Repeated nuclear translocations underlie photoreceptor positioning and lamination of the outer nuclear layer in the mammalian retina

Abstract

In development, almost all stratified neurons must migrate from their birthplace to the appropriate neural layer. Photoreceptors reside in the most apical layer of the retina, near their place of birth. Whether photoreceptors require migratory events for fine-positioning and/or retention within this layer is not well understood. Here, we show that photoreceptor nuclei of the developing mouse retina cyclically exhibit rapid, dynein-1-dependent translocation toward the apical surface, before moving more slowly in the basal direction, likely due to passive displacement by neighboring retinal nuclei. Attenuating dynein 1 function in rod photoreceptors results in their ectopic basal displacement into the outer plexiform layer and inner nuclear layer. Synapse formation is also compromised in these displaced cells. We propose that repeated, apically directed nuclear translocation events are necessary to ensure retention of post-mitotic photoreceptors within the emerging outer nuclear layer during retinogenesis, which is critical for correct neuronal lamination.

Keywords: development; dynein; interkinetic nuclear migration; lamination; motor proteins; neuronal migration; retina; synaptogenesis; tissue differentiation; translocation.

Graphical abstract

Figure 1

Rod photoreceptor nuclei are motile during retinogenesis (A) Location of rods (green) in the developing and adult Nrl.GFP^+/+ retina. (B) Individual rod in P2 Nrl.GFP^+/+ retina. Soma is highlighted (white). (C) Schematic of retinal preparation for time-lapse live imaging to track rod nuclear motions (red dots) in Nrl.GFP^+/+ retinae. (D) Nucleus (red dot) of a segmented rod (green) migrating apico-basally in P3 Nrl.GFP^+/+ retina (grayscale). (E) Overlaid example trajectories depicting apico-basal rod nuclear motility at P3. Basal ONL limit indicated (Ferguson et al., 2013). (F) Representative kinetics of apico-basal rod nuclear motility at P3. Nuclear position, velocity, and acceleration plotted against time. Real data points (position: black dots/red line; velocity and acceleration: gray lines) and moving average (position: blue; velocity and acceleration: black) are shown. Periods of apical- and basal-directed movement are shaded green and red, respectively. Local velocity minima indicate peak of rapid apically directed translocation phase (middle dotted line). Local acceleration minima/maxima indicate initiation and cessation of rapid apically directed translocation (outer dotted lines). (G) Overlaid apically directed (top) and basally directed (bottom) events from a single recording, normalized with respect to the onset of movement at P3 (black/green traces). Green traces show rapid apical movements according to threshold criteria (see STAR Methods). (H) Isolated, above-threshold rapid apical movements normalized as in (G). (I) Distribution of apico-basal starting positions of rapid apical nuclear translocations at P1–P10 normalized with respect to cumulative recording time (sum of trajectory durations) for each condition. (J) Velocity distribution of total (gray), rapid apical (green), and basal (red) rod nuclear movements at P3. The latter two are scaled up for clarity because component data points were only 2.2% and 30.1% of the number of total data points, respectively. (K) Total rod nuclear velocity distributions at P1–P10. (L) Mean squared displacement (MSD) profiles of total pooled rod nuclear translocations at P1–P10. Data show mean ± SEM. See also Figure S3A. (M) Coefficients of movement at P1–P10. Individual data points (blue) represent experimental repeats, with each containing a whole set of nuclear trajectories. Two-way ANOVA with post hoc permutation test. Scale bars, 25 μm (A) and 5 μm (B and D). n.s., not significant; ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001. Abbreviations: GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; NBL, neuroblastic layer; ONL, outer nuclear layer; OPL, outer plexiform layer. See Figure S1 for more information.

Figure 2

Rod nuclear motility does not require dynamic microtubule polymerization (A) Rod MT plus ends labeled with Nrl.EB3-tdTomato (red) in segmented Nrl.GFP^+ve rods (green) at P4 and P8. (B) Time-lapse series of dynamic MT plus ends in rods labeled with Nrl.EB3-tdTomato (red) at 4 days post-retinal electroporation at P0. White arrowheads show moving EB3-tdTomato foci; yellow arrowheads demarcate starting position in the peri-nuclear fork (green region of interest [RO) and growing axon (magenta ROI). (C) P3 Nrl.GFP^+/+ retina (green) following exposure for 10 h to 45 nM demecolcine leading to accumulation of PH3^+ve mitotic figures (red). (D) Time-lapse series of a segmented rod (green) nucleus (red dot) moving apico-basally in a P3 Nrl.GFP^+/+ retina (grayscale) exposed to demecolcine, added at 120 min. (E) Representative, overlaid apico-basal rod nuclear trajectories at P3 in the presence of demecolcine after 120 mins. (F–H) Effect of demecolcine versus DMSO on velocity distribution (F), MSD profiles (G), and coefficients of movement (H) for total rod nuclear movements. Data points represent experimental repeats; unpaired t test. (I–L) Effect of demecolcine versus DMSO on frequency of basal events (I), velocity distribution (J), MSD profiles (K), and MSD profile-derived quadratic coefficients (L) for basal movements. (M–P) Effect of demecolcine versus DMSO on frequency of apical events (M), velocity distribution (N), MSD profiles (O), and MSD profile-derived quadratic coefficients (P) for rapid apical nuclear translocations. Scale bars, 5 μm (A and B) and 10 μm (C and D). Unpaired t test; ^∗p < 0.05; ^∗∗p < 0.01. Data show mean ± SEM (G, K, and O). See also Figures S2B, S3B, and S3C.

Figure 3

Rod nuclear translocation does not require actomyosin constrictions (A) P3 Nrl.GFP^+/+ retina (green) following 10-h exposure to 25 μM blebbistatin resulting in accumulation of PH3^+ve mitotic figures (red). (B) Time-lapse series of a segmented rod (green) nucleus (red dot) migrating apico-basally in P3 Nrl.GFP^+/+ retina (grayscale) exposed to blebbistatin, added after 120 min. (C) Representative, overlaid apico-basal rod nuclear trajectories at P3 in the presence of blebbistatin after 120 mins. (D–F) Effect of Blebbistatin versus DMSO on velocity distribution (D), MSD profiles (E), and coefficients of movement for MSD profiles (F) of total rod nuclear movements. Data points represent experimental repeats, unpaired t test. (G–J) Effect of blebbistatin versus DMSO on frequency of basal events (G), velocity distribution (H), MSD profiles (I), and MSD-profile-derived quadratic coefficients (J) for basal movements. (K–N) Effect of blebbistatin versus DMSO on frequency of apical events (K), velocity distribution (L), MSD profiles (M), and MSD-profile-derived quadratic coefficients (N) for rapid apical nuclear translocations. Scale bars, 10 μm (A) and 5 μm (B). Unpaired t test; ^∗p < 0.05; ^∗∗p < 0.01. Data show mean ± SEM (E, I, and M). See also Figures S3B and S3G.

Figure 4

Dynein 1 mediates rapid apical translocation of photoreceptor nuclei (A) P3 Nrl.GFP^+/+ retina (green), counterstained with DAPI (blue) following 10-h exposure to 25 μM ciliobrevin D resulting in targeting defects of the ciliary transport protein IFT88 (red). (B) Time-lapse recording of a segmented rod (green) nucleus (red dot) migrating apico-basally in P3 Nrl.GFP^+/+ retina (grayscale) exposed to ciliobrevin D, added after a control period of 120 min. (C) Representative, overlaid apico-basal rod nuclear trajectories at P3 in presence of ciliobrevin D after 120 mins. (D) Representative, overlaid apico-basal trajectories at P3 following 30-min ciliobrevin D treatment and subsequent wash out (4 × 30 min). (E–G) Effect of ciliobrevin D versus DMSO on velocity distribution (E), MSD profiles (F), and coefficients of movement (G) for total rod nuclear movements, including following respective washouts. Data points represent experimental repeats. Unpaired t test. (H–K) Effect of ciliobrevin D versus DMSO on frequency of basal events (H), velocity distribution (I), MSD profiles (J), and MSD-profile-derived quadratic coefficients (K) for basal movements. (L–N) Effect of ciliobrevin D versus DMSO on frequency of apical events (L), velocity distribution (M), and MSD profiles (N). N.B. (N) reflect values from only n = 3 recorded rapid apical movements. Scale bars, 5 μm. Unpaired t tests. ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001; ^#, no statistical test performed due to insufficient data points. Data show mean ± SEM (F, J, and N). See also Figures S3B and S3D–S3F.

Figure 5

Short-term dynein 1 loss of function in rods results in impaired rapid apical nuclear translocation and basal displacement (A) Apico-basal positions of transfected rods expressing shCtrl/DsRed (red) or shDync1h1/EGFP (green) in Nrl.Cre^+/− retina following electroporation at P1 and 4 DIV. Arrowhead indicates OPL. See also Figures S5D and S5E. (B) qRT-PCR analysis of Dync1h1 expression in rods at 10-days post-in vivo administration of AAV2/8 shDync1h1/EGFP or AAV2/8 shCtrl/DsRed in P0–P1.5 Nrl.Cre^+/− mice. (C and D) Representative apico-basal nuclear trajectories of shCtrl/DsRed^+ve (C) and shDync1h1/EGFP^+ve (D) rods from time-lapse live-imaging experiments. Retinae were electroporated at P1 and cultured 6 DIV. (E–K) Instantaneous velocity distribution (E), MSD profiles (F), and coefficients of movement (G) of total rod nuclear movements in shDync1h1/EGFP^+ve versus shCtrl/DsRed^+ve cells. Normalized event count (H), velocity distribution (I), MSD profiles (J), and MSD-profile-derived quadratic coefficients (K) of basally directed rod nuclear movements. See also Figure S3H. (L–P) Normalized event count (L), velocity distribution (M), MSD profiles (N), and MSD-profile-derived quadratic coefficients (O) of rapid apically directed rod nuclear movements. N.B. (N) and (O) reflect values from n = 8 recorded rapid apical movements. See also Figure S3H. (P) Normalized apico-basal nuclear distribution relative to ONL thickness of shCtrl/DsRed^+ve (black) or shDync1h1/EGFP^+ve (red) rod cells following electroporation of P1 Nrl.Cre^+/− retinae and culturing for 4 DIV (fixed tissue). (Q and R) Apico-basal nuclear positions relative to the thickness of the ONL of rod cells expressing shCtrl/DsRed (Q) or shDync1h1/EGFP (R). Number of displaced rod PR nuclei per 1,000 μm² of retina. Scale bar, 25 μm (A). Mann-Whitney test, unpaired t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001.

Figure 6

Long-term dynein 1 loss of function in rods results in ectopically located photoreceptors and impaired retinal lamination (A) Virally transduced (AAV2/8) rods in 3-week-old Nrl.Cre^+/− retina expressing shCtrl/DsRed (top panel, red) or shDync1h1/EGFP (bottom panel, green). Virus administered at P1. (B) PKCα (green/red) in 3-week-old retina with virally transduced rods expressing shCtrl/DsRed (red) or shDync1h1/EGFP (green). (C) Representative vertical intensity line profile (position indicated by white arrowheads in B). (D) Number of displaced rod PR nuclei per 1,000 μm² of retina. (E) ONL thickness. (F) Apical process tracking of shCtrl/DsRed^+ve (red) or shDync1h1/EGFP^+ve (green) rods in Nrl.Cre^+/− retina following electroporation at P1 and culturing for 4 DIV; arrowheads indicate OPL. Only processes that extended from the soma to the apical limit of the ONL were tracked. (G) Basally displaced shDync1h1/EGFP^+ve rod PR. Native fluorescence signal and 3D surface rendering are shown. Arrowheads indicate OPL. (H) Apical process status in total (top) and displaced shDync1h1/EGFP^+ve rod population (bottom). (I) Apical processes of selected, displaced shDync1h1/EGFP^+ve rod cells (arrowheads) at 3 weeks post-transduction at P1. For some rods, apical processes were not reliably detectable (cyan arrowhead, right panel). (J) Apical process status among displaced, shDync1h1/EGFP^+ve rods. Scale bars, 25 μm (A, F, and I), 10 μm (G), 5 μm (B). Unpaired t test; ^∗∗∗p < 0.001.

Figure 7

Displaced rod photoreceptors form atypical synaptic contacts (A) Low magnification of AAV2/8 virally transduced rods in 3-week-old Nrl.Cre^+/− retina expressing shCtrl/DsRed (red) with pre-synaptic marker (ribeye, green) and PKCα (BCs; grayscale) immunolabeling. (B) High magnification of shCtrl/DsRed^+ve rods (red) and ribeye (green) relative to PKCα-labeled BCs (grayscale). Typical rod synaptic bouton shown in magnified panels (ribeye was 3D surface rendered and segmented in bottom panel). (C) Low magnification of virally transduced shDync1h1/EGFP^+ve rods (green) with Ribeye (red) and PKCα (grayscale) immunolabeling. (D) High magnification of rod soma (green) displaced in INL exhibiting atypical perinuclear ribeye (red; magenta arrowheads). Rod BCs were stained for PKCα (grayscale). ROIs show horseshoe-shaped (1) and punctated ribeye (2) with 3D surface rendering (red). (E) Schematic representation of cellular and synaptic organization between shCtrl/DsRed^+ve and shDync1h1/EGFP^+ve rods and rod BCs including ribeye (green). (F) Number of ribeye foci per cell. Data points represent individual cells. (G) Ribeye shape (horseshoe versus punctate) frequency. (H) Ribeye location (bouton/process versus soma) frequency. (I) AAV2/8 shCtrl/DsRed-treated retina (red) stained for ribeye (grayscale) and mGluR6 (green) showing correct synaptic labeling (inserted panel shows high magnification of ROI). (J–N) shDync1h1/EGFP-treated retinae (green) stained for ribeye (grayscale) and mGluR6 (red) exhibiting different categories of pre/post-synaptic marker labeling, as follows: horseshoe-shaped ribeye/mGluR6 opposition (J), punctate ribeye/mGluR6 apposition (K), unapposed horseshoe-shaped ribeye (L), unapposed punctate ribeye (M), and completely absent ribeye (N) (inserts show high magnification of ROIs; note that only the first panels show representative examples of the intended categories). (O) Probability of ribeye/mGluR6 apposition in shCtrl/DsRed^+ve versus shDync1h1/EGFP^+ve cells. For shDync1h1/EGFP^+ve cells, a further distinction was made between total, horseshoe-shaped, and punctated ribeye foci. (P) shDync1h1/EGFP^+ve cells remaining within the ONL exhibit normal synaptic structures (correct number, location, and shape of ribeye foci, as well as mGluR6 opposition). Boxed regions of interest were magnified to highlight representative cell (green outline). (Q) Synapse status (correct number, location, and shape of ribeye foci, as well as mGluR6 opposition) in shCtrl/DsRed^+ve and non-displaced shDync1h1/EGFP^+ve rod cells. Scale bars, 10 μm (A and C), 5 μm (B, D, I–N, and P), and 1 μm (I–N and P magnified panels). Mann-Whitney test (F–H), one-way ANOVA (O). ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001.