Chapter 15. Exploring Fungal Activity with Confocal and Multiphoton Microscopy

DK3133_C015.fm Page 308 Tuesday, April 19, 2005 9:56 AM



308



15.2



Czymmek



CONFOCAL MICROSCOPY



In 1955, Marvin Minsky invented the confocal microscope, which was called a doublefocusing stage scanning microscope. As a student at Harvard, Minsky used a zirconium

arc light source, two objectives with conjugal focal points, a scanning stage, and a surplus

military radar scope to generate the first scanning confocal image (Minsky, 1988). In 1960,

the first laser was patented and built. However, nearly a decade had passed before the first

published report of a laser scanning microscope that incorporated a 5-mW HeNe laser

with an x-y-z scanning objective lens (Davidovits and Egger, 1969). Pioneering work

continued in several labs (Brakenhoff et al., 1979; Wilson and Sheppard, 1984) with the

arrival of the first commercial instruments in the mid-1980s. Ultimately, a series of

improved laser and computer capabilities, as well as many other technical advances, led

to development of the highly versatile and powerful confocal systems of today.

Confocal microscopy has been described as one of the most significant additions to

the field of microscopy in the last century (Blancaflor and Gilroy, 2000). Clearly, its impact

has permeated virtually every aspect of biological imaging. Although there are variations

in how a confocal image is generated, by far the most common approach is to use laser

light focused to a point in a sample (Figure 15.1a). Because a significant number of detailed

published reports on the principles of confocal microscopy are readily available, an indepth review of the inside workings of a confocal microscope will not be provided here.

For more comprehensive treatments on the subject, see the following recommended reading: a compilation of patents and publications related to confocal microscopy (Masters,

1996), an excellent evaluation of a variety of critical parameters for optimizing confocal

performance (Zucker and Price, 2001), and the Handbook of Biological Confocal Microscopy (Pawley, 1995). Reviews on confocal microscopy specifically addressing applications

in mycological research, including basic theoretical considerations of lateral and axial

resolution, optical section thickness, and z-sectioning, are also available (Kwon et al.,

1993; Czymmek et al., 1994).



15.3



MULTIPHOTON MICROSCOPY



Multiphoton fluorescence microscopy is a powerful technology that enables the acquisition

of optical sections without the use of the pinhole aperture typically used for confocal

microscopy. Because this technology is relatively new, some emphasis will be placed on

aspects of image formation and how multiphoton microscopy compares with confocal

microscopy. The two-photon principle was first described by Goeppert-Mayer (1931). The

potential of this concept for imaging was initially reported by Sheppard and Kompfner

(1978) and finally realized 12 years later in cultured cells stained with Hoechst 33258

(Denk et al., 1990). The two-photon effect occurs when a fluorescent molecule simultaneously absorbs two photons, producing an electronic transition from the ground to excited

state equal to two times the energy of each incident photon (Figure 15.2). For example,

the simultaneous absorption of two red photons (each 980 nm) can yield an electronic

transition equivalent to a single blue photon (490 nm).

Although the excitation properties of a fluorescent molecule are different for singlephoton and multiphoton events, the emission spectrum remains unchanged, regardless of

how the electronic transition occurred (Xu and Webb, 1996; Xu et al., 1996). Under specific

imaging conditions, a multiphoton effect can be induced with more than two photons (e.g.,

three-photon excitation; Maiti et al., 1997). For the sake of brevity, this discussion will



DK3133_C015.fm Page 309 Tuesday, April 19, 2005 9:56 AM



Exploring Fungal Activity with Confocal and Multiphoton Microscopy

(a)



309



(b)

Confocal



Multiphoton

PMT

detector



PMT

detector

Pinhole



Collector

lens



Collector

lens



Laser

exitation



Dichroic

mirror



Laser

exitation



Objective

lens



Objective

lens



Sample



Dichroic

mirror



Sample



PMT detector



Figure 15.1 Simplified light path of laser scanning confocal (a) and multiphoton (b) microscopes.

(a) Laser light focused via an objective lens to a diffraction-limited spot in a labeled sample will

excite fluorescence. Any fluorescence generated from the spot returning through the objective lens

is then focused via a collector lens to a conjugal focal point at a pinhole (gray-filled solid lines).

The out-of-focus fluorescence signal generated by the cone of light above or below the focal plane

(dashed lines), as well as scattered light, will not come into focus at the pinhole and will be blocked

from the detector. Because the laser is scanned in a raster pattern across the sample (using linear

mirror galvanometers), a digital image is created with each pixel representing a single coordinate

during the scan. The resulting image is an optical section. (b) Multiphoton imaging is achieved using

mode-locked near-infrared lasers that produce short-duration pulses (~100 femtoseconds) and high

peak powers. The rapidly pulsed near-infrared light is focused to a diffraction-limited spot by a

high-numerical-aperture objective lens. Because the multiphoton effect has a quadratic dependency

on illumination intensity, the likelihood of such an event occurring outside the objective focal spot

decreases rapidly along the beam path. Hence, an inherent confocal image is generated by raster

scanning the spot across the specimen of interest. Because the fluorescence signal emanates from

only a small volume (defined by the NA of the objective lens), no pinhole is required and detectors

can be positioned in a variety of locations, including the transmitted light path.



be limited to two-photon excitation, and two- or three-photon excitation will be referred

to collectively as multiphoton. For a two-photon event to occur, a significant density of

incident photons is required. This typically is achieved using ultrafast, mode-locked nearinfrared lasers (e.g., Ti:sapphire, 100- to 200-femtosecond pulse duration, 76-MHz repetition rate, one pulse every 13.2 nsec). Under the appropriate conditions, these lasers

produce short-duration pulses with the high peak power required for a multiphoton effect

and an average power low enough to make specimen damage negligible. Such lasers are

typically tunable over a range of ~700 to 1000 nm, which permits optimal wavelength

selection to elicit an efficient multiphoton effect. The rapidly pulsed near-infrared light is

focused to a diffraction-limited spot by a high-numerical-aperture (NA) objective lens.



DK3133_C015.fm Page 310 Tuesday, April 19, 2005 9:56 AM



310



Czymmek



Single photon



Two photon



Excitation Emission



Excitation Emission



Figure 15.2 Jablonksi diagram illustrating the absorption and emission of light, comparing an

electronic transition with single-photon and two-photon events. When a fluorophore absorbs the

appropriate wavelength of light, it is excited from the ground state to a higher-energy level. With

fluorescence, when the molecule returns to the ground state, a photon having lower energy than the

incident photon is emitted. In this diagram, a two-photon electronic transition required the instantaneous absorption of two lower-energy photons to achieve the same electronic transition as the

higher-energy single-photon event.



The actual volume of this spot can be as small as 0.1 femtoliters (µm3) for a high-NA

(1.4) lens (Denk et al., 1995). Because the multiphoton effect has a quadratic dependency

on illumination intensity, the likelihood of such an event occurring outside the objective

focal spot decreases rapidly along the beam path (z-axis). Hence, an inherently confocal

image is generated by raster scanning the laser spot across the specimen.

Hybrid confocal/multiphoton systems share many optical components (compare

Figure 15.1a and 15.1b). Fluorescence generated using multiphoton excitation may be

collected following the same light path as the confocal signal, but the pinhole typically

would be opened to its widest position, maximizing signal collection. This is referred to

as descanned detection because the emission signal passes back through the same scanning

mirrors used for laser excitation. Alternatively, external detectors may be placed in appropriate positions to collect more light in a nondescanned configuration. In Figure 15.1b, a

detector is placed below the sample in the transmitted light path. Such a configuration

would potentially improve overall sensitivity by minimizing the number of optical elements

in the collection pathway. However, this approach is also more likely to have problems

with stray light from room lights, monitors, etc., and the microscope must be shielded as

much as possible. In order to circumvent this, a relatively simple system modification for

nondescanned detectors, based on the principle of phase-sensitive demodulation of the

pulsed laser, permits clean separation of multiphoton images in the presence of ambient

room light (Fisher et al., 2002).

In specific situations, multiphoton microscopy has several advantages over conventional fluorescence and confocal imaging:



DK3133_C015.fm Page 311 Tuesday, April 19, 2005 9:56 AM



Exploring Fungal Activity with Confocal and Multiphoton Microscopy

(a)



311



(b)



Figure 15.3 Comparison of confocal (a) and multiphoton (b) imaging within a thick specimen.

XZ images of the same FITC-lysozyme-loaded 100-µm-diameter sepharose bead using confocal and

multiphoton microscopy illustrated the improved signal collection deep within the sample using

multiphoton microscopy. The top of the image was the direction from which the incident laser

originated. (Images provided courtesy of S. Dziennik and A. Lenhoff.)



1.



2.

3.

4.



(a)



Under the appropriate power (usually 3 to 5 mW at the specimen), near-infrared

light is far less damaging to many living samples than visible and UV light

(Stelzer et al., 1994).

Lower-energy near-infrared light is scattered less; thus, imaging deeper into

highly scattering thick specimens is possible (Figure 15.3 and Figure 15.4).

No confocal pinhole is required; therefore, the emission signal collection is

more efficient.

Photobleaching is restricted to the focal plane rather than throughout the beam

path, as with confocal microscopy (Figure 15.5).



(b)



Figure 15.4 Comparison of confocal (a) and multiphoton (b) imaging within a thick fungal

specimen. (a) Fusarium oxysporum hyphae expressing ZsGreen fluorescent protein can be imaged

100 µm into nutrient agar using confocal microscopy. However, regions of blurriness due to scattered

light effects can be seen. This phenomenon is most often observed when overlying hyphae are

present. (b) Multiphoton image at the same optical plane as image A demonstrating improved imaging

deep within the specimen.



DK3133_C015.fm Page 312 Tuesday, April 19, 2005 9:56 AM



312



Czymmek



(a)



(b)



Figure 15.5 Comparison of confocal (a) and multiphoton (b) imaging after photobleaching. XZ

images of the same fluorescein isothiocyanate (FITC)-filled 100-µm-diameter sepharose beads photobleached using confocal and multiphoton microscopy. In confocal mode (a), photobleaching a single

plane (arrow) within the sepharose bead demonstrated that the fluorophore was considerably bleached

above and below the plane of focus. In multiphoton mode (b), photobleaching was clearly restricted

to the narrow plane of focus (arrow). (Images provided courtesy of S. Dziennik and A. Lenhoff.)



5.

6.

7.



Phototoxic effects are essentially limited to the focal plane.

Inner filter effects due to absorption of incident laser light by fluorescent molecules before the plane of focus are eliminated.

Multiphoton excitation can provide highly localized spatial control (on the order

of femtoliters) for fluorescence recovery after photobleaching (FRAP), photoactivation, and uncaging experiments.



Even with its many advantages, a number of practical considerations must be weighed

when choosing multiphoton microscopy over confocal or other conventional imaging

methods, namely, whether the imaging requirements include one or more of the advantages described above. If this is the case, a good knowledge of other important factors

involved with generating optimal multiphoton image formation is beneficial and will be

described further.

One distinct difference between confocal and multiphoton microscopy is reduced

resolution. Principally, due to lower-energy near infrared radiation (NIR) light forming a

larger diffraction-limited spot at the objective focal plane, the resolution of two-photon

vs. one-photon (confocal microscopy) excitation for the same fluorophore results in an

approximately twofold decrease in resolution for two-photon excitation (Denk et al., 1995;

Wolleschensky et al., 1998). However, when a confocal aperture is used in conjunction

with multiphoton excitation, a nearly 50% improvement in multiphoton resolution can be

achieved (Stelzer et al., 1994). Although resolution enhancement is gained by incorporating

a pinhole, usable signal is rejected and system sensitivity is sacrificed.

Multiphoton systems, like UV confocal microscopes, are expensive primarily due

to the need for specialized high-power lasers. They also require specific optical coatings

to allow efficient laser transmission at NIR wavelengths throughout the system and a

collimation lens to bring the NIR laser coincident with any visible lasers when performing

simultaneous multiphoton and confocal imaging. Although some lenses have been

designed with exceptional NIR throughput, many standard objective lenses have been

optimized for visible or UV transmission and require significantly more laser power to



DK3133_C015.fm Page 313 Tuesday, April 19, 2005 9:56 AM



Exploring Fungal Activity with Confocal and Multiphoton Microscopy



313



achieve a multiphoton effect. Fortunately, most direct coupled lasers have sufficient power

at the lower and upper ends of the tuning curve, where the lower laser output can be more

challenging. Fiber-optic laser delivery, which has shown great utility with continuous

wave-visible and UV lasers, poses special difficulties with multiphoton systems. This is

in large part due to decreased peak power, a restricted range of tunable wavelengths, and

sensitivity to misalignment. Some progress has been made in this area (Helmchen et al.,

2002); however, the vast majority of multiphoton systems are direct coupled, facilitating

ease of use. With the above-mentioned laser throughput issues, it is advisable to empirically

evaluate available lenses for a given system and specimen. Although not directly a throughput concern when tuning at NIR wavelengths, water absorption peaks at 760, 820, and

900 nm, and strongly between 920 and 980 nm, make the attainment of mode lock (pulsed

laser light) problematic unless the laser cavity is purged with dry nitrogen gas. In addition,

significant absorption is still possible in the aqueous sample environment, resulting in

increased and potentially harmful localized heating effects (Konig et al., 1996). Such

heating effects may also occur with increased laser excitation powers and with NIRabsorbing molecules or media (e.g., melanin and salts). Tirlapur and Konig (1999) used

this heating phenomenon to their advantage in Arabidopsis root meristematic cells by

focusing the multiphoton NIR laser to the plasma membrane of individual cells, thus

increasing subsequent uptake of propidium iodide. Any cell that was coupled to the targeted

cell would also incorporate the dye, providing some information about which cells in the

meristem were linked cytoplasmically.

A major optical effect that is not a concern with confocal microscopy is related to

pulsed laser light. When a light pulse propagates through optical elements (e.g., the

objective lens), the higher-frequency components travel more slowly than the lower-frequency components and the pulse becomes “chirped” or frequency swept (Helmchen et

al., 2001). This phenomenon is also known as group velocity dispersion, and the result is

a temporal spreading of the pulse that can cause a considerable reduction in peak intensity,

and hence the multiphoton effect (Fan et al., 1999). The dispersion properties of a given

objective lens (and other optical components) can result in significant pulse broadening,

which can be compensated for if the system is configured to do so (Guild et al., 1997).

The ability to tune most multiphoton lasers is advantageous, since the optimal

excitation wavelength for a given fluorescent molecule can be selected, as opposed to

single-photon confocal, where at best several discrete wavelengths are available. However,

one common difficulty associated with multiphoton microscopy is determining the optimal

excitation wavelength. This problem can be exacerbated by the fact that some fluorescent

molecules have poor multiphoton absorption cross sections. The multiphoton absorption

cross section (expressed in Goppert-Mayer (GM) units: 1 GM = 10–50 cm4 sec/photon)

(Kiskin et al., 2002) indicates the absorption maximum and how well a fluorophore is

excited at a given wavelength. Ideally, the simple doubling of the one-photon excitation

maximum would yield the appropriate two-photon maximum, and this is a good approximation for a number of fluorescent probes, such as rhodamine B, fluorescein, DiI, and

lucifer yellow (Xu and Webb, 1996). However, in many cases, this rule does not apply

due to vibronic coupling (Xu et al., 1996). Recent developments in automated tunable

lasers will permit the rapid generation of two-photon excitation characteristics of individual

fluorescent molecules with reasonable accuracy, eliminating this as a complicating factor

when using new or uncharacterized fluorophores or unknown tissue-related fluorescence.

In addition, laser operation will be greatly simplified, as manually tuned lasers require

specific, in-depth training for proper use.

The fact that multiphoton excitation can interact as a single-photon event in a discrete

volume presents many intriguing possibilities for uncaging and photoactivation techniques.



DK3133_C015.fm Page 314 Tuesday, April 19, 2005 9:56 AM



314



Czymmek



Photoactivatable green fluorescent protein (GFP) (PA-GFP) is basically nonfluorescent

when excited with 488-nm laser light, but become 100 times more fluorescent after

irradiation with UV light (Patterson and Lippincott-Schwartz, 2002). PA-GFP molecules

can be easily and specifically photoactivated in user-defined areas of the cell. Spatial and

temporal changes in tagged molecules can be followed after activation. To date, no

published reports have demonstrated PA-GFP with multiphoton excitation. However, my

lab, as well as others (George Patterson, personal communication), has successfully activated PA-GFP in mammalian cells, and it is likely that published reports will soon follow.

Successful uncaging of caged compounds has been an area of sporadic success. Caged

compounds that would normally be readily cleaved upon UV irradiation are poorly released

with the appropriate multiphoton excitation. Efforts have been made to find improved

multiphoton caged compounds (Albota et al., 1998). However, Kiskin et al. (2002) reported

that caged compounds would require a 31-GM cross section to be physiologically useful,

and that current caged compounds are only a few GM at best; hence, the laser powers

required to uncage existing compounds would be harmful to cells. With that said, photoactivation and photolysis of caged compounds using the precision of multiphoton excitation, if and when available, would provide added capabilities that cannot be realized to

the same extent with confocal microscopy.

The usefulness of multiphoton microscopy for reducing photobleaching varies. Even

though photobleaching is restricted to the objective focal plain with multiphoton imaging

(Figure 15.5b), bleaching rates at higher-incident laser powers can greatly exceed those

observed with confocal microscopy (Patterson and Piston, 2000). However, the authors

point out that this effect is greatly reduced and very manageable when using the lower

powers typically used to image live cells. Ultimately, when imaging near the coverslip,

where scattered light events are less pronounced, confocal microscopy would likely be

the best choice for many fungal samples. However, imaging deeper into highly scattering

tissues can result in a rapid drop in the signal-to-noise ratio, with a concomitant increase

in spherical aberration for confocal imaging. In such cases, the image degradation noted

in confocal microscopy can be alleviated with multiphoton microscopy, markedly improving image quality (Figure 15.4 here and Figure 2 in Howard, 2001).



15.4



SPECTRAL IMAGING



Recently, spectral confocal/multiphoton microscopes have become commercially available. The ability to derive spectral information from samples has significant implications

for fungal research. There are several approaches to obtain spectral data from a sample

using a grating, prism, or series of long-pass and band-pass filters (Haraguchi et al., 2002).

Spectral imaging is not limited to confocal or multiphoton microscopy and can be used

in conjunction with conventional fluorescence as well, resulting in defined spectral regions

collected by a single detector or array of detectors. A specified total range of collected

fluorescence (for example, 500 to 600 nm) could consist of 10 images, each representing

10-nm windows of the spectrum. Intensity values at any given pixel or group of pixels

can be plotted over the selected range, providing an emission curve. In the case of multiple

probes, defining a fingerprint for each individual fluorophore in a single-label experiment

results in a reference spectrum that can be used to cleanly separate even closely overlapping fluorophores by a process called linear unmixing (Hiraoka et al., 2002). Figure 15.6

illustrates an example of raw data collected from a mixture of three strains of Fusarium,

each transformed to express a different reef coral fluorescent protein. In this case, the



DK3133_C015.fm Page 315 Tuesday, April 19, 2005 9:56 AM



Exploring Fungal Activity with Confocal and Multiphoton Microscopy

473



483



494



505



516



526



537



548



558



569



580



315



590



Figure 15.6 Spectral confocal microscopy of fluorescent proteins. Cytoplasmic expression of the

fluorescent proteins AmCyan, ZsGreen, and ZsYellow resulted in significant fluorescence emission

overlap that was readily separated in germinating conidia of Fusarium using spectral confocal and

linear unmixing techniques. This data set was acquired on an LSM510 META confocal microscope

using 12 channels (10.2-nm lambda width per channel) of a 32-channel photomultiple tube (PMT)

array ranging from 473 to 590 nm (channel center points). The arrows depict spores expressing

AmCyan (473 nm), ZsGreen (505 nm), and ZsYellow (558 nm). Note the intensity changes for these

spores as you compare adjacent images in the lambda series. See Figure 15.7 for umixed results.



emissions from the three fluorescent proteins overlap significantly, making it problematic,

if not impossible, to cleanly separate using conventional filters (Bourett et al., 2002), but

were easily separated following linear unmixing (Figure 15.7a). Although it is preferred

to select fluorophores that avoid such spectral overlap, this is not always possible. For

example, the fluorescent probes Calcofluor for chitin and 4′,6-diamidino-2-phenylindole

(DAPI) for nuclei (or GFP and YFP) are extremely useful fungal cellular markers that

have overlapping excitation and emission characteristics that could be cleanly separated

by this approach. Another notable advantage of spectral imaging is that a reference

spectrum from tissue autofluorescence (e.g., wood and other plant tissues or growth media)

can be acquired and separated from the desired fluorescent probe signal. The autofluorescent signal need not be discarded because it can provide details about the surrounding

sample. Autofluorescence is often broad spectrum, making it very difficult to eliminate

by standard filters. A spectral scan of several species of cultured hyphae (Baschien et al.,

2001) proved invaluable for selecting probes whose emission wavelengths avoided insidious autofluorescence with in situ hybridization experiments. Spectral imaging has also

been shown to be advantageous when using techniques such as fluorescence resonance

energy transfer (FRET) for imaging calcium in mammalian cells with cameleon-2

(Hiraoka et al., 2002). Typically, FRET pairs are spectrally close and suffer from bleed-



DK3133_C015.fm Page 316 Tuesday, April 19, 2005 9:56 AM



316

(a)



Czymmek

(c)

0

Y (µm) 40

80

120

160

200

0

16 0

Z (µm)



50



100



150



200 250

X (µm)



300



350



400



100



150



200 250

X (µm)



300



350



400



100



150



200 250

X (µm)



300



350



400



(d)

(b)



0

Y (µm) 40

80

120

160

200

0

16 0

Z (µm)



50



(e)

0

Y (µm) 40

80

120

160

200

0

16



0

Z (µm)



50



Figure 15.7 (See color insert following p. 460.) (a) Spectral confocal techniques allowed clear

separation of closely overlapping fluorescent molecules such as AmCyan (blue), ZsGreen (green),

and ZsYellow (red) following linear unmixing. (Images provided courtesy of K. Czymmek, T.

Bourett, J. Sweigard, and R. Howard.) (b) In vivo constitutive cytoplasmic expression of the reef

coral fluorescent protein ZsGreen in Fusarium oxysporum was used to monitor disease progression

in Arabidopsis during root infection. (Images provided courtesy of K. Czymmek, J. Sweigard,

M. Fogg, and S. Kang.) (c–e) Four-dimensional (three-dimensions over time) series of cytoplasmic-expressing ZsGreen and AsRed Fusarium hyphae as they interact in culture. These threedimensional stacks were selected from a four-dimensional data set (T = 0, 32, and 72 min) that

monitored hyphal growth every 8 min over a 3-h period. (Images provided courtesy of V. Cooke

and K. Czymmek.)



through problems, necessitating specific postacquisition corrections. By virtue of linear

unmixing spectral sequences, clean separation of donor and acceptor fluorescence can be

faithfully obtained, minimizing these bleed-through artifacts. To date, there are only a

few examples of spectral confocal/multiphoton microscopy applied to mycological problems (Baschien et al., 2001; Bourett et al., 2002) — a reflection of the very recent

introduction of this technology. There is little doubt that multispectral imaging will

become invaluable as a tool for clean separation of probed fungal structures and byproducts from their complex environment.



DK3133_C015.fm Page 317 Tuesday, April 19, 2005 9:56 AM



Exploring Fungal Activity with Confocal and Multiphoton Microscopy



15.5



317



SUPPORTING TECHNOLOGIES



The state-of-the-art confocal/multiphoton microscopy is also tied to concomitant advances

in fluorescence-based technologies. One such example, Alexa Fluor dyes, is significantly

brighter and far more photostable than previous generations of fluorophores (PanchukVoloshina et al., 1999). This is typically an advantage; however, it can sometimes be

problematic for photobleaching or FRET studies because the dyes can be very difficult to

photobleach. Another technology, called Zenon, allows several primary antibodies against

the same species (e.g., mouse) to be easily applied in a single step for multilabel experiments

in a direct labeling approach (Molecular Probes, Eugene, OR). Quantum dots are semiconductor nanocrystals that pose interesting possibilities for a host of applications. These small

particles, on the order of a few nanometers in size, have very discrete emission wavelengths

depending on subtle differences in size (Seydel, 2003). This feature alone would lead to

very exciting potential for multilabeled experiments, as one could conceive of labeling six

or more targets in the same cell simultaneously. But this is not the only useful feature of

quantum dots; they are also extremely resistant to photobleaching and exceptionally bright

(Larson et al., 2003), making them well suited to experiments where sensitivity is an issue

(e.g., in situ hybridization localization of low-abundance molecules). In addition, it has

been reported that quantum dots have a two-photon cross section of 47,000 GM, which is

two to three orders of magnitude greater than any existing fluorescent probe (Larson et al.,

2003). This is incredibly bright and means that individual dots could likely be seen at very

low laser energy input and conceivably imaged with low-power objective lenses. Although

only a few spectral versions of quantum dots are commercially available (specially treated

to remain stable in the hydrophilic environment typically found in biological systems), it

is expected that a growing repertoire will be offered in the near future. It is less clear how

these could be used in living cells, but they may have potential for studying endocytosis.

Other advanced fluorescence-based imaging techniques are readily adaptable to laser

scanning confocal microscopy, such as FRET (Bastiaens and Jovin, 1997; Gadella et al.,

1999), fluorescence lifetime imaging microscopy (FLIM) (Pepperkok et al., 1999), fluorescence correlation spectroscopy (FCS) (Schwille et al., 1999), fluorescence recovery

after photobleaching (FRAP) (Siggia et al., 2000; Brandizzi et al., 2002), fluorescence

loss in photobleaching (FLIP) (Judd et al., 2003), fluorescence localization after photobleaching (FLAP) (Dunn et al., 2002), photoactivatable GFP (Patterson and LippincottSchwartz, 2002), and kindling fluorescent protein (KFP) (Chudakov et al., 2003). Many

of these techniques (e.g., FRAP, FLIP, FLAP, KFP) can be used with modern confocal/multiphoton systems without additional or expensive hardware upgrades.



15.6



MYCOLOGICAL APPLICATIONS



A rapidly growing body of literature clearly illustrates the power of optical sectioning

approaches for investigating fungi in vitro, as well as with interactions in exceptionally

multifaceted and diverse environmental situations. The relative ease in which discrete, highresolution image planes can be acquired, increased accessibility to confocal microscopes

in many institutions, and the amenability of fungi to optical microscopy in general have

played important roles in their popularity and success. From single optical sections of fixed

material to multidimensional, multichannel, or multispectral imaging of living hyphae,

confocal and multiphoton microscopies have proven to be invaluable as research tools. The

remainder of this chapter, although not exhaustive and primarily focused on filamentous



DK3133_C015.fm Page 318 Tuesday, April 19, 2005 9:56 AM



318



Czymmek



fungi, will explore the major areas in which investigators have already used the sophisticated capabilities of multiphoton/confocal microscopy in a multitude of applications.

15.6.1

Immunofluorescence

Some of the earliest reports of confocal microscopy for fungal applications used immunofluorescence techniques (Hardham et al., 1991; Kwon et al., 1991) and have been previously

reviewed (Kwon et al., 1993; Czymmek et al., 1994). These reports serve as useful starting

points if considering fluorescence-based antibody localization in fungi. Due to the availability and relative ease in which suitable antibodies can be created and applied for immunofluorescence microscopy, this methodology has become relatively commonplace in the

literature. For example, antibodies were raised against a series of cellulolytic enzymes, and

their distribution in cultured Volariella hyphae was documented via confocal microscopy

(Cai et al., 1999). Wood autofluorescence, which posed a significant challenge when using

conventional microscopy, due to the obscured visibility of labeled structures, easily delineated antibody-labeled Ophiostoma hyphae in thick-sectioned pine wood samples (Xiao et

al., 1999). Another interesting approach was the correlation of the opportunistic pathogenic

yeast Candida albicans (loaded with FITC) and actin in human cell lines (Tsarfaty et al.,

2000). In this case, the authors used colocalization analysis and optical sectioning in the zaxis to illustrate that actin specifically assembled near Candida interaction sites.

Antibodies raised against any number of antigens (in conjunction with a fluorescent

tag) can be used to localize specific cellular proteins of interest. However, due to their

size, unless the antigen is on the cell surface or otherwise introduced (e.g., microinjection),

employing antibodies for immunofluorescence typically requires fixed and permeabilized

cells. A number of protocols exist for antibody labeling of fungal samples. The method

of choice is influenced by a variety of factors, including the particular antigen, subcellular

compartment, and variations in cell wall biochemistry. A simple approach using freezesubstitution fixation and methacrylate de-embedment, thus avoiding sectioning and wall

digesting enzymes in filamentous fungi, has been described (Bourett et al., 1998). A major

advantage with this technique is that whole-mount fungi can be imaged in three dimensions

using confocal microscopy with high-fidelity cryogenic preservation. Others have used a

slightly modified version of this method to monitor microtubule and nuclear behavior in

ropy-1 mutants of Neurospora crassa with similar success (Riquelme et al., 2002), or just

freeze-substitution and methanol fixation to visualize the tubulin and actin cytoskeleton

in the chytridiomycete Allomyces (McDaniel and Roberson, 2000). Regardless of the

permeabilization tactic used, immunofluorescence is now routine and a core competency

of confocal/multiphoton microscopy.

15.6.2

In Situ Hybridization

Although examples of immunolabeled proteins are abundant in the fungal literature, very

little data are available using in situ hybridization for nucleic acid detection. Li et al.

(1997) quantified fluorescence of Aureobasidium pullulans on slides and leaf surfaces

counterstained with propidium iodide; Baschien et al. (2001) were successful with hybridization of freshwater fungi; and Sterflinger and Hain (1999) demonstrated hybridization

in black yeast and meristematic fungi. Major causes for the scarcity of hybridization data

include lack of sensitivity, difficulty with reliable fungal cell wall permeabilization, and

RNA degradation by wall digesting enzyme cocktails, as alluded to previously (Bourett

et al., 1998). When optimized and broadly applicable approaches become available for

fungi (e.g., using peptide nucleic acid probes or polymerase chain reaction (PCR)), hybridization methods are likely to derive similar benefits with confocal/multiphoton microscopy

as have other fluorescence probe techniques.