Synthesis and evaluation of D-thioluciferin, a bioluminescent 6’ -thio analog of D-luciferin

All known light-emitting firefly-bioluminescent luciferin analogs are either derived from the 6’ -hydroxy- and/or 6’ -aminoluciferin. We report the synthesis of D-thioluciferin, a 6’ -thio analog or isostere of D-luciferin, starting from p -aminothiophenol, using a unique thioacrylate-S -protecting-group strategy. Upon treatment of D-thioluciferin with purified Photinus pyralis (Ppy) luciferase (Luc), a bioluminescence emission with a red-shift λ max relative to D-luciferin was observed. It was also shown that disulphide and sulphide analogs of D-thioluciferin did not produce similar bioluminescences relative to D-thioluciferin when treated with Ppy Luc under standard conditions, thus, providing a foundation for the development of D-thioluciferin based probes based on disulphide reduction and S -dealkylation .


D-Luciferin
is the light-emitting molecule responsible for the bioluminescence observed in the American firefly Photinus pyralis (scheme 1). 1 Biosynthesis of D-L-luciferin starts with quinone followed by the addition of two mol equivalents of L-cysteine with concomitant loss of CO2. 2 The enzyme-controlled stereochemical inversion of L-luciferin to D-luciferin occurs by activation with CoA by virtue of a thioester conjugate at the carboxyl moiety. 3The latter process can be considered a natural light switch that is utilised as a means of communication for the firefly.Both D-and L-luciferin reacts in the presence of O2, ATP, Mg 2+ and the luciferase enzyme, but only D-luciferin (1) produces a yellow/green light.There is speculation that the firefly does not waste the resulting oxyluciferin and can use luciferin-regenerating enzyme (LRE) to produce the 6hydroxy-1,3-benzothiazole-2-carbonitrile once again.How, or even if, the oxyluciferin is recycled, is still under investigation.The chirality at the carboxyl group in natural firefly luciferin is of the S form, as it was established by the early chemical synthesis of D-luciferin (1) from 6-hydroxy-1,3-benzothiazole-2-carbonitrile and D-cysteine.L-Luciferin has the R form and is not used for the luminescence reaction by firefly luciferase. 4hus, firefly luciferase oxidizes only D-luciferin, a specificity which has been exploited in gene reporter as well as cell viability assays based on ATP production. 5heme 1. Biosynthesis of D-luciferin (1) in the firefly, Photinus pyralis.
Modifications to the natural substrate have resulted in new luminogenic substrates with often improved properties, which have been exploited in the development of sensitive luciferin-based probes for invivo imaging, also known as "caged luciferins". 6Most of these probes, however, are based on the release of either natural D-luciferin (1) or a 6'-amino analog, D-6-aminoluciferin (2). 7,8 he natural 6-OH and synthetic 6-NH2 bioluminescent substrates have limited bioanalytical applications, particularly in terms of coupling bioluminescence activity directly with sulfur biology.The development of thiol-sensing technologies has recently become an area of increased interest because of the biological importance of thiol-containing molecules such as cysteine (Cys), homocysteine (Hcy) and glutathione (GSH). 9] It is known that HC≡C−EWG compounds with strong EWGs such as SO2R and CO2R can react at ambient temperatures with relatively very weak nucleophiles, in the presence of suitable catalysts. 12In the latter report, pyrrolidine-mediated deprotection of thiolacrylates has been demonstrated in organic medium.The reactions of Cys with HC≡C−CONHR, which are poorer Michael acceptors, proved to be sufficiently quick, complete, and Z-stereoselective in aqueous media. 13The kinetics of thioacrylate protection of cysteine and its release under physiological conditions have been investigated. 14Notably, peptides modified by terminal alkynones could be converted back into the unmodified peptides by treatment with thiophenol and free cysteine under mild reaction conditions.
Using the latter synthetic approach requires a p-aminothiophenol with an ideal sulfur-protecting group that could withstand the subsequent reaction conditions.Moreover, preparation of luciferin derivatives requires the use of a palladium (II) catalyst, which is expensive to use and can be poisoned by free thiolcontaining reagents.There is, therefore, a need for novel synthetic methods of producing D-thioluciferins and their derivatives.We utilised an expedient synthetic approach for D-thioluciferin (3) (Jardine et al.PCT/IB2018/055542) based on the established preparation of both D-luciferin (1) and D-aminoluciferin (2). 8ost thioether-sulfur protecting groups are either too labile (e.g., trityl) or too stable (e.g., alkyl) for consideration under the latter synthetic strategy. 16In addition, thioesters could not be considered as protecting groups since the cross-coupling reaction that produces benzothiazoles from thioanilides also produces benzothiofurans.8] This methodology proceeds via the 6-allylthiobenzothiazole (IV), followed by periodate oxidation, to give the allylic sulfoxide (VII), which then rearranges to an intermediate allylic sulfenate that is subsequently cleaved by triphenyl phosphite, as a reductant, to give the target intermediate 6-mercapto-1,3-benzothiazole-2-carbonitrile (VIII).Facile addition of D-cysteine gave Dthioluciferin (3).
would release D-thioluciferin (3) via the addition of 2 mol equivalents D-cysteine.The thiazoline ring formation is known to be a facile, near quantitative, reaction.The regioselective addition of one mol of cysteine to complete the carboxy-thiazoline ring to give the thioacrylate protected thioluciferin (IX) and, subsequently, another mol equivalent of D-cysteine to effect a thia-Michael addition, followed by a retro-Michael reaction, resulting in the release of D-thioluciferin (3).Gratifyingly, the thioacrylate (IVd) was well tolerated by the coupling reaction, which was interesting because there are not many reported options for thiol-protecting groups which are both easily removed and stable to palladium-mediated chemistry.
The unsaturated vinyl sulfide units (IVd or IX) could be cleaved by an addition/elimination mechanism by treatment with a thiol (RSH) (Scheme 3).Notably, D-cysteine reacts regioselectively at the nitrile group when limited to 1 mol equivalent.Thiol-sensitive or "caged" D-thioluciferin (IX) could essentially be deprotected with any biothiol (RSH) in aqueous medium.Shiu et al. investigated the modification of cysteinecontaining peptides in which they found that thiol-protection of cysteine using electron-deficient alkynes favoured formation of the Z-isomer. 14Accordingly, peptides modified by terminal alkynones could be converted back into the unmodified peptides by treatment with thiols under mild reaction conditions.The driving force for the reaction is the elimination of the more stable thiolate anion of D-thioluciferin.Scheme 3. Mechanism of thiol mediated thioacrylate (IX) deprotection and simultaneous D-thioluciferin (3) synthesis.

Spectroscopic characterization
In the presence of ATP, D-luciferin ( 1) is oxidized by luciferase to generate oxyluciferin, thereby, resulting in production of bioluminescence and loss of fluorescence proportional to the concentration of ATP.The emission wavelength of bioluminescence for D-thioluciferin (3) was then evaluated (Figure 1).It exhibited a red-shifted light emission (599 nm) when treated with purified firefly luciferase (luc) expressed from E. coli, relative to (1) (557 nm) and D-aminoluciferin (2) (593 nm) The efficiency of a bioluminescent reaction is determined by the product of the quantum yield and reaction rate.Thus, quantitative analysis and knowledge of the quantum yield and the reaction kinetics are important.The emission intensity increased, as expected, with increasing concentrations of (3) when treated with purified luciferase under standard conditions (Figure 2a).No emission was observed for the pure enzyme in the absence of D-luciferin (1) (control 1) as well as for the pure substrate in the absence of the enzyme (control 2).The burst-kinetics profile of (3) (Figure 2b) was like that reported for both (1) and (2).A rapidinjection experiment was performed in which the light output for the reaction over time was recorded.As with all known luciferins, (3) gave a robust initial burst of light followed by sustained light output of much lower intensity (Figure 2b).This trend is consistent with that previously reported for (1) and ( 2) where rapid decay in emission intensity post-burst corresponds to product inhibition. 21The lower emission intensity of (3) relative to the natural substrate (1) should, however, not be a deterrent for its applications in bioluminescence imaging.Such applications rely purely on light generated from the enzyme-substrate reaction and, as a result, generally have good sensitivity.Notably, (3) displayed >100-fold emission over the background.
The relative luminescence-emission intensity of natural (1) (Figure 2c) was, however, 100-fold greater than both (2) and (3) when treated with purified firefly luciferase as compared with the corresponding negative controls (substrates in the absence of luciferase) (Figure 2d).As reported for (2), (3) was also found to have a 100-fold less intense emission signal when compared to (1).The reduction in light output could be due to substrate (3) combined with a luciferase light-emitting reaction having a lower quantum yield or because of differences in the rate of oxyluciferin production.
To lay the groundwork for biothiol-specific biosensing applications, the thioacrylate sulphide (IX) and the D-thioluciferin homodisulphide (3') (Figure 2e), prepared from an iodine oxidation, were also evaluated for the bioluminescent reaction.Neither produced a bioluminescent signal comparable to D-thioluciferin (3).
Importantly, for the purpose of thiol sensing, the luminescence output for (3) was 90-fold greater than its S-protected-thioacrylate (IX), and 2.5-fold greater than the D-thioluciferin homodisulphide (3'), when treated with luciferase under physiological conditions (Figure 2d).It was demonstrated that neither pure luciferase nor pure D-thioluciferin thioacrylate (IX, control 1) emitted light.It was also demonstrated that, when a 0.1 µM thioacrylate solution was treated with luciferase in enzyme buffer, the luminescence output remained negligible.This reinforces that the thioacrylate (IX) of ( 3) is, indeed, not a substrate for luciferasemediated bioluminescence, and, perhaps by extension, that all sulphides of (3) are inactive, as is the case with (1) and its 6'-O-alkyl analogues and the previously reported D-luciferin-6'-sulphides. 15 Negative controls (Figure 2d) contained the homodisulphide (3') and (3), respectively.In both cases, in the absence of luciferase, a small degree of luminescence was detected (4% above background).The luminescence increased significantly when (3) and its disulphide were treated with luciferase to a final enzyme concentration of 10 nM.Notably, the homodisulphide (3') treated with luciferase emitted a degree of bioluminescence relative to the corresponding control.This effect can be ascribed to the reduction of the non-bioluminescent D-thioluciferin homodisulphide (3') to the bioluminescence-active-free (3) by the reducing agent in the enzyme buffer, namely DTT.Kinetic data of the D-thioluciferin homodisulphide (3'), however, did not show an increase in bioluminescence over time as one would expect if the free thiol were constantly being formed via disulphide reduction.Instead, the degree of bioluminescence was observed to be decreasing over time, and the rate of decrease in bioluminescence was comparable to that of (3).This could indicate that the reduction with DTT had occurred relatively quickly, generating a fixed amount of (3) which was not replenished via further disulphide reduction.As a result, the free thiol displayed five-fold greater luminescence than the disulphide, which is indeed a promising result for future redox based sensing applications.
From the kinetic assays, it was also observed that D-thioluciferin (3)'s rate of decay in bioluminescence emission, when treated with luciferase under standard conditions, was reduced when compared to that of Dluciferin (1) and D-aminoluciferin (2).This was a particularly attractive discovery since luciferins are generally known not to have a very stable bioluminescence output and, therefore, require constant re-supply or readministration.These bioluminescent properties are a good starting point for D-thioluciferin (3)-based bioluminescence imaging, despite the lower bioluminescence output relative to D-luciferin (1).
It has been reported that size and hydrophobicity at the C-6 position influence the quantum yield of cyclic amino-luciferins. 22Other factors include pH and the microenvironment in the enzyme active site.Substitution of the 6'-oxygen in D-luciferin (1) with a nitrogen or sulphur resulted in a weakening of bioluminescent intensity, a phenomenon that is not fully understood yet, but might involve bivalent metal ions, which are a cofactor in enzymatic reactions of firefly bioluminescence.
Interestingly, the 6′-methylthio-luciferin reported by Miller et al. 15 proved not to be a substrate for luciferase, whereas 6'-N-alkylated-aminoluciferins were better in vivo substrates for bioluminescence experiments than (2) itself. 23The absorbance of the 6′-methylthio-luciferin is slightly red-shifted compared to (1), which is opposite to the trend observed with (3).The fluorescence of the S-methyl-thioluciferin is blueshifted ca.40 nm, and there was a reduction of the fluorescence quantum yield.In addition to its bioluminescent emission, however, (3) was found to have a strong fluorescence emission, while its protected thioacrylate (IX) was only weakly fluorescent, after excitation at a range of wavelengths (360-520 nm) (Figure S1-S5).Thus, the latter molecules provide further opportunities for imaging applications, the most obvious of which relate to sulphide deprotection and disulphide reduction.Since it was recently reported that 6'-sulphides could be potential inhibitors for the WT luciferase enzyme, the thioacrylate-protected D-thioluciferin (IX) was further evaluated as an inhibitor of luciferase where it was shown to be strongly inhibitory (Figure S6).The sulphide's inhibition of luciferase could similarly be used to inform the design of D-thioluciferin (3)-based probes.

Kinetics
Using a plot of initial rates, the apparent Km of D-thioluciferin (3) was calculated using the Km's for D-luciferin (1) and D-aminoluciferin (2) as references.The apparent Km was then calculated as 0.09801 µM, which was on the same order as that previously calculated for (2) (0.39-0.69 µM) and related analogues (Figure S7), and consistent with the 0.16 µM reported by Pirrung et al. 15,21 The Km was, surprisingly, much lower than that of the native substrate, (1) (8.3 µM), despite the lower emission intensity at the same concentration (Table S8). 21- 22The latter result, along with the fact that the substrate (3) combined with the luciferase light-emitting reaction has a lower quantum yield compared to (1), may shed some light on the bioluminescence activity of (3).

Conclusions
The synthesis of D-thioluciferin and the S-protected-thioacrylate have now paved the way for the development of novel biothiol-relevant applications.The kinetics of D-thioluciferin release, in the case of Sprotected-thioacrylate, is expected to be more favourable compared to thiophenol and needs to be evaluated further under physiological conditions.The lower Km and longer, red shift of λmax relative to D-luciferin and Daminoluciferin make D-thioluciferin a promising bioluminogenic candidate whose properties and applications should be further explored.Moreover, thioluciferin provides a unique handle that readily allows for bioluminescence to be coupled to biologically relevant sulphur chemistry.By exploiting the difference in bioluminescent activity of thioluciferin and its oxidised forms, e.g., disulphide, one can envisage several potential applications which should be investigated further.The S-protected-thioacrylate can be utilized as a general thiol sensor.Furthermore, the established synthetic methodology for unsymmetrical disulfides would allow for the synthesis of disulfide reductase substrates that would release the bioluminescent D-thioluciferin molecule upon enzyme cleavage.This may then lead to the development of new bioluminescent sensors based on the D-thioluciferin molecule.

Experimental Section
General.All reactions were carried out in oven-dried glassware under an inert nitrogen atmosphere, unless otherwise stated.Reagents were obtained from commercial sources (Sigma-Aldrich, Merck) and used as received unless otherwise stated.Solvents were evaporated under reduced pressure at 40 °C using a Buchi Rotavapor, unless otherwise stated.Aqueous solutions were prepared using distilled water.All reactions were monitored by TLC using aluminum-backed Merck silica-gel 60 F254 plates, and compounds were visualised on TLC under a UV-lamp and/or sprayed with a 2.5% solution of p-anisaldehyde in a mixture of sulfuric acid and ethanol (1:10 v/v), iodine vapour or ceric ammonium sulphate solution, and then heated using a 1600 W heat gun.Normal-phase column chromatography was carried out using silica-gel (Fluka Silica Gel 60, 40-63 microns), and compounds eluted with the appropriate solvent mixtures.All compounds were dried under vacuum before yields were determined and spectroscopic analyses performed.Purity was determined by analytical chromatography using an Agilent HPLC 1260 equipped with an Agilent infinity diode array detector (DAD) 1260 UV-Vis detector, with an absorption wavelength range of 210 -640 nm.The compounds were eluted using a mixture of 10 mM NH4OAc/H2O and 10 mM NH4OAc/MeOH at a flow rate of 0.9 mL.min−1 (10% NH4OAc/MeOH between 0 and 1 min, 10 -95% NH4OAc/MeOH between 1 and 3 min, 95% NH4OAc/MeOH between 3 and 5 min).

Figure 2 .
Figure 2. a) Graph of luciferase luminescence at a final enzyme concentration of 10 nM, at varying D thioluciferin concentrations 1 min post-enzyme addition (control 1 is the emission recorded for the enzyme solution in the absence of substrate (1) and control 2 is the recorded emission for substrate (2) in the absence of the enzyme).b) Burst kinetics profile of purified 10 nM luciferase treated with 100 µM D thioluciferin.c) Relative luminescence emission intensity of the core luciferins (6-hydroxyl, 6-amino and 6-thiol at 0.1 µM substrate concentration and at a final luciferase concentration of 10 nM.Control 3 is the recorded emission for substrate (3) in the absence of the enzyme.d) Luminescence output of 0.1 µM of protected D-thioluciferin thioacrylate (IX)(sulphide), D-thioluciferin homodisulphide (3')(disulphide), and free D-thioluciferin (3)(thiol) at a final luciferase concentration of 10 nM (Control readings were recorded for substrates in the absence of the luc enzyme).The Relative Light Units (RLUs) were determined in triplicate and are represented as the mean ± SEM. e) Thiol-sensitive thioacrylate protected D-thioluciferin (IX) probe and redox reaction of D-thioluciferin (3).