Abstract
Abstract: :
Purpose: The rod single photon response (SPR) exhibits a high degree of reproducibility not predicted by a single molecular decay process [Baylor et al., 1979; Rieke & Baylor, 1998 (RB98); Whitlock & Lamb, 1999 (WL99)]. We developed a stochastic molecular model of rod phototransduction to test theories of SPR reproducibility, with emphasis on the roles of multiple (n) sequential phosphorylation (Pn) of activated rhodopsin (Rn*), capping of Rn* by arrestin (Arr), and Ca feedback (WL99). The model also can simulate point mutations, knockouts, and biochemical manipulations. Methods: Monte-Carlo simulations of dim-flash responses were run with Poisson arrival of photons. Model assumptions: Rn* competitively binds with inactive G protein (G), rhodopsin kinase (RK) or Arr; Rn* activity is terminated upon Arr-capping; G affinity for Rn* decreases exponentially with n while the affinity of Arr for Rn* increases linearly with n (Gibson et al., 2000); affinity of RK for Rn* decreases exponentially with n; the rate of PDE* recovery (τPDE) is rate-limiting; Ca feedback occurs at guanylate cyclase only. Model variations: Arr caps Rn* only after the final Pn ("late-quenching"); reduced τPDE so that Rn* lifetime (TR*) controls recovery; add Ca feedback onto TR* via RK. Results: (1) Gibson et al.'s scheme of sequential Pn and Arr-capping does reduce the variability of R* lifetime, of SPR kinetics, and of SPR amplitude. However, (2) empirical variability of SPR amplitude and kinetics cannot be achieved simultaneously, even with 7 Pn sites. (3) Late quenching works, but only if there are ≥4 functional Pn sites and the affinities of RK for Rn* and G for Rn* are proportional for all n. Cyclase feedback alone: (4) reduces the variability of SPR amplitudes, but not sufficiently, (5) increases tlife (inferred TR*, from fitting SPRs; WL99), and decreases tlife variability, even though this feedback cannot affect actual TR*. (6) Ca-feedback onto TR* has negligible effect on SPR reproducibility. (7) With TR* rate-limiting, SPR waveforms are abnormally sustained. Conclusions: Inactivation of R* by Pn and late Arr-capping can account for SPR reproducibility, but only with constraints on the relative affinities of RK and G for Rn*. However, no model variations tested account for other key data (e.g., G gain manipulations of RB98).
Keywords: 364 computational modeling • 555 retina: distal(photoreceptors, horizontal cells, bipolar cells) • 517 photoreceptors