UNIT 14.11 Basic Confocal Microscopy Carolyn L. Smith 1 1 National Institute of Neurological Disorders and Stroke, Bethesda, Maryland ABSTRACT This unit introduces the reader to the basic principles of confocal microscopy and the design and capabilities of current confocal microscopes. The advantages and disadvantages of confocal microscopy compared to other techniques for fluorescence imaging are described. There are also practical guidelines for sample preparation and optimization of imaging parameters, as well as examples of some of the applications of confocal microscopy. Curr. Protoc. Mol. Biol. 81:14.11.1-14.11.18. C 2008 by John Wiley & Sons, Inc. Keywords: confocal microscopy fluorescence imaging resolution three-dimensional reconstruction INTRODUCTION Confocal microscopy is a powerful tool for visualizing fluorescent specimens. The prin- cipal advantage of confocal microscopy over conventional wide-field microscopy is that it can reveal the three-dimensional structure of the specimen. Fluorescent specimens viewed with a conventional wide-field fluorescent mi- croscope appear blurry and lack contrast be- cause fluorophores throughout the entire depth of the specimen are illuminated, and fluores- cence signals are collected not only from the plane of focus but also from areas above and below. A confocal microscope selectively col- lects light from a thin (<1 µm) optical section at the plane of focus in the specimen (Fig. 14.11.1). Structures within the focal plane ap- pear more sharply defined than with a conven- tional microscope because there is essentially no flare of light from out-of-focus areas. A three-dimensional view of the specimen can be reconstructed from a series of optical sec- tions at different depths (Fig. 14.11.2). The capability for optical sectioning makes confocal microscopy well suited for study- ing the structure and function of cells using immunofluorescence reagents (Fig. 14.11.1C; UNIT 14.6), organic dyes (Fig. 14.11.1C), flu- orescent fusion proteins (Fig. 14.11.1D; Fig. 14.11.2; UNIT 9.7C), quantum dots (Michalet et al., 2005), and fluorescence in situ hy- bridization (FISH; UNIT 14.7). Confocal imag- ing in living specimens is feasible, making it possible to study dynamic processes such as gene expression, cytoskeletal assembly and turnover, chromosome dynamics (Bystricky et al., 2005), and molecular binding interac- tions (Sprague and McNally, 2005). Confo- cal microscopy also is useful for visualizing cells in situ in fixed (Fig. 14.11.1A) and liv- ing specimens. The maximum depth into the specimen at which images can be captured de- pends on the transparency of the specimen, characteristics of the objective, and excitation wavelength. Single-photon excitation with vis- ible wavelengths (450 to 650 nm) can typically penetrate up to 50 to 200 µm. Multiphoton ex- citation with infrared wavelengths (>700 nm) can penetrate deeper, up to 1 mm in some circumstances. Several types of confocal microscopes are available. The most common type is the laser scanning confocal microscope (LSCM), which captures images by scanning the specimen with a focused beam of light from a laser and collecting the emitted fluorescence signals with a photodetector. LSCMs sometimes are referred to as “spot-scanning” confocal micro- scopes, to distinguish them from microscopes that scan the specimen with a slit of light (slit- scanning) or multiple spots of light (spinning- disk or Nipkow disk). Spot-scanning LSCMs have slower image acquisition rates than slit- scanning or spinning-disk microscopes (<1 frame/sec versus 30 frames/sec or higher). However, they are more versatile in a num- ber of ways. They can accommodate lasers with a wide range of wavelengths (from the UV to the infrared) and can be configured to image multiple fluorophores either simul- taneously or sequentially. Some include spec- tral detectors that can capture the entire spec- trum of the fluorescence emitted at each pixel in the image. The most sophisticated LSCMs Current Protocols in Molecular Biology 14.11.1-14.11.18, January 2008 Published online January 2008 in Wiley Interscience (www.interscience.wiley.com). DOI: 10.1002/0471142727.mb1411s81 Copyright C 2008 John Wiley & Sons, Inc. In Situ Hybridization and Immunohisto- chemistry 14.11.1 Supplement 81
Transcript
UNIT 14.11Basic Confocal Microscopy
Carolyn L. Smith1
1National Institute of Neurological Disorders and Stroke, Bethesda, Maryland
INTRODUCTIONConfocal microscopy is a powerful tool for
visualizing fluorescent specimens. The prin-cipal advantage of confocal microscopy overconventional wide-field microscopy is that itcan reveal the three-dimensional structure ofthe specimen. Fluorescent specimens viewedwith a conventional wide-field fluorescent mi-croscope appear blurry and lack contrast be-cause fluorophores throughout the entire depthof the specimen are illuminated, and fluores-cence signals are collected not only from theplane of focus but also from areas above andbelow. A confocal microscope selectively col-lects light from a thin (<1 µm) optical sectionat the plane of focus in the specimen (Fig.14.11.1). Structures within the focal plane ap-pear more sharply defined than with a conven-tional microscope because there is essentiallyno flare of light from out-of-focus areas. Athree-dimensional view of the specimen canbe reconstructed from a series of optical sec-tions at different depths (Fig. 14.11.2).
The capability for optical sectioning makesconfocal microscopy well suited for study-ing the structure and function of cells usingimmunofluorescence reagents (Fig. 14.11.1C;UNIT 14.6), organic dyes (Fig. 14.11.1C), flu-orescent fusion proteins (Fig. 14.11.1D; Fig.14.11.2; UNIT 9.7C), quantum dots (Michaletet al., 2005), and fluorescence in situ hy-bridization (FISH; UNIT 14.7). Confocal imag-ing in living specimens is feasible, makingit possible to study dynamic processes suchas gene expression, cytoskeletal assembly andturnover, chromosome dynamics (Bystrickyet al., 2005), and molecular binding interac-
tions (Sprague and McNally, 2005). Confo-cal microscopy also is useful for visualizingcells in situ in fixed (Fig. 14.11.1A) and liv-ing specimens. The maximum depth into thespecimen at which images can be captured de-pends on the transparency of the specimen,characteristics of the objective, and excitationwavelength. Single-photon excitation with vis-ible wavelengths (450 to 650 nm) can typicallypenetrate up to 50 to 200 µm. Multiphoton ex-citation with infrared wavelengths (>700 nm)can penetrate deeper, up to 1 mm in somecircumstances.
Several types of confocal microscopes areavailable. The most common type is the laserscanning confocal microscope (LSCM), whichcaptures images by scanning the specimenwith a focused beam of light from a laserand collecting the emitted fluorescence signalswith a photodetector. LSCMs sometimes arereferred to as “spot-scanning” confocal micro-scopes, to distinguish them from microscopesthat scan the specimen with a slit of light (slit-scanning) or multiple spots of light (spinning-disk or Nipkow disk). Spot-scanning LSCMshave slower image acquisition rates than slit-scanning or spinning-disk microscopes (<1frame/sec versus 30 frames/sec or higher).However, they are more versatile in a num-ber of ways. They can accommodate laserswith a wide range of wavelengths (from theUV to the infrared) and can be configuredto image multiple fluorophores either simul-taneously or sequentially. Some include spec-tral detectors that can capture the entire spec-trum of the fluorescence emitted at each pixelin the image. The most sophisticated LSCMs
Supplement 81 Current Protocols in Molecular Biology
Figure 14.11.1 Applications of laser scanning microscopy. (A,B) Imaging in thick specimens.Neurons in a Drosophila embryo were immunolabeled with antibodies against three different tran-scription factors (images provided by Dr. Ward Odenwald of the National Institutes of Health,Bethesda, Md.; reproduced from Kamabadur et al., 1998, by permission of Cold Spring HarborLaboratory Press). (A) A single optical section (∼2.5 µm) captured with a 25×, 0.8-NA objective.Labeled neurons in the plane of focus appear sharply defined, while those outside it are not visual-ized. (B) A maximum projection of 65 optical sections collected at 2-µm intervals in the z axis. (C)Imaging intracellular structures. Dissociated rat fibroblasts were immunolabeled with anti-tubulinantibodies to visualize microtubules (green), and stained with fluorescent probes for mitochondria(Mitotracker, red) and DNA (DAPI, blue). The image is a projection of 20 optical sections (0.3-µmintervals) captured with a 100×, 1.4-NA objective. (D) Measuring molecular mobility in living cells.In a living fibroblast expressing a Golgi membrane protein (galactosyltransferase) fused to GFP(S65T), GFP fluorescence (green) is localized in the Golgi complex, shown superimposed on aDIC image of the cell. After the first image was collected, the boxed region (yellow) was scannedwith full laser power to photobleach the GFP in the boxed area. The second image was col-lected 2 sec later. Subsequent images (not illustrated) showed that the GFP-galactosyltransferaserapidly diffused back into the photobleached area. Images were captured with an LSM410 (CarlZeiss, Inc.). For the color version of this figure go to http://www.currentprotocols.com.
allow the user to control the illumination wave-length and intensity on a microsecond timescale. This feature makes it possible to performexperiments that require selectively illuminat-ing fluorophores in a defined region of inter-est in order to photobleach (Fig. 14.11.1D) orphotoactivate them. Measurement of fluores-cence recovery after photobleach (FRAP) orfluorescence loss in photobleach (FLIP) canprovide information about molecular mobilityand binding (Cole et al., 1996; McNally andSmith, 2002; Lippincott-Schwartz et al., 2003;
Sprague and McNally, 2005). Photosensitivemolecules include certain fluorescent proteins(e.g., see Patterson and Lippincott-Schwartz,2002, for the fluorescent protein PaGFP, andAndo et al., 2002, for the photosensitive pro-tein Kaede), “caged” molecules such as cagedCa2+ chelators, neurotransmitters, and secondmessengers (Nerbonne, 1996). Confocal mi-croscopy also can be used to measure flu-orescence resonance energy transfer (FRET;Wouters and Bastiaens, 2000).
In SituHybridizationandImmunohisto-chemistry
14.11.3
Current Protocols in Molecular Biology Supplement 81
Figure 14.11.2 Three-dimensional imaging in living specimens. Comparison of water- and oil-immersion objectives. Living yeast cells expressing a GFP construct that targets the mitochondrialmatrix were embedded in an aqueous solution with 0.2% agarose and visualized with (A,B) aC-APO 63×, 1.2-NA water-immersion objective or (E,F) a Plan Apochromat 100×, 1.4-NA oil-immersion objective. The images show xy (A,E) and yz (B,F) projections of stacks of 40 imagescollected at 0.2-µm intervals along the optical axis. The xy projections appear sharper than theyz projections because the resolution is higher in the focal plane of the objective than along theoptical axis. (C,G) yz projections of images of 0.19-µm fluorescent beads embedded in an aqueoussolution with 2% agarose and captured with a 63× water-immersion (C) or 100× oil-immersion (G)objective. (D,H) Intensity profiles along the horizontal and vertical axes of the beads. A 63×, 1.2-NAwater-immersion objective (D) provides better axial resolution than a 100×, 1.4-NA oil-immersionobjective (H) in specimens in an aqueous solution. Scale bars = 5 µm (A,B,E,F); 0.5 µm (C,G).Images were captured with an LSM510 laser scanning confocal microscope (Carl Zeiss, Inc.).
The purpose of this unit is to provide back-ground information and practical tips for op-timizing confocal imaging. The first section(Basis of Optical Sectioning) explains thebasic principle of confocal imaging as im-plemented in a LSCM. The second section(Configuration of an LSCM) describes thecomponents and light path in a typical LSCMand compares this with the light paths of spot-scanning microscopes and a new type of slit-scanning confocal microscope. The third sec-tion (Practical Guidelines) provides guidelinesfor preparing specimens and configuring thecritical parameters for confocal imaging. TheCommentary provides references to sources ofadditional information.
BASIS OF OPTICAL SECTIONINGConfocal microscopes accomplish optical
sectioning by scanning the specimen with afocused beam of light and collecting the fluo-rescence signals emitted by the specimen via apinhole aperture. The pinhole aperture blockssignals from out-of-focus areas of the speci-men whereas light from the focal plane passes
through the pinhole to reach the detector. Thephysical basis of optical sectioning is illus-trated in Figure 14.11.3. The microscope ob-jective focuses light from a point source (alaser) to a diffraction-limited spot in the spec-imen. The irradiation is most intense at the fo-cal spot, but areas of the specimen above andbelow the focal spot are also illuminated. Fluo-rescent molecules excited by the incident lightemit fluorescence in all directions. The objec-tive captures a portion of the emitted light. Theobjective projects light from the focal spot inthe specimen to a conjugate spot in an “image”plane. The pinhole aperture is positioned inthe image plane so as to be centered on thisspot. The light that passes through the apertureis detected by a photomultiplier tube (PMT).Light from out-of-focus areas of the specimenis spread out at the image plane and is largelyblocked by the pinhole aperture.
The diameter of the pinhole determines howmuch of the fluorescence emitted by the illu-minated cone in the specimen is detected, aswell as the thickness of the optical section.From wave optics, it is known that a point lightsource in the plane of focus of an objective
Basic ConfocalMicroscopy
14.11.4
Supplement 81 Current Protocols in Molecular Biology
Figure 14.11.3 The basis of optical sectioning. Illumination from a point light source is reflectedby a dichroic mirror into the back aperture of a microscope objective. The objective lens focusesthe light to a diffraction-limited spot within the specimen. Fluorophores at the focal spot and withinthe cones of illumination above and below it are excited, emitting fluorescence in all directions.The fluorescence captured by the objective passes through the dichroic mirror because the fluo-rescence is at a longer wavelength than the excitation. The confocal pinhole allows fluorescencefrom the focal spot to reach the photodetector and blocks fluorescence from out-of-focus areas.Redrawn from Shotton (1993).
produces a three-dimensional diffraction pat-tern in the image plane. The cross-section atthe image plane is an Airy disk, a circulardiffraction pattern with a bright central region.The radius of the bright central region of theAiry disk in the reference frame of the spec-imen is given by RAiry = 0.61λ/NA, where λ
is the emission wavelength and NA is the nu-merical aperture of the objective (Inoue andSpring, 1997). At the image plane, the ra-dius of the central region is RAiry multipliedby the magnification at that plane (Wilson,1995).
Adjustment of the pinhole to a diameterslightly less than the diameter of the centralregion of the Airy disk allows most of thelight from the focal point to reach the detec-tor and reduces the background from out-of-focus areas by ∼1000-fold relative to wide-field microscopy (Sandison et al., 1995). Theseparation of the in-focus signal from the out-
of-focus background achieved by a properlyadjusted pinhole is the principal advantage ofconfocal microscopy for examination of thickspecimens.
Point illumination and the presence of apinhole in the detection light path also produceimproved lateral and axial resolution relativeto conventional microscopy (Table 14.11.1).The actual extent of improvement depends onthe size of the pinhole. Near-maximal axialresolution is obtained with a pinhole radius of∼0.7 × RAiry, whereas optimal lateral resolu-tion is obtained with a pinhole smaller than0.3 × RAiry (Wilson, 1995). However, a pin-hole smaller than ∼0.7 × RAiry significantlyreduces the total signal, a sacrifice that maynot be worth the gain in resolution, espe-cially when imaging dim samples. In fluores-cence imaging, resolution is also influencedby the emission and excitation wavelengths(Table 14.11.1).
In SituHybridizationandImmunohisto-chemistry
14.11.5
Current Protocols in Molecular Biology Supplement 81
Table 14.11.1 Theoretical Resolutions of Confocal and Conventional Microscopesa
Objective
10×, 0.4 NA, air 40×, 0.85 NA, air 60×, 1.4 NA, oil
680 1.04 8.50 0.49 1.88 0.32 1.22aData reprinted from Brelje et al. (1993) by permission of Academic Press.bλex and λem, excitation and emission wavelengths (nm).
CONFIGURATION OF AN LSCMConfocal microscopes use lasers for illu-
mination because they provide intense exci-tation within a narrow range of wavelengths.Mixed krypton-argon gas lasers are popularfor multicolor confocal microscopy becausethey emit at three wavelengths (488, 568, and647 nm) that excite many commonly usedfluorophores—e.g., fluorescein (FITC), rho-damine, Cy3, Cy5, Alexa 488/555/568/647,green fluorescent protein (GFP), and red flu-orescent protein (mRFP or DsRed). The dis-advantage of krypton-argon lasers is that theirlife spans are short (∼2000 hr). Another way toachieve multiwavelength excitation is to com-bine the outputs of multiple lasers. Many of theconfocal microscopes currently on the marketcombine an argon laser (488 nm) with a greenhelium-neon (HeNe) laser (543 or 594 nm) anda red HeNe laser (633 nm). The argon laser alsomay provide 458- and 514-nm lines, which canbe used to excite the cyan and yellow variantsof GFP (CFP and YFP). Some confocal mi-croscopes can accommodate a 405-nm diodelaser. The 405-nm laser is more optimal for ex-citation of CFP than the 458-nm line of the ar-gon laser, and also excites photosensitive GFP(PaGFP). It can even be used to visualize someUV fluorophores such as DAPI and HoechstDNA dyes, although 405 nm is not the opti-mal excitation wavelength for these dyes. UVargon lasers (351/364 nm) also are available.Inclusion of a 405-nm or UV argon laser addsconsiderably to the cost of the confocal mi-croscope system due to the requirement foradditional optical components to handle thesewavelengths.
The light path in a simple confocal micro-scope is illustrated in Figure 14.11.4. The out-put of the laser (or the combined output of mul-tiple lasers) is reflected into the optical axis ofthe microscope by the primary dichroic beamsplitter (splitter 1 in Fig. 14.11.4). Wavelength-selection filters are inserted into the lightpath to block specific laser lines, and neutral-density filters may be inserted to attenuatethe illumination. In current, high-end confocalsystems, the line selection and neutral-densityfilters have been replaced with an electroni-cally controlled acousto-optical tunable filter(AOTF). An AOTF can alter its transmissioncharacteristics to allow selected wavelengthsto pass, while completely blocking others. AnAOTF also provides precise control over theattenuation of the individual laser beams.
The scanner deflects the laser beam intothe objective at varying angles in order to scanthe laser beam across the specimen. Severaldifferent technologies for scanning have beendevised. The most common method employsa pair of galvanometer mirrors. One mirror os-cillates rapidly to excite sequential spots alongthe x-axis of the specimen, and the second mir-ror oscillates more slowly to move the illumi-nation from line to line in the y-axis.
The fluorescence emissions that are col-lected by the objective follow the reversepath through the scanner to the primarydichroic beam splitter, and are thereby“descanned” (Fig. 14.11.4). The fluorescencesignals (which are at a longer wavelength thanthe excitation due to the Stokes shift; are trans-mitted through the beam splitter. To simul-taneously image fluorescence from multiple
Basic ConfocalMicroscopy
14.11.6
Supplement 81 Current Protocols in Molecular Biology
Figure 14.11.4 The light path of a laser-scanning confocal microscope (LSCM) set up for simul-taneous imaging of FITC and lissamine rhodamine. The 488- and 568-nm lines of a krypton-argonlaser are reflected by dichroic beam splitter 1 into the optical axis of the microscope. The beam isreflected by a mirror into the microscope objective, which focuses the beam to a diffraction-limitedspot in the specimen. The scanner consists of a pair of galvanometer mirrors that deflect the laserbeams so as to scan the spot across the specimen in a raster pattern. Fluorescence emitted aseach point is illuminated travels the reverse path through the scanning system. The FITC fluores-cence (peak at 520 nm) and lissamine rhodamine fluorescence (peak at 590 nm) pass throughdichroic beam splitter 1 to dichroic beam splitter 2, which transmits the lissamine rhodamine fluo-rescence to photomultiplier tube 1 and reflects the FITC fluorescence to photomultiplier tube 2. Avariable pinhole in front of each photodetector blocks light from out-of-focus areas of the specimenwhile allowing light from the focal plane to reach the detector.
fluorophores requires selection of a primarydichroic beam splitter that reflects each of therequired excitation wavelengths and transmitsthe emissions of all of the fluorophores. Sec-ondary dichroic mirrors split the fluorescenceemissions from different fluorophores for de-tection by separate detectors. Emission filtersare inserted in the light path to the detectors(Fig. 14.11.4) to block back-scattered exci-tation light and to reduce bleed-through ofsignals between channels. Current high-endconfocal microscopes use more sophisticatedtechnology for emission discrimination; de-scriptions of the designs of specific systemsare available from the vendors.
The fluorescence captured by the objectivefocuses to a stationary spot (Airy disk) in theimage plane (Fig. 14.11.3). The pinhole aper-ture is positioned in the image plane so as to becentered on the Airy disk. The diameter of thepinhole aperture can be adjusted to allow op-timization for different Airy-disk sizes, whichvary with the objective’s numerical apertureand the emission wavelength. Adjustment of
the diameter of the pinhole to a value of 0.7to 1.0 Airy unit allows most of the in-focuslight to reach the detector and blocks most ofthe out-of-focus light. In systems with a sep-arate pinhole aperture for each detector, thepinhole apertures are located immediately infront of the detectors. Incorporation of a sep-arate pinhole for each detector allows the userto optimize the pinhole settings for differentwavelengths.
The photodetectors in LSCMs are photo-multiplier tubes (PMTs), which generate elec-trons at a rate proportional to the intensity ofthe incoming fluorescence signal (Art, 2006).The PMT output is converted to a digital imagethat can be displayed on a computer monitorand stored as a digital file for later analysis.Digitization may be at 8-bit (256 gray lev-els), 10-bit (1024 gray levels), or 12-bit (4096gray levels) resolution . Conofcal microscopestypically have two to four PMTs for reflectedlight/epifluorescence imaging and may have,in addition, a photodetector for transmittedlight.
In SituHybridizationandImmunohisto-chemistry
14.11.7
Current Protocols in Molecular Biology Supplement 81
In spinning-disk confocal microscopes, theillumination from a laser or white light sourcepasses through pinholes in the Nipkow diskso as to excite fluorescence at multiple sites(∼1000) within the specimen. The disk re-volves rapidly (1000 to 5000 rpm), causing theilluminating spots to sweep across the spec-imen as uniformly spaced scan lines (Inoueand Inoue, 2002; Toomre and Pawley, 2006).Fluorescence emitted by the specimen thatis collected by the objective returns throughthe same pinholes in the Nipkow disk thatprovided the excitation light before it is de-tected by a full-field CCD camera. In thisway, point light sources and detector pinholesto block out-of-focus fluorescence are pro-vided, with the advantage of higher collec-tion speeds than a spot scanner. Drawbacksof this approach include decreased illumina-tion to the specimen from light loss throughthe pinholes, and the inability to change thepinhole diameter. This means that, unlike thecase with a spot scanner, optimal confocal-ity is achieved only for one objective magni-fication, and the thickness of the optical sec-tion cannot be changed. Although the reducedspecimen illumination generates a smaller flu-orescent signal, scientific-grade CCD camerashave significantly higher quantum efficienciesthan the PMTs used for fluorescence detectionin LSCMs and are able to detect these levelsof fluorescence more than adequately. Fluores-cent specimens have been reported to undergoless photobleaching during examination with aspinning-disk confocal microscope than witha LSCM. The lower rate of photobleaching isthought to be due to the lower illuminationlevels (Inoue and Inoue, 2002).
A new type of slit-scanning confocal mi-croscope (LSM 5 Live; Carl Zeiss, Inc.) hasrecently been introduced that allows images tobe acquired at rates as fast as or faster thancan be achieved with a spinning-disk confo-cal microscope, and with as low or lower ratesof photobleaching. The system adopts princi-ples from both the spot scanner and the spin-ning disk in that it uses a single scanning gal-vanometer to move an illumination line that iscombined with a sensitive single-line CCD de-tector. The point source of light from the laseris optically converted to a narrow line, whichis reflected onto the specimen by a novel beamsplitter consisting of a mirrored line on trans-parent glass. The line illumination is scannedacross the specimen. The emitted fluorescencefrom the specimen that is collected by the ob-jective passes through the beam splitter andis detected by a linear CCD detector. A slit
aperture in front of the detector blocks out-of-focus light, analogous to the pinhole aperturein an LSCM. The LSM 5 Live has somewhatpoorer resolution than a spot-scanning LSCM,but can capture images much more rapidly.
PRACTICAL GUIDELINES
Sample PreparationThe preeminent goal in preparing samples
for imaging with a confocal microscope is tomaximize the fluorescence signals while pre-serving the three-dimensional structure of thespecimen. Ideally, the sample should be lessthan ∼50 µm in thickness, although thickersamples can be visualized. Guidelines forpreparing fixed and living samples are de-scribed below.
FixationA standard fixative for fluorescence mi-
croscopy is 2% to 4% formaldehyde inPBS. Formaldehyde penetrates cells rapidlyand preserves the antigen-recognition sitesfor many antibodies. However, formaldehydecross-links proteins slowly and may causevesiculation of membranes. Some commercialpreparations of formaldehyde (formalin) con-tain methanol, which shrinks cells. Techniquesfor optimizing formaldehyde fixation are de-scribed by Bacallao et al. (2006). Fixativescontaining a small amount of glutaraldehyde(0.125% to 0.25%) in addition to formalde-hyde preserve cellular morphology better, butglutaraldehyde destroys the epitopes for someantibodies. Glutaraldehyde fixation inducesautofluorescence but autofluorescence can bereduced by treating the sample after fixationwith NaBH4 (1 mg/ml in PBS, pH 8.0, usingtwo treatments of 5 min each for dissociatedcells, longer for thicker samples). An alter-native procedure for preparing specimens forimmunocytochemistry is to immerse them incold (−20◦C) methanol or acetone but fixationby this method causes severe shrinkage.
Choices of fluorophoresThe choice of fluorophores should take into
account the available laser lines and the detec-tor channels of the confocal microscope. Ex-citation is most efficient at wavelengths nearthe peak of the excitation spectrum of the flu-orophore, but a precise match is not required.For experiments that require imaging mul-tiple fluorophores with standard photodetec-tors (PMTs), it is best to select fluorophoresthat are excited by different laser lines, in
Basic ConfocalMicroscopy
14.11.8
Supplement 81 Current Protocols in Molecular Biology
Figure 14.11.5 Excitation spectra of representative fluorophores and emission wavelengths oflasers for confocal microscopy. The graph at the top shows the excitation spectra of Marina Blue,Alexa 488, Alexa 555, and Alexa 647 (Molecular Probes). The emission wavelengths of laserscommonly used for confocal microscopy are shown below. Data for the excitation spectra are fromMolecular Probes.
order to minimize spectral crossover (bleed-through) between the channels (Fig. 14.11.5).Excitation and emission spectra for many flu-orophores are available via the Internet (seeInternet Resources). A recommended combi-nation of fluorophores for excitation at 405 nm,488 nm, 543/561 nm, and 633 nm is MarinaBlue, Alexa 488, Alexa 555, and Alexa 647 (allavailable from Molecular Probes/Invitrogen).The nucleic acid stain DAPI can be excitedby illumination at 405 nm, although ultravio-let excitation (350 nm) is more optimal. Thecyanine dyes Cy2, Cy3, and Cy5 (availablefrom Jackson ImmunoResearch Laboratories)are also suitable for confocal microscopy. Formultiwavelength imaging with a spectral de-
tector and spectral unmixing, it is important toselect fluorophores that have distinct emissionspectra, but there is no advantage in using flu-orophores that have differing excitation spec-tra. Indeed, it is best to use fluorophores thathave similar excitation maxima, so that theycan be excited with a single laser line reflectedinto the microscope with a single-wavelengthdichroic mirror. Other important criteria toconsider in selecting fluorophores for confocalmicroscopy are the quantum efficiencies andrates of photobleaching. In addition, the stain-ing protocol should be designed to producesimilar signal intensities in each channel. Formore information on fluorophores for confo-cal imaging, see Tsien et al. (2006), Giepmans
In SituHybridizationandImmunohisto-chemistry
14.11.9
Current Protocols in Molecular Biology Supplement 81
et al. (2006), and the Molecular ExpressionsWeb site (see Internet Resources).
Control samplesConfocal microscopes rely on electronic
image enhancement techniques that can makeeven dim autofluorescence signals or nonspe-cific background staining look bright. In or-der to distinguish a real signal from back-ground, it is essential to prepare and examineappropriate control samples. For immunoflu-orescence experiments with one primary an-tibody, the appropriate control samples areunstained specimens and specimens treatedwith the secondary antibody but no primaryantibody. Other control experiments may berequired to verify the specificity of label-ing. Experiments with primary and secondaryantibodies require additional controls to testwhether the secondary antibodies cross-reactwith the “wrong” primary antibody. Singlystained samples also should be prepared andimaged to determine the extent of spectralcross-over between the channels.
Mounting the specimenSelection of a mounting medium should
take into account the type of microscope ob-jective that will be used to observe the spec-imen (see section on Microscope objectives).In order for an objective to perform optimally,the mounting medium should have the same
refractive index (RI) as the objective immer-sion medium. Mismatches in the refractive in-dices produce spherical aberration leading toloss of light at the detector, as well as de-creased z-axis resolution and incorrect depthdiscrimination. Image deterioration caused byspherical aberration increases with depth intothe specimen; therefore, matching the immer-sion and mounting medium refractive indicesis particularly important for thick specimens.The refractive indices of some commonly usedmounting media are listed in Table 14.11.2.
Mounting media that have refractive in-dices close to that of immersion oil (n = 1.51)include DPX (n = 1.5; ProSciTech) and Per-mount (n = 1.52; ProSciTech). However, spec-imens must be dehydrated prior to mount-ing in these media, and dehydration causesshrinkage and distortion. Moreover, some flu-orophores cannot withstand dehydration. Cellsretain their three dimensional shapes whenthey are kept in physiological saline (PBS) ora mixture of PBS and glycerol (Bacallao et al.,2006). If the specimen is to be mounted undera coverglass, it may be necessary to support thecoverglass to avoid damaging the specimen.
Addition of an antioxidant (antifade agent)to the mounting medium helps to alleviate pho-tobleaching of synthetic fluorophores such asthose used for immunocytochemistry. One ofthe best antifade agents is 100 mg/ml 1,4-diazabicyclo[2,2,2]octane (DABCO; Sigma;Bacallao et al., 2006). n-Propyl gallate (Giloh
Table 14.11.2 Refractive Indexes of Common Immersion and Mounting Media
0.25% PPD/0.0025% DABCO/5% n-propyl gallate in glycerol 1.473a
VectaShield (Vector Labs) 1.458a
Slow Fade (Molecular Probes) 1.415b
ProLong (Molecular Probes) 1.3865b,c
aData from Bacallao et al. (2006).bData from Molecular Probes.cRI for liquid medium. (RI for solidified medium will be higher.)
Basic ConfocalMicroscopy
14.11.10
Supplement 81 Current Protocols in Molecular Biology
and Sedat, 1982) and p-phenylenediamine(PPD; Johnson et al., 1982) are also effec-tive antifade agents, but the former may causedimming of the fluorescence while the lattermay damage the specimen (Bacallao et al.,2006). A wide variety of mounting media isavailable from commercial sources (Biomeda,Electron Microscopy Sciences, ProSciTech,Molecular Probes, Vector Laboratories), andmany of these contain antifade agents. It iswise to check with the fluorophore provider forrecommendations about which mounting me-dia and antifade agents to use. Antioxidantsdo not reduce photobleaching of fluorescentproteins.
Living specimensMicroscopy on living specimens grown
in vitro is most conveniently performed withan inverted microscope, because the speci-mens can be viewed through the bottom ofthe culture chamber and the top can be openedfor access. To allow imaging with an oil- orwater-immersion objective, the culture cham-ber substrate should be a coverglass. The cov-erglass can be coated with poly-L-lysine (usinga 1 mg/ml solution; Sigma) to promote adhe-sion of the specimens. Nonadherent specimenscan be immobilized by embedding them in athin layer of low-melting-point agar (0.2%).Culture chambers with coverglass substratescan be made from standard plastic petri dishesby boring holes in their bottoms and affixingcoverglasses to the holes with Silgard (Dow-Corning). Culture chambers with coverglasssubstrates are also available from commercialsources (Labtek coverglass chamber, FisherScientific; MatTek glass-bottom culture dish,MakTek Corp). Alternatively, cells may begrown on a coverglass that can be mountedin a chamber for observation on a microscope.A simple chamber can be constructed from agasket cut from a sheet of silicon rubber ora soft plastic ruler and affixed to a glass mi-croscope slide with silicon grease. The wellformed by the gasket is filled with medium,and then the coverglass with cells attached issealed onto the well. More elaborate cham-bers, some having built-in heaters and/or portsfor perfusion, are available from commercialsources (see Internet Resources for a list ofsuppliers).
Specimens that need to be kept warm dur-ing observation pose a particular challenge be-cause temperature fluctuations can make it dif-ficult to maintain focus. Probably the best wayto keep specimens warm is to place the entiremicroscope in a temperature-controlled en-
closure. Alternative strategies include warm-ing the microscope stage with heated air (us-ing an air stream incubator or hair dryer)or infrared lamps, or using a temperature-controlled specimen chamber (Terasaki andDailey, 1995; Dailey et al., 2006). If an oil-or water-immersion objective is used, heatingthe objective helps to maintain the specimenat the desired temperature. Microscope enclo-sures, stage warmers, temperature-controlledchambers, and objective heaters are availablefrom suppliers of microscopes and microscopeaccessories.
Living specimens should be kept in amedium that is buffered to maintain the correctpH. Many commonly used culture media arebuffered with bicarbonate and require an atmo-sphere with 5% to 10% CO2 to maintain thecorrect pH. For microscopy, it is more conve-nient to use a buffer that maintains the correctpH in air. Many types of cells can be main-tained for several hours in a balanced saline so-lution or culture medium that is buffered withHEPES (10 to 20 mM). Use of a medium thatcontains phenol red should be avoided becausephenol red adds background fluorescence andcan produce oxygen radicals when exposedto intense illumination. Addition of 0.3 U/mlOxyrase (Oxyrase, Inc.) to the medium canhelp to alleviate photobleaching of syn-thetic fluorophores (Waterman-Storer et al.,1993).
Optimizing Imaging Parameters
Microscope objectivesHigh-NA objectives are optimal for fluores-
cence microscopy because they collect morelight than low-NA objectives (brightness isproportional to NA4). Oil-immersion objec-tives have the highest numerical apertures (NA= 1.4 or 1.46). However, oil-immersion ob-jectives have short working distances (100 to200 µm). Moreover, they work optimally onlywith specimens mounted in a medium with arefractive index the same as that of immersionoil (n = 1.51). Mismatch of the refractive in-dices leads to a deterioration of image qualitythat becomes increasingly severe with depthinto the specimen (Fig. 14.11.2). When a high-NA oil objective is used to image a specimenmounted in an aqueous medium, image qual-ity and signal brightness decline noticeably atdistances of 5 to 10 µm from the coverglass.Mismatch of the refractive indices also leadsto spatial distortion in the z-axis. The actualmovement of the focal plane in the specimen(ds) produced by a movement of the objective
In SituHybridizationandImmunohisto-chemistry
14.11.11
Current Protocols in Molecular Biology Supplement 81
(dobj) depends on the ratio of the refractive in-dices: ds/dobj = ns/nobj (Majlof and Forsgren,2002).
A water-immersion objective is useful forimaging living specimens that are more thana few microns thick (Fig. 14.11.2). Water-immersion objectives with numerical aper-tures of 1.2 are available. These objectives aredesigned for viewing specimens mounted un-der a coverglass (0.17 µm; no. 1.5) and havefairly short working distances (130 to 220 µm).“Dipping” objectives, which are intended foruse without a coverglass, have lower numeri-cal apertures (NA = 0.9) and longer workingdistances (1 to 2 mm).
Objectives differ in their transmission ef-ficiency and degree of correction for spheri-cal and chromatic aberration and flatness offield. Plan Apochromat objectives provide theflattest fields of view and color correction forthree wavelengths. Plan Apochromat objec-tives generally transmit efficiently throughoutthe visible spectrum (400 nm to 700 nm), butmay transmit poorly in the UV (<400 nm)or infrared (>700 nm; Keller, 1995, 2006).Some objectives that are less highly corrected(Fluar, Plan NeoFluar, Plan Fluor) providehigher transmission at visible, UV, and in-frared wavelengths. For additional informa-tion about objectives for confocal microscopy,see Salmon and Canman (1998), Keller (1995,2006), Benham (2002), and the MolecularExpressions Web site (see Internet Resources).
Pinhole sizeAs explained in the section on the Basis of
Optical Sectioning, the size of the detector pin-hole has a critical influence on image quality.A pinhole with a diameter slightly less thanor equal to the diameter of the bright centralregion of the Airy disk will let most of thelight from the plane of focus reach the detec-tor, while blocking most of the out-of-focusflare. The lateral resolution will be ∼10% bet-ter than that obtainable by conventional mi-croscopy with the same optics (Centonze andPawley, 1995, 2006), although not as good ascan be achieved with a smaller pinhole. Lat-eral resolution continues to improve as pin-hole radius is decreased down to a pinholesize of ∼0.2 × RAiry but a pinhole this smallexcludes ∼95% of the signal (Wilson, 1995).Axial resolution improves as pinhole size de-creases, down to ∼0.7× RAiry then levels off.The best trade-off between signal intensity andresolution will depend on the characteristics ofthe sample and the required resolution.
Scan zoomThe scan zoom determines the dimensions
of the area in the specimen that is scanned.Increasing the zoom reduces the dimensionsof the scan area. The pixel number remainsthe same; consequently, individual pixels rep-resent a smaller area. For example, the scanarea at zoom 2 is one quarter the scan area atzoom 1, and the pixel dimensions are half aslarge in each dimension. That is, if the pixeldimensions represent 0.25 µm × 0.25 µmat zoom 1, then dimensions are 0.125 ×0.125 µm at zoom 2.
For each objective, there is an optimalzoom setting that yields pixel dimensionssmall enough to take advantage of the fullresolution of the objective, but large enoughto avoid oversampling. In order for the min-imum resolvable entity to be visible on thedisplay monitor, the pixel dimension needs tobe smaller than (less than one-half) the opticalresolution. However, if the pixel size is madetoo small by using a higher-than-optimal zoomfactor, the specimen is subjected to more irra-diation than necessary, with an increased riskof photobleaching. The rate of photobleach-ing increases proportionally to the square ofthe zoom factor (Centonze and Pawley, 1995,2006). A guideline for selecting an appropriatezoom factor derived from information theory(the Nyquist Sampling Theorem) states thatthe pixel dimensions should be equal to theoptical resolution divided by 2.3 (see Webband Dorey, 1995; Pawley, 2006). The opti-cal resolution in confocal imaging dependson the numerical aperture of the objective,the refractive index of the immersion medium,the excitation and emission wavelengths, andthe diameter of the pinhole aperture. Valuescalculated for different objectives and wave-lengths using the point-spread functions (PSF)for wide-field and confocal microscopy aregiven in Table 14.11.1. The lateral resolutionfor confocal microscopy can be approximatedby: Reselx,y confocal = 0.4λ/NA (see Webb andDorey, 1995). This equation assumes the use ofan infinitesimal pinhole; Reselx,y will be largerwith a pinhole of 0.7 to 1 Airy unit.
z-axis sectioning intervalIn order to study the three-dimensional
structure of a specimen, a series of imagesare captured at fixed intervals throughout theentire depth of the specimen. The intervalneeded between focal planes to achieve op-timal resolution in the z-axis is not as smallas the x,y pixel dimensions because the axial
Basic ConfocalMicroscopy
14.11.12
Supplement 81 Current Protocols in Molecular Biology
resolution is poorer than the lateral resolu-tion (see Table 14.11.1). The optimal inter-val (according to the Nyquist Sampling The-orem) is equal to the axial resolution dividedby 2.3. The axial resolution for an objectivein confocal imaging can be approximated by:Reselz confocal = 1.4λn/NA2, where n is the re-fractive index (see Webb and Dorey, 1995).Collecting images at shorter intervals resultsin oversampling, with an increased risk ofphotobleaching.
with illumination intensity up to a level atwhich emission saturates. Optimal signal-to-background and signal-to-noise ratios are ob-tained with illumination levels well below sat-uration (Tsien et al., 2006). The illuminationintensity on a laser-scanning microscope canbe adjusted by operating the laser at submax-imal power and by inserting neutral-densityfilters into the light path or varying the trans-mission through the AOTF. In general, the bestimages are obtained with illumination levelsthat are as high as possible without producingunacceptable rates of photobleaching.
PMT black level and gainThe contrast and background of confocal
images are determined by the gain and offsetsettings of the PMT amplifiers. To obtain max-imal information, the offset and gain should beadjusted to take advantage of the full dynamicrange of the PMTs. The appropriate offset set-ting can be found by scanning while the lightpath to the PMT is blocked. The image thatappears on the display monitor should be justbarely brighter than the background, which isblack (gray level = 0). To set the gain, scan thespecimen and adjust the gain so that the bright-est pixel in the image is slightly below white(gray level = 255, for 8-bit images). Ensuringthat all signals fall within the dynamic rangeof the PMT is especially important for quan-titative imaging experiments. Confocal imag-ing software typically includes a pseudocolorimage display mode (“range indicator”) thatfacilitates selection of appropriate offset andgain settings by highlighting pixels with in-tensity values near 0 or 255.
Reducing noiseConfocal images are inherently noisy due
to the statistics of photon emission and limitedquantum efficiency of the detectors (Pawley,2006). Improved signal-to-noise ratios can be
attained by scanning the specimen at a slowerrate or by scanning multiple times and av-eraging the signals. Current LSCMs allowindividual lines in the image to be scannedrepeatedly and averaged. Line averaging gen-erally produces sharper images than frame av-eraging (which averages full frames) becausethere is less risk of blurring due to movementsor changes in the specimen. Averaging reducesnoise by a factor of (1/
√n, where n is the
number of frames or lines averaged), but morescans will result in more bleaching. Residualnoise can be reduced by image processing (seebelow).
Imaging multiple fluorophoresConfocal microscopes can typically be con-
figured to capture images of two or more fluo-rophores simultaneously or sequentially. Eachapproach has advantages and disadvantages.For simultaneous imaging, the specimen isscanned with all of the required excitationwavelengths, and the emissions of the differ-ent fluorophores are split for detection by sep-arate photodetectors (Fig. 14.11.4). The draw-back of this approach is that spectral crossoverbetween channels may occur if the emissionspectra of the fluorophores overlap. If each flu-orophore is excited by only one laser line, thenexciting them sequentially will avoid spectralcrossover. The disadvantage of sequential ex-citation is that there may be misalignment ofthe signals in different channels, particularly ifthe specimen is alive and moving. A third wayof imaging multiple fluorophores is availablein confocal systems in which the laser excita-tion is controlled with an AOTF; such systemscan scan each line of the specimen sequentiallywith different excitation wavelengths with atime delay between scan lines of less than amillisecond. Line-by-line wavelength switch-ing provides rapid acquisition of fluorescencesignals from each spot in the specimen whileavoiding the spectral crossover between chan-nels that may occur when the fluorophores areexcited simultaneously.
Image displayConfocal images are typically displayed as
8-bit grayscale or 24-bit RGB (red/green/blue)color images. Each channel of an RGB im-age can represent a different fluorophore(Fig. 14.11.1A to C). Color mixtures indicatecolocalization of fluorophores within a pixel.A RGB fluorescence image can be mergedwith a grayscale transmitted-light image byadding the transmitted-light image to eachchannel of the RGB image (Fig. 14.11.1D;
In SituHybridizationandImmunohisto-chemistry
14.11.13
Current Protocols in Molecular Biology Supplement 81
fluorescence in green channel merged withDIC image).
The three-dimensional data set obtained bycapturing a series of optical sections throughthe specimen can be used to compute viewsof the specimen from different viewing an-gles. Commercial confocal microscopes typ-ically include the capability to generate or-thogonal views of the specimen (xy, xz, andyz) and may permit views from arbitrary an-gles. An xy projection or “z-series projection”is a two-dimensional display formed by merg-ing multiple image planes (Fig. 14.11.1B,C,Fig. 14.11.2A,E). The most common typeof projection is a “maximum” projection inwhich each pixel represents the intensity ofthe brightest pixel in the z-axis. Another typeof projection, referred to as “surface render,”displays the most superficial pixels with inten-sities above a defined threshold. Projectionsalso can be created for different viewing angles(Fig. 14.11.2B,F). Projections of the specimenfrom different viewing angles can be combinedto create an animation in which the specimenappears to rotate in space. Such animationsgive the viewer a striking impression of thethree-dimensional geometry of the specimen.Generating two projections at azimuths differ-ing by 4◦ to 10◦ creates stereo pairs that canbe visualized with a stereo viewer or color-coded and merged to form a stereo anaglyph.Volume-rendering imaging software is avail-able that provides additional options for three-dimensional visualization and measurements(see Internet Resources).
Image processingConfocal images can be enhanced by
an image processing technique known as“deconvolution” or “restoration” (Cannellet al., 2006; Holmes et al., 2006). The purposeof deconvolution is to remove the blur causedby diffraction. By measuring or approximat-ing the point spread function of the optics, onecan mathematically “work backwards” to ac-count for the distorting effects of the optics onthe true image. Deconvolution improves con-trast and resolution and reduces noise. Severalalgorithms for deconvolution have been de-vised that differ in computational intensity andthe extent to which they are quantitative (i.e.,retain intensity information). The importanceof eliminating noise for quantitative imagingwith confocal microscopy is discussed in de-tail by Pawley (2006). Although deconvolutionis the most robust method for reducing noise,it is computationally intensive. An alternative
and quicker method for reducing noise is toperform a linear deblurring operation (Russ,2002).
COMMENTARYEffective use of a confocal microscope re-
quires understanding of the principles of im-age formation and knowledge of how to set upand use a microscope. Salmon and Canman(1998) describe the components of a light mi-croscope and provide protocols for setting upa microscope for transmitted light and epiflu-orescence imaging. This unit also lists refer-ences to literature on light microscopy. TheHandbook of Biological Confocal Microscopy(Pawley, 1995, 2006) is a comprehensive ref-erence book on confocal microscopy. It in-cludes chapters on fundamental principles, in-strumentation, image acquisition and display,sample preparation, and much more. ConfocalMicroscopy (Wilson, 1990) provides a thor-ough discussion of the principles behind con-focal imaging. Cell Biological Applicationsof Confocal Microscopy (Matsumoto, 2002)discusses the performance of different typesof confocal microscopes and contains prac-tical information about common applicationssuch as imaging immunofluorescence and cal-cium ion indicators. Additional applicationsare described in Confocal and Two Photon Mi-croscopy: Foundations, Applications and Ad-vances (Diaspro, 2002). The Molecular Ex-pressions Web site also is an excellent sourceof information about light microscopy, in-cluding confocal microscopy (see Internet Re-sources).
Confocal microscopy is only one of sev-eral available techniques for capturing opticalsections in fluorescent specimens. An alterna-tive to confocal microscopy is computational“deconvolution” of images captured by wide-field epifluorescence microscopy (McNallyet al., 1999; Boccacci and Bertero, 2002).Computational deconvolution makes use of allof the fluorescence captured by the objective,in contrast to confocal microscopy, which dis-cards fluorescence from out-of-focus areas. Inaddition, wide-field microscopy can employCCD cameras that have higher quantum effi-ciencies than the photodetectors used for con-focal microscopy. For these reasons, wide-field microscopy and deconvolution can besuperior to confocal microscopy for imagingdim specimens or specimens that are suscepti-ble to photobleaching or photodamage. How-ever, computational deconvolution of imagesis time-consuming and does not work well in
Basic ConfocalMicroscopy
14.11.14
Supplement 81 Current Protocols in Molecular Biology
specimens with high levels of dispersed fluo-rescence. Confocal microscopy allows directvisualization of optical sections and is appli-cable to a wider range of specimens.
Another technique for confocal imagingtakes advantage of the optical phenomenonknown as multiphoton excitation. Multiphotonmicroscopy allows deeper penetration into tis-sue than either wide-field microscopy or con-ventional (single-photon) microscopy, and isparticularly useful for imaging in thick speci-mens such as tissue slices or multicellular or-ganisms. However, this method suffers fromloss of resolution due to the longer illumi-nation wavelength and absence of a detectorpinhole.
TroubleshootingTest samples are useful for monitoring the
performance of a confocal microscope. A mi-crometer slide should be used to check the spa-tial calibration of each objective. Fluorescentmicrospheres with mixtures of fluorophores(FluoSpheres; Molecular Probes) are usefulfor checking the x,y and axial alignment ofimages acquired at different excitation wave-lengths. Misalignment of the images in the xyplane may indicate that the pinholes are notcentered or that the lasers need to be aligned;misalignment in the z axis may be due to in-correct setting of a collimating lens, misalign-ment of pinholes, or chromatic aberration inthe objective. The optical resolution of the mi-croscope can be measured by capturing imagesof submicroscopic (<0.2 nm) fluorescent mi-crospheres (Fig. 14.11.2C,G). The images ofthe microspheres should be radially symmet-rical in the xy plane and elliptical in the z-axis(Fig. 14.11.2C). Horizontal and axial resolu-tions are defined by the full width at half-maximal intensity (FWHM) of intensity pro-files along the horizontal and vertical axes ofthe beads (Fig. 14.11.2D,H).
ages of fluorescent structures in thick speci-mens (Figs. 14.11.1 and 14.11.2). The max-imum depth at which adequate images canbe obtained depends on the objective and theoptical properties of the specimen. With ahigh-NA immersion objective, it may be pos-sible to capture images at depths of >100µm in a specimen that is transparent andnot heavily stained (Centoze and Pawley,1995). However, if the specimen scatterslight, both the illumination intensity and the
proportion of the emitted fluorescence thatis captured by the objective decline with in-creasing focal depth. Mismatch of the re-fractive indices of the immersion mediumand specimen will further reduce the depthat which adequate signal can be obtained.A low-NA objective can capture images atgreater depths but provides much poorer axialresolution.
A three-dimensional reconstruction of thespecimen can be generated from a series ofoptical sections at appropriately spaced inter-vals along the optical axis. The reconstructioncan be viewed from any angle, but the viewalong the optical axis of the objective will ap-pear sharper than off-axial views, because thelateral resolution of the objective is better thanthe axial resolution (Fig. 14.11.2). The axialdistortion can be corrected by computationaldeconvolution (Wouterlood, 2005).
Confocal imaging in living specimens isfeasible although care must be taken to avoidphototoxicity and photodamage. Robust flu-orophores such as the fluorescent proteinsEGFP and EYFP can be imaged hundreds oftimes with minimal photobleaching and noapparent phototoxicity, provided that the il-lumination is kept at a low level. Syntheticfluorophores, such as organelle-specific dyes(Molecular Probes), are generally more pho-tosensitive, although in some applications therate of bleaching can be reduced by the addi-tion of Oxyrase. The maximum rate at whichimages can be collected will depend on thescan speed, resolution, and area. Typical scantimes for a 512 × 512 image with a spot-scanning LSCM are 1 to 4 sec/frame.
A common application of confocal mi-croscopy is to determine the relative dis-tributions and extent of co-localization ofmolecules tagged with different fluorophores(Brelje et al., 2002). As many as four dif-ferent fluorophores can be discriminated ona confocal microscope with laser excitationat 350/405, 488, 546/568, and 633/647 nmand standard photodetectors, provided thatthe excitation spectra of the fluorophores arewell separated and matched to the laser lines.Confocal microscopes with spectral detectorscan discriminate larger combinations of fluo-rophores on the basis of their emission spectra.Spectral detection and linear unmixing allowsdiscrimination of fluorophores with highlyoverlapping emission spectra, such as GFP andYFP (Dickinson et al., 2001).
Confocal microscopy also is well suited forvisualizing fluorescent proteins such as GFP,
In SituHybridizationandImmunohisto-chemistry
14.11.15
Current Protocols in Molecular Biology Supplement 81
cyan FP (CFP), yellow FP (YFP), and DsRed.CFP and YFP can be visualized with minimalcross-talk between channels on a confocal mi-croscope with 405-nm and 514-nm excitation.The overlap of the excitation and emissionspectra of CFP and GFP or GFP and DsRedmay result in cross-talk between channels inexperiments with these combinations of fluo-rophores. New fluorescent proteins have beendeveloped that are less prone to dimerizationand provide more optimal combinations formulti-color imaging (Shaner et al., 2004). Flu-orescent proteins have also been incorporatedinto biochemical reporters for measuring in-tracellular calcium, kinase activity, and othersignaling molecules (Zhang et al., 2002).
Photosensitive fluorescent proteins areavailable that undergo a change in spectralproperties upon photoactivation. Photactivat-able GFP (PaGFP; Patterson and Lippincott-Schwartz, 2002) exhibits little fluorescenceunder 488-nm illumination prior to activation,but undergoes a 100-fold increase in fluores-cence after photoactivation at 400 to 430 nm.An LSCM with a 405- or 413-nm laser canbe used to photoactivate PaGFP within a user-defined region of interest within the specimen,and thereby selectively “highlight” GFP flu-orescence within that region. The activatedGFP retains its fluorescence indefinitely and,importantly, manifests these properties evenwhen fused to another protein. PaGFP fusionproteins provide a useful tool for studyingthe intracellular dynamics of proteins and or-ganelles (Karbowski et al., 2004).
LSCMs that incorporate an AOTF to con-trol the illumination wavelength and intensitycan be configured to perform various typesof photobleaching experiments. Measurementof fluorescence recovery after photobleach(FRAP) or fluorescence loss in photobleach(FLIP) can provide information about molec-ular mobility and binding (Cole et al., 1996;McNally and Smith, 2002; Lippincott-Schwartz et al., 2003). In FRAP, fluorescencein a small region of the specimen is pho-tobleached by scanning with high-intensityillumination, and recovery of fluorescenceinto the bleached area is then monitored byscanning with low-intensity illumination (Fig.14.11.1D). The rate of return of fluorescentmolecules into the bleached area may be gov-erned by diffusion, binding interactions withother molecules, or a combination of both, andappropriate mathematical models have beendeveloped to analyze these responses (Spragueand McNally, 2005). In FLIP, a region of the
specimen is photobleached several times witha delay between the bleach scans, and imagesare collected during this process to monitor thedistributions of bleached and nonbleached flu-orescent molecules. Observation of FLIP canshow whether there is exchange of fluorescentmolecules between two compartments of a cellor whether a fluorescent structure is a singleorganelle or a network of contiguous but inde-pendent organelles (Cole et al., 1996).
The spatial precision by which two fluo-rophores can be said to co-localize on thebasis of light microscopy is limited by theoptical resolution (∼0.2 µm in the xy planeand 0.6 µm in the z-axis). The phenomenonof fluorescence resonance energy transfer(FRET) can potentially reveal whether twofluorophores are within <10 nm proximity.FRET (Wouters and Bastiaens, 2000) is thenonradiative transfer of energy from a fluores-cent donor molecule to an acceptor molecule.Energy transfer occurs only if the moleculesare within a distance of less than ∼10 nm,and only if the emission spectrum of the donoroverlaps the excitation spectrum of the accep-tor. One application of FRET is to determinewhether two populations of molecules undergobinding interactions. One population is labeledwith donor fluorophores (e.g., CFP) and thesecond is labeled with acceptor fluorophores(e.g., YFP). Several techniques for measuringFRET have been devised (Jares-Erijman andJovin, 2003) and many of these can be carriedout with current LSCMs.
Current LSCMs are much superior to theirpredecessors in sensitivity, speed of image ac-quisition, and versatility. Although they areexpensive ($200,000 to $600,000) and requirecostly service contracts to ensure optimal per-formance, their many benefits justify thesecosts.
ACKNOWLEDGEMENTSThe author wishes to thank Dr. James
Galbraith (National Institutes of Health,Bethesda, Md.) for helpful comments on themanuscript. This research was supported bythe Intramural Research Program of the Na-tional Institute of Neurological Diseases andStroke at the National Institutes of Health,Bethesda, Md.
LITERATURE CITEDAndo, R., Hama, H., Yamamoto-Hino, M., Mizuno,
H., and Miyawaki, A. 2002. An optical markerbased on the UV-induced green-to-red photo-conversion of a fluorescent protein. Proc. Natl.Acad. Sci. U.S.A. 99:12651-12656.
Basic ConfocalMicroscopy
14.11.16
Supplement 81 Current Protocols in Molecular Biology
Art, J. 2006. Photon detectors for confocal mi-croscopy. In Handbook of Biological ConfocalMicroscopy, 3rd ed. (J. Pawley, ed.) pp. 251-264.Springer, New York.
Bacallao, R., Sohrab, S., and Phillips, C. 2006.Guiding principles of specimen preservation forconfocal fluorescence microscopy. In Handbookof Biological Confocal Microscopy, 3rd ed.(J. Pawley, ed.) pp. 368-380. Springer, NewYork.
Benham, G. 2002. Practical aspects of objectivelens selection for confocal and multiphotondigital imaging techniques. In Cell BiologicalApplications of Confocal Microscopy, 2nd ed.(B. Matsumoto, ed.) pp. 247-300. AcademicPress, San Diego, Calif.
Boccacci, P. and Bertero, M. 2002. Image-restoration methods: Basics and algorithms. InConfocal and Two-Photon Microscopy: Founda-tions, Applications, and Advances (A. Diaspro,A., ed.), pp. 253-269. John Wiley & Sons,Hoboken, N.J.
Brelje, T.C., Wessendorf, M.W., and Sorenson,R.L. 1993. Multicolor laser scanning confocalimmunofluorescence microscopy: Practical ap-plications and limitations. In Cell BiologicalApplications of Confocal Microscopy (B.Matsumoto, ed.) pp. 98-182. Academic Press,San Diego, Calif.
Brelje, T.C., Wessendorf, M.W., and Sorenson,R.L. 2002. Multicolor laser scanning confo-cal immunofluorescence microscopy: Practicalapplications and limitations. In Cell BiologicalApplications of Confocal Microscopy, 2nd ed.(B. Matsumoto, ed.), pp. 166-244. AcademicPress, San Diego, Calif.
Bystricky, K., Laroche, T., van Houwe, G.,Blaszczyk, M., and Gasser, S.M. 2005. Chro-mosome looping in yeast: Telomere pairing andcoordinated movement reflect anchoring effi-ciency and territorial organization. J. Cell Biol.168:375-387.
Canell, M.B., McMorland, A., and Soeller, C. 2006.Image enhancement by deconvolution. In Hand-book of Biological Confocal Microscopy, 3rded. (J. Pawley, ed.) pp. 488-500. Springer, NewYork.
Centonze, V. and Pawley, J. 1995. Tutorial on prac-tical confocal microscopy and use of the con-focal test specimen. In Handbook of BiologicalConfocal Microscopy, 2nd ed. (J. Pawley, ed.)pp. 549-570. Plenum, New York.
Centonze, V. and Pawley, J. 2006. Tutorial on practi-cal confocal microscopy and use of the confocaltest specimen. In Handbook of Biological Con-focal Microscopy, 3rd ed. (J. Pawley, ed.) pp.627-649. Springer, New York.
Cole, N., Smith, C., Sciaky, N., Terasaki, M.,Edidin, M., and Lippincott-Schwartz, J. 1996.Diffusional mobility of Golgi proteins in mem-branes of living cells. Science 237:797-801.
Dailey, M.E., Manders, E., Soll, D.R., and Terasaki,M. 2006. Confocal microscopy of living cells. InHandbook of Biological Confocal Microscopy,
3rd ed. (J. Pawley, ed.) pp. 381-403. Springer,New York.
Diaspro, A. (ed.) 2002. Confocal and Two PhotonMicroscopy: Foundations, Applications andAdvances. Wiley-Liss, New York.
Dickinson, M.E., Bearman, G., Tille, S., Lansford,R., and Fraser, S.E. 2001. Multi-spectralimaging and llinear unmixing add a wholenew dimension to laser scanning fluorescencemicroscopy. Biotechniques 31:1274-1278.
Giepmans, B.N.G., Adams, S.R., Ellisman, M.H.,and Tsien, R.Y. 2006. The fluorescent tool-box for assessing protein location and function.Science 312:217-224.
Giloh, H. and Sedat, J.W. 1982. Fluorescencemicroscopy: Reduced photobleaching of rho-damine and fluorescein protein conjugates byn-propyl gallate. Science 217:1252-1255.
Holmes, T.J., Biggs, D., and Abu-Tarif, A. 2006.Blind deconvolution. In Handbook of BiologicalConfocal Microscopy, 3rd ed. (J. Pawley, ed.)pp. 468-487. Springer, New York.
Inoue, S. and Inoue, T. 2002. Direct-View High-Speed Confocal Scanner: The CSU-10. InCell Biological Applications of Confocal Mi-croscopy, 2nd ed. (B. Matsumoto, ed.) pp. 88-128. Academic Press, San Diego, Calif.
Inoue, S. and Spring, K.R. 1997. Video Microscopy:The Fundamentals, 2nd ed. Plenum, New York.
Jares-Erijman, E.A. and Jovin, T.M. 2003. FRETimaging. Nat. Biotechnol. 21:1387-1395.
Johnson, G.D., Davidson, R.S., McNamee, K.C.,Russell, G., Goodwin, D., and Holborow,E.J. 1982. Fading of immunofluorescence dur-ing microscopy: A study of the phenomenonand its remedy. J. Immunol. Methods 55:231-242.
Kamabadur, R., Koizumi, K., Stivers, C., Nagle, J.,Poole, S., and Odenwald, W. 1998. Regulation ofPOU genes by castor and hunchback establisheslayered compartments in the Drosophila CNS.Genes Dev. 12:246-260.
Karbowski, M., Arnoult, D., Chen, H., Chan, D.C.,Smith, C.L., and Youle, R.J. 2004. Quantitationof mitochondrial dynamics by photolabeling ofindividual organelles shows that mitochondrialfusion is blocked during the Bax activation phaseof apoptosis. J. Cell Biol. 164:493-499.
Keller, E. 1995. Objective lenses for confocal mi-croscopy. In Handbook of Biological ConfocalMicroscopy, 2nd ed. (J. Pawley, ed.) pp. 111-126. Plenum, New York.
Keller, E. 2006. Objective lenses for confocal mi-croscopy. In Handbook of Biological ConfocalMicroscopy, 3rd ed. (J. Pawley, ed.) pp. 145-161.Springer, New York.
Lippincott-Schwartz, J., Altan-Bonnet, N., andPatterson, G.H. 2003. Photobleaching and pho-toactivation: Following protein dynamics inliving cells. Nat. Cell Biol. Suppl:S7-S14.
Majlof, L. and Forsgren, P. 2002. Confocal mi-croscopy: Important considerations for accu-rate imaging. In Cell Biological Applications of
In SituHybridizationandImmunohisto-chemistry
14.11.17
Current Protocols in Molecular Biology Supplement 81
Confocal Microscopy, 2nd ed. (B. Matsumoto,ed.) pp. 149-164. Academic Press, San Diego,Calif.
Matsumoto, B. (ed.) 2002. Cell Biological Appli-cations of Confocal Microscopy. 2nd ed. Aca-demic Press, London.
McNally, J.G. and Smith, C.L. 2002. Photobleach-ing by confocal microscopy. In Confocal andTwo Photon Microscopy: Foundations, Appli-cations and Advances (A. Diaspro, ed.) pp. 525-538. Wiley-Liss, New York.
Michalet, X., Pinaud, F.F., Bentolila, L.A., Tsay,J.M., Doose, S., Li, J.J., Sundaresan, G., Wu,A.M., Gambhir, S.S., and Weiss, S. 2005. Quan-tum dots for live cells, in vivo imaging, and di-agnostics. Science 307:538-544.
Russ, J. 2002. The Image Processing Handbook,4th ed. CRC Press, Boca Raton, Fla.
Salmon, E.D. and Canman, J.C. 1998. Proper align-ment and adjustment of the light microscope.Curr. Protoc. Cell. Biol. 0:4.1.1-4.1.26.
Sandison, D.R., Williams, R.M., Wells, K.S.,Strickler, J., and Webb, W.W. 1995. Quantita-tive fluorescence confocal laser scanning mi-croscopy. In Handbook of Biological ConfocalMicroscopy, 2nd ed. (J. Pawley, ed.) pp. 39-54.Plenum, New York.
Shaner, N.C., Campbell, R.E., Steinbach, P.A.,Giepmans, B.N., Palmer, A.E., and Tsien, R.Y.2004. Improved monomeric red, orange and yel-low fluorescent proteins derived from Disco-soma sp. Red fluorescent protein. Nat. Biotech-nol. 22:1567-1572.
Shotton, D.M. 1993. Electronic acquisition of lightmicroscope images. In Electronic Light Mi-croscopy (D.M. Shotton, ed.) pp. 1-38. Wiley-Liss, New York.
Sprague, B.L. and McNally, J.G. 2005. FRAP anal-ysis of binding: Proper and fitting. Trends CellBiol. 15:84-91.
Terasaki, M. and Dailey, M.E. 1995. Confocal mi-croscopy of living cells. In Handbook of Biolog-ical Confocal Microscopy, 2nd ed. (J. Pawley,ed.) pp. 327-346. Plenum, New York.
Toomre, D. and Pawley, J.B. 2006. Disk-scanningconfocal microscopy. In Handbook of Biolog-ical Confocal Microscopy, 3rd ed. (J. Pawley,ed.) pp. 221-238. Springer, New York.
Tsien, R.Y., Ernst, L., and Waggoner, A. 2006. Fluo-rophores for confocal microscopy: Photophysicsand photochemistry. In Handbook of BiologicalConfocal Microscopy, 3rd ed. (J. Pawley, ed.)pp. 338-352. Springer, New York.
Waterman-Storer, C.M., Sanger, J.W., and Sanger,J.M. 1993. Dynamics of organelles in the mi-totic spindles of living cells: Membrane and mi-crotubule interactions. Cell Motil. Cytoskeleton26:19-39.
Webb, R.H. and Dorey, C.K. 1995. The pixilatedimage. In Handbook of Biological Confocal Mi-croscopy, 2nd ed. (J. Pawley, ed.) pp. 55-68.Plenum, New York.
Wilson, T. (ed.) 1990. Confocal Microscopy. Aca-demic Press, London.
Wilson, T. 1995. The role of the pinhole in confo-cal imaging system. In Handbook of BiologicalConfocal Microscopy, 2nd ed. (J. Pawley, ed.)pp. 167-182. Plenum, New York.
Wouterlood, F.G. 2005. 3-D reconstruction of neu-rons from multichannel confocal laser scanningimage series. Curr. Protoc. Neurosci. 32:2.8.1-2.2.8.16.
Wouters, F.S. and Bastiaens, P.I.H. 2000. Imag-ing protein-protein interactions by fluorescenceresonance energy transfer (FRET) microscopy.Curr. Protoc. Cell Biol. 7:17.1.1-17.1.15.
Zhang, J., Campbell, R.E., Ting, A., and Tsien,R.Y. 2002. Creating new fluorescent probes forcell biology. Nature Rev. Mol. Cell Biol. 3:906-918.
KEY REFERENCESInoue and Spring, 1997. See above.Covers the basics of light microscopy, video mi-croscopy, and much more.
Matsumoto, 2002. See above.Good source of practical information about confo-cal imaging.
Pawley, 1995, 2006. See above.Two editions of comprehensive reference book onconfocal microscopy.
Russ, 2002. See above.Guide to digital image processing.
INTERNET RESOURCEShttp://rsb.info.nih.gov/ijImageJ is a public domain image analysis pro-gram developed by W. Rasband (Research ServicesBranch, National Institute of Mental Health, NIH)for operating systems running Java (including Win-dows/PC and OSX/Macintosh). ImageJ has manyuseful tools for analysis of confocal images.
Basic ConfocalMicroscopy
14.11.18
Supplement 81 Current Protocols in Molecular Biology
A manual written by Tony Collins that describesthe use of ImageJ to visualize and analyze confocalimages.
http://www.molecularexpressions.comThe Molecular Expressions Web site is a rich sourceof information about all aspects of light microscopy,including confocal microscopy. It includes sectionson the basic principles of confocal imaging, instru-mentation, sample preparation, and choices of flu-orophores. An interactive tutorial “Choosing flu-orophore combinations for confocal microscopy”allows the user to determine the extent of spec-tral crossover that will occur when imaging differ-ent combinations of fluorophores with specific laserlines and filter sets.
Vendors of Confocal MicroscopesThese provide product descriptions, manuals, tuto-rials and literature.
Sources of chambers for maintaining living speci-mens during observation by microscopy.
http://listserv.buffalo.edu/user/sub.htmlMany topics of interest to confocal microscopistsare discussed on the confocal listserver operatedby the listserver at the University at Buffalo. Tosubscribe to the list, go to the URL and type “con-focal” in the box that asks which list one wishes tojoin.
