Measurement of volume injected into individual cells by quantitative
fluorescence microscopy
GRETA MARLENE LEE*
The Department of Zoology, Duke University, Durham, NC 27706, USA
* Present address: Department of Cell Biology and Anatomy, The University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA
Summary
Pressure microinjection is frequently used to introduce substances into mammalian cells, but precise
quantitation of the volume injected into individual
cells has been difficult. A simple and reliable
procedure for determining the volume injected was
developed in order to determine what intracellular
concentration of AMP-PNP was necessary to inhibit
specific cellular processes. The technique uses fluorescent Lucifer Yellow-labeled dextrans in the
microinjection buffer and quantitative fluorescence
microscopy to measure the fluorescence intensity of
the injected cell. The volume injected is computed
from a standard curve derived from the volume and
Introduction
Pressure microinjection has become an established technique for introducing substances into cells (Celis, 1984;
Feramisco and Welch, 1986; Graessmann and Graessmann, 1986). Interpretation of the cellular effects of such
microinjections is improved by knowledge of the resulting intracellular concentration of the introduced substance. Since the intracellular concentration of the
injected substance depends on the volume injected as well
as on the pipet concentration, several methods have been
developed for determining the volume injected. Most
methods yield an average value. One method uses injection of
I-labeled bovine serum albumin into many
cells. The average radioactivity per cell gives an estimate
of the volume injected into each cell (Stacey and Allfrey,
1976; Zavortink et al. 1983). A second method measures
the size of vesicles that occasionally form upon microinjection (Izant, 1983). With a third method, called
picospritzing, the volume ejected is controlled by applying a known pressure for a set time to a pre-calibrated
pipet (McCaman et al. 1977; Sakai et al. 1979). Because
of the requirement for either high pressure or long
injection times, this method is primarily used with large
neurons.
With any of the methods described above, fluctuations
in the volume delivered can be produced by slight
changes in pressure or by partial clogging of the pipet.
Therefore, direct measurement of the volume injected
Journal of Cell Science 94, 443-447 (1989)
Printed in Great Britain © The Company of Biologists Limited 1989
fluorescence of spherical, microscopic droplets of
Lucifer Yellow dextran solution. The droplets are
ejected from a micropipet into immersion oil where
they sink to rest on a siliconized coverslip. For the
measurement of fluorescence, an inexpensive
photomultiplier system that is attached to a fluorescence microscope is described. The potential uses
of this method for other microassays are discussed.
Key words: microinjection, Lucifer Yellow dextran, cultured
cells.
into each cell is preferable. One approach is to include a
fluorescent dye in the injection buffer and then measure
the resulting fluorescence in the cell by quantitative
fluorescence microscopy (Keith et al. 1983; Seeley et al.
1983). One can then determine the volume injected by
comparing the intracellular fluorescence intensity with
the fluorescence of a known volume of dye (Keith et al.
1983).
In previous experiments, a hemacytometer was used to
obtain standard volumes of fluorescent solution (Keith et
al. 1983). In the hemacytometer, solution depth is
uniform; the area (and hence the volume) measured is
varied by changing the size of illumination and measuring
apertures. However, the hemacytometer is awkward and
inaccurate when used with an inverted microscope,
which is preferable for microinjection of tissue culture
cells.
I have developed an alternative method for producing
known volumes of solution for fluorescence measurements that is well suited for use on the inverted microscope. This method uses spherical droplets of fluorescent
solution ejected into immersion oil. After the droplets
sink and come to rest without flattening on a siliconized
coverslip, their diameter is measured using an ocular
micrometer, and their fluorescence is measured with the
same optics used to inject and measure cells.
The fluorescence intensity of microscope specimens
can be measured either directly with a photomultiplier
(Bohm, 1972) or indirectly with a computer after digitiza443
tion of the signal from a video camera (DiGuiseppi et al.
1985; Tanasugarn et al. 1984). Either way, the cost of
commercial equipment can range from $3000 for photometers designed primarily for applications in astronomy to over $50 000 for systems intended specifically for
fluorescence analysis. As an alternative, the system described here uses a high-quality photomultiplier tube and
was built for less than $1000.
The fluorescence technique for determining volume
injected into individual cells and the photomultiplier
measurement system was applied in a study to evaluate
the effects of AMP-PNP (adenylyl-imidodiphosphate) on
cellular processes during mitosis (Lee, 1989). This technique allowed dose-response curves to be produced as
the volume injected was well correlated with effects on
cellular processes such as rate of spindle elongation and
incidence of saltatory movements.
comparable to the range of fluorescence in injected cells, and at
the same time gave droplets with diameters in a range that could
be measured accurately with an ocular micrometer. The droplets were ejected from a micropipet into Cargille Type FF
immersion oil (R. P. Cargille Laboratories, Inc., Cedar Grove,
NJ) contained in a well slide (see Fig. 2). Release of the
droplets from the pipet tip was aided by briskly moving the oil
relative to the stationary, submerged pipet tip; the microscope
stage controls were used to shift the well slide. By varying the
ejection pressure and rate of movement of the stage, rows of
droplets of various sizes were produced. The droplets sank in
To power supply
and multimeter
To high-voltage
power supply
Materials and methods
Cell culture techniques and preparation of chambers for microinjection were as described in an accompanying report (Lee,
1989). Pipets for microinjection were pulled from 1.0 mm
borosilicate glass capillaries with an internal filament (no.
1B100F-6, World Precision Instruments, Inc., New Haven,
CT) using a Kopf vertical pipet puller (David Kopf Instruments, Tujunga, CA). The pipets were backfilled by placing the
untapered end in solution that had been centrifuged for 2 min at
900 £ in order to remove particulate contaminates that might
clog the pipet. 1% Lucifer Yellow-labeled dextrans (10000
average molecular weight, Molecular Probes, Eugene, OR)
were microinjected in a buffer of lOOmM-Pipes, 0.1 mM-MgCl2,
pH7.3, withKOH.
All cell observations, microinjections and measurements of
fluorescence intensity were made with an inverted Zeiss IM epifluorescence microscope equipped with X40/0.75 NA phasecontrast optics, a 50 W or 100 W mercury lamp, and the Zeiss
filter combination for fluorescein (BP485/20-LP520). A Zeiss
photochanger was used to attach the photomultiplier (described
below) to the microscope and to direct light to the oculars, the
photomultiplier or the video camera as required.
Fluorescence measurements were made using a head-on
photomultiplier (to R1104, Hamamatsu Corp., Bridgewater,
NJ) with a D type socket (no. E990), which was housed in a
modified aluminum pipe (see Fig. 1 for a diagram). The output
of the photomultiplier fed directly into an operational amplifier
(ICL7650 chopper stabilized, Intersil, Inc.) that was assembled
on perforated circuit board (see the 1985 Hamamatsu publication 'Photomultipliers' for a circuit diagram). A ±15Vd.c.
power supply (Lambda Electric, Melville, NY) powered both
the high-voltage power supply for the photomultiplier (Hamamatsu no. C-2456) and, after reduction via 5 V negative
(uA7905C) and positive (uA7805C) voltage regulators (Texas
Instruments, Dallas, TX), powered the operational amplifier.
Output of the high-voltage power supply was adjusted by a 10turn, 10 kQ potentiometer. For the measurements described
here, the photomultiplier was operated at 475 V. The operational amplifier converted the milliamp output of the photomultiplier into volts, which were then displayed by a Fluke
digital multimeter (John Fluke Mfr, Everett, WA).
Determination of the amount of fluorescence for a given
volume was done using droplets of the injected solution diluted
1:10 with distilled water. This dye concentration gave a range
of fluorescence intensity for droplets of varying size that was
444
G. M. Lee
Magnetic shield
Plastic ring
Zeiss attachment
ring
Fig. 1. Diagram of photomultiplier tube in housing with
electrical connections shown. The photomultiplier tube is
sleeved in and supported by a magnetic shield case
(Hamamatsu E989-03) lined with plastic foam. Thick
composite plastic rings are used to support the socket and
magnetic shield. The wires from the socket extend upwards
through a hole in the center of the operational amplifier
circuit board to allow easy connection with the amplifier. A
high-voltage plug (UG-932/U) connects the high-voltage
power supply via coaxial cable (RG-59/U) with the
photomultiplier tube socket. A four-pin plug connects the
operational amplifier with the power supply and output (see
text for description). Both plugs are made light-tight with the
housing by O-rings. A Zeiss attachment ring is used to
provide firm attachment to the microscope.
Fig. 2. The procedure for making droplets for the generation
of a standard curve. A. The well slide is formed by mounting
a no. 1, 25 mm 2 , siliconized coverslip over a 22 mm hole
bored into a standard glass microscope slide. Coverslips are
siliconized by soaking in Sigmacote (Sigma Chemical Co.) for
5 min followed by four rinses in distilled water and baking for
2h at 180°C. After gentle cleaning with distilled water and a
Kimwipe, the coverslip is attached with silicon vacuum
grease and sealed with clear nail polish. B. The well in the
slide is filled with Cargille Type FF immersion oil. (Type
DF has the wrong density; the droplets rise instead of
sinking.) C. A micropipet is used to eject droplets of Lucifer
Yellow dextran solution into the oil.
the oil and came to rest on the siliconized coverslip without
discernible flattening. Using bright-field optics and Kohler
illumination, the diameter of the droplets was determined with
a calibrated ocular micrometer; droplet volume was then
calculated from droplet diameter using the formula for a sphere.
Prior to a fluorescence measurement, the specimen, either a
cell or a droplet, was centered in the field of view using phasecontrast or bright-field optics. With the excitation light on, the
field diaphragm for the illuminating beam was closed until only
the cell or droplet of interest was visible. Cells and droplets
were exposed to the excitation light as briefly as possible. After
measuring the fluorescence of the specimen, a background
measurement of fluorescence was made adjacent to the specimen using the same focal level and field diaphragm opening.
For the cells, the area for a background measurement included
noninjected cells. (Note: their was no autofluorescence detectable in the living cells by eye or by the photomultiplier at the
sensitivity level used for these measurements.) A specimen's net
fluorescence was then obtained by subtracting the background
fluorescence from the measured specimen fluorescence (Bohm,
1972; Colemane/a/. 1981).
A standard curve of fluorescence intensity versus volume was
produced from the droplets. Using linear regression analysis
and correcting for the 1: 10 dilution, the volume injected into
the cell was computed from its fluorescence intensity. Droplets
were made and measured immediately before or after each
experimental session to control for variations in lamp output.
several ways. A representative standard curve relating the
volume of drops of Lucifer Yellow dextran solution to
fluorescence intensity is shown in Fig. 3. For the standard curves, correlation coefficients determined by linear
regression ranged from 0.980 to 0.999. Consistent net
(specimen fluorescence minus background fluorescence)
values were obtained when the same droplet was
measured several times, either with the same or with
different openings of the field diaphragm. The net
fluorescence intensity measured for a droplet of given size
and dye concentration equaled that for a droplet of twice
the volume but half the dye concentration.
The ability to determine the volume of a substance
injected into individual cells is valuable for determining
precise dose-response relationships for a variety of
injected substances such as antibodies, antisense RNA,
calcium, etc. This value of the technique is demonstrated
by the experimental results in Fig. 4. In Fig. 4A, spindle
elongation for all cells injected with AMP-PNP is presented as an average without regard to the volume of
AMP-PNP injected. In Fig. 4B, the same data are
presented, but in this case the separate curves distinguish
groups of cells that were injected with different volumes.
It is only when data can be compared on the basis of
differences in injection volume (as in Fig. 4B) that it
becomes apparent that AMP-PNP has a dose-dependent
effect on spindle elongation. The absence of any spindle
elongation with high injection volumes (Fig. 4B) was
entirely masked by the averaging of data as in Fig. 4A.
For additional details regarding these data and for other
experimental results, see Lee (1989, accompanying
article).
To ensure that fluorescence measurements accurately
represented injected volume, several important factors
were considered. (1) The choice of the fluorochrome,
Lucifer Yellow attached to dextran, was based on certain
desirable features: (a) the quantum yield of Lucifer
Yellow is independent of changes in pH or osmolarity
(Stewart, 1978); (b) dextrans of low molecular weight
diffuse uniformly throughout the cell; and (c) attachment
to the dextran prevents the Lucifer Yellow from being
pumped out of the cell (Steinberg et al. 1987). (2) The
concentration of the fluorochrome was carefully chosen
so as to provide sufficient fluorescence intensity for
0.300
0.250
0.000
Results and discussion
The reproducibility of the fluorescence measurement
technique for volume determination was examined in
10
15
20 25 30
Volume (pi)
35
40
Fig. 3. A standard curve of volume versus fluorescence
intensity for droplets of 0.1 % Lucifer Yellow dextran
solution.
Measurement of volume injected
445
10
0
Time (min)
Fig. 4. Anaphase spindle elongation in PtKl cells injected with 10 mM-AMP-PNP in late metaphase. Spindle length is given in
j.im on the y-axis. A. Average spindle elongation for all injected cells grouped together ( • ) Sham injected; (A) injected. B.
Average spindle elongation for cells grouped by volume injected ( • ) Sham injected; (A) 0.1-0.4pi; ( • ) 0.5—0.9 pi;
1.0—1.9 pi. The standard errors in B are greater than in A because of the smaller sample size of each group.
measurement in the volume range compatible with cell
health (see below), but at the same time, low enough to
be nontoxic. (3) To reduce the effect of regional variations in cell thickness, the fluorescence intensity of the
whole cell was measured. (4) To avoid possible damage to
the cells by exposure to the excitation light, the fluorescence was usually measured only after other measurements (e.g. rate of mitotic spindle elongation) were
completed. However, this precaution may not be necessary: cells injected with Lucifer Yellow dextrans progressed normally through anaphase after measurement of
their fluorescence.
For the experiments reported here and in the accompanying paper (Lee, 1989), the calculated range of
volumes injected into Ptk cells was 0.1 to 1.9 pi. These
values are within the range of previously published values
for volumes injected into various tissue culture cells:
0.01 pi per fibroblast cell (Graessmann and Graessmann,
1986), 0.1 pi per PtK or BS-C cell (Zavortink et al. 1983;
Izant et al. 1983), and 2.8pi per gerbil fibroma cell
(Keith et al. 1983). The volumes injected are usually
reported as being within 5-20% of the cell volume
(Keith et al. 1983; Stacey and Allfrey, 1976; Zavortink et
al. 1983). Only Keith et al. (1983) reported a cell volume
and gave the method for determining it. They determined cell volume of gerbil fibroma cells to be approximately 20 pi using the diameter of round, detached cells,
and thus an injection volume of 2.8 pi would be 14% of
the cell volume. Using a similar technique, I computed
the average cell volume of PtK cells to be 10 pi. Cells
injected with 10 % of this volume showed no indication of
damage from the injection and most cells injected with
greater than 20 % of this volume with control solution
completed anaphase normally.
Since the volume of the cell also affects intracellular
446
G. M. Lee
concentration of the injected substance, variations in cell
size can add additional variation to the data collected.
Ideally, the volume of every cell would be measured in
addition to the volume injected. Current methods to
determine cell volume include measuring the diameter of
round cells or measuring the area and thickness of flat
cells by quantitative interference microscopy (Ross,
1967) or stereo microscopy (Griffiths et al. 1989). These
approaches are not practical for every cell. Flat cells do
not round up on cue, and measuring cell thickness
requires specialized equipment and the interpretation is
complex (Griffiths et al. 1989; Ross, 1967). My alternative was to simply select cells that appeared uniform in
size. By this means, I was able to obtain fairly consistent
results as shown in Fig. 4B. The criteria used for
selection are discussed further (Lee, 1989).
Measuring the fluorescence of microscopic droplets of
solution may have diverse applications. The technique
could be used as a method of generating standard curves
for other quantitative fluorescence applications. For
instance, standard curves based on solutions of known
DNA concentration plus DAPI, a fluorochrome that
binds stoichiometrically to DNA (Coleman et al. 1981),
could serve as an alternative to the biological standards
currently in use for estimation of intracellular DNA
content (Coleman et al. 1981; Lee et al. 1984). Another
possible use of this technique is in microassays. In a
microassay, a droplet having a known concentration of
one reactant (e.g. ATP) would be formed and its diameter measured; then a second reactant (e.g. luciferase)
would be added; finally, the resulting fluorescence and
the diameter of the final droplet would be measured. In
this way, one could generate standard curves for measurement of substances in cells or in solutions where only
minute quantities are available.
This work was done in the laboratory of R. Bruce Nicklas.
Special appreciation is given to him for his support and
excellent advice. I am grateful to Carl Mills for his help with the
construction of the photomultiplier. I thank Donna Maroni for
editorial assistance. This work was supported by NIH GM13745 to R.B.N. and by NIH training grant no. 5T32GM07184
toG.M.L.
References
BOHM, N. (1972). Fluorescence cytophotometric determination of
DNA. In Techniques of Biochemical and Biophysical Morphology,
vol. 1 (ed. D. Click and R.M. Rosenbaum), pp. 89-141. New
York: John Wiley and Sons.
CELIS, J. E. (1984). Microinjection of somatic cells with
micropipettes: comparison with other transfer techniques.
Riochem.J. 223, 281-291.
COl-EMAN, A. W., MAGUIRE, M. J. AND COLEMAN, J. R. (1981).
Mithramycin- and 4'-6-diamidino-2-phenylindole (DAPI)-DNA
staining for fluorescence microspectrophotometnc measurement of
DNA in nuclei, plastids, and virus particles. J'. Histochem.
Cytochem. 29, 959-968.
DlGulSEPPI, J., INMAN, R., ISHIHARA, A., JACOBSON, K. AND
HERMAN, B. (1985). Applications of fluorescence microscopy to
problems in cell biology. Biotechniques 3, 394-403.
FERAMISCO, J. R. AND WELCH, W. J. (1986). Modulation of cellular
activities via microinjection into living cells. In Microinjection and
Organelle Transplantation Techniques (ed. J. E. Celis, A.
Graessmann and A. Loyter), pp. 39-58. New York: Academic
Press.
GRAESSMANN, M. AND GRAESSMANN, A. (1986). Microinjection of
tissue culture cells using glass capillaries: methods. In
Microinjection and Organelle Transplantation Techniques (ed. J.E.
Celis, A. Graessmann and A. Loyter), pp. 3-13. New York:
Academic Press.
GRIFFITHS, G., FULLER, S. D., BACK, R., HOLLINSHEAD, M.,
PFEIFFER, S. AND SIMONS, K. (1989). The dynamic nature of the
Golgi complex. J Cell Biol. 108,277-297.
IZANT, J. G. (1983). The role of calcium ions during mitosis:
calcium participates in the anaphase trigger. Chromosoma 88, 1-10.
KEITH, C. H., D I PAOLA, M., MAXFIELD, F. R. AND SHELANSKI, M.
L. (1983). Microinjection of Ca2+-calmodulin causes a localized
depolymerization of microtubules. J. Cell Biol. 97, 1918-1924.
LEE, G. M. (1989). Characterization of mitotic motors by their
relative sensitivity to AMP-PNP. J. Cell Sci. 94, 425-441.
LEE, G. M., THORNTHWAITE, J. T. AND RASCH, E. M. (1984).
Picogram per cell determination of DNA by flow cytofluorometry.
Analyt. Biochem. 137, 221-226.
MCCAMAN, R. E., MCKENNA, D. G. AND ONO, J. K. (1977). A
pressure system for intracellular and extracellular ejections of
picohter volumes. Brain Res. 136, 141-147.
Ross, K. F. A. (1967). Phase Contrast and Interference Microscopy
for Cell Biologist, pp. 128-134. London: Edward Arnold
(Publishers) Ltd.
SAKAI, M., SWARTZ, B. E. AND WOODY, C. D. (1979). Controlled
micro release of pharmacological agents: measurements of volume
ejected ;;; vittv through fine tipped glass microelectrodes by
pressure. Neuwphannacology 18, 209-213.
SEELEY, P. J., KEITH, C. H., SHELANSKI, M. L. AND GREENE, L. A.
(1983). Pressure microinjection of nerve growth factor and antinerve growth factor into the nucleus and cytoplasm: lack of effects
on neurite outgrowth from pheochromocytoma cells. J. Seurosci.
3, 1488-1494.
STACEY, D. W. AND ALLFREY, V. G. (1976). Microinjection studies
of duck globin messenger RNA translation in human and avian
cells. Cell 9, 725-732.
STEINBERG, T. H., NEWMAN, A. S., SWANSON, J. A. AND
SILVERSTEIN, S. C. (1987). Macrophages possess probenecidmhibitable organic anion transporters that remove fluorescent dyes
from the cytoplasmic matrix. J. Cell Biol. 105, 2695-2702.
STEWART, W. W. (1978). Functional connections between cells as
revealed by dye-coupling with a highly fluorescent napthalimide
tracer. Cell 14, 741-759.
TANASUGARN, L., MCNEIL, P., REYNOLDS, G. T. AND TAYLOR, D.
L. (1984). Microspectrofluorometry by digital image processing:
measurement of cytoplasmic pH.,7. Cell Biol. 98, 717-724.
ZAVORTINK, M., WELSH, M. J. AND MCINTOSH, J. R. (1983). The
distribution of calmodulm in living mitotic cells, lixpl Cell Res.
149, 375-385.
(Received 15 May 1989 - Accepted, in revised fonn, 2 August 19S9)
Measurement of volume injected
447