Full text of DOI:10.1007/s003400000454

Appl. Phys. B 71, 747–753 (2000) / Digital Object Identifier (DOI) 10.1007/s003400000454
Applied Physics B
Lasers
and Optics
Optical tweezers for confocal microscopy
A. Hoffmann∗, G. Meyer zu Hörste, G. Pilarczyk, S. Monajembashi, V. Uhl, K.O. Greulich
Institut für Molekulare Biotechnologie e.V. (IMB), Abt. Einzelzell- und Einzelmolekültechniken, Beutenbergstrasse 11, 07745 Jena, Germany
(Fax: +49-3641/656-410, E-mail: hoan@imb-jena.de)
Received: 24 March 2000/Revised version: 23 June 2000/Published online: 11 October 2000 – Springer-Verlag 2000
Abstract. In confocal laser scanning microscopes (CLSMs),
lasers can be used for image formation as well as tools for
the manipulation of microscopic objects. In the latter case,
in addition to the imaging lasers, the light of an extra laser
has to be focused into the object plane of the CLSM, for
example as optical tweezers. Imaging as well as trapping
by optical tweezers can be done using the same objective
lens. In this case, z-sectioning for 3D imaging shifts the op-
tical tweezers with the focal plane of the objective along the
optical axis, so that a trapped object remains positioned in
the focal plane. Consequently, 3D imaging of trapped ob-
jects is impossible without further measures. We present an
experimental set-up keeping the axial trapping position of
the optical tweezers at its intended position whilst the focal
plane can be axially shifted over a distance of about 15 µm.
It is based on fast-moving correctional optics synchronized
with the objective movement. First examples of application
are the 3D imaging of chloroplasts of Elodea densa (Cana-
dian waterweed) in a vigorous cytoplasmic streaming and
the displacement of zymogen granules in pancreatic cancer
cells (AR42 J).
Optical tweezers are based on forces due to radiation
pressure, with the gradient force as dominant force compon-
ent [8]. The gradient force pulls dielectric particles into the
high-intensity gradient region of an infrared laser beam fo-
cused by an objective with high numerical aperture. Thus,
optical forces can be used to fix and move dielectric particles
and furthermore to precisely measure forces in the piconew-
ton range [9–11]. Optical tweezers have now been widely
used in biology and physics. They offer the opportunity for
subcellular manipulation and studying cellular processes that
require precise positioning of cell organelles [12]. Examples
include the manipulation of cell nuclei as well as organelles
in plant cells and the displacement of chromosomes [13–15].
Confocal laser scanning microscopes (CLSMs) assem-
ble a number of two-dimensional optical sections to fi-
nally reconstitute a three-dimensional image (z-sectioning).
Using these sectioning capabilities combined with the 3D-
manipulation possibilities of the optical tweezers, CLSMs are
excellent imaging tools for 3D micromanipulation [16, 17].
Particularly in CLSMs, where the scan time per frame is of
the order of seconds, optical tweezers can alleviate the serious
problem of motional blurring of moving objects by spatially
fixing them, even inside living cells.
PACS: 87.64.-t; 87.16.-b
So far, attempts to use laser microtools in CLSMs have
been only moderately successful. When imaging as well as
trapping by optical tweezers are done using the same objec-
tive lens, a conflict arises between the 3D imaging and the
axial positioning of the focus of the laser used for optical
tweezers [17]: whenever the microscope objective is moved
along the optical axis for z-sectioning, the focus of the opti-
cal tweezers is shifted with the focal plane of the objective.
So a trapped object remains positioned in focus while its for-
mer surroundings go into defocus. Consequently, 3D imaging
of trapped objects as well as imaging of a trapped object
in its surroundings for 3D-position detection are impossible
without further measures. In the following we present a new
experimental set-up keeping the trapping position of the opti-
cal tweezers at its intended position whilst the focal plane can
be axially shifted, allowing 3D imaging of trapped objects.
As a first example confocal imaging of chloroplasts in
leaf cells of Elodea densa (Canadian waterweed) possess-
Laser light can be easily focused into a spot of submicrome-
ter dimensions in a microscope. When an object is transparent
for a given wavelength, even focusing into the object’s inte-
rior is possible. Since light is a carrier of energy that may be
converted into heat, high temperatures up to nanoplasmas can
be induced in the laser-light focus. Since, in addition, light is
a carrier of momentum, forces can be exerted too. This has
been exploited in conventional microscopes to use laser light
first for submicrometer machining with laser microbeams, see
for example [1–3], and then for complete micromanipulation
by light with microbeams and optical tweezers in conven-
tional microscopes [4–6]. Within the last few years the use
of optical tweezers for photonic force microscopy opened up
new field of applications with optical tweezers [7].
∗Corresponding author.

748
ing a vigorous cytoplasmic streaming is shown. Furthermore,
a perspective for an application with pancreatic cancer cells
(AR42 J) is given.
(MOVTEC Stütz & Wacht GmbH, Pforzheim, Germany).
The motor is adjusted to 2000 steps/revolution, resulting
in a stage shift of 250 nm/step, a maximum velocity of
10000 steps/s, and an acceleration of 20000 steps/s2. A con-
vex lens (“L”, f = 40 mm), added externally into the path of
the laser beam, focuses the light emerging from the fiber into
the epi-fluorescence illumination path (“FIP”) where another
convex lens ( f = 150 mm) is located. The single-mode fiber
together with the external lens present a suitable beam ex-
pander, because of the divergence of the beam emerging from
the fiber, reducing the number of optical elements required for
an adequate expansion of the laser beam. Finally, the dichroic
mirrors located in the fluorescence reflector slide (“FRS”) dir-
ect the laser light to the back aperture of the objective (“O”)
that has to be entirely illuminated by the infrared laser light
to obtain an almost diffraction-limited beam for good trap-
ping efficiency [6, 18]. For this purpose, the existing mirrors
for epi-fluorescence illumination are replaced by mirrors be-
ing additionally reflective for the infrared laser light used
for the optical tweezers (AHF Analysentechnik, Tübingen,
Germany). For the use of the optical tweezers in laser scan-
ning mode, yet another dielectric mirror (Laser Optik GmbH,
Garbsen, Germany) is placed on the laser scanning position
of the fluorescence reflector slide that otherwise would be
empty. It reflects the infrared laser light and is simultaneously
transparent for the scanning lasers coming from below and
the emitted fluorescence. Hence, the complete imaging capa-
bilities of the original LSM 510 are maintained.
For the determination of the relative objective position,
internal signals controlling the motor that shifts the micro-
scope objective are registered. This registration is carried out
by a PC equipped with a data-acquisition board and a stepper
board (PCI-6023E and PCI-Step-20X, National Instruments,
München, Germany) that is used for controlling the stepper-
motor position of the translation stage. A program routine
developed with the software LabVIEW 5.1 (National Instru-
ments, München, Germany) controls the signal registration
and determines the relative objective position. Mediated by
a calibration curve (see results and Sect. 1.2), this program
simultaneously calculates and controls the stage position as
a function of the relative objective position, as illustrated
in Fig. 1b.
1 Materials and methods
1.1 Experimental set-up
The experimental set-up is based on an inverted confocal
laser scanning microscope (LSM 510, Carl Zeiss Jena GmbH,
Jena, Germany) equipped with an argon-ion laser (458 nm,
488 nm) and a green HeNe laser (543 nm) as scanning lasers.
The LSM 510 carries out z-sectioning for 3D imaging by
moving the objective along the optical axis (z-direction)
in order to scan optical sections of an object at different
z-positions.
In our experimental set-up (shown in Fig. 1a) a Nd:YAG
laser (1064 nm, max. power 11 W; T40-Z-106C, Spectra
Physics, Darmstadt, Germany) running in cw mode is used
for the optical tweezers. A single-mode fiber (BTO Bungert
GmbH, Weil der Stadt, Germany) directs the laser light from
the Nd:YAG laser to the rear side of the microscope. The
fiber output efficiency with the single-mode fiber coupler
used (THORLABS, Grünberg, Germany) is about 60%. The
fiber end is fixed on a translation stage (“TS” in Fig. 1a)
that can be displaced in the axial direction. The stage has
a travel length of 10 mm (LINOS Photonics GmbH, Göt-
tingen, Germany) and is equipped with a stepper motor
1.2 Calibration
Calibration curves were determined for the oil-immersion ob-
jectives Zeiss Plan-Neofluar 100×/1.3 oil and 63×/1.25 oil
(Carl Zeiss Jena GmbH, Jena, Germany). We used a dielec-
tric mirror with high reflectance at 1064 nm (Laser Optik
GmbH, Garbsen, Germany) placed like a cover glass. The sig-
nal emerging from the laser light used for optical tweezers
that is reflected by the mirror was confocally detected with
the scanning lasers turned off. Figure 2a schematically shows
the optical path of the laser-tweezers system (components as
described in Fig. 1a) and the confocal imaging system. The
imaging system consists of a dichroic mirror (“M”), trans-
mitting the light of the scanning lasers coming from below
(not shown) and reflecting the signal coming from above, and
a convex lens focusing the signal on the pinhole (“P”) placed
in front of the detector (“D”). In spite of the dielectric mir-
ror in the fluorescence reflector slide (“FRS”) being highly
Fig. 1a,b. Experimental set-up. a Integration of the Nd:YAG laser used
for optical tweezers in the LSM 510. b Synchronization of axial objective
movement and movement of optical tweezers

749
Fig. 2a–c. Calibration. a Schematic illustration
of the laser-tweezers system and the imaging
system. b Translation-stage position as a func-
tion of the relative objective position at max-
imum signal intensity for the objectives Plan-
Neofluar 100 ×/1.3 oil (lower curve) and 63 ×
/1.25 oil (upper curve) c Power of the laser
light used for the optical tweezers emerging
from the objectives Plan-Neofluar 100×/1.3 oil
(lower curve) and 63×/1.25 oil (upper curve)
as a function of the translation-stage position
reflective at 1064 nm, the reflection signal reaching the de-
tector was sufficient using a laser power of about 140 mW to
290 mW and 110 mW to 170 mW emerging from the 100×
and 63× objective, respectively. Figure 2a shows a particu-
lar position of the translation stage that results in a divergent
laser beam entering the back aperture of the objective. The
situation for two different axial objective-lens positions is il-
lustrated by a solid and a dotted line and pointed out in the
two figures showing an enlarged view of the field including
the objective lens and the dielectric mirror. In the left-hand
view, due to the divergent beam, the focus of the optical
tweezers is situated above the focal plane of the objective. Be-
cause the reflected signal (dotted lines) is coming from out
of the focal plane, the pinhole suppresses this signal. In the
right-hand view the objective lens is shifted, placing the re-
flected focus of the optical tweezers (solid lines) in the focal
plane of the objective. Here the signal is focused onto the pin-
hole resulting in a maximum signal.
test objects. These polystyrene beads are ring-stained, i.e.,
only their surface is stained by a dye, combined with a sec-
ond fluorescent color throughout the bead. We used only the
surface stain that was excited with the 488-nm line of the
argon-ion scanning laser. Fluorescence was detected above
505 nm using a long-pass filter. For microscopy the micro-
spheres were suspended 1 : 10 in distilled water with a final
concentration of about 50 beads/µl. A drop of several mi-
croliters was put on a cover glass and covered by a small
cap to prevent evaporation. The microspheres were trapped at
a power of the optical tweezers of about 100 mW to 200 mW.
We determined the ring dimension of the individual cross
sections of a microsphere with the software LSM 510 2.01
(Carl Zeiss Jena GmbH, Jena, Germany) using the profile
function. This function displays the intensity in an optical
section along an individually drawn straight line. Drawing
this line across a ring through its center, two intensity maxima
can be seen. The distance between these maxima was taken
as the ring dimension. To reduce the random error accom-
panying this kind of determination, the ring dimension was
determined by drawing the line four times across each ring
through its center, each line forming an angle of 45◦ with the
former line.
The power of the optical tweezers emerging from the ob-
jective was recorded by placing a power meter (407 A, Spec-
tra Physics, Darmstadt, Germany) just behind the particular
objective close to its focus.
1.3 Fluorescent microspheres
1.4 Elodea densa
Fluorescent microspheres with a diameter of 15 µm (varia-
tion of bead diameter < 3%) (FocalCheck F-7237, Molecular
Probes via Mo Bi Tec, Göttingen, Germany) were used as
Elodea densa (Canadian waterweed) was commercially ob-
tained. For investigations with the CLSM, individual leaves

750
were freshly excised from a plant and mounted between two
cover glasses with some tap water. Quiet cells were directly
illuminated by the halogen lamp (“HAL” in Fig. 1a) to induce
or accelerate their cytoplasmic streaming. For confocal imag-
ing the chloroplasts were excited with the 488-nm line of the
argon-ion scanning laser and showed an autofluorescence de-
tected with a long-pass filter above 505 nm. Trapping of the
chloroplasts was done using a power of the optical tweezers
of about 100 mW to 200 mW. With this power no cell damage
was observed.
/1.25 (upper curve). The highest relative objective positions
at the stage position 10 mm are arbitrarily set to zero. With
increasing stage position (increasing distance between fiber
end and lens “L” in Fig. 1a) the objective position at max-
imum signal is nonlinearly shifted upward, meaning that the
focus of the optical tweezers is shifted downward twice the
distance of the objective shift due to reflection. Thus, mul-
tiplying the relative objective position in Fig. 2b by a factor
of 2 reveals the final calibration curve giving the translation-
stage position that maintains the focus position of the optical
tweezers as a function of the relative axial objective position.
Hence, a compensation of an objective motion to maintain the
relative axial focus position of the optical tweezers appears
possible over a z-distance of about 20 µm for the 100× ob-
jective and of about 50 µm for the 63× objective, restricted
by the travel length of the translation stage. Since different
objects are held at slightly different heights with respect to
the focus of the optical tweezers [19, 20], the zero point of
the calibration curve will finally be given at the stage position
where the center of a particular trapped object is placed in the
focal plane of the objective. The particular zero position can
be simply adjusted within the developed controlling program.
Besides this, we determined the power of the laser light
used for optical tweezers emerging from the objective as
a function of the stage position (see Sect. 1.2). Figure 2c
shows the power of the optical tweezers given in percent of
the laser-output power for the 100× objective (lower curve)
and the 63× objective (upper curve). Obviously, a linear rela-
tionship exists between the power of the optical tweezers and
the stage position, being maximum at the smallest distance
between the fiber end and the external lens. Thus, trapping ef-
ficiency decreases with increasing stage position, i.e., when
the focus of the optical tweezers is shifted downward.
Furthermore, to compensate the axial focal-plane motion,
the actual z-position of the objective has to be available with-
out delay at any time. Otherwise the trapped object is still
moving with the focus of the optical tweezers when the next
optical section is scanned during z-sectioning. We could cal-
culate the actual relative objective position from the signals
controlling the motor that shifts the microscope objective (for
more details see Sect. 1.1 and Fig. 1b). Although translation-
stage motion starts almost immediately with axial objective
motion (after about 20 ms), the final stage position is reached
some 100 ms after objective positioning is completed. Since
this time lag is situated in the time lag between the record-
ing of two optical slices during z-sectioning, it has no visible
effect.
1.5 AR42J cells
AR42J cells (ATCC, Manasses, USA) were cultivated at
37 ◦C and 5% CO2 in RPMI medium with 10% FCS
(SIGMA, Deisenhofen, Germany). For microscopy, cells
grew on cover glasses under the same conditions for two
up to five days. Staining was performed with cSNARF-AM
(Molecular Probes via Mo Bi Tec, Göttingen, Germany) at
a concentration of 10 µg/ml in RPMI medium without FCS.
After incubation over 5 to 15 min the cells were rinsed with
RPMI medium without FCS. Staining and microscopy were
performed in a thermostat (Bachofer, Reutlingen, Germany)
at 37 ◦C. Fluorescence was excited using the 488-nm and
543-nm lines of the scanning lasers. Emission was detected
with two band-pass filters (565–595 nm, 625–656 nm).
2 Results
2.1 Compensated motion of the focus of the optical tweezers
For the 3D imaging of objects fixed by optical tweezers we
started with the following approach: by varying the distance
between the fiber end fixed on the motorized translation stage
and lens “L”, by moving the fiber end along the optical axis
(see Sect. 1.1 and Fig. 1a for more details), the divergence
of the beam entering the back aperture of the objective is
changed. In this way the focus of the optical tweezers can be
effectively shifted perpendicular to the focal plane (along the
z-axis). This allows the compensation of an axial focal-plane
motion by a simultaneous reverse motion of the focus of the
optical tweezers in such a way that a trapped object maintains
positioned on its intended position in the specimen.
For the realization of this approach it is necessary to deter-
mine a calibration curve giving the translation-stage position
as a function of the axial objective position in such a way
as to maintain the trapping position of the optical tweezers
when the objective is moved. This was done using a dielectric
mirror highly reflective at 1064 nm as a specimen reflecting
the laser light used for optical tweezers (see Sect. 1.2 and
Fig. 2a for more details). Here, using z-sectioning in laser
scanning mode the z-position of the objective showing the
optical section with the maximum intensity of the reflected
signal was determined for different stage positions with an
average accuracy of 250 nm. On these particular objective
positions the focus of the optical tweezers was situated in
the focal plane of the objective (see Fig. 2a: maximum sig-
nal). The results are displayed in Fig. 2b showing the stage
positions as a function of the determined relative objective
positions at maximum signal intensity for the objectives Plan-
Neofluar 100 × /1.3 (lower curve) and Plan-Neofluar 63 ×
2.2 Test of the compensation
Surface-stained microspheres with a diameter of 15 µm were
used as test objects. When a drop of these polystyrene beads
suspended in distilled water was put on a cover glass, the
beads dropped onto the surface within a few minutes. It
turned out to be hard to trap them once they were situated on
the cover-glass surface. So they were trapped by turning on
the optical tweezers when a bead, moving to the cover glass,
reached a position shortly under the final trapping position.
Above this position the bead was pushed away as radiation
pressure countered the restoring force from the intensity gra-
dient. Due to the surface stain (see Sect. 1.3), these beads,

751
Zero position is given at the objective position with maximum
ring dimension, i.e., at the ring center. This curve and the
x − z view in Fig. 3a represent the reference used for the z-
sectioning of microspheres trapped by the optical tweezers.
For this we started with the objective Plan-Neofluar 100×
/1.3 oil using the lower curve shown in Fig. 2b. The x − z
view obtained from a z-sectioning of a microsphere trapped
at a distance of about 20 µm to 30 µm from the cover-glass
surface is shown in Fig. 3c. Compared with the reference in
Fig. 3a it clearly appears compressed in the z-direction due
to ignoring the multiplication factor 2. The determined ring
dimensions as a function of the relative objective position
together with the reference curve are shown in Fig. 3d. Com-
paring both it turns out that a multiplication factor of about
2.2 instead of 2 would have been relevant under the par-
ticular conditions. Equivalently, a factor of about 2.5 turned
out for the objective Plan-Neofluar × 63/1.25 oil (results not
shown). Considering these new factors to get the final calibra-
tion curves, the results obtained for microspheres trapped by
the optical tweezers are shown in Fig. 3e–h: Figs. 3e and 3g
show x − z views taken with the 100× and 63× objectives,
respectively. The respective ring dimensions as a function
of the relative objective position together with the reference
curves are shown in Figs. 3f and 3h. Now the x − z views of
the trapped microspheres and the determined ring dimensions
match well with the reference ones over the z-dimension of
15 µm that could be tested with the microspheres.
2.3 CLSM imaging of moving chloroplasts of Elodea densa
We imaged chloroplasts (5 µm to 10 µm in diameter) in
leaf cells of Elodea densa (Canadian waterweed) in order to
demonstrate the optical tweezers for the 3D imaging of mov-
ing objects in cell biology. The cell cytoplasm containing the
chloroplasts is limited to a thin layer surrounding a large cen-
tral vacuole. Cytoplasmic streaming is seen in most cells as
a cyclic movement along the cell walls with streaming rates of
up to 12 µm/s [21]. It has been known for a long time that vis-
ible light can accelerate the cytoplasmic streaming in the cells
that are illuminated directly [22]. We used this effect to in-
duce or accelerate a vigorous cytoplasmic streaming in quiet
cells (see Sect. 1.4).
Due to long image formation times of up to several sec-
onds in laser-scanning mode, the chloroplasts following the
vigorous cytoplasmic streaming could not be imaged with-
out motional blurring as shown in Fig. 4a, using a scan time
per frame of 15.7 s (1024×1024 pixels, average 2). With the
optical tweezers an individual moving chloroplast could be
fixed and shifted to the chloroplast-free cell center for undis-
turbed imaging, as shown in Fig. 4b. Using the optical tweez-
ers for 3D imaging, Fig. 4c shows several optical sections of
a trapped chloroplast spaced at z-intervals of 450 nm.
Fig. 3a–h. Compensation test with 15-µm microspheres. a x − z view of
an individual microsphere lying on the cover-glass surface in water (Plan-
Neofluar 100×/1.3 oil, 1024 ×1024 pixels, interval 300 nm). b Reference
curve: Dependence of the ring dimension of the cross section (x − y views)
on the relative objective position. Data from four microspheres, taken as the
microsphere in a , were averaged and fitted by a polynomial of third order.
c x − z view of a microsphere trapped by the optical tweezers using the
lower curve of Fig. 2b (Plan-Neofluar 100 ×/1.3 oil, 1024 ×1024 pixels,
interval 450 nm). d Ring dimensions (dots) of the cross sections (x − y
views) of the microsphere in c as a function of the relative objective position
compared with the reference curve of b (solid line). Each diameter was de-
termined four times and the data are fitted by a polynomial of third order.
e Same as c, but using lower curve of Fig. 2b multiplied by a factor of 2.2
(interval 300 nm). f Same as d, but ring dimensions of the microsphere in e.
g Same as c, but using the upper curve of Fig. 2b multiplied by a factor of
2.5 (Plan-Neofluar 63×/1.25 oil, scale bar 5 µm). h ring dimensions of the
microsphere in g compared with the reference curve of the 63 ×/1.25 oil
objective
when viewed in cross section (x − y view) with the CLSM,
appeared as fluorescent rings of varying dimensions depend-
ing on the focal-plane position. Using z-sectioning a typical
vertical view (x − z view) of an individual bead situated on
the cover-glass surface in water is shown in Fig. 3a. The sec-
ond ring appearing in the lower part of this view is caused by
a beam splitter the fluorescence signal has to pass through.
We determined the ring dimensions of the individual cross
sections as a function of the relative objective position as
explained in Sect. 1.3. After averaging the data from four dif-
ferent microspheres the curve shown in Fig. 3b was obtained.
2.4 Dislocation of zymogen granules in AR42J cells
Pancreatic acinar cells are the starting point of pancreatitis,
the self-digestion of the pancreas, and sometimes of pan-
creatic cancer. The cells possess zymogen granules (ZGs)
containing proenzymes that get acidified to maintain the
inactivity of the proenzymes and change their subcellular lo-
calization during a maturation process [23, 24]. To examine

752
Fig. 4a–c. CLSM imaging of chloro-
plasts of Elodea densa. a A group of
chloroplasts moving vigorously in the cy-
toplasmic streaming (1024 ×1024 pixels,
scan time 15.7 s). b A single chloroplast
fixed by the optical tweezers (1024 ×
1024 pixels, scan time 15.7 s). c Cross
sections (x − y views) of
plast fixed by the optical tweezers
with
a chloro-
a
z-spacing of 450 µm (512 ×
512 pixels, scan time per frame 1.9 s).
(Plan-Neofluar 100 ×/1.3 oil, scale bars
5 µm)
while these objects are three-dimensionally imaged by the
laser scanning microscope, for example in studies on the ef-
fects of intracellular microenvironment on organelle shape.
An alternative attempt to solve the problem of the
movement of the focus of optical tweezers during three-
dimensional image acquisition in CLSMs has been published
in the literature [16, 17]. Here, in an inverted CLSM, an ad-
ditional high numerical aperture objective placed opposite to
the imaging one has been used to focus the laser used for the
optical tweezers into the specimen. In this way the optical
paths of the confocal imaging system and the laser tweezers
system are completely separated and the two devices can op-
erate independently. A disadvantage is the limited specimen
thickness of about 50 µm due to the rapidly degrading quality
of the confocal imaging, and degrading trapping efficiency of
the optical tweezers with increasing penetration depth in the
specimen arising from spherical aberration [16, 26, 27]. Fur-
thermore, the use of devices like a microinjection apparatus or
a thermostat is strongly restricted or even impossible because
the specimen has to be placed between two cover glasses
when using an oil-immersion objective as additional objec-
tive. These problems are not relevant with our set-up where
the imaging lasers of the laser scanning microscope and the
laser used for the optical tweezers are focused independently
using the same objective lens.
The test of the compensation using surface-stained micro-
spheres resulted in a multiplication factor of the calibration
curve differing from the theoretical factor 2. Factors of 2.2
and 2.5 for the Plan-Neofluar objectives 100×/1.3 oil and
63 × /1.25 oil, respectively, turned out to be relevant. This
is because we only studied the focus position of the op-
tical tweezers using a mirror for the determination of the
calibration curves. Actually, the axial trapping position of
an object in the optical tweezers is relevant and should be
studied. This was done by using surface-stained microspheres
as test objects. However, the trapping position can be influ-
enced by different factors. For example, a changing distance
between the focus of the optical tweezers and its trapping pos-
ition is most likely when the focus of the optical tweezers
is shifted perpendicularly to the focal plane of the objective.
Another problem ignored using a mirror for the calibration
arises from optical effects such as spherical aberration. Spher-
ical aberration is induced by mismatches in the refractive
index between cover glass and mounting medium, resulting
in paraxial rays are focused differently from peripheral ones.
Fig. 5a,b. Displacement of zymogen granules (ZGs) in AR42J cells by opti-
cal tweezers. a A part of an AR42J cell with three visible ZGs, one marked
by a ring. b The marked ZG after dislocation as indicated by rings and an
arrow. (Plan-Neofluar 100×/1.3 oil, 1024 ×1024 pixels, scale bar 5 µm)
the regulation of the ZG acidification in AR42J cells, derived
from pancreatic acinar cells [25], as a function of the ZG
position in the cell, we started to dislocate ZGs by the op-
tical tweezers and to observe intracellular pH with the dye
cSNARF-AM in the CLSM. Figure 5 shows the dislocation of
an individual ZG in an AR42J cell using a power of the opti-
cal tweezers of about 120 mW. Because the optical tweezers
is always located in the frame center, the specimen has to be
moved for the dislocation of the trapped ZG. Thus, the cell
appears shifted with respect to the frame after ZG displace-
ment in Fig. 5b. The observable morphology change during
displacement has to be the subject of further experiments to
distinguish between effects caused by the light used for the
optical tweezers and displacement effects.
The compensated motion of the focus of the optical
tweezers will be necessary in future studies for the correct po-
sitional information of the ZG at subcellular level before and
after dislocation.
3 Discussion
The experimental set-up presented here expands the field
of applications of optical tweezers in confocal laser scan-
ning microscopy. It is now possible to image objects three-
dimensionally, such as chloroplasts in a cytoplasmic stream-
ing, without distortion due to object movement during the
image-acquisition time. In other applications optical tweez-
ers can exert a tension on objects to maintain a dislocation

753
The result is a blurring of the intensity distribution of the focal
region. Furthermore, the focus is situated closer to the cover
glass [27, 28]. Hence, leaving the cover-glass surface results
in a loss of resolution, intensity, and a focal shift. This affects
the confocal imaging [15, 16] as well as the optical tweez-
ers [17] and its trapping position. First studies with micro-
spheres trapped by the optical tweezers without compensation
showed a change of the axial position of a trapped micro-
sphere with respect to the focal plane of the objective, when
the penetration depth was increased (results not shown). Here,
spherical aberration may have affected the trapping position
of the optical tweezers differently from the focal-plane pos-
ition of the objective due to the different wavelengths used,
depending on the penetration depth in the specimen. Because
this observation may have the consequence of calibration
curves being dependent on the penetration depth, the effect
of spherical aberration will be the subject of future studies.
Since different objects have different shapes and optical prop-
erties resulting in different trapping positions in the optical
tweezers, it further will have to be tested if they need separate
calibration curves.
4. K.O. Greulich, U. Bauder, S. Monajembashi, N. Ponelies, S. Seeger,
J. Wolfrum: Labor 2000, 36 (1989)
5. R. Wiegand Steubing, S. Cheng, W.H. Wright, Y. Numajiri, M.W. Berns:
Cytometry 12, 505 (1991)
6. K.O. Greulich: Micromanipulation by Light in Biology and Medicine:
The Laser Microbeam and Optical Tweezers (Birkenhäuser, Basel,
Wien, Boston 1999)
7. A. Pralle, E.-L. Florin, E. Stelzer, J.K.H. Hörber: Single Mol. I, 2, 129
(2000)
8. A. Ashkin, J.M. Dziedzic, T. Yamane: Nature 330, 769 (1987)
9. R.M. Simmons, J.T. Finer, S. Chu, J.A. Spudich: Biophys. J. 70, 1813
(1996)
10. M.D. Wang, H. Yin, R. Landick, J. Gelles, S.M. Block: Biophys. J. 72,
1335 (1997)
11. A. Kusumi, Y. Sako, T. Fujiwara, M. Tomishige: Methods Cell Biol. 55,
173 (1998)
12. H. Felgner, F. Grolig, O. Müller, M. Schliwa: Methods Cell Biol. 55,
195 (1998)
13. A. Ashkin, J.M. Dziedzic: Proc. Natl. Acad. Sci. USA 86, 7914 (1989)
14. G. Leitz, G. Weber, S. Seeger, K.O. Greulich: Physiol. Chem. Phys.
Med. NMR 26, 69 (1994)
15. M.W. Berns, W.H. Wright, B.J. Tromberg, G.A. Profeta, J.J. Andrews,
R.J. Walter: Proc. Natl. Acad. Sci. USA 86, 4539 (1989)
16. K. Visscher, G.J. Brakenhoff: Cytometry 12, 486 (1991)
17. K. Visscher, G.J. Brakenhoff, J.J. Krol: Cytometry 14, 105 (1993)
18. A. Ashkin, J.M. Dziedzic: United States Patent Number 4893886 (16
Jan. 1990)
19. T. Wohland, A. Rosin, E.H.K. Stelzer: Optik 102, 181 (1996)
20. A.D. Mehta, J.T. Finer, J.A. Spudich: Methods Cell Biol. 55, 47
(1998)
21. J. Forde, M.W. Steer: Can. J. Bot. 54, 2688 (1976)
22. L.V. Heilbrunn, F. Weber: Protoplasmotologia Band VIII (Springer,
Wien 1959)
Acknowledgements. We gratefully acknowledge the financial support of the
Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie
BMBF (grant No. AZ 11672) and of the Carl Zeiss Jena GmbH (Jena,
Germany). We further wish to thank Dr. R. Wendenburg and R. Wolleschen-
sky (Carl Zeiss Jena GmbH, Germany) for technical support as well as
Dr. L. Wollweber (IMB Jena) for ideas and discussions and B. Lanick (IMB
Jena) for cultivating the cells.
23. P.M. Motta, G. Macchiarelli, S.A. Nottola, S. Correr: Microsc. Res.
Tech. 37, 384 (1997)
24. V. Iwanji, J.D. Jamieson: J. Cell Biol. 95, 734 (1982)
25. J. Christophe: Am. J. Physiol. 266, G963 (1994)
26. S. Hell, G. Reiner, C. Cremer, E.H.K. Stelzer: J. Microsc. 169, 391
(1993)
References
1. K.O. Greulich, J. Wolfrum: Ber. Bunsen-Ges. Phys. Chem. 93, 245
(1989)
2. K.O. Greulich, J. Wolfrum: SPIE 1225, 174 (1990)
3. M.W. Berns, W.H. Wright, R. Wiegand Steubing: Int. Rev. Cytol. 129,
1 (1991)
27. J.B. Pawley: Handbook of Biological Confocal Microscopy (Plenum,
New York, London 1995)
28. P.C. Ke, M. Gu: J. Mod. Opt. 45, 2159 (1998)