Engineering of a red-light-activated human cAMP/cGMP-specific phosphodiesterase

Abstract

Sensory photoreceptors elicit vital physiological adaptations in response to incident light. As light-regulated actuators, photoreceptors underpin optogenetics, which denotes the noninvasive, reversible, and spatiotemporally precise perturbation by light of living cells and organisms. Of particular versatility, naturally occurring photoactivated adenylate cyclases promote the synthesis of the second messenger cAMP under blue light. Here, we have engineered a light-activated phosphodiesterase (LAPD) with complementary light sensitivity and catalytic activity by recombining the photosensor module of Deinococcus radiodurans bacterial phytochrome with the effector module of Homo sapiens phosphodiesterase 2A. Upon red-light absorption, LAPD up-regulates hydrolysis of cAMP and cGMP by up to sixfold, whereas far-red light can be used to down-regulate activity. LAPD also mediates light-activated cAMP and cGMP hydrolysis in eukaryotic cell cultures and in zebrafish embryos; crucially, the biliverdin chromophore of LAPD is available endogenously and does not need to be provided exogenously. LAPD thus establishes a new optogenetic modality that permits light control over diverse cAMP/cGMP-mediated physiological processes. Because red light penetrates tissue more deeply than light of shorter wavelengths, LAPD appears particularly attractive for studies in living organisms.

Fig. 1.

Engineering of the red-light–activated PDE LAPD. (A) Both HsPDE2A [Left, Protein Data Bank (PDB) ID code 3IBJ] and PaBPhy (Center, PDB ID code 3C2W) adopt dimeric structures with extended α-helical interfaces; the regulatory GAF-A and GAF-B domains in PDE2A correspond structurally to the GAF and PHY domains of BPhy. (Right) Chimeric PDEs, shown as a model, derive their regulatory PAS-GAF-PHY tandem from BPhy and their catalytic domain from PDE2A. (B) The biliverdin chromophore of BPhys undergoes photoisomerization around the C15 = C16 double bond. The 15Z isomer gives rise to the red-light–absorbing state Pr, and the 15E isomer gives rise to the far-red-light–absorbing Pfr state. (C) Structures of HsPDE2A and PaBPhy were superposed with respect to their GAF-B and PHY domains, respectively. A helix emanating from the C terminus of the PaBPhy PHY domain (red) spatially overlaps with a helix emanating from the N terminus of the HsPDE2A catalytic domain (light pink). The residue leucine 553 at which the PDE domain is fused to the PAS-GAF-PHY photosensor in LAPD+2 is highlighted in deep purple. (D) Absorption spectra of LAPD in its dark-adapted (black line) and red-light–adapted (dashed line) states. Vertical lines indicate peak wavelengths of light sources used in this study. a.u., absorption units.

Fig. 2.

Red-light–activated cGMP hydrolysis by LAPD. (A) Aliquots were taken from a reaction mix containing 5 nM red-light–adapted LAPD and 100 μM cGMP after 15 s (solid line), 30 s (dashed line), 60 s (dotted line), and 120 s (dashed/dotted line), and were analyzed by HPLC. Over time, the cGMP peak decreases and the GMP peak concomitantly increases (arrows). (B) Integration of peak areas from HPLC analysis yields initial reaction velocities, v₀, at concentrations of 50 μM (▲), 100 μM (●), and 500 μM (■) cGMP. LAPD activity is enhanced under 690-nm light (open symbols) compared with dark conditions (closed symbols). Data are mean ± SD of two measurements; lines denote linear fits to determine initial v₀ values. (C) LAPD displays Michaelis–Menten kinetics in both darkness (●, solid line) and under 690-nm light (Δ, dashed line). The v₀ values were determined by least-squares fitting and are reported as mean ± asymptotic SE. The v_max is increased from 2.5 ± 0.4 μM⋅min⁻¹ (nM LAPD)⁻¹ in the dark to 15.1 ± 0.3 μM⋅min⁻¹ (nM LAPD)⁻¹ under 690-nm light; the K_m values amount to 440 ± 140 μM in the dark and to 340 ± 20 μM under 690-nm light. Lines denote fits to hyperbolic functions. (D) Influence of different light qualities on LAPD activity was studied at an initial cGMP concentration of 1 mM. LAPD was subjected to different light regimes: dark; 690-nm light; 850-nm light; 850-nm light followed by 690-nm light; 690-nm light followed by 850-nm light for 1, 2, and 4 min; 690-nm light followed by dark incubation for 1, 2, and 4 min; 455-nm light; and broadband white light. The rightmost column (apo) represents the activity of LAPD C24A, which lacks the biliverdin chromophore. Catalytic turnover was normalized to the value obtained for 690-nm light. Data are mean ± SD of four measurements.

Fig. 3.

Linker-length dependence of catalytic activity and light regulation in LAPD. Initial reaction velocities in the dark (black bars) or under 690-nm light (white bars) at 1 mM cGMP for LAPD variants differing in the length of the linker between the photosensor and effector modules. Data were determined by least-squares fitting and are reported as mean ± asymptotic SE; asterisks denote activities below the detection limit.

Fig. 4.

LAPD mediates light-activated cGMP hydrolysis in CHO cells. (A) LAPD reporter cell line was stimulated with different amounts of ANP in the dark (●) or light (Δ). The luminescence signal in the light is suppressed compared with the dark. Luminescence data are reported as relative light units (r.l.u.) and represent mean ± SEM of six to seven measurements; lines denote fits to hyperbolic functions. (B) Addition of the PDE2 inhibitor BAY 60-7550 specifically inhibits the catalytic activity of LAPD under both dark and light conditions. Data are mean ± SEM of six to seven measurements; lines denote fits to hyperbolic functions.

Fig. 5.

LAPD mediates light-activated cAMP hydrolysis in zebrafish embryos. Zebrafish embryos were injected at the one-cell stage. At 1 d postfertilization, embryos were incubated with forskolin and illuminated with red or IR light; whole-body cAMP levels were then determined by ELISA. (A) Under red light, LAPD expression resulted in suppression of cAMP levels by 40% in comparison to uninjected control embryos (two-tailed Welch’s t test, P < 10⁻⁵). Data are mean ± SEM of 23 LAPD-injected and 22 uninjected cohorts, where each cohort comprised 15 individual embryos. (B) Under IR light, cAMP levels are the same in LAPD-injected and uninjected control embryos. Data are mean ± SEM of 16 cohorts each.