LE3

A RTIC

of theInternalBodin ATP
InquiryFigure9.15:ls theBotation
Synthase Responsible forATPSynthesis?

Introduction-The Article and PhenomenonUnder Study

ATP is made from ADP and inorganic phosphateusing the energy from proton flow through the membrane protein compler
known as ATP synthase.A particular subunit of the protein complex acts as a rotary motor in the production of AIP. Inquirr
Figure 9. 15 in Campbell/ReeceBiology, Eighth Edition, showshow investigatorsprovided direct evidencefor this mechanism
of ATP synthesisby meansof rotation in the following article:
Yoshida,andKazuhikoKinositaJr.,
KengoAdachi,Hiroyuki Noji, RyoheiYasuda,Masasuke
HiroyasuItoh,Akira Takahashi,
MechanicallydrivenATPsynthesis
by F1-AIPase,Nature421:465168(29 January 2004).

for ReadingThisArticle
Guiding Questions
A. AbouttheArticle
1. What is the name of the first authorof this researchteam?

2. Most of the authorsof this paper work in major researchlaboratoriesin what nation?What is the name of the author af-
filiated with the Cold Spring Harbor Laboratory in the United States?

3 . What is the nameof thej ournalin which this article was published?This paperwas publishedin what month and yearI
1-l Inquiry in Action: Interpreting Scientific Papers

l. This research was supported in part by grants from what organization? (See Acknowledgements, p. a68.)

5. The afticle's abstract(p. 465, first paragraph,in bold print) summarizesthe major contributions of the study.Write the
one sentencefrom the abstractthat statesthe basic concept for which the study provides direct evidence.

6. Specializedvocabulary:Write a brief definition of eachterm.

chemiluminescence

hydrolysis

luciferase

photon

B. About the Study

7. Was this an in vivo study (in intact cells, in living organisms)or an in vitro study (literally, "in glass,"meaningin test
tubes or in laboratorv containers)?

8. The action of isolatedF1 parliclesof AIP synthaseappearedto be reversible.(a) The centralsubunitrotatesanticlock-

wise (that is, counterclockwise)when what happens-A or B? (A) AIP is synthesized;(B) ATP is hydrolyzed(broken
down to ADp and inorganic P). (b) The central subunit rotatescloclcwisewhen what happens-A or B? (A) ATP is syn-
thesized;(B) ATP is hydrolyzed(brokendown to ADP and inorganicP).

9. The researchersmanipulated a single subunit of the AIP synthasecomplex, the y (gamma) subunit' Prior to this study,
what aspectof the y subunitin Fl reversalhad not yet beentested?
 Article 3: Inquiry Figure 9.15

10. One end of the Fl test complex was attachedto a glassslide by modifying two cysteineamino acidsof one end of the
complex. The other end of the F1 test complex, the y end, was attachedto a magnetic bead using a specific attachment
protein (streptavidin).Examine the experimental setupin Figure | (p. 466).Although the drawing in Figure 1a is not to
scale,which is the larger element, the Fl protein complex or the magnetic bead?How did the investigatorsmanipulate
this systemto causethe y end of the complex to rotate?

11. The investigatorsnoted that during ATP hydrolysis,the complex rotated in one direction.What reactiondid they hy-
pothesizewould occur if the complexwas rotatedin the oppositedirection?

12. The investigatorsused photons of light produced in a chemiluminescentluciferin-luciferase system as an indicator of

ATP synthesis.When AIP is present,this enzyme systemwill hydrolyze AIP to produceADP and inorganic phosphate
and releasephotons of light. (Thus, it serves as a "reporler enzyme," as mentioned in Inquiry Figure 9.15 of
Campbell/ReeceBiology,Eighth Edition.) What doesthe y-axis on the left side of Figures3a and 3b represent?This is
an indirect measureof the rate of which rotation-driven process-AlP synthesisor ATP hydrolysis?

13. Figure 3a showsthe numberof photonsdetectedwith respectto rotation of the y-subunitof the ATP synthase.N, S, and
H refer to what three rotation states?In Figure 3a, the relative number of photons detectedis highest at which type of
rotation-N. S. or H?

14. In the control experimentsgraphed in Figure 3b, the broken lines show the lowest number of photons detected.What
was different aboutthoseexperimentsto producesuch low numbers?Why did this omissionserveas a control for the
experiment?
16 Inquirl- in Action: Interpreting ScientificPapers
C. General Conclusionsand Extensionsof the Work
15. What new information about the role of the y sub-
unit in AIP synthasedid this researchproject eluci-
date? How does the subunit serve as a reversible
motor?
16. What future tasks do the authorswish to do once
they understandthis motor mechanismfully?
letters to nature
Mechanically driuenAIP
synffie$isby F1-ATPase
Hiruyasu ltohl'2, Akira Takahashl3, Kengo Adaofil4, Hiroyuki l{oil5,
Rydrei Yasuda6, Masasuke YchidaT & lhanhiko Kinoslta Jl
rTsukubaResearchLaboratory,Hattdmatsu Photonics KK, and2CILEST
"Creation and application of soft nano-machinz,the lryperfunctionalmolecular
machine" Team lj+, Tokodai, Tsukuba3a0-2635,Japan
3
SystemDivision, HamamatsuPhotonics KK, Joko,Hamamatsu 431-3103,Japan
"Cmter for Integrative Bioscience,Okazaki National ResearchInstihttes,
Okamki 444-8585, Japan
5-
'lnsritute of Indwtial Stience,Univqsity ofTolcyo,Takyo 153-85a5,Japan
6Cold
Spring Harbar Laboratory CoId Spring Harbor, New York 11724 tlSA
7ERATO "ATP System", 580a4 Nagatsuu, Yokohama226-a026,
lapan
ATB the main biological energy currency, is synthesized &om
ADP and inorganic phosphate by AIP synthase in an energy-
requiring reactionl-3. The F1 portion of AIP synthase, also
known as F1-ATPase, functions as a rotary molecular motor:
irc vitra its 1-subunit rotatesa against the surrounding o3$3
subunitss, hydrolysing ATP in three separate catalytic sites on
the $-subunits. It is widely believed that reverse rotation of tfre
1-subunit, driven by proton flow through the associated Fo
portion of ATP synthase, leads to ATP synthesis in biological
systemsl-3'5'7.Here we present direct evidence for the chemical
synthesis of AIP driven by mechanical €nergy. We attached a
magnetic bead to the 1-subunit of isolated F1 on a glass surface,
and rotated the bead using electrical magnets. Rotation in the
appropriate direction resulted in the appearance of AfP in the
medium as detected by the luciferase-luciferin reaction. This
shows that a vectorial force (torque) working at one particular
point on a protein machine can influence a chemical reaction
occurring in physically remote catalytic sites, driving tlre reac-
tion far from equilitrrium.
When isolated F1 hydrolyses AIB its central 1-subunit rotates
anticioclrsdseawhen viewed from above in Fig. 1a,with an efficiency
of chemical-to-mechanical energy conversion approaching 1000/o
{ref. 8). The purpose of this study w'as to show that the chemo-
mechanical coupling in the F1 motor is completely reversible, and
that reversal is achieved by manipulating a single variable-that is,
the rotary angle of the 1-subunit" Any molecular machine would be
reversible if one could manipulate all constituent atoms at will.
Whether one or a few thermodynamic handles exist in a chemo-
mechanical molecular machine such that its operation can be
controlled through that handle in both directions is aa important
but unresolved issue. For example, rvhether one can synthesizeATP
by pulling back a iinear molecular motor such as myosin or
kinesin-and if so, where to pull-is unknown. Reversal of the
whole AIP synthase is well documentedl'e, including the demon-
stration of 1-subunit reorientation under sytrthesis conditionsto,
but whether or not the 1-subunit angle servesas a single handle for
F1 reversalhas not been tested.
To prove this reversibility, we used a3$31, the minimal subcom-
plex of F1 that shows AlP-catalysed rotationa. The subcomplex was
attached to a glasssurface through histidine residues engineered at
the amino terminus of the $-subunits, and a magnetic bead coated
with streptavidin was attached to the 1-subunit, which had been
biotinylated at two engineered cysteines (Fig. 1a). The beads were
rotated with magnets (Fig. 1b-d) in a medium containing ADP and
phosphate as substrates and the luciferin-luciferase syst"mlr'r2,
which emits a photon when it captures and hydrolyses ATP. The
initial idea was to count thesechemiluminescentphotons (Fig. lb);
however, background luminescence originating from contaminant
AIP present in ADP evenafter purification was a problem. Thus, the

letters to nature
.,,olumeof medium per active F1 molecule had to be small'
First, we tried to reduce the volume by making microdroplets in
oii (Fig. 2a). Figure 2b shows data fiom a 4 x 4 array of droplets in
one chimber. The beadsin droplets were rotated at 10Hz alternately
for 5 min each in the direction of hydrolysis (anticlockwise when
viewed from top irl Fig. 1a) and synthesis(clockwise)' As seenin
Fig. 2b, 14 out of 16 droplet cuwes showedthe M-shaped pattem of
photon counts exPectedfor the sequenceofrotation direction when
ihe overall decline was taken into account. The decline was due to
the gradual disappearanceof the aqueous phase into oii: although
we iaturated the oil with water before the experiment, droplets
tended to shrink crver time. For random photon counts, the
probability of observing a sianting M shape is 8 '. The probability
ofobserving14ormore M shapesoutof 16is 2 x 10-11.The dataset
shown in Fig. 2b thus strongly indicates mechanical synthesis' Th-e
experiment is extremely difficult (at most a ferv beads rotate in each
droplet), however, and we have obtained only a few more data sets
that contained severalM-shaped patterns.
We thus tried to increase the number of rotating beads in an
ordinary obserwation chamber (Fig. 1c) by infusing a concentrated
solution ofbeads carrying F1.To aliow F1 to rotate in the proper
direction, we derivatized only the bottom surface of the chamber
with nickel nitrilo-triacetic acid (Ni2+-NTA), which would specifi-
cally bind the $-histidines. In control experiments done in 4mM
AIP (no ADP) and without magnets,we found in the fieid of view of
1.0x 105pmt us many as 480 | 70 F1 molecules rotating anti-
clockwise at the bottom {three chambers). However, the high
density of beads resulted in nonspecific binding to the ceiling,
where 100 + 20 beads rotated clockvrise (as viewed from above the
chamber). We also tested in the AIP medium rvhether forced
rotation by external magnets would damage F1. After confirming
AlP-<lriven rotation, we turned on the magrets and applied several
bursts of hundreds of revolutions at l0 Hz in both directions'When
Magn€tie bead
$lrertavidjn *:
Gal}i1lcompiex
li:_lis,
boMaonet
%8
g6pp16._=-.--i1-";''n:-
ffi
Ob je ctivele n s._| I
-F- I
l'ij.
r.i"*
Ph+tcncounlingcaFr€ra;ncoalinghouse
Figu re1 Exp erin r ent als et - up' a, Bas ic des ign' The s t r u c t u r e s o {F 1 a n d s t r e p t a v i d i n a r e N i 2 *- N T A v l a s a p p | i e d o n | y t o th e b o tto m su a ce '
fromref.2landrcL22.respectively.Thebeadisnottoscaleandtheorientationof
isunknown.
streptavidin b, Sideview of the system.
optical c, Observation chamDer.
Article 3: Inquirl'Figure 9.15 li
the magnets were tumed off' AlP-driven rotation resumed'
Si'nthesiswas shown in the ADP-luciferase medium by accumu-
iating chemiluminescence photons over a series of 5-min intervals
in wfrich the magnetic field was rotated at 10Hz in either direction
or tumed off. all seriesproduced the expectedpattern (Fig' 3a):
higher photon counts during ciockwise rotation (S, synthesis) than
cluling anticlockwiserotation (H, hydrolysis) or no rotation (N)'
This graph compiles ali data taken in consecutive experiments
(aUoui half of the experiments failed at some point, for exainple
during chamber preparation or because of a large focus drift, and
did nJt produce data)' Mechanical synthesisin the flat chamber was
reproducible, although variation among data rvas stiil large'
We note that irl most curves shown in Fig. 3a including the total
counts, counts during anticlockwise rotation (H) are higher than
those at no rotation (N). This is due to the presenceof F1 at the
ceiling of the chamber as stated above' For these upside-down F1
mole.-.,I.s, anticlockwise bead rotation will drive ATP s1'nthesis'
Indeed, when we flipped the chamber upside down after obtaining
the unbrokenblue ii"e in Fig. 3a' the count Patternwasreversed,as
shown by the broken blue line. Another reason for the higher counts
durilg anticlockwise rotation was that luciferase did not consume
a1lof ihe newly synthesizedAIP in 5 min: as shown by the unbroken
lines in Fig. 3b, the luminescence at no rotation lvas high after ATP
synthesis it the arrows. Luminescence decay after AIP mixing was
sirown to involve a comPonent with a lifetime of about 3 min (see
Supplementary Information). Thking aii of the above points into
aciount, we consider that mechanical synthesis has been conclu-
sively demonstrated.
Figure 4 shows the effect of rotary speed on the efficiency of
ATP synthesis. The synthesis rate apparently saturated above 3 Hz
(fig. +U). This was becauselarger beadsor bead aggregatesfailed to
rotite at high speeds,as confirmed by direct obserwation (uncou-
pling betwJen magnet and bead rotations)' Calibration of the
i'-i
i:
i, j.
d
1
I
lffase
E+
._-> Frce€ssing
system
ontheceiling.d,Topvlewofthernagnets
NATURE VOL427 29 IANUARY 2004 twnaturc'com/Earure
@2004 NaturePublishingGrouP
1E Inquiry in,\cfion : Interpreting Scientific Papers

letters to nature
Sample d.oF,ets i\ergi't 3 um. dJanelH 30 Fml
otr oa:te'ned area coated 4,th P'i, ' NTA 20Sx103
30
o

*zs 150
o
n
Er o tr
q E
100e
E 15 G
b
b10
E 50
z-

b ?sF
0

E 150
's bss
s
+
6 rcO 30

E E
*zs
d

a
Ezo
g
0
E1s
Figure2 Rotationalsynthesisinmicrodroplets. chamber
a, Observation 0nthesilanized
h {n
bottom coverglass,4 x 4 spotsseparatedby-50 pmandmeasuring 20-30pmujere ts t f#[."*;;;:
E
derivatizedwithNi2+NTA, whichrendered thespotshydrophilic,Anoillayer witha
z- : ;i-:=i:EES;lt ,
thicknessof .'3mmwasplacedontop Usingaglassmicropipette,weformedadroplet c

thesolutionwithonecontaining N H SH S H N S H S H N ]
of - 1pl conlainingF, oneachspotandthenreplaced
beads, ADP, phosphate andthe chemiiuminescencesystem b,Simultaneous observation
Thickmagenta
of 16 droplets, curvewithopencircles ihesumofallunbroken Figure3 Rotational
represenls synthesis in a flatchamber. Eachsymbol shows thenumber of
cuTvest errorbarsrepresent + 2q whereo istheexpected s d, forphotonstatlstics photonsdetectedover5mininanimageareaof 7,8x 104pm2 N,norotation;H,10-Hz
(square rooi0f thetotalcount).Thebrokencurveat thebottom represents a controlin rotationinthehydrolysis direction (forF1onthebottom ofthechamber); S,10 Hzrotation
whichanareaoutside wasimaged.
thedroplets SeeSupplementary Informationfor inthesynthesis direction. a, Results from12chambers, distinguishedbydifferent
details. colours. Broken bluecurve repfesents results
obtainedafterthechamber gtvingrisetothe
unbroken blue curve was flipped upside do\.n. Thickmagenta curuewith open circles
represents thesumof allunbroken curves;err0rbarsrepresent + 4o (>99,99%
confidence). Forclarity, somecun/es havebeen shifted withint 1,000counts
verlically
photon-to-ATP ratio (seeSupplementaryInformation) indicated a b, Control experiments. Unbroken curves showresults without
either orwith
rotation
ivnthesis rate of about five molecules of ATP per second under imposed clockwise rotation forsynthesis at 10 Hz(arrows).Broken curuesshow
rotation at 3Hz, as compared with the nine molecules of AIP per experiments carried outasina exceptthat phosphate wasomitted fromthemedium. The
second expectedfor the one AIP molecule per 120' schemeu. counts inihebroken curves arelowin comparison to theothers because phosphate
Although Fig. 4 represents our best data so far and variation tended toincrease thebackground, apparently byincreasingthepodion ofluciferaseihat
arnong chambers is large (Fig. 3a), we anticipate that the coupling reacted ATP
v,rith extremely slowly $ee Supplementary Information)
between mechanical rotation of the 1-subunit and chemical syn-
thesis is tight, at least at 1ow speeds.The significant synthesisat low
speeds (Fig. 4), which were much lower than the maximal rotary
speed of 130H2 of this motor during AIP hydrolysisr3,merits Information). If F1, or the motor enzyme myosin, is mixed with
attention. Becausethe slow rotation was at a constant speed,it is high concentrationsofADP and phosphatein the absenceofATR
likely that chemical reactions were at quasi-equilibrium at all angles' some AIP is spontaneously formed on the enzyme without input of
Rotation by ATP hydrolysis can also be made slow and at a constant energvlt-tu. This, however, is a dead-end reaction and the AIP that
speed by attaching a long actin filament to the 1-subunit8, again has been formed is not availabie in the medium: releaseof the tightly
suggestingquasi-equilibrium. The implication is that ATP slmthesis bound ATP requires an external supply of energyt'lt.
in-F1 proceeds as a straightforward reversal of the hydrolysis Here we have demonstratedrepetitivesynthesisby F1 (-10'ATP
reaction, tracking the same reaction pathway in the opposite molecules in 5min), leading to aPpearanceof the product in the
direction. On that pathway, both hydrolysis and slnthesis reactions medium. To our knowledge, this is the first accomplishinent of
are controlled by one mechanical handie' the rotary angle of the
rotary artificial chemical s1'nthesisby a vectorial force (although nature
1-subunit. The situation contrasts with the more complex presumably has been doing this for millions of years)' Pressure
motor of bacterial flageila, where rotational directions can be
could also shift a chemical equilibrium by actirg on substrates(and
switched without reversing the proton motive forcetn.
solvent); however, in our experiments the chemical equilibrium
AIP synthesis is a chemical reaction that is energetically uphill,
per seis on the side of almost complete hydrolysis, and a force on a
requiring 80-100 pN nm ofenergy under physiologicalconditions'?'
In the experiments shown here, contaminant AIP amounted to poilt remote &om substrates counters the hydrolysis reaction and
to the point offavouring s1'nthesis'Because
about I irM, impiying that roughly 30 pN nm of fiee energy was pushes the equilibrium
needed per molecule of ATP synthesized (see Supplementary r.vestill have to rely on nature's nanomachine, the F1 motor' our

NAIURE VOL 427 29.IANUARY2004 M.nature.com/nature @2004 Nature Publishing Group

letters to nature
a Microscopy
(Zeiss) Luminescenceita;
The chmber wro placed on an inverted ICM 450 microscope
(OlmPusl and
9s o collected with an oi1 immersion x60 obiective, numerical aperture 1.45
deflectedwithaprismtoafactorymadesideport(Fig 1b) Thebeamt'astbcusedrdtha
g,25 (V8070L' 64 \210:
o
o
ED Plan x2 objective (Nilon) onto a cooied photon counting amera
Photonicsi'
12 0 C4566 equipped with an Argus 50 image processingsystem; Hamamatsu
2 whichrelordedcentroidpositionsofincomingphorons(countsintheduk:-l5s'o\tr
pairs of cuslom
€rs the whole image plane). To rotate the beads,we placed thtee opposing
m"de"L".to*"gnetswithaPemailoycoreonthechamber(Fig 1d) Thethreepairswere
fie1d lin
b10 activated with a custom circuit 120'out ofphce to Produce a rotating magnetic
either direction).
E5
z Received 5 Man accePted 31 October 2003; doii1o 10l8lna1ure02212'
Biod1m 66'7lJJ19
bo I. Boyer P. D. 'Ite ATP synthase a spleedid nolecular machiae Auuu Rer'
it997).
2. Kinosib,K.Jr,Yduda,R,Noii'H.&Adachi,KAtoktnolecularmotortiat'3rworkatleat1009t
tro efficiency. Plril. Trare. R- Soc.Lond. B 355, 173-189 (2000)
d maflellous rottry ergine ofthe cejl Naare
u 3 . Yoshida, M., Munenrli, E & Hisaboti' T AlP sFrthu+a
o
o Rev.Mol. Ce\| Biol. 2, 669-677 (2001)
.1. Noji,H.,Yauda,R,Yoshida,M &Kinosita'K lrDiredobsetrationoftherotationofFl-ATPase
o \dtarp J86, 2qa 102 ( lo97l.
G lV,LufteLR &walkerI E Structureat2SAr6olutionofFl ATPse
5, Abrahms,l.P,L6ligA
(laq4)
ftom boune hean milo.hondn& Ndtore J70 02l b28
6. Boyer,PD.&KohlbrennelWE inEnetgyCouplirginPhorosythais(edsSelman'B R &
!
E
z
'a Selman Reimer S ) 231-240 (Elsevier Asterdm' 1981)
machines of living cells'
7. ooswa F. & Hayshi, S The loose coupling mechanism in molecular
n!
-o 2468 l0 Adr.BioPhy\.22, r51 183 (1986)
Rate of rotatim (Hz) 8.l-auda,R,Noji,H.,Kinosita,KIr&lbshida'M,FlAlPaseisahighlyef6'ientnolecularmotorthat
rotatd with dis.reie 120" stePs.Cell 93, 1I 17 1121 (198)
Figure4 Dependence 0fsyntresis ontherateof magnet
efficiency a, Photons 9. Tirdna,P,Samoray,D.&Grdber,PH+/AIPratioofProtontansportcouPledAlPslnthesrsand
rotation.
.1 .1 pm2in5 minperiods atvari6usrotaryspeeds hy&olysis etalysed bv CF6F1-liposoma. E&8O J 22' 41v126 12003)
overanimage
detected areao{ .0 x 05
slnthme during
(shownin Hz;negativenumbers indicate inthehydro|ysis
rotation b, Photon 10. Zhou, y, Dmcan, T. NI. & cross, R, L. submit lobtion in Esdlerichia aliFoFl NP
direction)' (1997)'
oxidative phosphoryl ation Proc. Ndl A.ad. Sct' t/SA 94, 10583 10587
increments From
rotation.
cluring thephoton count foreach in
rotation
synthesis a, the md
11. M.Elroy, W' D., Seligs, H H &White, E. H. Mechanism ofbioluninescence"hedilminscence
averago counts
no-ro.tation
of theacijacent wassubtracted.Syrnbolsrntheparentheses entfoefundionintheoridationoffireflyi\cifern Photuchm'Photobiol l0' 153 170(1959)

showthelasthvomeasurements {0rsynthesis forthesepresumably 12.Hattoi,N,,Kairyma,N,MaedaM.&Mutakami,S

in a; lowvalues Mutantluciferaseegmsftomfiteflidwith

increcedrci$anetobenzalkoniumcbloride Biosci Biotahnol Biochem 66'2587 2593Q0A2)

reflect deterioration.
sample RNlulionofdi$inctroiationalsubsteps
13.YsudaR,Noii,H,YoshidaM.,Kirosiu'K h&itoh'H
(200 1) '
by submillisecond kinetic anallais of F1 AlPase Nature 410, B98-904
T2' 19 54 (2003)
14. Ber$ H, C. The rotary motot ofbaderial flagelTaAnau' Ra Biochm
primary goal is to understand fully how it works,and thereby to 15. Yoshida. M. The snthesis of en4me bound AIP by the F1 AIPse fiom
the thernophilic badenun
907-912 (1983)
PS3 ir 50o/odimethyls|;lfarjde Biochm. Biophys Res'Commun l14'
exploit the mechanism in artificial ways. ImProving the present soluble F1-AlPm
16. Sal(amoto, I. Effea of dimethylsulfoxide on ATP slntheis by mitochon&ial
system for more quantitative assays,such as the-torque and speed I. Biochm.96, 183 187 (1984)
d.p.nd.n.. of the coupling efficiency, that can be compared with 17. wolcoft, R. G & Bo1-er,P D The reversal ofthe myosin and adon)'osin
ATPase redions md the fie

theoretical predictionst'le should be our next task' The key is to enaSv of AIP binding to myosin Biotleu Biophys Res'Contuut 57' 7og
716 ( ra74)

obtair magnetic beads that are small and uniform in size' n 18, MmnheqH.G.,Schenck,H &Goody,R s SynrhesisofATPfiomADPmdinorgmicphosPhateat
(1974)
the myosin-subftagment I adive site Ezf, / Biachetu 48,287 295
slnthm N|tu|e 396' 279-282
19. Wang" H. & Ostel G Energy bansductiol in the F1 flotor ofAlP
Methods f1998).
is not irfuenced by subsitution
Materials 20. Oishi, N. & Sugi, H Ift riho AIP dePendent F anin sliding on myosin
(199'1)
I210C or removal ofbound nucleotid,e Biochitu BioPhys''4caa 1185' 346-349
A mutant q3031 subcomplex (comprising Cl93S o , Histo-S-, and S107C' I1-AlPase with
subunits; ;eierred to Fr in this paper), derived ftom the themophilic Bacillro strain 2L Menz, R. L, Wa.lker I E & Lsliq A G w' Strudue of bovine mitochondrial
^/ "s (ref 8) Streptavidin-coated nucleotide bound to all tluee catallti. sitsr inplicaions for tie nechoism of rotary catalysis Cell
PS3, wro biotinylated at the only cl'steineson the 1 subunit
magnetic beads (Seradln; nominally 0 7 pm) were lighdy centifuged to remove large 106,331-341 (2001).
R E shdural studis of the
beais and aggregates(but elimination nas incomplete) Biotinylated Ft
(400PM) w6 22. Freitag S., Tiong I L, K1umb, L, SbFos' P S & StenkmP'
(1997)
(N-morpholino)propanesulphonic streptavidin binding looP Ptoz;n S.j. 6' 1157-1166
ircrb"t"d-ithb""ds(-50pM)inbufferA(50mM3
with buffer
acid/KOH, 20 mM KrSOa 4 mM MgSOa, pH 7'6) for 10 min at 23 "C, wmhed
Kikkoman Co We
A, md concentrated with a magnet. Luciferaser2was a kind gift from Supplementary Inft rmaton accompmies the Paper on www'natrlre'Gfln/nature'
(App)ied
purified ADP (K-salt; Sigma) as described'oon a Poros HQIL coiumn
Photonia KK who
Biosystems). Acknowledgements We thmt T. Hayakawa md T' Hiruma of Hmmatsu
for the idea of using
allowed H.I. to work on this project for more than 6 learsj S Brenrer
chamber
0bservation microdropletsi M. Sugai for initial work; M. Shio, menbers ofthe
fomer CREST Team 1l and the
Mizuoo' K Suzuki' S Llchiyama
Forthechambershominlig. Ic,a32/.24mm2coverglasswasfunctionalizedbyasilane curent Kirosita and Yosh'ida labomtorie for help and advice; I
reacted first with and H N'fiyaiima for the magneic
coupling agent with m SH group (TSL 83801GE Toshiba Silicone)'and and Y. Mizuguchi for the photon-counting system; C Gosse
i paallel
ro rril maleimide Cr-NTe-1oojindo) and then with i0 mM NiCl, Two tweas; K. Abe and K. Rikukawa for nicroscopy; S Murakami for lucifercq and H' Umsaw
-g of Lumirror polyester film (TORAY) were placed on the covergis' and a
strips and M. Fukatsu for labomtory management. This rvork was supPofied
in Part by Gnnts-in-Aid
top to form a
t8 x ts mmz cot"rglus coated with hexamethyldisilazanewas placed on from the MinistryofEduetion, Cultue, SPo{is, Science andTqlhrologyof lapan' and Bunoughs
Ar (buftbr A
flow chmber. We inft*d -soo plt t1-conjugated magnetic beads in buffer Wellcome Fund (R.Y).
3 mgml-1 bovine serum albumin (BSA)) md incubated the chamber tbr
containing
10 pM
:O min at"z:'C. efter infusing buffer A: severaltimes, buffer A containing competingfinancial
1
BSA' together CompetingIntefestsstatsnent The authorsdedre that they haveno
luciferase, 1 mM luciferin, i0 mM K2Poa, 200 pM ADP and 1 5 mg ml
eventuallysere 1tr1erss.
with a small number of 3-pm DlnabeadsM-280 beads(Dynal) thatwould
upper coverglas
as spacers,wx infused. The spacerstrips were arefully removed.md rhe
the spacerbeads' Conespondence ud requsts for malerialsshouldbe addrssed to H I
*^ p."rrud ao reduce the chmber height to -3 pm, as determined by
(hiiloh@hpk.trc-net.cojP).
The chmter wc then sealed with wu and subjected to observation'

NAIURE VOL 427 29 .IANUARY2004 Mnature con/narure

464 @20M NaturePublishingGrouP