Axially-confined in vivo single-cell labeling by primed conversion using blue and red lasers with conventional confocal microscopes

Abstract

Green-to-red photoconvertible fluorescent proteins have been found to undergo efficient photoconversion by a new method termed primed conversion that uses dual wave-length illumination with blue and red/near-infrared light. By modifying a confocal laser-scanning microscope (CLSM) such that two laser beams only meet at the focal plane, confined photoconversion at the axial dimension has been achieved. The necessity of this custom modification to the CLSM, however, has precluded the wide-spread use of this method. Here, we investigated whether spatially-restricted primed conversion could be achieved with CLSM without any hardware modification. We found that the primed conversion of Dendra2 using a conventional CLSM with two visible lasers (473 nm and 635 nm) and a high NA objective lens (NA, 1.30) resulted in dramatic restriction of photoconversion volume: half-width half-maximum for the axial dimension was below 5 μm, which is comparable to the outcome of the original method that used the microscope modification. As a proof of this method's effectiveness, we used this technique in living zebrafish embryos and succeeded in revealing the complex anatomy of individual neurons packed between neighboring cells. Because unmodified CLSMs are widely available, this method can be widely applicable for labeling cells with single-cell resolution.

Keywords: Dendra2; axial resolution; confocal laser-scanning microscope; primed conversion; zebrafish.

Figure 1

Schematics of dual laser illumination for primed conversion. (a) Schematic of the technique originally reported by Dempsey et al. (2015). The author implemented a specially‐designed filter plate in the excitation light path of a CLSM. The filter plate consists of two semi‐circular optical filters transmittable for blue laser or infrared laser with a thick opaque separator in the middle. With this design, the two laser beams can be aligned to meet only in a defined 3D volume, allowing for efficient axial confinement of photoconversion. (b) Schematic of the technique reported in the current study. Samples are simultaneously illuminated by blue and red laser beams in the standard manner used in CLSM. Spatially‐confined primed conversion can be achieved with this simple method.

Figure 2

Blue/red dual‐laser illumination of Dendra2 in CLSMs leads to spatially‐confined primed conversion. In all experiments, GST‐Dendra2 fusion proteins embedded in a 6% polyacrylamid gel were photoconverted. In (a–d), a 60 × (silicone oil immersion; NA, 1.30) objective lens was used to view the region that was 20 μm away from the coverslip. In (e–g), three different objective lenses were used to view the same region. In (h) and (i), deeper regions in the Z dimension (50 μm) were viewed. (a, b) Photoconversion by spot illumination of 473/635 nm (a) and 405 nm (b). Top panels show X–Y images of red channel at the focal plane, while bottom panels show Y‐Z images. (c, d) Rectangular (31 μm × 3 μm) regions were first illuminated by 473/635 nm (c) or 405 nm (d). Then, the voxel images of red channel were projected to the X‐dimension. Y–Z sections of the projected images are shown in (c) and (d). The right panel for each figure shows the intensity profile in the Z dimension. Full‐width at half‐maximum (FWHM) was 4.2 μm for the illumination of 473/635 nm (c) and 12.1 μm for the illumination of 405 nm. (e‐g) The same experiments shown in (c) were carried out with three different objective lenses. (e) 40 × (silicone oil immersion; NA, 1.25) objective lens. FWHM was 7.3 μm. (f) 40 × (water immersion; NA, 1.15) objective lens. FWHM was 9.8 μm. (g) 30 × (silicone oil immersion; NA, 1.05) objective lens. FWHM was 11.7 μm. (h) The same experiment shown in (c) was carried out except that a deeper region (50 μm away from a coverslip) was viewed. FWHM was 7.8 μm. (i) The same experiment as in (h) was carried out except that 405 nm illumination was used. FWHM was more than 14.5 μm.

Figure 3

Spatially confined primed conversion enables individual neuron labeling in tightly packed neural clusters in living larval zebrafish. Photoconversion experiments were performed in the spinal cords of 3‐to‐4‐dpf larvae expressing Dendra2. (a‐d) Photoconversion of Dendra2 in Tg[HuC:Gal4]; Tg[UAS:Dendra2] with primed conversion (a and b) or conventional 405 nm illumination (c and d). (a and c) are stacked images of the X‐Y dimension, while (b and d) are stacked images of the X‐Z dimension. In each image, the left panel shows both green and red channels (red channel is shown in magenta), while the right panel shows only the red channel (shown in black‐and‐white). The primed conversion successfully highlights only the targeted cell (arrowheads in a and b). In contrast, with 405nm‐illumination, not only the targeted cell (arrowheads in c and d) but also cells located nearby (primarily those located in the Z dimension) express red Dendra2 (arrows in c and d). (e) Photoconversion of Dendra2 in Tg[vachta‐hs:Gal4]; Tg[UAS:Dendra2] with primed conversion. The morphology of an individual trunk motoneuron is clearly revealed in its entirety. (f) Photoconversion of Dendra2 in Tg[chx10:Gal4]; Tg[UAS:Dendra2] with a primed conversion. The morphology of an individual chx10‐positive neuron (a neuron whose axon descends on the same side of the spinal cord; Kimura et al., 2006) is clearly revealed. (g) Photoconversion of Dendra2 in Tg[eng1b‐hs:Gal4]; Tg[UAS:Dendra2] with primed conversion. The morphology of an individual eng1b‐positive neuron (a neuron whose axon ascends on the same side of the spinal cord; Higashijima et al., 2004a) is clearly revealed. Scale bars: 5 μm in a–d; 20 μm in e–g.