Introduction to Fluorescence Imaging of
Live Cells: An Annotated Checklist
It is becoming clear that most biological molecules
in living cells are in a highly dynamic state, changing
their interactions and spatial organizations in response
to signals. With recent advances in optics,
probe design, and photon detection, there are few
approaches as suitable as fluorescence imaging for
investigating dynamic events in living cells. However,
successful execution of fluorescence imaging relies
heavily on proper setup of facilities. The purpose of
this article is to provide a comprehensive list of the
equipment, with associated notes, for those setting up
new imaging facilities.
There are many options in setting up a fluorescence
imaging facility, each with its advantages and disadvantages.
The decision is usually dictated by a combination
of experimental goals and personal preferences.
While this article attempts to provide some useful
guidelines, it is important not to equip much beyond
what is necessary, as excess equipment wastes money,
creates confusion, and often becomes obsolete when
one finally finds the opportunity to use it.
An inverted microscope generally allows more flexibility
and workspace for the culture (see later) and
manipulation of live cells than an upright microscope.
In addition, while inverted stands by major manufacturers
give comparable performances, it is important
to take into consideration the feasibility, accessibility,
and convenience for special third-party accessories,
such as the culture chamber, micromanipulator, filter
wheels, and confocal optics to be used. Therefore,
choice of the stand should not be made until the design
of the rest of the system becomes largely clear.
One aspect particularly relevant to living cell
imaging is the stability of the microscope stand. Some
old microscopes are prone to stage drift, particularly
at an elevated temperature. They require either extensive
manual input or an autofocusing mechanism for
time-lapse imaging. However, it is often difficult to
obtain reliable information on the stability through
manufacturers or on-site demonstration, and experience
of colleagues is often the most reliable source.
Several current microscope designs incorporate
useful automatic features, such as lamp shutters (see
later) and motorized magnification and stage controls,
which may alleviate the need to incorporate thirdparty
components and facilitate automated multimode
Objective Lenses and Contrasting Method
A basic set of objectives consists of 10x, 40x dry, 40x
immersion, 100x immersion lenses and possibly a 60
or 63x immersion lens. Dry lenses are used primarily
for scanning the samples and do not have to be expensive.
However, immersion lenses should have as high
a numerical aperture and light transmission efficiency
as possible. Because images are typically collected near
the center of the field, lenses highly corrected for flat
field usually provide no detectable benefit and are
more costly and less light efficient than simpler lenses
such as Fluar lenses. All lenses for fluorescence imaging
should be checked upon delivery for the quality
of point-spread function, using fluorescent beads as
the sample (see later).
For most applications, phase-contrast optics should
suffice for scanning the sample and for collecting
paired fluorescence and transmission images. The
presence of quarter-wave plate in the phase lens does
cause some (~<5%) light loss, although it is usually not
serious enough to defeat the use of phase-contrast
optics. The alternative is DIC optics, which requires the repeated insertion and removal of an analyzer in
the optical path when one shifts between DIC and fluorescence
optics. This, and the higher cost, makes DIC
optics less desirable in most cases.
Unless the experiment involves high-resolution
transmission optics or dark-field optics, a condenser
with a long working distance should be used in conjunction
with an inverted stand.
Control of Projection Magnification
It is critical to choose an optimal magnification for
live cell imaging, balancing between signal intensity
(favored by low magnification) and resolution (favored
by high magnification). It is particularly important to
match the final magnification with the pixel size of
the detector (see later). Projection magnification may
be controlled conveniently in some microscopes by
switching the tube lens or by adding additional lenses
(referred to as Optovar for Zeiss microscopes). New
stands allow changes of magnification to be controlled
automatically in time-lapse imaging.
Epi-illuminator and Fluorescence Filter Sets
Some new epi-illumination systems provide a light
trap, which reduces the background stray light, and
standard Kohler illumination with both field and aperture
diaphragms. The aperture diaphragm may be
used to control the lamp intensity as well as the
angular span and depth of illumination. The light trap
is useful for single molecular imaging, which is typically
limited by the background.
There are several commercial sources of highquality
fluorescence filter sets. A handbook on the
selection of filter sets may be found at the Web site of
Chroma Optics. The main consideration is to balance
signal strength (favored by cuton/cutoff filters or wide
band-pass filters) against the reduction of signal
crossover from probes of different colors (favored by
narrow band-pass filters). The latter consideration is
particularly important when an intense, long wavelength
probe is used in conjunction with a weak, short
wavelength probe. In addition to standard filter sets,
multiband filters are now readily available that allow
simultaneous illumination and/or detection of multiple
Heater Filter or Heat Mirror
When imaging living cells it is critical to remove the
infrared component from the light source, as the fluorescence
filter set may not be able to block infrared
light. Failure to do so may cause not only heat damage
to the cell, but also high background with some
infrared-sensitive cameras. The filter (e.g., BG38) or
heat-reflecting mirror may be placed either in front of
the lamp or in the epi-illuminator. Attention should be
paid to the UV transmission of these filters if UV excitation
is to be used for imaging.
Lamps and Lamp Power Supplies
The system should include both a mercury arc lamp
and a 100-W quartz-halogen lamp for epi-illumination,
coupled through a selection mirror to the microscope.
Contrary to common practice, the most suitable lamp
for fluorescence imaging of live cells is often a 100-W
quartz-halogen lamp. Unless the experiment involves
single molecule or speckle imaging, quartz-halogen
lamps are much more cost effective and are sufficiently
intense for imaging most cellular structures while
minimizing radiation damage. They also allow easy
adjustment of the light intensity, using a variable, stabilized
DC power supply.
The mercury arc lamp should consist of a wellshielded
housing, power supply, and power supply
cable to minimize the potentially damaging electromagnetic
wave during ignition. Although some power
supplies allow adjustment of the light intensity, the
range of adjustment is limited. Therefore, it is often
necessary to attenuate the light from mercury arc
lamps using a set of neutral density filters to avoid
radiation damage. Alternatively, the intensity may be
controlled using the aperture diaphragm in the epiilluminator
as mentioned earlier.
Electronic shutters should be used to control fluorescence
excitation. These shutters should be used as
much as possible to minimize the duration of illumination
and should have the interface for computer
control during automated time-lapse recording.
Cooled CCD cameras are used for most fluorescent
imaging applications. In choosing a camera, important
parameters include quantum efficiency, noise level,
pixel size and full-well capacity, and scanning frequency.
In order to minimize the excitation light for
imaging live cells, the camera should be as "sensitive"
as possible, which generally means a high quantum
efficiency, low noise, large pixel size, and slow scan
rate. The sensitivity requirement must therefore be
balanced against the required resolution (favored by a
larger number of small pixels) and imaging rate
(favored by a higher scanning frequency).
Slow-scan CCD cameras are generally limited in
their frame rate. In addition, unless the sample is very
intense, the signal-to-noise ratio is poor under short exposures. This limits both their use for high-speed
imaging and the ease in focusing the images. Focusing
is facilitated with cameras using the shutterless, frametransfer
or interline CCDs, which are able to provide a
continuous stream of images at video rate in addition
to slow-scan digital signals.
Intensified CCD cameras are generally more suitable
for high-speed imaging, although usually with a
compromised quantum efficiency. Of particular interest
are cameras that use CCD chips with the new electron
amplification technology (e.g., photometric
Cascade and Andor Ixon cameras), which allow both
long-exposure and high-speed imaging at a high
Optical Coupling of Cameras
Coupling with the detector should be achieved with
as few lens elements as possible. It is recommended to
have several couplers with different magnification
factors. In conjunction with different objective lenses
and projection magnifications mentioned earlier, they
allow a wide range of magnifications for both lightlimiting
and high-resolution applications. The Nyquist
resolution criterion should be considered when choosing
the magnification: each pixel should correspond to
no more than half the required resolution limit on the
sample. Due to the diffraction limit of the microscope,
this distance needs not be smaller than 50 nm for most
applications. The actual area imaged onto each pixel
may be determined easily by taking an image of a scale
standard (see later).
Vibration Isolation Table
A full-fledged vibration isolation table is necessary
in adverse environment, e.g., in areas of heavy traffic
or in high-rise buildings, or for demanding experiments
of micromanipulations or single molecular
imaging. Simple isolation measures may suffice otherwise.
These include inner tires under the table or
rubber isolation pads (Edmond Scientific) or tennis
balls under a slab tabletop.
Motorized Stage and Focusing Control
A motorized XY stage is optional. It allows one to
monitor multiple cells in separate regions and may
increase the output greatly in time-lapse experiments.
A motorized focusing control is required for optical
sectioning, three-dimensional imaging, and automatic
focusing. However, many simple imaging experiments
may be better served without the complications of
motorized stage controls.
Computer Hardware and Software
A high-end personal computer is required not only
for image acquisition, but also for device control and
data analysis. The system should have a high-capacity
hard disk, a recordable CD/DVD drive for archiving
and porting data, and double monitors to accommodate
all the images and control windows. Before
making a decision on the software package, it is advisable
to prepare a list of application requirements, as
many features seen in software demonstrations are
visually dazzling but practically useless. The efficiency
of a package should be judged by counting the number
of mouse clicks or key strokes for setting up and triggering
the most frequently used functions.
As a minimum for live cell imaging, the program
should be able to control the camera, shutters, and
motorized devices, to perform time-lapse recording,
and to allow changes of recording parameters without
stopping the recording. In addition, the user should be
able to review dynamic processes as movies even
during the recording. The program should also be able
to perform automatic contrast enhancement (without
losing the original intensity values) and to save images
in a nonlossy file format with automatically generated
or manually entered file names.
There are a number of options available through
microscope manufacturers and independent companies
such as Bioptechs. Discussions of cell culture on
microscopes may be found at the Web site of Bioptechs
and in McKenna and Wang (1989).
There is no "ideal" culture device for all the applications.
The range of possible devices varies from a
heated stage, a heating collar for the objective lens, a
small heated culture chamber, to a large enclosure that
fits over the entire microscope. A heated stage is the
most convenient but the least functional, as the point
of observation is over an open area far away from the
heat source. A heating collar for the objective lens
applies heat much closer to the sample. However, it
works only in conjunction with immersion lenses and
has a limited area of heating. It is usually used in conjunction
with an additional chamber or enclosure.
Large microscope enclosures provide the most stable
temperature; however, it must be designed carefully
to provide convenient access. There are also a number
of perfusion and heated culture chambers for
microscopy. These chambers generally provide excellent
optical and culture conditions; however, the associated
wires and tubing may add to the inconvenience,
and the sealed environment may not be compatible
with micromanipulation experiments.
It is important to maintain the temperature stability
to within a fraction of a degree, as even minor drifts in
temperature can cause severe drifts in focusing and sample positioning. In addition to heating, it is important
to maintain the pH and osmolarity of the culture
medium. While this should not be a problem with
sealed chambers or perfusion chambers, open chambers
should be used in conjunction with either a CO2-
independent medium (e.g., L-15) or injection of CO2 into the incubator. It should be noted that HEPESbuffered
media only slow down pH drift and are not
suitable for long-term cultures by themselves. Osmolarity
may also be a serious problem with open dishes
in a heated environment. It may be controlled by
replacing the medium periodically or by covering the
medium with a layer of mineral oil (Sigma).
Several testing samples should be prepared for
characterizing the imaging system. First, a micrometern
scale is essential for all microscopy laboratories
for determining the final magnification. Second, a "flat
field" sample is prepared by spreading a drop of
appropriate fluorophores in 50% glycerol under a
coverslip. It is useful for checking the uniformity of
epi-fluorescence illumination and for setting the field
diaphragm. Also essential is a sample of fluorescent
beads. It is prepared by diluting 0.1-µm-diameter fluorescent
latex beads (Molecular Probes) by ~105 into
melt 1% agarose, and mounting a small volume
(~20-50µl) of the suspension on a heated glass slide
under a coverslip, before letting the sample cool down.
The bead sample is used for checking the combined
optical quality of the imaging system. Defects not
readily visible with cell samples are often recognized
easily when one examines the images of single beads.
Defocused beads should appear as radially symmetric
disks or concentric rings.
Additional equipment may be required for special
purposes. Multiwavelength imaging is often achieved
with filter wheels at the illumination and/or detection
optical path. In addition, devices such as Dual-View
and Quad-View from Optical Insights generate composite
images at different wavelengths. Spinning disk
confocal heads have been used extensively for imaging
fine structures in single cells. Its balance between light
efficiency and resolution proves particularly suitable
for cultured cells. Finally, total internal reflection fluorescence
optics is being used extensively for imaging
structures near the cell-glass interface, such as focal
adhesions. Details of these approaches and devices are
beyond the scope of this article; however, their potential
use should be considered when designing the
McKenna, N. M., and Wang, Y.-L. (1989). Culturing cells on the
microscope stage. Methods Cell Biol. 29, 295-305.