Synaptogenesis and synaptic protein localization in the postnatal development of rod bipolar cell dendrites in mouse retina

Abstract

Retinal responses to photons originate in rod photoreceptors and are transmitted to the ganglion cell output of the retina through the primary rod bipolar pathway. At the first synapse of this pathway, input from multiple rods is pooled into individual rod bipolar cells. This architecture is called convergence. Convergence serves to improve sensitivity of rod vision when photons are sparse. Establishment of convergence depends on the development of a proper complement of dendritic tips and transduction proteins in rod bipolar cells. How the dendrites of rod bipolar cells develop and contact the appropriate number of rods is unknown. To answer this question we visualized individual rod bipolar cells in mouse retina during postnatal development and quantified the number of dendritic tips, as well as the expression of transduction proteins within dendrites. Our findings show that the number of dendritic tips in rod bipolar cells increases monotonically during development. The number of tips at P21, P30, and P82 exceeds the previously reported rod convergence ratios, and the majority of these tips are proximal to a presynaptic rod release site, suggesting more rods provide input to a rod bipolar cell. We also show that dendritic transduction cascade members mGluR6 and TRPM1 appear in tips with different timelines. These finding suggest that (a) rod bipolar cell dendrites elaborate without pruning during development, (b) the convergence ratio between rods and rod bipolar cells may be higher than previously reported, and (c) mGluR6 and TRPM1 are trafficked independently during development.

Keywords: RRID:AB_2261205; RRID:SCR_002285; RRID:SCR_014237; RRID:SCR_014305; TRPM1; convergence; glutamate receptors; retina; rod bipolar cells; rod photoreceptors; vision.

FIGURE 1

Identification and analysis of dendrites in individual rod bipolar cells. (a) Side view of a rod bipolar cell in the Grm6-tdTomato mouse line at P30. Maximum intensity projection of a confocal image from a flat mount retina. Background fluorescence is from weakly labeled ON bipolar cells. Inset shows the bulbous axon terminal of the cell in the innermost portion of the inner plexiform layer (IPL). (b) En face view of the dendritic tree of the same cell. (c) Three-dimensional binary representation of the bipolar soma and dendrites from the confocal image shown in (b), referred to as ‘mask’ in the main text. (d) Dendritic skeleton generated from the 3D binary representation of the cell in (c). Skeleton consists of nodes (red) and segments (white). (e) Heat map of dendritic tree and soma in which z-position is indicated by color

FIGURE 2

Rod bipolar cell dendrites elaborate during postnatal development. (a–e) En face view of rod bipolar cell 3D binary mask at indicated postnatal days (P) with a corresponding dendritic contour map on the right. Somas in cell mask images appear different sizes because somas have been cut off at different planes to accentuate the dendritic trees. The soma is displayed in the dendritic contour map in cool colors. (f) Dendritic tip counts based on the skeletonization of the cell mask for each age. For filled circles: P8 =15.9 ±4.4, n =9 cells, n = 4 retinas; P14 =23.6 ±7.2, n =14 cells, n = 4 retinas; P21 =38.8 ±8.7, n =10 cells, n =3 retinas; P30 =56.3 ±3.8, n =13 cells, n =4 retinas; P82 =58.3 ± 10.9, n =12 cells, n =4 retinas. For filled triangles: P30 =58.3 ± 8.8, n =8 cells, n =1 retina (electron microscopy data, green triangle), P82 =74 ±7.1, n =3 cells, n = 2 retinas (super resolution confocal data, blue triangle). Different letters above time points represent a significant difference in the number of dendritic tips (e.g., a vs. b); same letters above time points represent no significant difference in the number of dendritic tips (e.g., a vs. a); mean ±SD, two-tailed post-hoc Tukey test, p <.0001 for all comparisons). Red line is a fit of the data with a saturating exponential function; time constant τ is in postnatal days. Filled triangles data excluded from the fit and offset for clarity

FIGURE 3

Number of mGluR6 puncta per rod bipolar cell increases during postnatal development. (a) Maximum intensity projection of mGluR6 puncta associated with a masked rod bipolar cell at P30; cell mask is not shown. (b) Each mGluR6 punctum of the masked cell was accounted for by a region of interest (ROI) whose size is representative for each age, for P30 ROI diameter =600 nm (see Methods). (c) Total mGluR6 puncta/cell at each developmental stage: P8 =23.6 ±3.8, n =9 cells, n = 4 retinas; P14 =42.2 ±7.5, n =14 cells, n = 4 retinas; P21 =52.3 ±9.7, n =10 cells, n = 3 retinas; P30 =52.2 ±8.7, n =13 cells, n = 4 retinas; P82 =52 ±8.7, n =12 cells, n = 4 retinas. Different letters above time points represent a significant difference in the number of mGluR6 puncta; same letters above time points represent no significant difference in the number of mGluR6 puncta (mean ±SD, two-tailed post-hoc Tukey test, p <.0001 for P8 vs. all, p <.05 for P14 vs. P21, P30, P82). Red line is a fit of the data with a saturating exponential function; time constant τ is in postnatal days

FIGURE 4

Determination of background noise, threshold, and tip localization for mGluR6 and TRPM1. (a) Example of a cell in which the 3D binary cell mask is collapsed into 2D (red). To determine the level of background noise for each protein, 2D masks were used to trace background segment ROIs from the center of the cell (black circle) to a termination within two or more dendritic arms (black lines 1 & 2). (b) Following the step in (a), background segments (red) were transposed onto the maximum intensity projection of the masked mGluR6 signal. (c) The same procedure was performed for masked TRPM1 (blue). (d) Intensity profile of background ROI segments for mGluR6 taken from (b) and concatenated along the x-axis. The regions of the segments that fall within the soma and the dendrites are outlined with rectangles. The dashed red line represents the mGluR6 background threshold, which was the average of the lowest 25% intensity values of mGluR6 in this cell. (e) Intensity profile of background ROI segments for TRPM1 taken from (c) and concatenated along the x-axis. The dashed blue line represents the TRPM1 background threshold selected for this cell obtained in the same way as (d). (f) To determine the staining of mGluR6 and TRPM1 within a dendritic ending, tip segment ROIs were traced over the 2D mask of the cell. Tip segments were drawn over the area where a dendritic process terminates (black lines drawn over 15 tips, which represent 25% of total tips analyzed for this cell). To ensure unbiased selection of dendritic tips, the tip segments were drawn at the same time as the background segments (in a) and without any reference to the masked mGluR6 or TRPM1. (g) Tip segments (red) transposed onto the maximum intensity projection of the masked mGluR6 signal. (h) Tip segments (blue) transposed onto the maximum intensity projection of the masked TRPM1. (i) Intensity profile of the tip ROI segments for mGluR6 taken from (g) and concatenated along the x-axis. Each peak represents the maximum intensity for mGluR6 within one tip. Dashed red line is the background threshold for mGluR6, as obtained from (d). Peak intensity values 2 standard deviations (SD) over the baseline are denoted with a red dot. (j) Intensity profile of the tip ROI segments for TRPM1 taken from (e) and concatenated along the x-axis. Dashed blue line is background threshold from (e) and peaks of 2SD over baseline are denoted with a blue dot. Arrows indicate tips without TRPM1 detected

FIGURE 5

Expression of mGluR6 and TRPM1 in rod bipolar cell dendrites increases with age. (a–e) Examples of mGluR6 and TRPM1 expression in rod bipolar cells at P8, P14, P21, P30, and P82. The mGluR6 and TRPM1 signal are overlaid in all images. (Right) Intensity profiles of tip line segments for each cell (20–25% of total tips analyzed) concatenated along the x-axis, obtained as in Figure 4. Peaks are maximum intensity values for mGluR6 (red) and TRPM1 (blue) within a tip. Dashed red and blue lines represent the baseline threshold for each signal. Peaks with values greater than 2SD of respective baseline are marked with a red and/or blue dot. (f) Summary for all tips analyzed at the different ages. Percentage of tips with both mGluR6 and TPRM1 (red bars, mean ±SD) were 45.5% ±19.9 at P8, n = 9 cells, n = 4 retinas; 58.1% ±23.1 at P14, n = 14 cells, n = 4 retinas; 70.3% ±10.4 at P21, n = 9 cells, n = 3 retinas; 67.9% ±18 at P30, n = 12 cells, n = 4 retinas; and 74.2% ±10.2 at P82, n = 9 cells, n = 4 retinas. For tips with mGluR6 and TRPM1, P8 versus P14 was not significant, P8 versus P21, P30, P82 had p <.05, two-tailed post-hoc Tukey test; different letter above red bars indicates significance for the measure between time points, same letter or no letter indicates no significance. Percentage of tips without mGluR6 (gray bars) were 30.1% ±21.3 at P8, 17.8% ±11.3 at P14, 18.6% ±8.6 at P21, 21% ±14.4 at P30, and 16.9% ±10.7 at P82. There was no significant difference for all comparisons between ages for tips without mGluR6. Percentage of tips without TRPM1 (white bars) were 42% ±11.7 at P8, 33% ±21.4 at P14, 22.7% ±7.9 at P21, 27.6% ± 16.7 at P30, and 17.2% ±6.9 at P82. Comparisons for P8 versus P21 (p <.05) and P8 versus P82 (p <.01) for percentage of tips without TRPM1 were significant, there was no significant difference for all other comparisons between ages for tips without TPRM1 (two-tailed post-hoc Tukey test)

FIGURE 6

mGluR6 and TRPM1 independently trafficked to rod bipolar cell dendritic tips. (a) Example of rod bipolar cell at P82 stained for mGluR6 and TRPM1 and displaying a tip (solid red outline within boxed region and dashed white outlines with arrows in b and c) without either mGluR6 (b) or TPRM1 (c). (d) Rod bipolar cell at P82 and a tip (solid red outline within boxed region and dashed white outlines with arrows in e and f) with mGluR6 (e) and without TRPM1 (f). (g) Rod bipolar cell at P82 and a tip (solid red outline within boxed region and dashed with outlines in h and i) without mGluR6 (h) and with TRPM1 (i). Saturating exponential fits and time constants for percentage of tips with mGluR6 and TRPM1 (black smooth line, black open circles, τ =7.4 days), percentage of tips with TRPM1 (blue smooth line, blue open circles, τ =7.5 days), and percentage of tips with mGluR6 (red smooth line, red open circles, τ =4.0 days). Tips with mGluR6 were observed earlier compared to tips with TRPM1

FIGURE 7

Localization of rod bipolar cell postsynaptic proteins and demonstration of rod divergence. (a) Deconvolved confocal image of P82 rod bipolar cell expressing tdTomato and stained for mGluR6, TRPM1, and Ribeye. (b) The same rod bipolar cell without synaptic protein staining and with the soma cutoff to highlight the dendritic tips. (c) Magnification of a single dendritic process (white dashed box in a and b) showing the tdTomato expression (top left), tdTomato and mGluR6 (top right), tdTomato and TRPM1 (bottom left), and tdTomato, mGluR6, TRPM1, and Ribeye (bottom right). Note that the rod bipolar cell process coincides with one punctum of the mGluR6/ TPRM1 ‘doublet’. (d) Confocal image of a tdTomato-expressing P82 rod bipolar cell. (e) PKCα staining showing all rod bipolar cells. (f) Overlay of the tdTomato and PKCα. (g) Cell from D at the level of dendritic tips. (h) Magnification of one dendritic tip (white dashed box in g) showing the tdTomato expression (top left); PKCα (top right); tdTomato, PKCα, and Ribeye (bottom left); and tdTomato, PKCα, Ribeye, and mGluR6 (bottom right). Example of divergence of the rod (Ribeye) to two different rod bipolar cell processes (white dashed outlines 1 and 2). Only one tdTomato rod bipolar cell process coincides with one of the PKCα-positive dendrites and one of the mGluR6 ‘doublet’ puncta

FIGURE 8

Apposition of rod terminals and rod bipolar cell dendritic tips. (a) Transparent 3D mask from a rod bipolar cell with all Ribeye within the field of view (P82). (b) A single dendritic tip traced to a ribbon and overlaid with mGluR6 signal. The tip occupies one of the two mGluR6 puncta. Red line represents the shortest distance (~480 nm) from the center of the masked tip to the center of the Ribeye staining used to determine apposition of rod bipolar cell tips and rod terminals. (Bottom) The same tip, ribbon and mGluR6 are rotated ~90° to highlight the position of the dendritic tip under the ribbon

FIGURE 9

Rod bipolar cells connect to the majority of available rods. (a) Masked (yellow) and total (green) mGluR6 for a P82 rod bipolar cell. Boxes 1 and 2 show instances of possible missed rod input, as judged by mGluR6 unoccupied by the mask. (b) Transparent rod bipolar cell mask overlaid with the masked and total mGluR6 with same boxes highlighted as in (a). (c) Mature rod bipolar cells rarely missed available rod input. An average of 2 ‘missed’ rod inputs per cell (open triangles) were found at P30 and P82. Only single points analyzed at the earlier ages (P8–21)