Signal transduction in light-oxygen-voltage receptors lacking the active-site glutamine

Abstract

In nature as in biotechnology, light-oxygen-voltage photoreceptors perceive blue light to elicit spatiotemporally defined cellular responses. Photon absorption drives thioadduct formation between a conserved cysteine and the flavin chromophore. An equally conserved, proximal glutamine processes the resultant flavin protonation into downstream hydrogen-bond rearrangements. Here, we report that this glutamine, long deemed essential, is generally dispensable. In its absence, several light-oxygen-voltage receptors invariably retained productive, if often attenuated, signaling responses. Structures of a light-oxygen-voltage paradigm at around 1 Å resolution revealed highly similar light-induced conformational changes, irrespective of whether the glutamine is present. Naturally occurring, glutamine-deficient light-oxygen-voltage receptors likely serve as bona fide photoreceptors, as we showcase for a diguanylate cyclase. We propose that without the glutamine, water molecules transiently approach the chromophore and thus propagate flavin protonation downstream. Signaling without glutamine appears intrinsic to light-oxygen-voltage receptors, which pertains to biotechnological applications and suggests evolutionary descendance from redox-active flavoproteins.

Fig. 1. Photochemistry of light-oxygen-voltage receptors and sequences of proteins under study.

a Photocycle of light-oxygen-voltage (LOV) receptors. Absorption of blue light by the dark-adapted state (D₄₅₀) prompts the LOV receptor to traverse short-lived excited singlet (S₁) and triplet (T₁) states before assuming the light-adapted state (S₃₉₀), which is characterized by a thioadduct between the flavin atom C4a and the sidechain of a conserved cysteine. Adduct formation goes along with protonation of the N5 atom which entails changes in hydrogen bonding within the LOV receptor, particularly of a conserved glutamine residue situated in strand Iβ of an antiparallel β-pleated sheet. The light-adapted state passively decays to the dark-adapted state over a matter of seconds to hours, depending on the flavin surroundings. b Multiple sequence alignment of A. sativa phototropin 1 LOV2 (AsLOV2), B. subtilis YtvA LOV (BsYtvA), N. multipartita PAL LOV (NmPAL), and N. crassa Vivid LOV (NcVVD). The secondary structure, as observed in AsLOV2, is indicated on top, with α helices in tan and β strands in blue. Residues strongly conserved across LOV receptors are highlighted by arrows and gray shading.

Fig. 2. Activity and light response of YF1 variants.

a The net kinase activity of the variants was assessed in the pDusk-DsRed setup, where alongside the response regulator BjFixJ, YF1 drives the expression of the red-fluorescent reporter DsRed in blue-light-repressed manner. b Normalized DsRed fluorescence of E. coli cultures harboring pDusk plasmids encoding different YF1 variants. Cells were cultivated in darkness (black dots and gray bars) or under constant blue light (white dots and blue bars). Data represent mean ± s.d. of three biologically independent replicates. c Absorbance spectra of YF1 Q123L (solid lines) in its dark-adapted (black) and light-adapted states (blue), compared to the corresponding spectra of YF1 (dotted lines). d Schematic of the coupled fluorescence anisotropy assay to probe YF1 activity. Once phosphorylated in light-dependent manner (see a), BjFixJ homodimerizes and binds to its cognate DNA operator sequence. Said operator is embedded in a TAMRA-labeled double-stranded DNA molecule, and BjFixJ binding can be detected as an increase in fluorescence anisotropy due to decelerated rotational tumbling. e YF1 (black dots), YF1 Q123L (white diamonds), YF1 Q123H (yellow triangles), or YF1 Q123P (red squares) were incubated in darkness together with BjFixJ and the TAMRA-labeled DNA. At time zero, the reaction was initiated by ATP addition and fluorescence anisotropy was recorded for 30 min. Samples were then illuminated for 30 s with blue light (blue bar), and the measurement continued. All experiments were repeated at least twice with similar results.

Fig. 3. Activity and light response of NmPAL variants.

a By embedding a specific aptamer (denoted 04.17) near the Shine-Dalgarno sequence (SD) of an mRNA encoding DsRed, the expression of the fluorescent reporter can be modulated with NmPAL as a function of blue light. In darkness, NmPAL shows little affinity for the aptamer, and expression ensues. Under blue light, NmPAL binds and thus attenuates expression. b E. coli cultures harboring different NmPAL variants and the reporter system depicted in panel a were cultivated in darkness (black dots and gray bars) or under blue light (white dots and blue bars). Normalized DsRed fluorescence represents mean ± s.d. of at least three biologically independent samples. c NmPAL variants were expressed in HeLa cells to translationally repress expression of a luciferase reporter, conceptually similar to the setup shown in panel a but using the 53.19 aptamer. Bars represent the mean ± s.d. of luminescence acquired for six biologically independent samples incubated in darkness (black dots and gray bars) or under blue light (white dots and blue bars). UTR5 refers to the intact reporter system giving rise to NmPAL-mediated light responses; in M21, NmPAL binding is disrupted by a mutation in the target aptamer, and light responsiveness is abolished. As positive and negative controls, luciferase was constitutively expressed (Luc) or left out altogether (-Luc). d Ratio of the luminescence values obtained under light and dark conditions. e The interaction of wild-type NmPAL with the TAMRA-labeled 04.17 aptamer was assessed in its dark-adapted (black dots) and light-adapted states (white dots) by fluorescence anisotropy. The line represents a fit to a single-site binding isotherm. f As in e but for NmPAL Q347L. Experiments in panels e and f were repeated twice with similar results.

Fig. 4. Light response of AsLOV2 variants.

a Far-UV circular dichroism (CD) spectra of AsLOV2 in its dark-adapted (black) and light-adapted states (blue), and after dark recovery (gray dotted). b Recovery reaction of AsLOV2 following blue-light exposure, as monitored by the CD signal at (220 ± 5) nm. Data were fitted to a single-exponential decay (red line), yielding a recovery rate constant k₋₁ of (1.43 ± 0.05) × 10⁻² s⁻¹. c As a but for AsLOV2 Q513L. d As panel b but for AsLOV2 Q513L, with k₋₁ amounting to (6.61 ± 0.15) × 10⁻⁴ s⁻¹. e As a but for AsLOV2 C450A:Q513D. f As a but for AsLOV2 C450A:Q513D ΔA’α ΔJα. Experiments were repeated at least twice with similar results.

Fig. 5. Structural analyses of AsLOV2 variants.

a Chromophore-binding pocket of wild-type AsLOV2 in its dark-adapted state as revealed by a 1.00 Å crystal structure. b Chromophore-binding pocket of wild-type AsLOV2 in its light-adapted state as revealed by a 1.09 Å crystal structure. c Chromophore-binding pocket of AsLOV2 Q513L in its dark-adapted state as revealed by a 0.90 Å crystal structure. d Chromophore-binding pocket of AsLOV2 Q513L in its light-adapted state as revealed by a 0.98 Å crystal structure. For clarity, helices Cα and Dα are not shown in panels a-d. The Jα helix is drawn in orange, and the flavin- mononucleotide cofactor and key amino acids are highlighted in stick representation. Minor conformations of residues and the flavin nucleotide are drawn in narrower diameter. The active-site cysteine 450 adopts two principal orientations, denoted ‘a’ and ‘b’. In the structures of dark-adapted AsLOV2 wild-type, dark-adapted Q513L, and light-adapted Q513L, orientation ‘b’ splits into two subpopulations with slightly different χ₁ angles. Dashed lines denote hydrogen bonds. e Water density in the interior of dark-adapted AsLOV2 Q513L derived from a 300 ns classical molecular dynamics simulation. The red mesh denotes a density level of 0.3 water molecules per ų. f As e but for light-adapted AsLOV2 Q513L. Corresponding simulations for AsLOV2 wild-type are provided in Suppl. Fig. 13a, b.

Fig. 6. Naturally occurring, glutamine-deficient LOV^ΔQ receptors.

a Sequence searches identify around 350 receptors that have homology to bona fide LOV receptors but lack the conserved active-site glutamine. The multiple sequence alignment shows AsLOV2 as a reference and four selected glutamine-deficient receptors. The sequence logo below the alignment was calculated for the entire set of glutamine-deficient LOV receptors (see Suppl. Data 3). Coloring, shading, and arrows as in Fig. 1, with the position of the conserved glutamine residue indicated by a red arrow. b Activity and light response of the LOV^ΔQ-GGDEF fragment of WP_140774521.1 were assessed in an E. coli reporter strain harboring a dgcE knockout and a translational fusion between the cyclic-di-GMP-controlled csgB and lacZ. Light-dependent diguanylate cyclase activity can hence be gauged by measuring β-galactosidase levels. c Bacteria expressing the wild-type LOV^ΔQ-GGDEF receptor or the M140Q variant were cultivated in darkness (black dots and gray bars), under blue light (white dots and blue bars), or under red light (red dots and bars). ‘−‘ refers to an empty-vector negative control, and ‘+’ denotes a strain expressing the major diguanylate cyclase DgcE that served as the positive control. β-galactosidase activity is reported in Miller units and represents mean ± s.d. of four biologically independent replicates. The experiment was repeated twice with a similar outcome.

Fig. 7. Signal transduction in light-oxygen-voltage (LOV) receptors lacking the conserved glutamine, exemplified for the A. sativa phototropin 1 LOV2 domain.

a Lewis formulae show the flavin- nucleotide chromophore and surrounding residues of the glutamine-deficient leucine variant in the dark-adapted (left) and light-adapted states (right). As revealed by X-ray crystallography (see Fig. 5), qualitatively similar structural responses to light-induced N5 protonation (see Fig. 1) are observed in both the absence of the conserved glutamine Q513 and in its presence (see Suppl. Fig. 15). Without the glutamine, water molecules might transiently enter the chromophore-binding pocket, thereby stand-in for the glutamine, and relay the signal as changes in hydrogen bonding to the Iβ and Aβ strands of the central β pleated sheet (involving residues N414 and 513). Notably, signals are thus also propagated to the LOV C terminus (D515) which is frequently engaged in signal transduction, and often exhibits a conserved DIT motif. b The observation that LOV receptors can transduce signals without either or both of their strongly conserved cysteine and glutamine residues suggests a potential origin from redox-active flavoproteins. LOV signal transduction in a primordial LOV ancestor lacking the Cys and Gln residues would have relied on flavin photoreduction to the NSQ radical and on water mediation. Both the Cys and Gln residues would be secondary acquisitions that minimize side reactions (Cys); enhance the fidelity of signal transduction (Cys and Gln); bathochromically shift the action spectrum (Gln); accelerate the dark recovery and thereby benefit temporal resolution (Cys and Gln); and render the signaling state less susceptible to the cellular environment (Cys). Note that we have no evidence in which sequential order the Gln and Cys residues may have been acquired.