Comparing confocal and two-photon Ca2+ imaging of thin low-scattering preparations

Abstract

Ca²⁺ imaging provides insight into biological processes ranging from subcellular dynamics to neural network activity. Two-photon microscopy has assumed a dominant role in Ca²⁺ imaging. The longer wavelength infra-red illumination undergoes less scattering, and absorption is confined to the focal plane. Two-photon imaging can thus penetrate thick tissue ∼10-fold more deeply than single-photon visible imaging to make two-photon microscopy an exceptionally powerful method for probing function in intact brain. However, two-photon excitation produces photobleaching and photodamage that increase very steeply with light intensity, limiting how strongly one can illuminate. In thin samples, illumination intensity can assume a dominant role in determining signal quality, raising the possibility that single-photon microscopy may be preferable. We therefore tested laser scanning single-photon and two-photon microscopy side by side with Ca²⁺ imaging in neuronal compartments at the surface of a brain slice. We optimized illumination intensity for each light source to obtain the brightest signal without photobleaching. Intracellular Ca²⁺ rises elicited by one action potential had twice the signal/noise ratio with confocal as with two-photon imaging in axons, were 31% higher in dendrites, and about the same in cell bodies. The superior performance of confocal imaging in finer neuronal processes likely reflects the dominance of shot noise when fluorescence is dim. Thus, when out-of-focus absorption and scattering are not issues, single-photon confocal imaging can yield better quality signals than two-photon microscopy.

Fig. 1

(A) Light paths for 2PM and confocal microscopy (black, excitation; gray, emission). For 2PM, 820-nm light is reflected by DM1 to the scanhead, into the turret, the objective, and to the sample. Nondescanned emitted light is reflected in the turret to PMT2. For confocal microscopy, 488-nm light was reflected by DM2, through DM1, into the scanhead, the empty turret, the objective, and sample. Emitted light was directed to the scanhead, through DM1 and DM2 to a pinhole (set to 200 μm), and to PMT1. In 2PM, the turret could also be selected to direct light to the scanhead and PMT1 (pinhole, 2 mm). (B) 2PM image of a granule cell soma filled with OGB1. (C1) An OGB1-filled axon from a Z-stack projection using 2PM. (C2) The same axon in confocal microscopy. (D1) A dendrite in 2PM and (D2) in confocal microscopy. All images normalized to their maxima.

Fig. 2

(A1) Confocal image of axon and bouton, with white line selected for scanning. (A2). Line scans show fluorescence versus time, increasing following an action potential (white arrow). With confocal microscopy, line-scanned fluorescence (arbitrary units) was averaged through a bouton (B1), dendrite (B2), or soma (B3) and plotted versus time. Action potentials were evoked at arrows. 2PM in the same bouton (C1), dendrite (C2), or soma (C3). Fluorescence rose after action potentials (arrows). The ΔF/F₀ scale applies to all panels. (D) Comparison of SNR between confocal microscopy and 2PM, with nondescanned and descanned 2PM detection (materials and methods; Fig. 1A). Recordings were made pairwise in the same compartment, alternating between confocal microscopy and 2PM (12 boutons, 11 dendrites, 11 somata; mean ± SE, ∗p < 0.05, ∗∗p < 0.01, one-tailed paired t-test).