStructure of Macromolecular ComplexesKeiji Takamoto and Mark R. Chance � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 251

v

Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

Contents ARI 12 April 2006 20:14

The ESCRT Complexes: Structure and Mechanism of aMembrane-Trafficking NetworkJames H. Hurley and Scott D. Emr � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 277

Ribosome Dynamics: Insights from Atomic Structure Modeling intoCryo-Electron Microscopy MapsKakoli Mitra and Joachim Frank � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 299

NMR Techniques for Very Large Proteins and RNAs in SolutionAndreas G. Tzakos, Christy R.R. Grace, Peter J. Lukavsky, and Roland Riek � � � � � � � � � � 319

Single-Molecule Analysis of RNA Polymerase TranscriptionLu Bai, Thomas J. Santangelo, and Michelle D. Wang � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 343

Quantitative Fluorescent Speckle Microscopy of CytoskeletonDynamicsGaudenz Danuser and Clare M. Waterman-Storer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 361

Water Mediation in Protein Folding and Molecular RecognitionYaakov Levy and José N. Onuchic � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 389

Continuous Membrane-Cytoskeleton Adhesion Requires ContinuousAccommodation to Lipid and Cytoskeleton DynamicsMichael P. Sheetz, Julia E. Sable, and Hans-Günther Döbereiner � � � � � � � � � � � � � � � � � � � � � � � 417

Cryo-Electron Microscopy of Spliceosomal ComponentsHolger Stark and Reinhard Lührmann � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 435

Mechanotransduction Involving Multimodular Proteins: ConvertingForce into Biochemical SignalsViola Vogel � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 459

INDEX

Subject Index � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 489

Cumulative Index of Contributing Authors, Volumes 31–35 � � � � � � � � � � � � � � � � � � � � � � � � � � � 509

Cumulative Index of Chapter Titles, Volumes 31–35 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 512

ERRATA

An online log of corrections to Annual Review of Biophysics and Biomolecular Structurechapters (if any, 1997 to the present) may be found athttp://biophys.annualreviews.org/errata.shtml

vi Contents

Ann

u. R

ev. B

ioph

ys. B

iom

ol. S

truc

t. 20

07.3

6:34

9-36

9. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

NIV

ER

SIT

Y O

F IL

LIN

OIS

on

05/0

4/07

. For

per

sona

l use

onl

y.

new fluorescent tools for watching nanometer-scale...

Documents