Illuminating the mechanism and allosteric behavior of NanoLuc luciferase

NanoLuc, a superior β-barrel fold luciferase, was engineered 10 years ago but the nature of its catalysis remains puzzling. Here experimental and computational techniques are combined, revealing that imidazopyrazinone luciferins bind to an intra-barrel catalytic site but also to an allosteric site shaped on the enzyme surface. Structurally, binding to the allosteric site prevents simultaneous binding to the catalytic site, and vice versa, through concerted conformational changes. We demonstrate that restructuration of the allosteric site can boost the luminescent reaction in the remote active site. Mechanistically, an intra-barrel arginine coordinates the imidazopyrazinone component of luciferin, which reacts with O2 via a radical charge-transfer mechanism, and then it also protonates the resulting excited amide product to form a light-emitting neutral species. Concomitantly, an aspartate, supported by two tyrosines, fine-tunes the blue color emitter to secure a high emission intensity. This information is critical to engineering the next-generation of ultrasensitive bioluminescent reporters.

Supplementary Table 2. Luciferase activities of NanoLuc and its mutants with CTZ and FMZ luciferins.Data indicate the average of relative luciferase activity (NanoLuc wild-type = 100%), the error of measurement is represented by the standard deviation (SD).     in a 250 mL round bottomed flask, 3-benzyl-2-hydrazinyl-5-phenylpyrazine (0.4 g, 1.45 mmol) and furan-2carbaldehyde (0.12 mL, 1.45 mmol) were dispersed in acetic acid (4 mL).The mixture was stirred at room temperature during 2 minutes.The resulting solid was re-dissolved in dichloromethane (40 mL) and cyanoborohydride (0.18 g, 2.9 mmol) was added.This was stirred at room temperature for 1 h.The solution was then dispersed in water and ethyl acetate, neutralized with 1 N NaOH (1 equivalent in regard with the acetic acid added).This was extracted with ethyl acetate thrice, the organic layer was washed with a saturated solution of sodium hydrogenocarbonate, distilled water, brine and dried over MgSO4.The solvent was removed under vacuum, to give the crude hydrazine (0.43 g) as a brown oil which was considered pure.

Supplementary Fig. 6 .
Solution structure of NanoLuc determined by SAXS.(a) Experimental SAXS scattering curve for NanoLuc (black dots) is shown against the calculated scattering curves for the NanoLuc monomer (cyan line), dimer (violet line), and tetramer (yellow line).The SAXS curves collected in absence of luciferin (top panel) and in presence of 4 molar excess of FMZ-luciferin (bottom panel) are shown.(b) Ab initio molecular envelope generated from SAXS data analysis.The molecular SAXS envelope of the NanoLuc monomer is shown in a semi-transparent grey color superposed on the NanoLuc monomer of the crystal structure represented as a cyan cartoon.
Confidence contour analysis of Michaelis constant Km, turnover number kcat and enzymeproduct complex dissociation constant Kp obtained as the result of the numerical analysis of kinetic data of NanoLuc catalyzed conversion of FMZ (a, b, c) and CTZ (d, e, f) and NanoLuc-Y94A catalyzed conversion of FMZ (g, h, i) andCTZ (j, k, l), confirming that the determined kinetic parameters are well defined and constrained by the collected kinetic data.The grey dashed lines represent the χ² threshold of 0.95.Numerical analysis of kinetic parameters of NanoLuc-D9R/K89R catalyzed luciferin conversion.(a) The reaction progress curves corresponding to cumulative luminescence production in time recorded upon mixing 0.01 μM -D9R/H57A/K89R with 0.50 (black), 0.24 (dark blue), 0.10 (green) and 0.06 (red) μM FMZ, with gain of the reader set to 1250.(b) The reaction progress curves obtained upon mixing 0.01 μM NanoLuc-D9R/K89R with 3.63 (black), 2.02 (dark blue), 0.97 (green), 0.44 (red), 0.24 (yellow) and 0.09 (magenta) μM CTZ, with gain of the reader set to 1000.Each trace represents an average of three repetitions.The thicker lines represent the experimental data, thinner lines represent the best fit using reaction model in Scheme 1 (see Methods section Measurement of steady-state kinetic parameters of luciferase reaction).
Numerical analysis of kinetic parameters of NanoLuc D9R/H57A/K89R-catalyzed luciferin conversion.(a) The reaction progress curves corresponding to cumulative luminescence production in time recorded upon mixing 0.01 μM NanoLuc-D9R/H57A/K89R with 0.98 (black), 0.34 (blue), 0.23 (green), 0.14 (red), 0.08 (yellow) and 0.04 (magenta) μM FMZ, with gain of the reader set to 1250.(b) The reaction progress curves obtained upon mixing 0.01 μM NanoLuc-D9R/H57A/K89R with 3.69 (black), 0.99 (blue), 0.50 (green) and 0.12 (red) μM CTZ, with gain of the reader set to 1000.Each trace represents an average of three repetitions.The thicker lines represent the experimental data, thinner lines represent the best fit using reaction model in Scheme 1 (see Methods section Measurement of steadystate kinetic parameters of luciferase reaction).Confidence contour analysis of Michaelis constant Km, turnover number kcat and enzymeproduct complex dissociation constant Kp obtained as the result of the numerical analysis of kinetic data of NanoLuc D9R/K89R catalyzed conversion of FMZ (a, b, c) and CTZ (d, e, f) and NanoLuc-D9R/H57A/K89R catalyzed conversion of FMZ (g, h, i) and CTZ (j, k, l), confirming that the determined kinetic parameters are well defined and constrained by the collected kinetic data.The grey dashed lines represent the χ² threshold of 0.95.Docking of FMZ to different structures of monomeric NanoLuc using different-sized docking grids: from x = 60 Å, y = 50 Å, z = 46 Å covering the whole protein (blind docking) to gradually smaller grids constraining the docking into the β-barrel of NanoLuc.

Table 6 .
Dimer dissociation kinetics of NanoLuc dimer with and without FMZ.The kinetic parameters were calculated from the Markov state models.The kinetics were calculated between the macrostates with the lowest and highest mean RMSD.The standard deviations were obtained from bootstrapping of random 80 % of the data 100 times.ΔG 0the free energy of the associated state; kasrate constant of association; kdisrate constant of dissociation; kdis/kasthe ratio of kdis to kas; KDdissociation constant Supplementary Fig.30.Equilibrium probability of associated dimer macrostates from simulation of dimeric NanoLuc with (left column) and without FMZ (right column).The associated macrostate corresponds to the lowest-RMSD macrostate.The error bars show the standard deviation from bootstrapping calculation with a random 80 % of the data repeated 100 times.This plot shows that the associated dimer is significantly more probable when FMZ is present in the simulation. *

Table 8 .
List of used primers in RT-PCR experiments.