Latent luciferase activity in the fruit fly revealed by a synthetic luciferin

Significance

Few organisms are capable of glowing in the dark. Perhaps the best known example is the firefly, a beetle that can emit flashes of yellow-green light when the small molecule d-luciferin is oxidized by the enzyme firefly luciferase. Here we show that a homologous enzyme in fruit flies is also capable of bioluminescence, but only when treated with a rigid synthetic analog of d-luciferin. The discovery of hidden luciferase activity in a nonluminescent insect suggests that the inherent chemistry required for bioluminescence is more common than previously thought, and that the evolution of new enzymatic activities can be directed by the appearance of a new substrate even in the absence of mutation.

Abstract

Beetle luciferases are thought to have evolved from fatty acyl-CoA synthetases present in all insects. Both classes of enzymes activate fatty acids with ATP to form acyl-adenylate intermediates, but only luciferases can activate and oxidize d-luciferin to emit light. Here we show that the Drosophila fatty acyl-CoA synthetase CG6178, which cannot use d-luciferin as a substrate, is able to catalyze light emission from the synthetic luciferin analog CycLuc2. Bioluminescence can be detected from the purified protein, live Drosophila Schneider 2 cells, and from mammalian cells transfected with CG6178. Thus, the nonluminescent fruit fly possesses an inherent capacity for bioluminescence that is only revealed upon treatment with a xenobiotic molecule. This result expands the scope of bioluminescence and demonstrates that the introduction of a new substrate can unmask latent enzymatic activity that differs significantly from an enzyme’s normal function without requiring mutation.

Bioluminescence in insects is almost exclusively confined to a small subset of beetles, including click beetles (1), railroad worm beetle larvae (2), and perhaps the best known example, the firefly Photinus pyralis (3). However, all insects express long-chain fatty acyl-CoA synthetases (ACSLs) that share high homology to beetle luciferases and are hypothesized to be their evolutionary antecedents (4–6). These two classes of enzymes are both members of the adenylate-forming superfamily (7) and share the ability to make AMP esters of fatty acids as well as the ability to displace the AMP ester with CoASH (Fig. 1) (8). Beetle luciferases differ from other insect ACSLs in their ability to chemically generate light by adenylating and oxidizing d-luciferin, a small molecule naturally found in bioluminescent beetles. How this additional activity developed is unknown, although weak bioluminescence has been reported by treating a beetle ACSL with d-luciferin (6, 9).

Fig. 1.

Firefly luciferase and long-chain fatty acyl-CoA synthetases catalyze similar two-step mechanisms. (A) Firefly luciferase catalyzes the formation of an activated AMP ester of its native substrate, d-luciferin. Subsequent oxidation within the luciferase binding pocket generates an excited-state oxyluciferin molecule that is responsible for light emission. (B) Long-chain fatty acyl-CoA synthetases catalyze the formation of activated AMP esters from long-chain fatty acids such as arachidonic acid. AMP is then displaced by CoASH to form the fatty acyl-CoA product.

We have previously found that mutation of firefly luciferase can improve light emission from synthetic luciferin substrates while concurrently reducing light emission from d-luciferin (10). This suggested that the requirements for d-luciferin bioluminescence and for bioluminescence with synthetic luciferin substrates are not the same in mutant luciferases and, by extension, in luciferase homologs. Consequently, we reasoned that even though insect ACSLs from nonbioluminescent organisms outside the order of beetles fail to emit light with d-luciferin, this does not necessarily mean that they are incapable of luciferase activity. Clearly, the catalytic machinery needed to form AMP esters from carboxylic acids is present in ACSLs (Fig. 1), potentially allowing access to an adenylate of a luciferin analog. Furthermore, oxygen has ready access to ligand binding sites in proteins (11), and luciferin active esters in basic DMSO are known to be readily oxidized to generate an excited-state molecule that can emit light (12, 13). We therefore speculated that ACSLs lack luciferase activity with d-luciferin because d-luciferin is a poor ligand for ACSLs and/or binds in a geometry that is not conducive to adenylation. If this hypothesis were true, treatment with a suitable synthetic luciferin substrate that possessed higher affinity and/or conformational rigidity could potentially reveal latent luciferase activity in an ACSL.

Results

To test the idea that ACSLs could have latent luciferase activity, we turned to the Drosophila fatty acyl-CoA synthetase CG6178 (14) (Fig. S1). The fruit fly Drosophila melanogaster is a widely used insect model organism from the order Diptera. With the exception of fungus gnats from the Mycetophilidae family (15), no members of this order of insects are bioluminescent. Furthermore, none of the Diptera expresses a beetle-like luciferase, and CG6178 has been shown to lack luciferase activity with d-luciferin (14). We therefore incubated purified CG6178 protein with a panel of synthetic luciferins that we previously designed to emit red light with firefly luciferase (Fig. 2) (10, 16). Strikingly, the rigid luciferin substrate CycLuc2 revealed latent luciferase activity in CG6178. The peak emission wavelength is in the red (610 nm), nearly identical to that of CycLuc2 with firefly luciferase, and consistent with the predicted effect of the luciferin structure on its photophysical properties (Fig. S2) (16). By contrast, no light emission was observed after treatment of CG6178 with d-luciferin or 6′-aminoluciferin. CycLuc1—differing by a single methyl group from CycLuc2—is a much weaker light emitter with CG6178, as is the dialkylated but less rigid substrate 6′-Me₂NLH₂. Only the d-enantiomer of CycLuc2 results in bioluminescence (Fig. S3), consistent with the stereoselective oxidation of d-luciferin observed with firefly luciferase (17), where oxygen has access to only one side of the binding pocket (3, 18). The addition of CoASH, frequently used as an additive in luciferase assays (3), significantly reduces the light emission observed from CG6178 (Fig. S3).

Fig. 2.

CG6178 is a latent luciferase when treated with the synthetic luciferin CycLuc2. (A) Chemical structures of d-luciferin and all synthetic luciferins. (B) Photon flux from CG6178 (20 nM) treated with the indicated substrate (125 μM). The assay was performed in triplicate and is represented on a log scale as the mean ± SEM and compared by t test to a no substrate control. ns, not statistically significant; ***P < 0.001, ****P < 0.0001.

To emit light, the luciferin substrate must be converted to an active ester and then oxidized to the excited-state oxyluciferin. Using radiolabeled ATP, Oba et al. (14) found that CG6178 can adenylate fatty acids but fails to form the adenylate of d-luciferin. This could reflect a lack of binding by d-luciferin, or the inability of CG6178 to catalyze the formation of the respective AMP ester. To clarify the basis for this defect, we measured light emission from CycLuc2 in the presence of d-luciferin. We found that d-luciferin is a competitive inhibitor of CycLuc2-mediated light emission with a K_i value of 25.7 ± 4.5 μM (Table S1). Surprisingly, the K_m for CycLuc2 with CG6178 is 13.8 ± 1.9 μM, similar to that of d-luciferin, and much higher than the submicromolar K_m of CycLuc2 with firefly luciferase (10). Thus, CG6178 binds both d-luciferin and CycLuc2 with similar midmicromolar affinity but is unable to subsequently adenylate d-luciferin to form LH₂-AMP. Consistent with the role of CG6178 as a long-chain fatty acyl-CoA synthetase, the long-chain fatty acids palmitic acid, oleic acid, and linolenic acid were all competitive inhibitors of light emission with K_i values of 2–4 μM (Table S1). The medium-chain caprylic (octanoic) acid was a weaker inhibitor (13 μM), and the short-chain acetic acid had a calculated K_i value of >1 mM (Table S1).

Light emission from firefly luciferase in vitro is typically characterized by a “burst” phase, where a high initial rate of photon emission is achieved in the first few seconds, followed by a reduction in the rate of photon flux in a subsequent “glow” phase of much longer duration (3). The basis for this behavior has not been fully elucidated, but it is generally thought to be due to rapid formation of the excited-state oxyluciferin, followed by slow dissociation of the products after initial photon emission (3). CG6178 lacks this characteristic burst phase but proceeds directly to the glow phase (Fig. 3), suggesting that the maximal rate of light emission is slower than the dissociation rate of the products and, unsurprisingly, slower than firefly luciferase. Light emission is dependent on the presence of ATP, because only a very weak signal is observed without it, probably owing to residual levels in the protein prep of CG6178 (Fig. S4). Total integrated light output over 2 min for 100 μM CycLuc2 with CG6178 was 0.11% of 100 μM d-luciferin with firefly luciferase, and 2.5% of 100 μM CycLuc2 with firefly luciferase. When integrating the signal emitted during the second minute (e.g., after the luciferase burst), the relative emission from CycLuc2 with CG6178 increased to 0.14% of firefly luciferase with d-luciferin and only 6.7-fold less than luciferase with CycLuc2 (Fig. 3).

Fig. 3.

Burst kinetics profiles of d-luciferin, CycLuc2, and their respective adenylates. Purified FLuc, CG6178, or enzyme buffer with or without BSA was rapidly injected into (A) LH₂, LH₂-AMP or (B) CycLuc2, CycLuc2-AMP (100 μM final). Light emission was recorded every 0.2 s for 1 s preinjection and 120 s postinjection.

We next synthesized the adenylates of d-luciferin and CycLuc2 (LH₂-AMP and CycLuc2-AMP) to determine the effect of bypassing adenylation on the rate of photon flux from CG6178. Treatment of CG6178 with CycLuc2-AMP led to rapid and robust light emission at a rate that was 4.5-fold higher than CycLuc2 alone (Fig. 3). This suggests that adenylation of CycLuc2 is the rate-limiting step in the formation of the excited-state oxyluciferin (19). Consistent with this observation, bypassing the prohibitive adenylation step for d-luciferin by supplying LH₂-AMP also allowed measurable light emission, albeit at a slower rate than CycLuc2. Both adenylates also displayed weak but measurable background bioluminescence in the presence of a high concentration (7,500 nM) of BSA (20), but CycLuc2-AMP emission increased 2,100-fold upon the addition of catalytic quantities of CG6178 (40 nM), whereas LH₂-AMP emission increased 40-fold (Fig. 3).

In principle, the presence of a latent luciferase in fruit flies means that these insects could be rendered bioluminescent if treated with CycLuc2. However, we were unable to detect bioluminescence from fruit flies fed food containing 100 μM CycLuc2. We therefore asked whether bioluminescence could be detected from the endogenous expression of CG6178 in Drosophila Schneider 2 (S2) cells (21), where compound access and cell number can both be readily controlled. Both live and lysed Drosophila S2 cells do, in fact, elicit a bioluminescent glow when treated with CycLuc2 (Fig. 4 and Fig. S5). Photon flux was linear with S2 cell number down to a detection limit of 5,000 cells at an average of 0.3 photons per second per cell (Fig. S6). No photon flux over background was observed when S2 cells were treated with d-luciferin (Fig. 4).

Fig. 4.

CG6178 bioluminescence is detected in both live S2 cells and live transfected CHO cells. (A) Plate image of 3.0 × 10⁵ live S2 and CHO cells treated with (1) no substrate or with 100 μM of (2) d-luciferin, (3) CycLuc2, (4) 6′-NH₂LH₂, (5) 6′-MeNHLH₂, (6) 6′-Me₂NLH₂, or (7) CycLuc1. (B) Quantified flux from live S2 and CHO cells. (C) Plate images of ∼8.0 × 10³ live CHO cells transiently transfected with FLuc or CG6178 after treatment as above. (D) Quantified flux from transfected CHO cells. All assays were performed in triplicate, represented as the mean ± SEM, and compared by t test. ns, not statistically significant; ***P < 0.001,****P < 0.0001.

Mammalian CHO cells did not emit light after treatment with any of the tested luciferins. However, transfection of CHO cells with CG6178 rendered them highly bioluminescent in the presence of CycLuc2. Overexpression from a CMV promoter rather than the endogenous Drosophila dHNF4 promoter (22) yielded bioluminescence of much greater intensity (∼800 photons per second per cell; Fig. 4 and Fig. S5). The luciferase activity exhibited the same selectivity for CycLuc2 as purified CG6178 and live S2 cells, and treatment of transfected CHO cells with d-luciferin did not result in light emission (Fig. 4). Photon flux from live CG6178-transfected CHO cells treated with 100 μM CycLuc2 was 63% of that from firefly luciferase-transfected CHO cells treated with 100 μM d-luciferin (Fig. 4).

Discussion

It has long been surmised that beetle luciferases evolved from ACSLs. Here we have shown that an ACSL from a nonbioluminescent insect species outside the order of beetles is in fact a latent luciferase, capable of catalyzing light emission from a small molecule substrate. Bioluminescence can be selectively detected with the synthetic luciferin CycLuc2 in vitro and in live Drosophila S2 cells and CG6178-transfected mammalian CHO cells. d-luciferin is an inhibitor of CG6178 but is not a substrate for CG6178. Interestingly, both CG6178 and firefly luciferase are competitively inhibited by many medium- and long-chain fatty acids (C8–C20), but not all of these are good substrates for adenylation (e.g., palmitate) (23, 24). This suggests that subtle conformational differences can affect whether or not the carboxylate is able to react with ATP in the binding pocket to form the AMP ester. We therefore postulate that the unique rigid and asymmetric ring structure of CycLuc2 acts as a handle to help properly align the substrate within the CG6178 pocket, allowing adenylation to occur where it fails with d-luciferin.

According to our model, CG6178 possesses latent luciferase activity because its natural substrate promiscuity (24–27) allows the adenylation of CycLuc2, thereby enabling the intrinsic chemistry accessible to this adenylated intermediate to proceed within the protected confines of the enzyme pocket. Unlike fatty acids, CycLuc2 possesses a chromophore, and once activated to CycLuc2-AMP it can be oxidized to emit light. Although there is no known role for oxygen in the natural catalytic function of ACSLs, oxygen is ubiquitous and has ready access to hydrophobic pockets in proteins (11). Because of the spin-forbidden interaction of triplet oxygen with singlet-state molecules, oxygen is generally unreactive in the absence of an activating cofactor such as a transition metal or flavin, so its presence in protein active sites is usually unnoticed and inconsequential. One exception is the reduction of oxygen to superoxide by carbanions (28–30). Although alternative explanations for the mechanism of firefly luciferase have been offered (18), we propose that formation of the luciferin-AMP ester within the enzyme allows access to a resonance-stabilized carbanion that can reduce molecular dioxygen to superoxide by single-electron transfer and then react by subsequent recombination of the radical pair after spin inversion to form a peroxide (Fig. S7). The formation of a carbanion intermediate may be further facilitated in the enzyme by coordination of the substrate carbonyl oxygen to a conserved lysine residue (K443 in FLuc), which has been shown to be important for the oxidative reaction of LH₂-AMP with firefly luciferase (18, 31) and is found in all beetle luciferases and many ACSLs, including CG6178 (Fig. S1). Access to the luciferyl-AMP chemical intermediate thus opens up new chemical reactivity space that is directed by both the substrate and the enzyme.

The interplay between enzyme-directed substrate activation and the substrate-directed chemistry that ensues has significant implications for evolution and for the design of new enzymatic activities. In this case, a new overall catalytic function—light emission—is revealed simply upon the addition of a xenobiotic substrate. The selectivity of CG6178 for CycLuc2 over d-luciferin could potentially be exploited for the design of new substrate-selective luciferases (10), perhaps by combining features of both beetle luciferases and ACSLs. Furthermore, although we did not observe bioluminescence from the mammalian ACSLs in CHO cells, which have lower homology to firefly luciferase, mammalian ACS enzymes are known to adenylate xenobiotics such as ibuprofen (32, 33). We therefore expect that probing the intersection between the luminogenic chemistry of small-molecule luciferin analogs (16, 34–40) and the activation chemistry of existing adenylating enzymes (33) will reveal that latent luciferase activity is more common than previously thought.

Materials and Methods

General.

Chemicals for synthesis were obtained from Aldrich unless otherwise noted. d-luciferin was obtained from Anaspec and 6′-aminoluciferin was obtained from Marker Gene Technologies, Inc. CycLuc1, CycLuc2, 6′-MeNH-LH₂, and 6′-Me₂NLH₂ were synthesized as previously described (16). Protein concentrations were determined using Coomassie Plus (Thermo Scientific). Immobilized glutathione (Thermo Scientific) was used for GST-tagged protein purification. Unless otherwise stated, all protein purification steps were performed at 4 °C. Data were plotted and analyzed with GraphPad Prism 6.0. High-resolution mass spectral data were recorded on a Waters QTOF Premier spectrometer (University of Massachusetts Medical School Proteomics and Mass Spectrometry Facility). Small molecule absorbance was measured using a Cary 50 Bio UV-Visible spectrophotometer. Burst kinetics assays were performed on a Turner Biosystems 20/20ⁿ luminometer and reported as relative light units (RLU). Unless otherwise noted, all other bioluminescence assays were performed on a Xenogen IVIS-100. Data acquisition and analysis were performed with Living Image software and reported as radiance [photons per second per square centimeter per steradian (p/s/cm²/sr)] or total flux [photons per second (p/s)] for each region of interest corresponding to each well of the 96-well plate. All RP-HPLC was performed on an Agilent 1100 using a C18 column (Waters Atlantis 4.6 × 250 mm) at a flow rate of 1.0 mL/min using solvent A (0.1% formic acid in H₂O) and solvent B (0.1% formic acid in CH₃CN).

Protein Expression and Purification.

The Drosophila protein CG6178 was PCR-amplified from the Drosophila Gene Collection cDNA library (GM05240) (41) and cloned into the BamHI–NotI sites of pGEX6P-1. CG6178 and firefly luciferase were expressed and purified as GST-fusion proteins from the vector pGEX6P-1 as previously described (10). PreScission Protease (GE Healthcare) cleavage of the GST-fusion was used to elute the untagged protein (10).

Substrate Dose–Response Assays with Purified Protein.

Luminescence assays were initiated by adding 30 μL of 40 nM purified enzyme in enzyme buffer [20 mM Tris (pH 7.4), 0.1 mM EDTA, 1 mM (tris-(2-carboxyethyl)phosphine) (TCEP), and 0.8 mg/mL BSA] to 30 μL of 2× substrate in substrate buffer [20 mM Tris (pH 7.6), 0.1 mM EDTA, 8 mM MgSO₄, and 4 mM ATP] in a black 96-well plate (Costar 3915). Imaging was performed 1 min after enzyme addition, with final substrate concentrations ranging from 0.122 to 125 μM using the IVIS-100 as described above.

Bioluminescence Emission Scans.

Each purified enzyme in enzyme buffer was rapidly injected into a cuvette containing substrate in substrate buffer to a final enzyme concentration of 100 nM for luciferase and 1,000 nM for CG6178 and a final substrate concentration of 200 μM. The emission from 400 to 800 nm was recorded in a SPEX FluoroMax-3 fluorimeter with closed excitation slits 10 s after injection. Data are normalized to the peak emission intensity and reported as normalized emission.

Burst Kinetics Assays.

Using a Turner Biosystems 20/20ⁿ luminometer, 40 μL of purified enzyme in injection buffer [25 mM Tris (pH 7.7), 0.125 mM EDTA, 5 mM MgSO₄, 2.5 mM ATP, 0.625 mg/mL BSA, and 0.625 mM TCEP] was rapidly injected into a clear Eppendorf tube containing 10 μL substrate in 10 mM sodium acetate (pH 4.5) to final enzyme concentrations of 0.4 nM for luciferase and 40 nM for CG6178, a final substrate concentration of 100 μM for all substrates, and a final pH of 7.4. Measurements were taken every 0.2 s for 1 s preinjection and 120 s postinjection. Data acquisition was performed with SIS for 2020n v1.9.0 software. Data are reported as RLU integrated for each 0.2 s interval. Because photon flux for luciferase is linear with luciferase concentration over the 0.4–40 nM range, data for luciferase were multiplied by 100 to correct for the concentration difference for comparison with CG6178. To correct for the wavelength sensitivity of the PMT in the 20/20ⁿ, the emission intensities were also measured using the IVIS-100 as described above. Data from the IVIS and from the 20/20ⁿ at the 60-s time point were normalized to the WT + LH₂ value. The correction factor of CycLuc2 was calculated by dividing the normalized IVIS data by the normalized 20/20ⁿ data. All 20/20ⁿ data were then multiplied by this correction factor (underreported by 3.6-fold).

Cell Culture.

Drosophila S2 cells were grown at ambient temperature and were cultured in Schneider’s Drosophila Medium (Gibco) supplemented with 10% FBS and 100 U/mL penicillin/streptomycin (P/S). CHO-K1 cells were grown in a CO₂ incubator at 37 °C with 5% CO₂ and were cultured in F-12K Nutrient Mixture (Gibco) supplemented with 10% FBS and 100 U/mL P/S.

Live Cell Luminescence Assays.

S2 cells were washed with HBSS, scraped from the tissue culture dish, and suspended in Schneider’s medium. CHO cells were washed with HBSS, trypsinized, and suspended in F-12K medium. Both were centrifuged at 25 × g for 10 min to pellet the cells. Each was suspended in HBSS at a concentration of 6,000 cells per microliter and 50 μL per well was plated in 96-well black tissue culture-treated plates (3916; Costar). Luminescence assays were initiated by adding 50 μL of 2× substrate in HBSS at a final concentration of 100 μM. Imaging was performed 1 min after addition of substrate using the IVIS-100 as described above.

Lysed Cell Luminescence Assays.

S2 cells were washed with HBSS, scraped from the tissue culture dish, and suspended in Schneider’s medium. CHO cells were washed with HBSS, trypsinized, and suspended in F-12K medium. Both were centrifuged at 25 × g for 10 min to pellet the cells. Each was suspended in 1× Passive Lysis Buffer (Promega) at a concentration of 6,000 cells per microliter. Luminescence assays were initiated by adding 50 μL of 2× substrate in lysed cell substrate buffer [20 mM Tris (pH 7.4), 0.1 mM EDTA, 8 mM MgSO₄, 4 mM ATP, 1 mg/mL BSA, and 1mM TCEP] to 50 μL of lysate in a black 96-well plate (3915; Costar) with a final substrate concentration of 100 μM. Imaging was performed 1 min after addition of substrate using the IVIS-100 as described above.

Transfections.

CG6178 and firefly luciferase were cloned into the BamHI and NotI sites of pcDNA 3.1 and transfected into CHO-K1 cells for live and lysed cell experiments. Transient transfections were performed using Lipofectamine 2000 on cells plated at 60–75% confluency in 96-well black tissue culture-treated plates (3916; Costar) for live cell assays or six-well plates for lysed cell assays. For live cells, 0.075 mg DNA per well was transfected; for lysed cells, 2.25 mg DNA per well was transfected. Assays were performed in triplicate 24 h after transfection.

Transfected CHO Cell Luminescence Assays.

Transfected CHO cells were washed with HBSS. For live cell imaging, the cells in 96-well plates were incubated with 60 μL of 100 μM substrate in HBSS and imaging was performed 3 min after addition of substrate. Cells grown in six-well plates were first lysed for 20 min at room temperature with Passive Lysis Buffer (1 mL per well). Luminescence assays were initiated by adding 30 μL of lysate to 30 μL of 2× substrate in lysed cell substrate buffer in a black 96-well plate (3915; Costar). Imaging was performed 1 min after addition of lysate, at a final substrate concentration of 100 μM using the IVIS-100 as described above.

d-luciferyl-adenylate Synthesis.

The synthesis of d-LH₂-AMP was similar to a previously described method (42). Under an argon atmosphere, a solution of 100 mg (0.49 mmol) of dicyclohexylcarbodiimide in 0.8 mL dry DMSO was added to a solution of d-LH₂ (4.5 mg, 0.16 mmol) and adenosine-5′-monophosphate (15 mg, 0.043 mmol) in dry DMSO. The reaction mixture was mixed at room temperature for 10 min and 5 mL acetone was added to quench the reaction. The resulting white precipitate was collected by centrifugation, washed twice with 3 mL cold acetone, and extracted into 10 mM sodium acetate, pH 4.5, containing 40 mM sodium chloride (2 × 0.75 mL). d-LH₂-AMP was isolated from the pooled extracts by RP-HPLC (0–5 min, 15% B; 5–40 min, linear gradient to 40% B). d-LH₂-AMP eluted at 14.8 min. The concentration in eluent was determined spectrophotometrically using d-LH₂ as a reference [UV LH₂ (H₂O:CH₃CN, 75:25, vol/vol) λ_max 330 nm (ε = 16,800)] and the product was aliquoted, lyophilized to a solid, and stored at −20 °C. High-resolution MS-electrospray ionization (HRMS-ESI) [M-H]⁻ calculated for C₂₁H₁₉N₇O₉PS₂ was 608.0444; the value found was 608.0424.

CycLuc2-Adenylate Synthesis.

Under an argon atmosphere, a solution of 59 mg (0.0.29 mmol) of dicyclohexylcarbodiimide in 0.47 mL dry DMSO was added to a solution of CycLuc2 (3.0 mg, 0.009 mmol) and adenosine-5′-monophosphate (8.8 mg, 0.025 mmol) in dry DMSO. The reaction mixture was mixed at room temperature for 10 min and 3 mL acetone was added to quench the reaction. The resulting white precipitate was collected by centrifugation, washed twice with 2 mL cold acetone, and extracted into 10 mM sodium acetate, pH 4.5, containing 40 mM sodium chloride (2 × 0.45 mL). CycLuc2-AMP was isolated from the pooled extracts by RP-HPLC (0–5 min, 25% B; 5–40 min, linear gradient to 50% B). CycLuc2-AMP eluted at 11.8 min. The concentration in eluent was determined spectrophotometrically using CycLuc2 as a reference [UV CycLuc2 (H₂O:CH₃CN, 70:30, vol/vol) λ_max 396 nm (ε = 8,600)] and the product was aliquoted and lyophilized to a solid and stored at −20 °C. HRMS-ESI [M-H]⁻ calculated for C₂₄H₂₄N₈O₈PS₂ was 647.0899; the value found was 647.0896.