Measuring Serotonin Distribution in Live Cells with Three-Photon Excitation

See allHide authors and affiliations

Science  24 Jan 1997:
Vol. 275, Issue 5299, pp. 530-532
DOI: 10.1126/science.275.5299.530


Tryptophan and serotonin were imaged with infrared illumination by three-photon excitation (3PE) of their native ultraviolet (UV) fluorescence. This technique, established by 3PE cross section measurements of tryptophan and the monoamines serotonin and dopamine, circumvents the limitations imposed by photodamage, scattering, and indiscriminate background encountered in other UV microscopies. Three-dimensionally resolved images are presented along with measurements of the serotonin concentration (∼50 mM) and content (up to ∼5 × 108 molecules) of individual secretory granules.

Neurotransmitter granules have typically been studied either with various imaging techniques (1, 2) that do not directly detect the granular content, or with chemical or electrical assays (3, 4) that identify the granule contents but can probe only the extracellular medium. Thus, it has not been possible to determine neurotransmitter concentration or total neurotransmitter content of individual granules in intact cells. As a solution, we have excited the native UV fluorescence of these molecules by simultaneous absorption of three infrared photons, which accesses shorter wavelength UV transitions in living cells than conventional or two-photon microscopy (5).

When subjected to a high-intensity irradiation at wavelength λ, a molecule that ordinarily absorbs at λ/3 may exhibit fluorescence at ≳λ/3 by a three-photon absorption mechanism (6). The average fluorescence photon count rate F measured from three-photon excitation depends on the three-photon molecular absorption cross section σ3 [unit: (length)6 (time)2 (photon)−2], the instantaneous intensity I, the fluorescence quantum efficiency Q, the detection efficiency K, the wavelength λ, and the concentration C (number of molecules per unit volume). Analogous with two-photon excitation (7), F is determined by the product of I3 and C, integrated over space (r) and time (t′):

Embedded Image (1)

For excitation of a homogeneous dye solution with a focused and pulsed laser beam with a gaussian temporal and spatial profile, the integral yields (SI units)

Embedded Image Embedded Image (2)

where Pav is the average power, f is the pulse repetition rate of the laser, τ is the full-width-at-half-maximum (FWHM) pulsewidth, and ω0 is the beam waist (radius of the 1/e2 intensity contour at the focal plane). This expression reveals the importance of minimizing the focused beam waist and reducing the pulsewidth for maximizing the fluorescence at a given average excitation power. Because λ/ω0 is nearly independent of λ for a focused Gaussian beam, a measurement of F as a function of λ yields the dependence of σ3 on λ. The results of such measurements for tryptophan (Trp), serotonin, and dopamine between 710 and 886 nm are shown in Fig. 1. The shape of the three-photon excitation (3PE) spectrum of Trp differs from the linear (one photon, σ1) spectrum (8). For example, σ1(270 nm)1(237 nm) = 1:5, whereas in the measured three-photon spectrum the corresponding ratio is σ3(810 nm)3(710 nm) = 1:30. For each species, the fluorescence count rate at low powers (5 to 25 mW) shows a cubic dependence on excitation power (9), as expected for 3PE (Eq. 2).

Fig. 1.

Three-photon absorption spectra of Trp (open circles, values multiplied by 2), serotonin (filled circles), and dopamine (filled triangles, values multiplied by 4) (17). Relative errors in the data are <10% of the respective values. The absolute value of the three-photon fluorescence action cross section (Qσ3, see Eq. 2) for Trp at 710 nm is estimated to be 2 × 10−96 m6 s2 photon−2, which can be used as a scaling factor for the graph. Because of uncertainties involved in estimating the collection efficiency (K in Eq. 2), the absolute value has a large uncertainty (∼5×); au, arbitrary unit.

These cross section measurements show how 3PE fluorescence can be used to image the distribution of proteins and neurotransmitters in cellular environments at reasonably benign laser intensities. We built a three-photon UV fluorescence microscopy apparatus based on an existing laser-scanning microscope capable of imaging with two- (5) or three-photon (10) excitation of visible emission dyes. We collected 3PE UV fluorescence images of rat basophilic leukemia (RBL-2H3) cells that were incubated in 250 μM serotonin for 6 hours. Serotonin is actively transported into these cells, and subsequently into secretory granules. It is released in response to immunogenic stimulation by substances in the extracellular environment (11). The images contain sets of ∼20 different horizontal planes (“optical slices”) through the cell bodies in increments of 1.3 μm. A 3PE fluorescence image of a single optical slice 3 μm above the cover slip is shown in Fig. 2A. The corresponding intrinsic fluorescence image from a control culture of cells that lacked serotonin in the incubation medium is shown in Fig. 2B. Bright punctate features are evident only in the cells that were loaded with serotonin. The spectral characteristics of the punctate fluorescence and its absence in the control cells indicate that the fluorescence is generated by serotonin (or potentially a serotonin metabolite) sequestered in granules. Substantial (∼75%) disappearance of the punctate features after immunogenic stimulation was also observed (9) and supports this conclusion. The observed granules may represent both individual vesicles and vesicular clusters that are not separately resolved in the image (12). Figure 3 shows the front and the top view of a three-dimensionally reconstructed image obtained from a set of serotonin-loaded RBL cells (13). The three-dimensional image contains the information required to calculate the volume of each of the observed serotonin granules. Such measurements performed with ∼400 cells yielded a volume distribution histogram (Fig. 4) for these granules (14).

Fig. 2.

Images of RBL cells obtained from 3PE fluorescence of serotonin and other cellular components (18). (A) Cells incubated in 250 μM serotonin for 6 hours (imaged at a horizontal focal plane 3 μm above the cover slip.) (B) Control cells without serotonin incubation. Punctate fluorescence in (A) is caused by serotonin granules.

Fig. 3.

Three-dimensional image of a set of serotonin-loaded RBL cells. The excitation wavelength was 700 nm, and the average power at the sample ∼25 mW. The image is rendered by reconstruction of 20 optical slices. This reconstruction draws isointensity surfaces (white) around regions of equal or higher brightness to demarcate the granules. The red hue represents intensity below the granular cutoff intensity, and thus the density of red overlapping an object provides a depth cue. Box width is 80 μm.

Fig. 4.

Histograms of the distribution of granule volume. Cells were incubated in serotonin for 6 hours. (Inset) Histogram obtained from control cells. The histograms represent the average distributions obtained from ∼400 randomly chosen cells. Serotonin-incubated cells can have 0 to ∼20 recognizable serotonin granules per cell, with a wide variability observed even among neighboring cells. The small (0 to 1 μm3) granules counted in control cells indicate the level of background in these measurements.

Calibration of the granular fluorescence with solutions of known concentrations yields a value of ∼50 mM for the concentration of serotonin in the larger (>1-μm diameter) granules (15). The concentration calibration, together with our volume measurements, provides a measure of the total serotonin content of individual granules or cells. For example, the largest 5% of the granules have an average volume of 4 μm3 (Fig. 4) and thus contain, on average, ∼108 serotonin molecules per granule.

Three-photon excitation thus enables quantitative imaging of cellular processes by harnessing the UV fluorescence of native molecules. This technique offers a necessary complement to extracellular secretion assays such as carbon-fiber amperometry (4), providing the ability to measure the distribution and content of neurotransmitter granules within viable cells.

Note added in proof: We have recently learned that Tan et al. (16) have reported qualitative detection of a general brightness increase of astrocytes incubated in serotonin.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
View Abstract

Navigate This Article