Impairment of visual function in Gtγ-deficient mice. A, Spatial contrast sensitivity functions (CSFs) of wild type (left, n = 3), Gngt1^−/− (middle, n = 3), and Gnat1^−/− (right, n = 3) mice. Temporal frequency (f_t) was fixed at its optimal value of 0.75 Hz for all light intensities. To determine the maximal contrast sensitivity under each condition, data were fitted with a mouse contrast sensitivity model (Umino et al., 2008). All data points below unity indicate no detectable optomotor responses. B, Averaged amplitudes of spatial CSFs as functions of background light intensity. All values are means ± SEM (n = 3 for all groups). C, Families of ERG responses from wild type (left) and Gngt1^−/− (right) animals. D, Intensity-response relations for scotopic a-waves (left) and b-waves (right). Data were fitted with hyperbolic functions that yielded scotopic a-wave half-saturating light intensities of 0.39 ± 0.08 cd·s m⁻² (wild type, n = 7) and 13.0 ± 3.1 cd·s m⁻² (Gngt1^−/−, n = 7), and a-wave maximum amplitudes of 382 ± 41 µV (wild type, n = 7) and 207 ± 15 µV (Gngt1^−/−, n = 7). Fitting the b-wave data yielded half-saturating light intensities of 0.009 ± 0.002 cd·s m⁻² (wild type, n = 7) and 0.34 ± 0.04 cd·s m⁻² (Gngt1^−/−, n = 7), and b-wave maximum amplitudes of 912 ± 69 µV (wild type, n = 7) and 821 ± 50 µV (Gngt1^−/−, n = 7). Values are means ± SEM.

Interaction of Gtα with R* and analysis of purified Gt for the presence of several Gγ subunits. A, Gtα interactions with photoactivated ROS disk membranes. (Top) representative Western blot of the membrane (M) and soluble (S) fractions using anti-Gtα antibodies. (Bottom) Corresponding densitometry results for the Gtα bands. Experimental conditions were: 3 µM rhodopsin, medium ionic strength buffer, 100% bleach, 4 °C. All values are means ± SEM (n = 3). B, Western blot analysis of Gt purified from photoactivated washed Gngt1^−/− ROS disk membranes. Shown are the starting (crude, left) ROS sample (0.6 µg) and the final purified Gt sample (right). Top portion was stained with anti-Gtα antibody; bottom portions were stained with anti-Gγ₃, anti-Gγ₅ and anti-Gγ₂ antibodies.

Genetic, morphological, and biochemical characterization of Gngt1^−/− mice. A, Schematic representation of wild type (WT) and knockout (KO) alleles. WT gene exons 1–3 are shown by tall white boxes. The Gtγ protein coding region in exons 2 and 3 is shown in black. SA is the 1.6 Kb short homology arm. LA is the 7.3 Kb long homology arm. The 3.4 Kb region of the Gngt1 gene encompassing exons 1, 2, and 3 was replaced by 1.8 Kb Neo cassette. For PCR genotyping, a 460 bp DNA fragment in the WT allele was amplified by WTi1 and WTi2 primers, and a 1.8 kb fragment in the KO allele was amplified by A1 and N1 primers (not shown). Arrows show the position of the primers. B, Immunostaining of wild type and Gngt1^−/− retinas with anti-Gtγ antibody: immunogold staining with silver enhancement, LR White embedment, Toluidine Blue counterstaining (scale bar, 20 µm). C, Average number of rows of ONL nuclei (means ± SD; n = 15) as a function of age. Cyan circles (Deltagen) represent comparative data reconstituted from Fig. 2 in Lobanova et al., (2008). D, Immunoblotting of whole retina extracts and purified rod outer segment (ROS) disk membranes from Gngt1^−/− retinas using anti-opsin and anti-cytochrome c antibodies. The lack of cyt c in ROS membranes demonstrates that they were not contaminated by the rod inner segment (RIS) material. E–H, Expression of retinal proteins determined by quantitative immunoblotting: rhodopsin in the retina (E); Gtγ (F), Gtβ (G), and Gtα (H) in ROS disk membranes isolated from Gngt1^−/− retinas. Levels of rhodopsin and transducin α and β subunits in whole mouse retinas and ROS disk preparations were determined based on quantitative calibrations with highly purified bovine rhodopsin and transducin standards. The data represent means ± SD from three independent experiments. I, Effects of Gtγ deletion on the expression of phototransduction proteins in rods. Whole retina and ROS samples were prepared from two-month old wild type, Gngt1^+/−, and Gngt1^−/− mice. The sample rhodopsin level was used as a loading control for quantification of phototransduction proteins. Similar results were obtained when total protein was used as the loading control (not shown).

Light responses of control and Gtγ-deficient mouse rods. A, Representative families of flash responses from two-month-old wild type (left), Gngt1^+/− (middle), and Gngt1^−/− (right) mouse rods. Test-flashes of 500 nm light with intensities of 5, 15, 39, 125, 444, and 1406 photons µm⁻² (for wild type and Gngt1^+/− rods), or 444, 1406, 4630, 14670, 40440 and 128160 photons µm⁻² (for Gngt1^−/− rods) were delivered at time 0. Red traces show responses to identical light intensity (1406 photons µm⁻²). B, Normalized averaged intensity-response functions. Data were fitted with saturating exponential functions that yielded half-saturating light intensities of 93, 130 and 8408 photons µm⁻² for wild type (n = 50), Gngt1^+/− (n = 27), and Gngt1^−/− (n = 41) mouse rods, respectively (see Table 1). Error bars (SEM) are smaller than the symbol size. C, Phototransduction cascade amplification in mouse rods. Population-averaged dim flash responses to light intensities corresponding to 15 photons µm⁻² for wild type and Gngt1^+/− rods and 1406 photons µm⁻² for Gngt1^−/− rods were normalized to their corresponding maximum dark currents, r_max. Then the Gngt1^+/− and Gngt1^−/− responses were scaled to make the initial parts of all three responses to coincide. Correspondingly scaled light intensities were 1:1:0.025 (wild type:Gngt1^+/−:Gngt1^−/−). A slight response shift of 3 ms (Gngt1^+/−) and 5 ms (Gngt1^−/−) to longer times compared to wild type rods was necessary, mostly caused by the low-pass filtering of the recordings.

Phototransduction cascade inactivation in control and Gtγ-deficient mouse rods. A, Normalized population-averaged dim flash responses (to light intensities of 15 photons µm⁻² for wild type and Gngt1^+/− rods and 1406 photons µm⁻² for Gngt1^−/− rods) demonstrating the accelerated photoresponse inactivation in Gγ-deficient rods. B, Determination of the dominant recovery time constant (τ_D) from a series of supersaturating flashes. Linear fits throughout the data yielded τ_D-values indicated in Table 1. Values are means ± SEM. C, Simulations of wild type and Gngt1^−/− rod responses with a mathematical model of mouse rod phototransduction (Kuzmin, 2004). Grey traces show model responses superimposed on experimental curves. For parameters of fitting, see Tables 1 and 2.