| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
INNOVATIVE METHODOLOGY
1Laboratory of Neural Systems and 2Research Engineering Core Laboratory, Memorial Sloan-Kettering Cancer Center, New York, New York 10021
Submitted 29 January 2004; accepted in final form 25 February 2004
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
) and the discovery of genetically encodable fluorophores (Chalfie et al. 1994; Prasher et al. 1992; Shimomura et al. 1962)—has led to a renaissance of optical microscopy in biological research. Two-photon laser scanning microscopy (2PM), the prototypical nonlinear imaging method, now routinely provides diffraction-limited images of biological structures deep within living tissues (Denk and Svoboda 1997; Denk et al. 1990, 1995; Helmchen and Denk 2002). It is ideally complemented by genetically encoded fluorophores [such as the green fluorescent protein (GFP) and its spectral variants and orthologs] that label these structures noninvasively. Expressed under the control of defined promoters or attached to defined cellular proteins through gene fusion, these fluorophores serve as convenient intravital markers for following the fates of cells within organisms or of subcellular structures within cells (Chalfie et al. 1994; Lichtman and Fraser 2001; Zhang et al. 2002).
Powerful variations on the theme of "genes and photons" are genetically encoded optical probes that signal not only the anatomical positions of tagged cells or proteins but also their physiological states (Zemelman and Miesenböck 2001; Zhang et al. 2002): changes in protein contacts or conformations (Baird et al. 1999; De Angelis et al. 1998; Heim and Tsien 1996), shifts in the concentrations of intracellular ions (Kuner and Augustine 2000; Miesenböck et al. 1998; Miyawaki et al. 1997) and small-molecule messengers (Oancea et al. 1998; Zaccolo et al. 2000), membrane depolarization (Siegel and Isacoff 1997), or the controlled release of secretory content (Miesenböck et al. 1998). The optical detection of fast, transient physiological signals, however, poses a particular challenge for scanning microscopies that sample each point in space only briefly and intermittently: the rate at which each pixel can be revisited determines whether the temporal resolution of the optical recording can keep pace with the relevant physiology. Two principal factors determine this rate. First, and fundamentally, the power of the optical signal dictates the length of the observation period (pixel dwell time) required to estimate the signal reliably. The second, and often limiting, factor is speed restrictions on the scanning optics themselves. Most scan engines rely on the movement of reflecting surfaces with inert masses to steer the excitation light beam. If scan mirror control involves a linear command function with feedback, the operating frequency of the mirrors is typically 
500 scan lines or 1 full-frame image/s. Higher scan frequencies can be attained when the mirrors are driven at their resonant frequency (Bacskai et al. 1995; Fan et al. 1999; Tsien and Bacskai 1995), but mirror deflections become sinusoidal instead of linear. This has the drawbacks that only a fraction of the working cycle of a full deflection period (the fraction in which the sine function is approximately linear) is available for data acquisition, or that significant postacquisition processing is required to linearize the image (Tsien and Bacskai 1995).
Alternatives to reflective scanning optics that have been used successfully in single-photon (i.e., confocal) laser-scanning microscopy are acousto-optic deflectors (AODs) (Draaijer and Houpt 1988). AODs are crystals onto which one or more radio frequency (RF) transducers are bonded. The transducers produce acoustic waves through the medium and thereby impress a periodic refractive index modulation on the initially homogeneous material. Due to the large difference in the velocities of sound and light, incident light "sees" this refractive index modulation as a stationary grating and is diffracted at a specific angle with respect to the acoustic wave. Because the deflection of the incident beam is proportional to the frequency of the acoustic wave (and thus the period of the refractive index modulation), a linear scan can be produced by ramping the frequency of the RF signal driving the transducers (Gottlieb et al. 1983).
Since they function without moving parts, AODs possess significant advantages in beam steering over systems using movable deflection mirrors. AODs achieve operating frequencies in the kilohertz range and can, depending on the number of scanned lines per frame, support image acquisition at video (i.e., 30 Hz) to kilohertz rates. Importantly, patterns of beam deflection other than a raster scan, including "random access" (i.e., motion of the beam through a set of predefined positions without an intervening sweep), are possible by applying the appropriate command functions to the RF transducers (Bullen et al. 1997). Because the analysis of physiological systems often concentrates on regions of interest, the ability to interrogate these regions selectively would increase the temporal resolution of the recording dramatically without compromising signal strength.
What has hindered the use of AODs in nonlinear imaging is a fundamental conflict between dispersive optics and the ultra-short light pulses required for nonlinear contrast generation: since ultra-short pulses are not monochromatic, dispersive optics such as AODs separate their spectral components both spatially and temporally (Denk et al. 1995). We (Ng et al. 2002) and others (Iyer et al. 2003; Lechleiter et al. 2002) have designed systems of corrective optics that compensate for the broadening of ultra-short pulses in space and time in an effort to enable the use of AODs for fast or random-access nonlinear imaging. Here we describe the principles of our design, detail its implementation on a widely available commercial instrument, and present examples of its use for optical imaging of neurons, using different nonlinear contrasts, in vitro and in vivo. We achieve video-rate or faster performance with minimally intrusive modifications that preserve the single-photon imaging capabilities of the instrument.
|
|
METHODS |
|---|
|
Light pulses for nonlinear imaging were generated by a mode-locked Ti:sapphire laser (Tsunami, Spectra-Physics, Mountain View, CA) pumped by a 10-W continuous-wave diode laser (Millenia Xs, Spectra-Physics). The Ti:sapphire laser was equipped with long-wavelength optics and configured for pulse widths of 70-fs full-width half-maximum (FWHM) at a repetition rate of 82 MHz. The output beam, whose spectral content was monitored continuously (IST-REES, Horseheads, NY), was coupled to the scanning optics via a series of tuned broadband laser mirrors (TLM2-910, S- and P-polarization versions as appropriate, CVI Laser, Albuquerque, NM).
The output beam of the Ti:sapphire laser is slightly divergent and astigmatic. To maximize throughput, the beam profile was re-shaped by inserting an astigmatic lens arrangement formed by a spherical lens with a focal length of 5,000 mm (PLCX-25.4-2575.0-C-633-1064, CVI Laser) and a cylindrical lens with a focal length of 2,000 mm (SCX-50.8-1000.0-UV-890, CVI Laser) into the beam path 1.1 m from the entrance of the scan head. The total path length from the laser output window to the scan head entrance, including the double-pass through the temporal corrective optics (Fig. 1), was 4.25 m.
Scanning optics
The scanner was a modified Oz confocal laser scan head (Thermo-Noran, Middleton, WI) on an Axioskop 2FS upright microscope (Zeiss, Oberkochen, Germany). The Oz scan engine uses an AOD to produce a fast x-axis raster and a conventional galvanometric scan mirror to generate the slow y-axis raster (Fig. 1). Because the scanning optics were designed for UV-visible light, they contained many elements that transmitted, reflected, or deflected infrared radiation very inefficiently. The most inefficient (but irreplaceable) of these was the AOD. Inversely proportional to the square of the wavelength (Gottlieb et al. 1983), the deflection efficiency of the AOD dropped to <20% in the near-infrared region; in other words, <20% of the incident optical power at 910 nm was deflected; the remainder was transmitted and absorbed by a specially added metal beam stop. It was therefore imperative to provide sufficient input power (typically 1.1 W at 910 nm) and optimize all other elements in the optical path. To that end, all lenses in the scan head were replaced with equivalent optics coated for visible–near-infrared transmission (custom ground to size with HEBBAR/073 coating, Melles Griot, Carlsbad, CA). Dielectric mirrors in the combined visible–near-infrared light path, including the galvanometer-driven scan mirror for the slow y-axis (Laser Scanning Keiser KG, Stallikon, Switzerland), were replaced with silver-coated mirrors (CVI Laser); mirrors exclusively in the near-infrared light path, including a mirror in the main dichroic wheel, were S- and P-polarization versions, as appropriate, of tuned broadband laser mirrors (TLM2-910, CVI Laser). The transmission efficiency of the microscope was optimized by replacing the factory-installed tube lens with a broadband visible–near-infrared–coated lens of the same focal length (Zeiss 164.5-mm focal length tube lens with Melles-Griot HEBBAR/073 coating). These modifications resulted in a measured throughput (407A power meter, Spectra-Physics) of 6.84% of the input at 910 nm to the back focal plane of the Zeiss 40x 0.8 W IR-Achroplan objective and of 5.31% to the specimen.
|
Temporal corrective optics
Since the scanning optics included highly dispersive elements, the temporal spread of the illumination pulse (Bor 1988) was corrected with a pulse compressor, which added an adjustable amount of negative group velocity dispersion ("prechirp") to the ultra-short pulse (Fork et al. 1984; Soeller and Cannell 1996). The compressor consisted of a pair of high-index (SF10) Brewster prisms (prisms 1 and 2 in Fig. 1; Spectra-Physics) in a folded arrangement with an interprism distance of 1.25 m that forced longer wavelengths to travel through a longer optical path than shorter wavelengths. On passing through the positively dispersing scanning optics, the negative prechirp was canceled, and the pulse emerged with all spectral components in phase. The degree of negative dispersion introduced into the optical path was adjusted with a micrometer-driven linear stage that moved one Brewster prism along its center axis. The adjustment was performed by minimizing the pulse width measured at the back focal plane of the objective, using relay mirrors and an autocorrelator (Model 409, Spectra-Physics). The uncompensated and compensated pulse widths at the back focal plane were 442 and 70.5 fs, respectively, that at the output window of the Ti:sapphire oscillator was 79.5 fs. The pulse compressor thus cancelled fully the temporal dispersion due to the scanning optics as well as the laser output window. If an autocorrelator is not available, a qualitative adjustment of the pulse compressor is possible by maximizing the signal amplitude of a fluorescent test specimen. The measured throughput of the pulse compressor at 910 nm was 88.29%.
Spatial corrective optics
A custom-fabricated isosceles Brewster prism with an apex angle of 58.4° (SF10 glass, broadband antireflective coating with 
1% reflection in the 840-920 nm band; CVI Laser) was inserted into the illumination light path prior to the AOD (prism 3 in Fig. 1; see also Fig. 2). The prism was designed to match the angular dispersion of the AOD at an acoustic operating frequency of 265 MHz (see RESULTS AND DISCUSSION). The frequency of 265 MHz was selected because it was found to provide the best deflection efficiency at 910 nm in the 200-MHz bandwidth of the AOD.
|
|
Emitted photons were detected by R3896 photomultiplier tubes (Hamamatsu Photonics, Hamamatsu City, Japan). Fluorescence was collected through a 40x 0.8-W IR-Achroplan objective (Zeiss) in the epi-illumination path (objective 1 in Fig. 1) and reflected off a dichromatic mirror (DM3 in Fig. 1; LWP-RU520-TP900, CVI Laser) through a custom opening in the microscope chassis to the detector assembly, which held a second dichromatic mirror (DM4 in Fig. 1) acting as a cold mirror (UV/Vis High Reflectivity, Chroma, Brattleboro, VT), a 2-mm colored glass filter (BP2 in Fig. 1; Schott BG39, Chroma), and a collector lens (Melles Griot). The detector assembly for second-harmonic light was inserted into the condenser carrier of the transmitted light path; it consisted of a 40x 0.8-W Achroplan objective (objective 2 in Fig. 1; Zeiss), a dichromatic mirror (DM5 in Fig. 1; 550DCXR, Chroma), a dielectric interference filter (BP3 in Fig. 1; D440/40, Chroma), and a cold mirror (DM6 in Fig. 1; UV/Vis High Reflectivity, Chroma).
Beam profiling
The in-scan and cross-scan diameters of the scanned beam were measured with a beam profiler (ProBeam, Photon, San Jose, CA) at a discrete set of deflection angles within the scanning range of the AOD. A 3,000-mm spherical lens (Melles Griot), placed in the infinity space of the microscope, was used to focus the beam onto the profiler.
Genetically encoded optical probes: expression, characterization, imaging
Superecliptic synapto-pHluorin (Miesenböck et al. 1998; Yuste et al. 2000) was expressed in the Or83b-positive subset of olfactory receptor neurons of Drosophila melanogaster by crossing P{w+;OR83b-GAL4}III driver and P{w+;UAS-spH}II responder strains (Ng et al. 2002). The antennal lobes were imaged through an opening in the cuticle that was superfused with a solution containing 5 mM Na-HEPES, pH 7.5, 115 mM NaCl, 5 mM KCl, 6 mM CaCl2, 1 mM MgCl2, 4 mM NaHCO3, 5 mM trehalose, 10 mM glucose, and 65 mM sucrose. Odor-evoked synaptic release was stimulated by directing a carrier airstream (0.6 l/min) via software-controlled solenoid valves (The Lee Company, Westbrook, CT) through a channel loaded with 0.2 µl of banana fragrance (Bell Flavors and Fragrances, Northbrook, IL).
A cDNA encoding the enhanced GFP (EGFP) was modified by PCR to introduce a 10-residue isoprenylation motif containing a CAAX box (–GCMSCKCVLS; Hancock et al. 1989) at the truncated C-terminus (
233-238) and/or a 13-residue internal lipidation motif between residues Ile-171 and Gly-174. The internal lipidation motif (–GGTKKFCGLCACP–) was selected by visually screening members of a mammalian expression library of SNAP-25-derived signals (Oyler et al. 1989; Veit et al. 1996) for plasma membrane localization.
To analyze posttranslational lipid modifications by phase separation (Bordier 1981), Xenopus oocytes injected with the indicated cRNAs were homogenized in 10 mM K-HEPES, pH 7.4, 250 mM sucrose, 2 mM MgCl2, 1 mM EGTA, 0.5 mM EDTA, and 1 mM DTT. The homogenates were extracted with 4% (wt/vol) Triton X-114 (Calbiochem, San Diego, CA) for 3 h at 4°C, clarified, clouded by a 10-min incubation at 30°C, and spun to separate aqueous and detergent phases (Bordier 1981); deacylation reactions (Dolci and Palade 1985; Schlesinger et al. 1980) involving 0.5 M hydroxylamine (Sigma, St. Louis, MO) at pH 7.4 were followed by a second phase separation reaction. Proteins in the aqueous and detergent fractions were separated by SDS-PAGE and transferred to nitrocellulose for chemiluminescent detection (ECL; Amersham Biosciences, Piscataway, NJ) with a polyclonal antiserum against GFP and secondary antibodies conjugated with horseradish peroxidase (Bio-Rad Laboratories, Hercules, CA).
For mass spectrometry, a detergent extract of material from 500 oocytes was treated with 50 mM 4-vinylpyridine (Sigma) at 12°C for 30 min to derivatize free cysteines (Sechi and Chait 1998). The modifying reagent was quenched with 167 mM reduced glutathione (Roche Applied Science, Indianapolis, IN) at 4°C overnight, and hydroxylamine deacylation (Dolci and Palade 1985; Schlesinger et al. 1980) and phase separation reactions (Bordier 1981) were performed as described above. EGFP in the detergent phase [adjusted to 50 mM Tris, pH 8.0, 100 mM NaCl, 10% (wt/vol) glycerol, 1 mM DTT] was collected by immunoprecipitation and isolated by SDS-PAGE. Unpolymerized residue leads to the formation of acrylamide adducts on unprotected cysteines during electrophoresis (Sechi and Chait 1998). Coomassie-stained bands were excised, fully alkylated with 1 M acrylamide, and digested in situ with 1 µg of sequencing-grade trypsin (Roche Applied Science). The resulting peptide mixture was loaded onto a microcolumn of 2 µl Poros R2 reverse-phase beads (Perseptive, Framingham, MA) and eluted in 4-µl steps with 16 and 30% acetonitrile in 0.1% formic acid. The eluates were analyzed by matrix-assisted laser-desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometry (REFLEX III, Bruker Daltonics, Billerica, MA) in the presence and absence of peptide calibrants.
Hippocampal neurons grown in dissociated cultures (Yuste et al. 2000) were stained with 50 µM FM 1-43 (Molecular Probes, Eugene, OR) in extracellular recording solution (25 mM Na-HEPES, pH 7.4, 119 mM NaCl, 2.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 30 mM glucose) and imaged within 10 min after dye application to minimize dye internalization or flip-flop. Alternatively, neurons were transfected with a calcium phosphate precipitate of a pCI-based plasmid (Promega, Madison, WI) encoding dually lipidated EGFP and imaged 3-8 days after transfection.
|
|
RESULTS |
|---|
|
To achieve the high photon flux densities necessary for nonlinear imaging while maintaining biologically tolerable average power levels, pulsed lasers with repetition rates of 100 MHz and pulse lengths of 100 fs to several picosecond are used as light sources (Denk et al. 1990). Due to their higher peak powers, femtosecond pulses (70 fs in the present case) are generally preferred for 2PM (but see also Koester et al. 1999). These pulses obey the inequality relationship established by the Heisenberg uncertainty principle, Et 
/2, which dictates a chromatic spread of 3.13 nm for a transform-limited 70-fs pulse centered at 910 nm. In practice, pulses are near transform-limited and span broader spectral ranges; the measured bandwidth in the present case was 17 nm FWHM at 910 nm.
An AOD operating under the Bragg condition, 
b = 
f/2v, deflects part of the incident optical field into the direction of double the angle of incidence (Gottlieb et al. 1983). Here, b (the so-called Bragg angle) is the angle of the incident as well as the deflected beam relative to the acoustic wavefront in the crystal, is the wavelength of the incident light, f is the frequency of the acoustic wave, and v is the speed of the acoustic wave in the AOD crystal. The small-angle approximation (sin 
) yields a linear relationship between the angle of deflection and the acoustic frequency
 | (1) |
of the deflected beam. Since the tolerance in the Bragg condition is small, the Bragg angle has to be matched for each acoustic frequency; during a scan, the angle of the incident optical field with respect to the acoustic field must therefore co-vary with the acoustic frequency. This is accomplished by using multiple RF transducers to introduce a frequency-dependent tilt into the acoustic field (Gottlieb et al. 1983).
Bragg diffraction of ultra-short laser pulses with broad spectral content leads to a spatial separation of their spectral components. Equation 1 shows that, irrespective of the acoustic frequency, the angle of the deflected beam is also a linear function of the wavelength of light, . The AOD therefore scans the different spectral component wavelengths of an ultra-short pulse by different amounts and hence in different directions. Longer wavelengths are deflected more than shorter ones for any given acoustic frequency, leading to an angular spread of the pulse. Angular dispersion causes the different spectral components of the excitation light pulse to be focused to different points in the specimen, resulting in a linearly varying lateral blur (Fig. 3A) that is minimal at one end of the scan and maximal at the opposite end. The situation is further aggravated because the spectral components of the pulse propagate at different velocities through the highly dispersive TeO2 crystal of the AOD (Bor 1988). This so-called group velocity dispersion leads to a temporal spread, with longer wavelengths leading the shorter ones. The combined result of the two effects—temporal and spatial dispersion—is a severe reduction of the photon flux density that renders the device ineffective for nonlinear imaging (Denk et al. 1995).
|
To estimate the error due to angular dispersion and the effectiveness of the corrective measures introduced below, we consider the resolution of the AOD in terms of its time-bandwidth product, N = 
f, where = D/v is the time taken by an acoustic wavefront to traverse the optical beam (D is the size of the beam in the scanned direction; v is the speed of sound in the AOD crystal) and f is the bandwidth of the acoustic field (Yariv 1991). The time-bandwidth product is equivalent to the ratio of the angular scan range to the angular width of the scanned beam: N = f = f/
. Here, f = (/v)f is the scan range derived from Eq. 1, and = /D is the far-field angular width of the beam. This ratio establishes the number of nonoverlapping angular widths (resolvable spots) within the scan range, independent of any subsequent optical elements. Throughout the following discussion, the scan range is therefore referred to in its fundamental input of acoustic frequency rather than its output of scanned angle.
The specifications of the AOD used in our instrument (fmin = 200 MHz, fmax = 400 MHz, f = 200 MHz, v = 4,322 ms–1, D = 12 mm) allow for a theoretical maximum of N = 555 resolved points. In practice, the working bandwidth at near-infrared wavelengths had to be limited to a smaller range of 60 MHz in which the Bragg condition could be satisfied. We centered this reduced working bandwidth at 265 MHz, the frequency determined empirically to cause the most efficient deflection of 910-nm light. The scanning resolution corresponding to the limited acoustic bandwidth was N = 167 points.
The angular spread, , caused by the spectral bandwidth of the pulsed beam at a particular scanning frequency, is given by the derivative of the angle of deflection with respect to the wavelength and obtained by differentiating Eq. 1
 | (2) |
 | (3) |
To compensate for the AOD-induced dispersion, a corrective prism was designed that would introduce a measure of angular dispersion into the beam that precisely canceled that due to the AOD at 265 MHz, the selected center frequency of the scan. The wavelength-dependent deflection angle 
of the pulsed beam by the prism was calculated as
 | (4) |
is the entrance angle of the beam with respect to the entrance surface normal, ' is the exit angle with respect to the exit surface normal, is the apex angle of the prism, and n() is the wavelength-dependent refractive index of the prism material. The index n() was obtained by substituting the coefficients for SF10 glass (Table 2) into the Sellmeir equation
 | (5) |
= 910 nm and bandwidth = 17 nm by the prism is
 | (6) |
and n() from Eqs. 4 and 5, and solving numerically for yielded an apex angle of 58.4° for the desired corrective prism. Placed into the optical path prior to the AOD (Figs. 1 and 2), this prism offset the angular spread of the AOD as predicted (Fig. 3B). However, since the amount of dispersion due to the AOD varied linearly across the acoustic bandwidth of the scan while that due to the prism remained fixed, the angular spread of the pulse was exactly matched only at 265 MHz but under- or overcorrected at all other frequencies, leaving the extremes of the 60-MHz scan with correction errors of 1.56 angular widths (Eq. 3).
|
Direct experimental measurements of beam diameters across the scanned range (Fig. 4) are in agreement with this analysis. The measurements were performed with the Ti:sapphire laser mode-locked to generate 70-fs pulses centered at 910 nm, or in continuous wave mode (CW) to produce essentially monochromatic 910-nm light. Figure 4 shows beam diameters in the scan range of 200-320 MHz, about twice the range used for imaging. As expected, the size of the CW beam was constant throughout the scan in both diameters. The beam was slightly elliptical, with the larger diameter in the in-scan than in the cross-scan direction. In contrast, the in-scan—but not the cross-scan—diameter of the mode-locked beam changed as a function of acoustic frequency. With the corrective prism in place, the minimum beam diameter was seen at a frequency of 255 MHz, identifying the point of full correction. At this point, the diameter of the pulsed beam was indistinguishable from that of the CW beam, indicating that the corrective optic fully compensated for the AOD-induced angular dispersion. The 10-MHz discrepancy between the observed (255 MHz) and the designed frequency for full dispersion correction (265 MHz) revealed a mounting imprecision that rotated the corrective prism by 2.3° relative to its intended orientation. Small rotations of the prism decrease or increase its dispersion by 4.3 MHz/° (Eqs. 4 and 6).
|
TWO-PHOTON FLUORESCENCE.
To test whether our instrument could capture dynamic signals in living neural tissue, we imaged 2P-excited fluorescence originating from a defined population of synapses in the brain of a fruit fly. The synapses were labeled with synapto-pHluorin, a genetically encoded probe of synaptic vesicle trafficking (Miesenböck et al. 1998; Ng et al. 2002; Yuste et al. 2000). Synapto-pHluorin consists of a pH-sensitive mutant of GFP, termed superecliptic pHluorin, and the membrane protein VAMP/synaptobrevin-2 (Baumert et al. 1989; Elferink et al. 1989), which attaches pHluorin to the inner surface of synaptic vesicles. The chromophore of superecliptic pHluorin titrates between a deprotonated, fluorescent, and a protonated, nonfluorescent state. The abundance of protons inside resting synaptic vesicles (pH 5.7) (Miesenböck et al. 1998) drives superecliptic pHluorin into the protonated state and quenches ("eclipses") its fluorescence. Exocytic events cause release of the accumulated protons and chromophore deprotonation, which is detected as a sudden increase in fluorescence (Miesenböck et al. 1998; Ng et al. 2002).
The expression of synapto-pHluorin was restricted, with the help of the OR83b promoter, to a defined subset of olfactory receptor neurons (Ng et al. 2002). The cell bodies of these neurons lie close to the sensory surface, to which they extend receptor-bearing dendrites (Hildebrand and Shepherd 1997). Binding of airborne odorants to these receptors activates a transduction cascade that leads to the depolarization of the olfactory receptor neuron (Prasad and Reed 1999) and the transmission of synaptic impulses to second-order neurons. Synaptic transmission between first- and second-order neurons takes place in the antennal lobe, which, like its vertebrate counterpart, the olfactory bulb, is partitioned into distinct anatomical regions, termed glomeruli (Hildebrand and Shepherd 1997). Because each glomerulus receives synaptic input from olfactory receptor neurons expressing only one or, at most, a few receptor genes (Gao et al. 2000; Vosshall et al. 2000), a one-to-one mapping links the activation of specific olfactory receptors to evoked neurotransmitter release in particular glomeruli (Ng et al. 2002).
Figure 5A shows optical sections through the left antennal lobe of a living fly, located 135 µm beneath the surface of the brain. Images were acquired at a sampling rate of 30 Hz, a digital resolution of 512 x 480 pixels, and a laser power of 34 mW at the sample. This power level was in 1.5-fold excess of that reported for all-reflective video-rate 2PM of cultured cells (Fan et al. 1999). The higher power requirement in the present case may simply be a consequence of the thicker specimen, or it may reflect distortions of the diffraction-limited wavefront after scanning the excitation light beam through the AOD. Such distortions would reduce the effective photon density at the focus and necessitate higher global power levels to achieve comparable 2P excitation efficiencies.
|
F) were computed. The data were normalized (F/F) to a moving-average background image to correct for the observed photobleaching rate of 1.1%/s. Every eighth frame in the video sequence is shown, corresponding to a displayed temporal resolution of 266 ms. Prominent responses to the odor stimulus, which consisted of a 2-s puff of banana fragrance (red bar), were seen in three of the eight glomeruli outlined in Fig. 5A. Within 100 ms after the onset of the odor pulse, several submicron-sized puncta of increased synapto-pHluorin fluorescence appeared. Judging from their size (1 µm; Fig. 5B), response homogeneity to multiple test odorants (Ng et al. 2002), and co-localization with presynaptic marker proteins (Ng et al. 2002), these puncta correspond—in all likelihood—to individual synaptic terminals. Their intensity continued to rise for 8-56 additional frames and peaked 400-1,680 ms after the fluorescence increase was first detected (Fig. 5, C and D). Different synaptic clusters within each responsive glomerulus became active at different times during the odor pulse, leading to asynchronous patterns of activation that are not apparent at slower image acquisition rates (Ng et al. 2002). Activity had spread to most synapses within the visible cross-section of each responsive glomerulus shortly before the stimulus was terminated (Fig. 5, A and D). Synapto-pHluorin fluorescence then declined as vesicles were internalized and re-acidified (Fig. 5, A, C, and D). Individual fluorescent puncta returned to baseline in an asynchronous manner that seemed to reverse the order of activation; often, the earliest responders were the last synapses to become re-eclipsed (Fig. 5A). SECOND-HARMONIC GENERATION.
Second-harmonic generation (2HG) is a nonlinear optical phenomenon whose quadratic dependence on the incident photon flux gives it the same intrinsic resolving power as 2P excitation. 2HG, however, provides a fundamentally different optical contrast. Whereas 2P excitation is an incoherent phenomenon involving photon absorption and re-emission, 2HG arises from coherent scattering of the incident optical field into a frequency-doubled "harmonic" field (Bloembergen 1965). Because 2HG is electric dipole forbidden in media with inversion symmetry, 2HG is generally observed only in highly anisotropic structures or at interfaces where this symmetry is broken (Bloembergen 1965; Shen 1989). The intercalation of lipophilic dyes into one leaflet of a biological membrane, for example, creates an oriented, asymmetrical chromophore array in which the phased second-harmonic emissions from individual dipoles can sum constructively (Ben-Oren et al. 1996; Bouevitch et al. 1993; Campagnola et al. 1999; Moreaux et al. 2000b, 2001).
The growing interest of neurophysiologists in 2HG stems from the observation that changes in transmembrane voltage can modulate the magnitude of the chromophore dipole moment induced by the driving optical field and thus the amplitude of the second-harmonic signal (Ben-Oren et al. 1996; Bouevitch et al. 1993; Campagnola et al. 1999; Dombeck et al. 2004; Pons et al. 2003). Since the voltage dependence of 2HG appears much steeper than that of fluorescent potentiometric dyes (Ben-Oren et al. 1996; Bouevitch et al. 1993; Campagnola et al. 1999), second-harmonic imaging is potentially superior to 2PM in detecting electrical activity in excitable cells. Bioelectric events, however, occur on millisecond timescales that call for significantly higher temporal resolution than that previously achieved in second-harmonic microscopy (Ben-Oren et al. 1996; Bouevitch et al. 1993; Campagnola et al. 1999; Moreaux et al. 2001; Nikolenko et al. 2003). The speed of AOD-based scanners could therefore prove especially valuable in this arena.
Simultaneously recorded second-harmonic and 2PM images of neurons stained by bulk application of 50 µM FM 1-43 are contrasted in Fig. 6, A and B. FM 1-43 is a lipophilic styryl dye that is virtually nonfluorescent in aqueous environments but increases its quantum yield dramatically on insertion into the outer leaflet of surface membranes (Betz et al. 1992). Its absorption spectrum, with a peak at 479 nm, resembles that of fluorescein and EGFP. Illumination of FM 1-43–stained membranes with 70-fs pulses of light centered at 890 nm generated both fluorescent photons at 530–590 nm and harmonic photons at exactly one-half the fundamental wavelength (445 nm). The wide spectral separation facilitated the detection of fluorescent and second-harmonic light, free of cross-contamination, in separate channels. Fluorescence was collected through the epi-illumination optical path; second-harmonic light, which propagates in the form of two off-axis lobes in the forward direction (Moreaux et al. 2000b, 2001), by a high-NA objective and detector assembly in lieu of the transmitted light condenser (see Fig. 1 and METHODS). Comparison of 2PM (Fig. 6A) and second-harmonic images (Fig. 6B) revealed the diagnostic hallmarks of 2HG: confinement of the signal to non-inversion symmetric structures (i.e., membranes) and an orientation dependence with respect to the driving optical field: maximal signal intensities were seen where the chromophore dipole vectors were aligned with the polarization vector of the laser beam. Because the dipole of a membrane-bound styryl dye lies roughly orthogonal to the plane of the membrane (Loew and Simpson 1981; Loew et al. 1979), maximal second-harmonic intensities coincided with regions where the polarization vector intersected a membrane at approximately a right angle (Fig. 6B).
|
; Lewis et al. 1999), it remains unclear how these rather disorderly arrangements of dipoles might meet the strict asymmetry and orientation requirements of 2HG. Our strategy, therefore, was to create a defined geometry of genetically encoded dipoles that would approximate that of a membrane labeled with a styryl dye. The key features of this arrangement are asymmetry (dye is present only in one membrane leaflet) and a membrane-orthogonal orientation of the chromophore dipole vector (Loew and Simpson 1981; Loew et al. 1979).
The electronic transition dipole moment of EGFP is thought to lie in the plane of the chromophore and subtend an angle of 13° with the vector that joins the phenolic and imidazolinone oxygens (Bublitz et al. 1998; Rosell and Boxer 2003); it thus points roughly perpendicular from the central axis of the cylinder-shaped molecule (Ormö et al. 1996; Yang et al. 1996) to the quadrangle formed by residues 146, 147, 204, and 205 in its outer shell (Fig. 7A). To bring this vector into a plane approximately orthogonal to the plane of the membrane, signals that directed the posttranslational addition of two lipid anchors, one at the base and the other at the top of the GFP cylinder, were added to the EGFP sequence (Fig. 7A). These anchors were designed to tie the GFP cylinder flush, with its axis membrane-parallel, against the cytoplasmic face of the plasma membrane. The placement of the lipid anchors was chosen to rotate the dipole vector toward the membrane and hence to replicate, as closely as physically possible, the membrane-orthogonal position of an intercalated styryl dye. One lipid anchor comprised a 10-residue isoprenylation and plasma membrane targeting signal from H-ras (Hancock et al. 1989); this anchor was appended to the C-terminally truncated (233-238) fluorescent core of EGFP. The second anchor was an internal lipidation motif derived from SNAP-25 (Oyler et al. 1989; Veit et al. 1996). It consisted of a 13-residue peptide containing three potentially palmitoylated cysteines (–GGTKKFCGLCACP–) that replaced the short loop connecting residues Ile-171 and Gly-174 in EGFP. These principles of membrane attachment are the same as those used to anchor many natural peripheral membrane proteins, such as heterotrimeric G protein subunits (Hamm and Gilchrist 1996) and H-ras (Hancock et al. 1989), to the cytoplasmic face of the plasma membrane (Solski et al. 1995): lipidated proteins retain some lateral mobility but cannot detach from or translocate across the bilayer.
|
), the distribution of each construct was examined by immunoblotting (Fig. 7B). While EGFP was found exclusively in the aqueous phase, 60% of palmitoyl-EGFP and >90% of prenyl-EGFP and dual-EGFP partitioned into the detergent phase, confirming the presence of lipid anchors.
To determine whether the palmitoylated population of dual-EGFP was also prenylated, its phase distribution was re-examined after treatment with hydroxylamine, a compound that hydrolyzes the thioester linkages of fatty acyl chains but leaves the thioether linkages of isoprenyl groups intact (Dolci and Palade 1985; Farnsworth et al. 1989; Schlesinger et al. 1980). Dual-EGFP remained in the detergent phase after removal of its putative fatty acyl anchors, establishing the simultaneous presence of hydroxylamine-resistant prenyl groups in the palmitoylated fraction (Fig. 7B).
The reverse experiment, to determine whether the prenylated fraction of dual-EGFP was also acylated, was performed by probing the occupancy of the potential palmitoyl acceptor sites directly. A Triton X-114 extract of cells expressing dual-EGFP was reacted with 4-vinylpyridine, an alkylating agent that modifies free cysteines (Sechi and Chait 1998) but spares residues protected by covalently attached fatty acids. Following hydrolysis of the protective acyl groups with hydroxylamine, the modified dual-EGFP was isolated by immunoprecipitation from the detergent phase and reacted with acrylamide to derivatize any now deacylated cysteines (Sechi and Chait 1998). Tryptic peptides containing the potential palmitoyl acceptor sites (a 15-mer and a 16-mer, identical down to a missed tryptic cleavage site in bold: KFCLGCACPGGVQLAR) were analyzed by mass spectrometry to count the number and type of cysteine substituents, bearing in mind that each acrylamide substituent marks a site originally occupied by an acyl chain. The masses of the two major peptide species, 1707.76 and 1835.86 daltons, matched those calculated for 15- and 16-mers with three acrylamide substituents, respectively (Fig. 8A). Smaller peaks at 1741.79 and 1775.82 daltons had the expected masses of 15-mers with two and one acrylamide substituent(s); 16-mers with fewer than three acrylamide substituents or peptides lacking acrylamide substituents altogether were not detected (Fig. 8A). This analysis implies that every prenylated dual-EGFP molecule was also palmitoylated: the majority at all three acceptor sites, and smaller subsets at two or only one site(s). Tryptic peptides generated from dual-EGFP expressed in E. coli, an organism that lacks the enzymatic apparatus for palmitoylation, served as negative controls. Only a 15-mer without acrylamide substituents was recovered in this case (Fig. 8B), suggesting that the 16-mer in the eukaryotic sample arose because palmitoylation partially masked a tryptic site.
|
Despite compelling evidence that both lipid anchors were functional in the majority of dual-EGFP molecules (Figs. 7B and 8A) and although the molecules formed mature chromophores and were correctly targeted (Fig. 6C), second-harmonic signals originating from the membrane-bound EGFP array proved elusive (Fig. 6D). Laser powers of up to 55 mW at the sample, extended signal averaging at maximal photomultiplier gain, and searches for electronic resonance enhancement of the second-harmonic signal (Campagnola et al. 2001; Heinz et al. 1982; Moreaux et al. 2001) by varying the fundamental wavelength between 870 and 920 nm were all ineffective, singly or in combination. The failure of the genetically encoded dipole array to generate detectable second-harmonic power suggested substantial differences of its structure to that of a dye-stained membrane, which it sought to emulate. These differences include, trivially, placement of the dipoles on opposite faces of the plasma membrane (extracellular for styryl dyes vs. intracellular for dual-EGFP) and different penetration depths (intercalation for styryl dyes vs. peripheral membrane association for dual-EGFP). None of these factors is a priori likely to affect 2HG, nor is a bulky peripheral membrane protein like dual-EGFP expected to translocate or flip-flop across the plasma membrane and thereby destroy the inversion asymmetry required for 2HG. Different types of membrane attachment may, however, restrict molecular motions to different degrees and cause more or less anisotropy in the orientational averages. By the same token, imperfect alignment of the chromophore dipole vector with the membrane-orthogonal axis of molecular rotation (Fig. 7A) might lead to a less anisotropic distribution of dipole moments.
Differences in chromophore structures might account for differences in second-harmonic signal strength as well. The nonlinear hyperpolarizability of a chromophore (which determines its 2HG efficiency; Campagnola et al. 2001; Shen 1989) is directly related to the magnitude of the dipole moment change between ground and excited states (Chemla and Zyss 1987; Moreaux et al. 2000a). Styrylpyridinum chromophores undergo large electronic redistributions on excitation that are reflected in dipole moment changes of 16 Debye (Loew et al. 1979). The anionic form of the p-hydroxybenzyledine-imidazolinium chromophore in EGFP exhibits a difference dipole moment of 7 Debye (Bublitz et al. 1998), which attests to a considerable but decidedly smaller charge shift. Like the orientational average discussed above, second-harmonic intensity scales quadratically with the nonlinear hyperpolarizability (Campagnola et al. 2001; Chemla and Zyss 1987; Moreaux et al. 2000a). Modest differences in several of these variables could thus combine to produce significant effects.
The amplitude of the second-harmonic signal can be resonance-enhanced by one to two orders of magnitude if the signal overlaps with an electronic absorption band (Campagnola et al. 2001; Heinz et al. 1982; Moreaux et al. 2001). Because the fundamental wavelengths used in our experiments (870-920 nm) were significantly shorter than that used in previous studies (1,064 nm; Khatchatouriants et al. 2000; Lewis et al. 1999), the specter arises that a difference in resonance enhancement might underlie the reported differences in signal strength. Two lines of evidence argue against this possibility. First, our fundamental wavelengths (870-920 nm) were selected to generate second-harmonic signals (435-460 nm) within the major absorption band of EGFP (425-500 nm); second-harmonic light generated at 1,064 nm, in contrast, appears off resonance because GFP does not absorb at 532 nm. Second, membranes decorated with FM 1-43, a synthetic dye whose absorption spectrum matches that of EGFP closely, generated bright, presumably resonance-enhanced, second-harmonic signals at a fundamental wavelength of 890 nm (Fig. 6B).
Perhaps most important, then, is the fact that the total radiated second-harmonic power is a quadratic function of the surface density of radiating dipoles (Campagnola et al. 2001; Moreaux et al. 2000a, 2001). (Fluorescent power, in contrast, is a linear function of the surface density of fluorophores.) It has been estimated that the density of styryl dyes in membranes averages 105 molecules/µm2 at a bulk concentration of 50 µM (Schote and Seelig 1998). The footprint of a membrane-parallel GFP cylinder, 4.2 x 2.4 nm (Ormö et al. 1996; Yang et al. 1996), would require that the cytoplasmic face of the plasma membrane be tiled seamlessly with GFP molecules to achieve comparable dipole densities. Steric hindrance may thus impose the most severe restriction on the utility of membrane-tethered GFP as a harmonic probe.
|
|
DISCUSSION |
|---|
|
). Genetic control over the sites of probe expression can single out functionally or anatomically defined groups of neurons for analysis; photons of light can address large numbers of such genetically highlighted neurons in parallel. The combination of "genes and photons" may thus offer a compromise between opposing demands of parallelism and selectivity that traditional experimental approaches to neural function have found difficult to reconcile.
Compared with electrical recordings, which remain the gold standard for analyzing neuronal activity, however, optical methods in general and scanning methods in particular have difficulty resolving events at physiologically relevant timescales. A number of strategies have therefore been devised to accelerate the acquisition of scanned images, singly or in combination: rapid movement of the exciting light beam with consequently reduced pixel dwell times (the strategy pursued here as well as by Fan et al. 1999); simultaneous scanning of multiple foci (Bewersdorf et al. 1998; Buist et al. 1998); and reductions of the scanned area, to a single line in the extreme (Denk and Svoboda 1997). While each of these strategies possesses advantages and disadvantages, the ultimate limit to fast imaging stems from the photon statistics of the optical signal itself, which set the minimal photon counting time per pixel and determine whether useful information can be extracted at high sampling rates.
Fluorescent signals emitted by micron-sized structures in the intact brain, such as synaptic terminals expressing synapto-pHluorin, carry sufficient optical power to register in pixel dwell times as short as 100 ns. These pixel dwell times translate to a theoretical maximum of 10 million image points that can be surveyed per second. Standard-sized images of 640 x 480 pixels can be scanned at video rate and smaller areas at correspondingly higher frequencies. Biological phenomena that would have remained undetected at conventional sampling speeds emerge clearly at the increased temporal bandwidth. Time-resolved fluorescence measurements at multiple synaptic sites, for instance, have revealed complex dynamics of evoked neurotransmitter release in a genetically homogeneous population of olfactory receptor neurons (Fig. 5).
Stringent anisotropy and dipole density requirements make 2HG a much more demanding contrast mode than fluorescence for genetically encoded membrane probes. Even if a preferred chromophore orientation can be forced with multiple membrane tethers, the protein shells surrounding the popular encodable chromophores may simply be too bulky to achieve the dipole densities required for efficient 2HG (Fig. 6). Since these shells are essential elements of tightly folded single-domain structures (Ormö et al. 1996; Yang et al. 1996), there is little hope that protein engineering could result in more compact, densely packed arrangements. To date, the only protein structures known to give rise to intense second-harmonic signals are semi-crystalline assemblies of structural proteins (Campagnola et al. 2002; Dombeck et al. 2003; Roth and Freund 1981; Zipfel et al. 2003; Zoumi et al. 2002). Whether the principles responsible for 2HG in these structures (which are currently only incompletely understood) can be duplicated in or at biological membranes, and importantly, whether these structures are capable of voltage-modulated 2HG, remains to be seen.
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: G. Miesenböck, Memorial-Sloan-Kettering Cancer Center, Box 205, 1275 York Ave., New York, NY 10021 (E-mail: g-miesenboeck{at}ski.mskcc.org).
|
|
REFERENCES |
|---|
|
Baird GS, Zacharias DA, and Tsien RY. Circular permutation and receptor insertion within green fluorescent proteins. Proc Natl Acad Sci USA 96: 11241–11246, 1999.
Baumert M, Maycox PR, Navone F, De Camilli P, and Jahn R. Synaptobrevin: an integral membrane protein of 18,000 daltons present in small synaptic vesicles of rat brain. EMBO J 8: 379–384, 1989.[Web of Science][Medline]
Ben-Oren I, Peleg G, Lewis A, Minke B, and Loew L. Infrared nonlinear optical measurements of membrane potential in photoreceptor cells. Biophys J 71: 1616–1620, 1996.[Web of Science][Medline]
Betz WJ, Mao F, and Bewick GS. Activity-dependent fluorescent staining and destaining of living vertebrate motor nerve terminals. J Neurosci 12: 363–375, 1992.[Abstract]
Bewersdorf J, Pick R, and Hell SW. Multifocal multiphoton microscopy. Opt Lett 23: 655–657, 1998.[CrossRef][Web of Science][Medline]
Bloembergen N. Nonlinear Optics. New York: Benjamin, 1965.
Bor Z. Distortion of femtosecond laser-pulses in lenses and lens systems. J Mod Opt 35: 1907–1918, 1988.[CrossRef]
Bordier C. Phase separation of integral membrane proteins in Triton X-114 solution. J Biol Chem 256: 1604–1607, 1981.
Bouevitch O, Lewis A, Pinevsky I, Wuskell JP, and Loew LM. Probing membrane potential with nonlinear optics. Biophys J 65: 672–679, 1993.[Web of Science][Medline]
Bublitz G, King BA, and Boxer SG. Electronic structure of the chromophore in green fluorescent protein (GFP). J Am Chem Soc 120: 9370–9371, 1998.[CrossRef][Web of Science]
Buist AH, Muller M, Squier J, and Brakenhoff GJ. Real time two-photon absorption microscopy using multi point excitation. J Microsc 192: 217–226, 1998.[CrossRef][Web of Science]
Bullen A, Patel SS, and Saggau P. High-speed, random-access fluorescence microscopy. I. High-resolution optical recording with voltage-sensitive dyes and ion indicators. Biophys J 73: 477–491, 1997.[Web of Science][Medline]
Campagnola PJ, Clark HA, Mohler WA, Lewis A, and Loew LM. Second-harmonic imaging microscopy of living cells. J Biomed Opt 6: 277–286, 2001.[CrossRef][Web of Science][Medline]
Campagnola PJ, Millard AC, Terasaki M, Hoppe PE, Malone CJ, and Mohler WA. Three-dimensional high-resolution second-harmonic generation imaging of endogenous structural proteins in biological tissues. Biophys J 82: 493–508, 2002.[Web of Science][Medline]
Campagnola PJ, Wei MD, Lewis A, and Loew LM. High-resolution nonlinear optical imaging of live cells by second harmonic generation. Biophys J 77: 3341–3349, 1999.[Web of Science][Medline]
Chalfie M, Tu Y, Euskirchen G, Ward WW, and Prasher DC. Green fluorescent protein as a marker for gene expression. Science 263: 802–805, 1994.
Chemla DS and Zyss J. Nonlinear Optical Properties of Organic Molecules and Crystals. Orlando, FL: Academic Press, 1987.
De Angelis DA, Miesenböck G, Zemelman BV, and Rothman JE. PRIM: proximity imaging of green fluorescent protein-tagged polypeptides. Proc Natl Acad Sci USA 95: 12312–12316, 1998.
Denk W, Piston DW, and Webb WW. Two-photon molecular excitation in laser-scanning microscopy. In: Handbook of Biological Confocal Microscopy, edited by Pawley JB. New York: Plenum, 1995, p. 445-458.
Denk W, Strickler JH, and Webb WW. Two-photon laser scanning fluorescence microscopy. Science 248: 73–76, 1990.
Denk W and Svoboda K. Photon upmanship: why multiphoton imaging is more than a gimmick. Neuron 18: 351–357, 1997.[CrossRef][Web of Science][Medline]
Dolci ED and Palade GE. The biosynthesis and fatty acid acylation of the murine erythrocyte sialoglycoproteins. J Biol Chem 260: 10728–10735, 1985.
Dombeck DA, Blanchard-Desce M, and Webb WW. Optical recording of action potentials with second-harmonic generation microscopy. J Neurosci 24: 999–1003, 2004.
Dombeck DA, Kasischke KA, Vishwasrao HD, Ingelsson M, Hyman BT, and Webb WW. Uniform polarity microtubule assemblies imaged in native brain tissue by second-harmonic generation microscopy. Proc Natl Acad Sci USA 100: 7081–7086, 2003.
Draaijer A and Houpt PM. A standard video-rate confocal laser-scanning reflection and fluorescence microscope. Scanning 10: 139–145, 1988.[Web of Science]
Elferink LA, Trimble WS, and Scheller RH. Two vesicle-associated membrane protein genes are differentially expressed in the rat central nervous system. J Biol Chem 264: 11061–11064, 1989.
Fan GY, Fujisaki H, Miyawaki A, Tsay R, Tsien RY, and Ellisman MH. Video-rate scanning two-photon excitation fluorescence microscopy and ratio imaging with chameleons. Biophys J 76: 2412–2420, 1999.[Web of Science][Medline]
Farnsworth CC, Wolda SL, Gelb MH, and Glomset JA. Human lamin B contains a farnesylated cysteine residue. J Biol Chem 264: 20422–20429, 1989.
Fork RL, Martinez OE, and Gordon JP. Negative dispersion using pairs of prisms. Opt Lett 9: 150–152, 1984.[Medline]
Gao Q, Yuan B, and Chess A. Convergent projections of Drosophila olfactory neurons to specific glomeruli in the antennal lobe. Nat Neurosci 3: 780–785, 2000.[CrossRef][Web of Science][Medline]
Gottlieb M, Ireland CLM, and Ley JM. Electro-Optic and Acousto-Optic Scanning and Deflection. New York: Marcel Dekker, 1983.
Hamm HE and Gilchrist A. Heterotrimeric G proteins. Curr Opin Cell Biol 8: 189–196, 1996.[CrossRef][Web of Science][Medline]
Hancock JF, Magee AI, Childs JE, and Marshall CJ. All ras proteins are polyisoprenylated but only some are palmitoylated. Cell 57: 1167–1177, 1989.[CrossRef][Web of Science][Medline]
Heim R and Tsien RY. Engineering green fluorescent protein for improved brightness, longer wavelengths and fluorescence resonance energy transfer. Curr Biol 6: 178–182, 1996.[CrossRef][Web of Science][Medline]
Heinz TF, Chen CK, Ricard D, and Shen YR. Spectroscopy of molecular monolayers by resonant 2nd-harmonic generation. Phys Rev Lett 48: 478–481, 1982.[CrossRef][Web of Science]
Helmchen F and Denk W. New developments in multiphoton microscopy. Curr Opin Neurobiol 12: 593–601, 2002.[CrossRef][Web of Science][Medline]
Hildebrand JG and Shepherd GM. Mechanisms of olfactory discrimination: converging evidence for common principles across phyla. Annu Rev Neurosci 20: 595–631, 1997.[CrossRef][Web of Science][Medline]
Iyer V, Losavio BE, and Saggau P. Compensation of spatial and temporal dispersion for acousto-optic multiphoton laser-scanning microscopy. J Biomed Opt 8: 460–471, 2003.[CrossRef][Web of Science][Medline]
Khatchatouriants A, Lewis A, Rothman Z, Loew L, and Treinin M. GFP is a selective non-linear optical sensor of electrophysiological processes in Caenorhabditis elegans. Biophys J 79: 2345–2352, 2000.[Web of Science][Medline]
Koester HJ, Baur D, Uhl R, and Hell SW. Ca2+ fluorescence imaging with pico- and femtosecond two-photon excitation: signal and photodamage. Biophys J 77: 2226–2236, 1999.[Web of Science][Medline]
Kuner T and Augustine GJ. A genetically encoded ratiometric indicator for chloride: capturing chloride transients in cultured hippocampal neurons. Neuron 27: 447–459, 2000.[CrossRef][Web of Science][Medline]
Lechleiter JD, Lin DT, and Sieneart I. Multi-photon laser scanning microscopy using an acoustic optical deflector. Biophys J 83: 2292–2299, 2002.[Web of Science][Medline]
Lewis A, Khatchatouriants A, Treinin M, Chen Z, Peleg G, Friedman N, Bouevitch O, Rothman Z, Loew L, and Sheves M. Second harmonic generation of biological interfaces: probing the membrane protein bacteriorhodopsin and imaging membrane potential around GFP molecules at specific sites in neuronal cells of C. elegans. Chem Phys 245: 133–144, 1999.[CrossRef][Web of Science]
Lichtman JW and Fraser SE. The neuronal naturalist: watching neurons in their native habitat. Nat Neurosci 4, Suppl.: 1215-1220, 2001.[CrossRef][Web of Science][Medline]
Loew LM, Scully S, Simpson L, and Waggoner AS. Evidence for a charge-shift electrochromic mechanism in a probe of membrane potential. Nature 281: 497–499, 1979.[CrossRef][Medline]
Loew LM and Simpson LL. Charge-shift probes of membrane potential: a probable electrochromic mechanism for p-aminostyrylpyridinium probes on a hemispherical lipid bilayer. Biophys J 34: 353–365, 1981.[Web of Science][Medline]
Miesenböck G, De Angelis DA, and Rothman JE. Visualizing secretion and synaptic transmission with pH-sensitive green fluorescent proteins. Nature 394: 192–195, 1998.[CrossRef][Medline]
Miyawaki A, Llopis J, Heim R, McCaffery JM, Adams JA, Ikura M, and Tsien RY. Fluorescent indicators for Ca2+ based on green fluorescent proteins and calmodulin. Nature 388: 882–887, 1997.[CrossRef][Medline]
Moreaux L, Sandre O, Blanchard-Desce M, and Mertz J. Membrane imaging by simultaneous second-harmonic generation and two-photon microscopy. Opt Lett 25: 320–322, 2000a.[CrossRef][Web of Science][Medline]
Moreaux L, Sandre O, Charpak S, Blanchard-Desce M, and Mertz J. Coherent scattering in multi-harmonic light microscopy. Biophys J 80: 1568–1574, 2001.[Web of Science][Medline]
Moreaux L, Sandre O, and Mertz J. Membrane imaging by second-harmonic generation microscopy. J Opt Soc Am B 17: 1685–1694, 2000b.[CrossRef][Web of Science]
Ng M, Roorda RD, Lima SQ, Zemelman BV, Morcillo P, and Miesenböck G. Transmission of olfactory information between three populations of neurons in the antennal lobe of the fly. Neuron 36: 463–474, 2002.[CrossRef][Web of Science][Medline]
Nikolenko V, Nemet B, and Yuste R. A two-photon and second-harmonic microscope. Methods 30: 3–15, 2003.[CrossRef][Web of Science][Medline]
Oancea E, Teruel MN, Quest AF, and Meyer T. Green fluorescent protein (GFP)-tagged cysteine-rich domains from protein kinase C as fluorescent indicators for diacylglycerol signaling in living cells. J Cell Biol 140: 485–498, 1998.
Ormö M, Cubitt AB, Kallio K, Gross LA, Tsien RY, and Remington SJ. Crystal structure of the Aequorea victoria green fluorescent protein. Science 273: 1392–1395, 1996.[Abstract]
Oyler GA, Higgins GA, Hart RA, Battenberg E, Billingsley M, Bloom FE, and Wilson MC. The identification of a novel synaptosomal-associated protein, SNAP-25, differentially expressed by neuronal subpopulations. J Cell Biol 109: 3039–3052, 1989.
Pons T, Moreaux L, Mongin O, Blanchard-Desce M, and Mertz J. Mechanisms of membrane potential sensing with second-harmonic generation microscopy. J Biomed Opt 8: 428–431, 2003.[CrossRef][Web of Science][Medline]
Prasad BC and Reed RR. Chemosensation: molecular mechanisms in worms and mammals. Trends Genet 15: 150–153, 1999.[CrossRef][Web of Science][Medline]
Prasher DC, Eckenrode VK, Ward WW, Prendergast FG, and Cormier MJ. Primary structure of the Aequorea victoria green-fluorescent protein. Gene 111: 229–233, 1992.[CrossRef][Web of Science][Medline]
Rosell FI and Boxer SG. Polarized absorption spectra of green fluorescent protein single crystals: transition dipole moment directions. Biochemistry 42: 177–183, 2003.[CrossRef][Medline]
Roth S and Freund I. Optical second-harmonic scattering in rat-tail tendon. Biopolymers 20: 1271–1290, 1981.[CrossRef][Web of Science][Medline]
Schlesinger MJ, Magee AI, and Schmidt MF. Fatty acid acylation of proteins in cultured cells. J Biol Chem 255: 10021–10024, 1980.
Schote U and Seelig J. Interaction of the neuronal marker dye FM1-43 with lipid membranes. Thermodynamics and lipid ordering. Biochim Biophys Acta 1415: 135–146, 1998.[Medline]
Sechi S and Chait BT. Modification of cysteine residues by alkylation. A tool in peptide mapping and protein identification. Anal Chem 70: 5150–5158, 1998.[Medline]
Shen YR. Surface properties probed by second-harmonic and sum-frequency generation. Nature 337: 519–525, 1989.[CrossRef][Web of Science]
Shimomura O, Johnson FH, and Saiga Y. Extraction, purification, and properties of aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea. J Cell Comp Physiol 59: 223–239, 1962.[CrossRef][Web of Science][Medline]
Siegel MS and Isacoff EY. A genetically encoded optical probe of membrane voltage. Neuron 19: 735–741, 1997.[CrossRef][Web of Science][Medline]
Soeller C and Cannell MB. Construction of a two-photon microscope and optimisation of illumination pulse duration. Pfluegers Arch 432: 555–561, 1996.[CrossRef][Web of Science][Medline]
Solski PA, Quilliam LA, Coats SG, Der CJ, and Buss JE. Targeting proteins to membranes using signal sequences for lipid modification. Methods Enzymol 250: 435–454, 1995.[Web of Science][Medline]
Tsien RY and Bacskai BJ. Video-rate confocal microscopy. In: Handbook of Biological Confocal Microscopy, edited by Pawley JB. New York: Plenum, 1995, p. 459-478.
Veit M, Söllner TH, and Rothman JE. Multiple palmitoylation of synaptotagmin and the t-SNARE SNAP-25. FEBS Lett 385: 119–123, 1996.[CrossRef][Web of Science][Medline]
Vosshall LB, Wong AM, and Axel R. An olfactory sensory map in the fly brain. Cell 102: 147–159, 2000.[CrossRef][Web of Science][Medline]
Yang F, Moss LG, and Phillips GN Jr. The molecular structure of green fluorescent protein. Nat Biotechnol 14: 1246–1251, 1996.[CrossRef][Web of Science][Medline]
Yariv A. Optical Electronics. New York: Holt, Rinehart and Winston, 1991.
Yuste R, Miller RB, Holthoff K, Zhang S, and Miesenböck G. Synapto-pHluorins: chimeras between pH-sensitive mutants of green fluorescent protein and synaptic vesicle membrane proteins as reporters of neurotransmitter release. Methods Enzymol 327: 522–546, 2000.[Web of Science][Medline]
Zaccolo M, De Giorgi F, Cho CY, Feng L, Knapp T, Negulescu PA, Taylor SS, Tsien RY, and Pozzan T. A genetically encoded, fluorescent indicator for cyclic AMP in living cells. Nat Cell Biol 2: 25–29, 2000.[CrossRef][Web of Science][Medline]
Zemelman BV and Miesenböck G. Genetic schemes and schemata in neurophysiology. Curr Opin Neurobiol 11: 409–414, 2001.[CrossRef][Web of Science][Medline]
Zhang J, Campbell RE, Ting AY, and Tsien RY. Creating new fluorescent probes for cell biology. Nat Rev Mol Cell Biol 3: 906–918, 2002.[CrossRef][Web of Science][Medline]
Zipfel WR, Williams RM, Christie R, Nikitin AY, Hyman BT, and Webb WW. Live tissue intrinsic emission microscopy using multiphoton-excited native fluorescence and second harmonic generation. Proc Natl Acad Sci USA 100: 7075–7080, 2003.
Zoumi A, Yeh A, and Tromberg BJ. Imaging cells and extracellular matrix in vivo by using second-harmonic generation and two-photon excited fluorescence. Proc Natl Acad Sci USA 99: 11014–11019, 2002.
This article has been cited by other articles:
|
|
 |
|
  L. Sjulson and G. Miesenbock Rational Optimization and Imaging In Vivo of a Genetically Encoded Optical Voltage Reporter J. Neurosci., May 21, 2008; 28(21): 5582 - 5593. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Sjulson and G. Miesenbock Optical Recording of Action Potentials and Other Discrete Physiological Events: A Perspective from Signal Detection Theory Physiology, February 1, 2007; 22(1): 47 - 55. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Iyer, T. M. Hoogland, and P. Saggau Fast Functional Imaging of Single Neurons Using Random-Access Multiphoton (RAMP) Microscopy J Neurophysiol, January 1, 2006; 95(1): 535 - 545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Dombeck, L. Sacconi, M. Blanchard-Desce, and W. W. Webb Optical Recording of Fast Neuronal Membrane Potential Transients in Acute Mammalian Brain Slices by Second-Harmonic Generation Microscopy J Neurophysiol, November 1, 2005; 94(5): 3628 - 3636. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
