Incorporation of a FRET Pair within a Phosphonate Diester

Abstract

Cell-cleavable protecting groups are an effective tactic for construction of biological probes because such compounds can improve problems with instability, solubility, and cellular uptake. Incorporation of fluorescent groups in the protecting groups may afford useful probes of cellular functions, especially for payloads containing phosphonates that would be highly charged if not protected, but little is known about the steric or electronic factors that impede release of the payload. In this report we present a strategy for the synthesis of a coumarin fluorophore and a 4-((4-(dimethylamino)phenyl)diazenyl)benzoic acid (DABCYL) ester chromophore incorporated as a FRET pair within a single phosphonate. Such compounds were designed to deliver a BTN3A1 ligand payload to its intracellular receptor. Both final products and some synthetic intermediates were evaluated for their ability to undergo metabolic activation in γδ T cell functional assays, and for their photophysical properties by spectrophotometry. One phosphonate bearing a DABCYL acyloxyester and a novel tyramine-linked coumarin fluorophore exhibited strong, rapid, and potent cellular activity for γδ T cell stimulation and also showed FRET interactions. This strategy demonstrates that bioactivatable phosphonates containing FRET pairs can be utilized to develop probes to monitor cellular uptake of otherwise charged payloads.

Graphical Abstract

1. Introduction

The phosphonate functional group has long been viewed as intriguing for its contribution(s) to biological activity of small molecules. From one standpoint, it can be considered a mimic of the tetrahedral intermediate in carboxylic acid ester hydrolysis,^1–3 while a different perspective suggests that it can serve as a phosphate analogue with greater metabolic stability.⁴ An inherent drawback to the phosphonate group per se is its high negative charge at physiological pH. Thus, unless it is a substrate for an active transport system, it can be difficult for phosphonate anions to cross the cell membrane.⁵ One strategy to address this challenge requires preparation of a protected form, a phosphonate derivative that is sufficiently lipophilic and stable outside the cell to cross the cell membrane and then undergoes biochemical modification once inside the cell to release a charged cargo.^6–8 This strategy has dramatically increased the utility of phosphonate-containing nucleotide analogues, including the important medicines adefovir⁹ and tenofovir,¹⁰ but has been investigated less in other systems.^11–14

The specific phosphonate protecting group used in clinical agents often are determined as a result of trial and error through a somewhat limited range of phosphonate esters or phosphonamides.⁷ More varied structures might be included in a protected form once they have been demonstrated to allow payload release after cellular uptake, but at this time there is limited information available on whether steric or electronic factors impede release of the cargo. In general, fluorescent probes and specifically Förster resonance energy transfer (FRET) probes,¹⁵ are desired to increase assay throughput or sensitivity during drug development.^16–17 However, there is also limited information on incorporation of FRET pairs into phosphonate esters, and whether the geography of the phosphonate impacts energy transfer is unclear. To facilitate development of new probes, we have developed novel phosphonate derivatives that contain a potential FRET pair linked to a single phosphorus via ester bonds that might be enzymatically cleaved.

Our recent studies of γδ T cells have suggested that this system provides an excellent matrix for studies of phosphonate uptake, because charged forms of the phosphonate ligand are required for binding to an intracellular domain of the transmembrane protein BTN3A1.¹¹ Phosphonate derivatives that are not readily hydrolyzed within the cell have minimal activity while those that readily release a charged species stimulate T cell proliferation at nanomolar^{11, 18–19} and even sub-nanomolar^{12–13, 20} concentrations with minimal cytotoxicity to confound results. In this report we disclose the preparation of a new type of phosphonate diester as a potential phosphonate FRET probe. The target phosphonates were designed to carry both a fluorescent moiety and a dark quencher as ester substituents of a single phosphonate, and to release a BTN3A1 ligand payload once inside the cell. If successful, this strategy should trigger an immune response detectable through established bioassays with γδ T cells (Figure 1). Furthermore, studies of the fluorescent properties of the new structures are consistent with intramolecular FRET which spans the central phosphorus.

Figure 1.

Schematic representation of phosphoantigen probe uptake, payload release, BTN3A1 engagement, and the resulting response of γδ T cells (e.g. cytokine release).

2. Results

Ligands that function as BTN3A1 agonists, aka phosphoantigens (pAgs), are a class of small molecule organophosphorus compounds that stimulate proliferation of Vγ9Vδ2 T cells (γδ T cells).^21–23 The most potent natural pAg is (E)-4-hydroxy-3-methyl-but-2-enyl diphosphate (HMBPP, 1, Figure 2) which functions by binding directly to the intracellular domain of the transmembrane protein BTN3A1 and displays an EC₅₀ in our assays of ~0.5 nM.¹¹ The diphosphate HMBPP is the last intermediate in bacterial synthesis of isoprenoids that is not found in human metabolism. It can be converted metabolically in bacteria to isopentenyl diphosphate (IPP, 2) which also functions directly as a pAg but is far less potent. In humans, the diphosphate group is susceptible to hydrolysis which reduces activity. Due to the rapid metabolism of these specific diphosphates, studies quickly turned to the phosphonate analogue C-HMBP (3),²⁴ which is far more stable to metabolism but also less potent (EC₅₀ ~4 μM). Much of the ligand potency can be recovered through use of a protected form to facilitate cell entry, and the first such prodrug was the pivaloyloxymethyl (POM) derivative of C-HMBP (POM₂-C-HMBP, 4, EC₅₀ ~5.4 nM).¹¹ Clinical agents such as zoledronate (5) and risedronate (6) function indirectly as pAgs through inhibition of farnesyl diphosphate synthase, which increases cellular concentrations of IPP. Given this landscape, we chose to study new derivatives of the phosphonate 3 to determine if both a fluorescent moiety and a fluorescence quencher could be incorporated in one phosphorus center and demonstrate FRET interactions while retaining potent bioactivity. The size and hydrophobicity of both the fluorescent group and the quencher would be expected to facilitate cell entry by passive diffusion. Whether a phosphonate bearing these esters would demonstrate the strong biological activity that would indicate effective payload release was much less clear.

Figure 2.

Chemical structures of naturally occurring pAgs that are direct agonists of γδ T cells (HMBPP and IPP),¹¹ synthetic pAgs that are direct agonists of γδ T cells (C-HMBP²⁴ and its prodrug POM₂-C-HMBP¹¹), and indirect agonists of γδ T cells (zoledronate²¹ and risedronate¹¹).

Because it is considered a “dark quencher,” absorbing light over a range of wavelengths and releasing the absorbed energy without light emission,²⁵ we first chose to establish if a 4-((4-(dimethylamino)phenyl)diazenyl)benzoic acid (DABCYL)^26–27 moiety that was linked to the phosphonate via an acyloxymethyl linker could be biochemically cleaved and thus utilized as a component of a metabolically labile protecting group. Although there are examples of acyloxymethyl esters used to protect phosphonates, there are only a few examples of groups based on aryl carboxylic acids, most of those are derivatives of benzoic acid,^28–29 and only one incorporates a larger, bicyclic system.¹⁹ The DABCYL structure itself represents an even more complex scaffold that might not provide a substrate for the esterases that are presumed to cleave acyloxy esters.³⁰ Therefore, it was necessary initially to determine if an appropriate variation of this moiety would undergo metabolic cleavage.

To prepare a chloromethyl ester for eventual reaction with the payload, the known diazocoupling of p-aminobenzoic acid (7) with dimethylaniline (8)³¹ gave compound 9 as a red solid in an acceptable yield (59%). After treatment of this red solid with sodium hydroxide in methanol, a biphasic reaction with chloromethyl chlorosulfate³² in aqueous sodium bicarbonate and methylene chloride, gave the desired chloromethyl ester 10 as a red solid in excellent yield (Scheme 1).

Scheme 1.

Preparation of the chloromethyl ester 10

To prepare the phosphonate scaffold, treatment of cyclopropyl methyl ketone 11 with methylmagnesium bromide was used to obtain a homoallylic bromide, which was elaborated to phosphonate 12 by the Michaelis-Becker reaction (Scheme 2).²⁰ Subsequent allylic oxidation with SeO₂ provided the allylic alcohol 13 as the pure E-isomer.^33–34 Allowing ester 13 to react with the known phosphonate– 1,4-diazabicyclo[2.2. 2]octane (DABCO) salt¹⁹ provided the DABCYL derivative 14.

Scheme 2.

Preparation of the DABCYL ester 14.

Given the centrality of the DABCYL group to our plans for incorporation of a FRET pair within one phosphonate, compound 14 was immediately subjected to bioassays. When dissolved in a solution of pooled human plasma in PBS it was possible to observe biochemical cleavage of the DABCYL derivative 14 by LCMS (Table 1). More importantly, this compound displayed nanomolar potency for stimulation of γδ T cells with an EC₅₀ of 110 nM (Figure 3A). When compound 14 was loaded into K562 leukemia cells it also stimulated γδ T cells (Figure 3B). While the activity of compound 14 is modest compared to our most potent compounds, it compares favorably with other derivatives of this ligand that bear just one readily hydrolyzed group.^{11, 19} These experiments are proof-of-principle that show the DABCYL group does not interfere with cellular payload release or result in significant toxicity.

Table 1.

Plasma stability of assayed compounds.

Compound	Percent remaining at 24 hoursa
14	7.5 +/− 0.7
16	62 +/− 1
17	100 +/− 0
18	49 +/− 4
22	27 +/− 2
23	12 +/− 3

^aData shown is from two independent experiments (n=2).

Figure 3.

DABCYL-containing acyloxyalkyl moiety enables γδ T cell stimulation. (A) expansion of γδ T cells from human peripheral blood mononuclear cells (PBMCs) treated with compound 14. The percentage of cells at day 14 post-stimulation is displayed. NS, not stimulated. HMBPP or compound 4 at 100 nM were used as positive controls (n=3) (B) Response of purified γδ T cells to K562 cells loaded for 4 hours with compound 14. (n=3). Data in panels A and B was analyzed by non-linear regression as described in the Experimental section and relevant EC₅₀ values with 95% confidence intervals are reported in Table 3.

Once the metabolic lability of the DABCYL-derived acyloxy group was established, two structurally distinct coumarin variants were explored for preparation of a mixed aryl/acyloxy phosphonate diester. Coumarin was chosen for to its small molecular size and for emission properties which overlap the absorption of the DABCYL quencher. In the first system, a direct linkage of the coumarin moiety to the phosphorus head group was employed (Scheme 3). After phosphonate 12 was converted to the corresponding chloride through reaction with oxalyl chloride and DMF, treatment with 7-hydroxycoumarin provided the aryl ester 15. Selenium dioxide oxidation enabled formation of the mixed phosphonate ester 16. The diester 16 could be converted selectively to the mono salt 17 upon treatment with sodium iodide through a procedure parallel to one established in the literature.^{11, 13} Alternatively, treatment of the diester 15 with DABCO gave an ammonium salt, and subsequent reaction with the chloromethyl ester 10 gave the mixed ester 18.

Scheme 3.

Synthesis of mixed coumarin phosphonate esters.

Both phosphonates 16 and 17 displayed modest EC₅₀ values of 790 nM and 210 nM respectively. The mixed diester 18, where both ester groups can undergo hydrolysis, was found to be more active than either compound 16 or 17. Compound 18 had potency similar to other effective prodrug forms with an EC₅₀ of 0.50 nM (Figure 4A). The activity of the compounds in ELISA assays mirrored the activity in the proliferation assays (Figure 4B). Unfortunately, compounds 16 and 17 did not display the desired level of fluorescence intensity (not shown), potentially due to the direct bonding to the phosphonate. Furthermore, the liberated coumarin was not highly fluorescent under biological conditions due to pH-dependent quenching.

Figure 4.

Impact of coumarin esters on γδ T cell stimulation. (A) expansion of γδ T cells from human PBMCs treated with compounds 16-18. The percentage of cells at day 14 post-stimulation is displayed. NS, not stimulated. HMBPP and compound 4 at 100 nM were used as positive controls (n=3 independent experiments) (B) Response of purified γδ T cells to K562 cells loaded for 4 hours with compounds 16-18. (n=3). Data in panels A and B was analyzed by non-linear regression as described in the Experimental section and relevant EC₅₀ values with 95% confidence intervals are reported in Table 3.

To explore a different coumarin moiety we designed a tyramine based linker to a coumarin carboxylic acid. The coumarin component was prepared through a facile Knoevenagel condensation of 2,4–dihydroxybenzaldehyde with Meldrum’s acid to produce the desired carboxylic acid 19.³⁵ After this carboxylic acid was heated to reflux in neat thionyl chloride, subsequent reaction of the resulting acid chloride with tyramine provided the coumarin donor moiety 20 (Scheme 4).

Scheme 4.

Synthetic route to the mixed coumarin-DABCYL phosphonate diester 23

We then performed a titration to investigate a possible FRET interaction between the linked coumarin and the DABCYL group.³⁶ While maintaining a constant concentration of the FRET donor 20, increasing amounts of the acceptor 9 were added. We expected that if there were energy transfer occurring between the donor 20 and the quenching acceptor 9, this would follow a Stern-Volmer relationship where there would be a correlation between the increasing concentration of quencher 9 and the decreasing fluorescence intensity of the solution.³⁷ As expected, the titration followed this relationship, providing proof of concept that the novel coumarin 20 and DABCYL 9 behave as a FRET donor and acceptor respectively. We could observe this increase of DABYL concentration in the UV–vis spectrum by the increase of a peak in the 440 nm region while the absorbance at 340 nm, which corresponds to the coumarin, remained relatively unchanged (Figure S1). Importantly, the novel linker was effectively quenched by a 1:1 ratio of donor to acceptor (Figure 5). The quantum yield of compound 20 was calculated to be 0.03 in ethanol relative to the known quantum yield of anthracene of 0.27 under the same conditions.³⁸

Figure 5.

FRET titration emissions at a constant donor concentration. Donor compound 20 at 23 μM in tris buffer was mixed with varying concentrations of acceptor compound 9. Data is representative of 4 independent experiments (n=4).

After an intermolecular FRET interaction between compounds 20 and 9 was observed, incorporation of both groups within a single phosphonate was pursued (Scheme 4). To avoid potential benzylic oxidation of the tyramine derivative 20, the SeO₂ oxidation of the phosphonate ligand was conducted prior to its introduction and the resulting alcohol was protected as the acetate 21 to allow activation of the phosphonate.²⁰ After treatment of the acetate ester 21 with oxalyl chloride and DMF, the intermediate phosphonic acid chloride was treated immediately with the coumarin derivative 20 and triethylamine to obtain the mixed ester 22. Selective deprotection of the methyl ester of compound 22 upon reaction with DABCO was followed by reaction with the chloromethyl ester 10 under standard conditions, to provide the mixed diester 23 in modest yield.

Both the mixed coumarin/methyl ester 22 and the final mixed diester 23 were evaluated for their ability to stimulate γδ T cells. As expected, bioassays of phosphonate 22 did not indicate substantial stimulation of proliferation, and even the mixed diester 23 was slightly less active in its proliferative properties than compound 18. Nevertheless, compound 23 displayed a sub-nanomolar EC₅₀ value of 0.69 nM, consistent with metabolic cleavage of both phosphonate esters including the tyramine linker (Figure 6A). It also stimulated potent interferon γ production (Figure 6B).

Figure 6.

Impact of tyramine linker on γδ T cell stimulation. (A) expansion of γδ T cells from human PBMCs treated with compounds 22-23. The percentage of cells at day 14 post-stimulation is displayed. NS, not stimulated. HMBPP or compound 4 at 100 nM were used as positive controls (n=3) (B) Response of purified γδ T cells to K562 cells loaded for 4 hours with compounds 22-23. (n=3). Data in panels A and B was analyzed by non-linear regression as described in the Experimental section and relevant EC₅₀ values with 95% confidence intervals are reported in Table 3.

The UV-visible and fluorescent properties of the novel phosphonate diesters (22 and 23) and their aromatic components (10 and 20) were studied because specific solvents might perturb the fluorescence intensity. In general, higher emission intensities were observed in polar solvents with the exception of DMSO where significantly decreased emissions were observed (Figure S2, Table S1). We also compared the emissions of phosphonate diesters 22 and 23 in aqueous solutions. The emission of compound 23 was low relative to 22, as desired, suggesting the DABCYL group was effectively quenching the coumarin fluorescence (Figure 7).

Figure 7.

Emission spectra of compounds 22 and 23. Compounds at 23 μM in tris buffer were excited at 340 nm (n=1).

Because compound 23 functions as a bioactive phosphonate FRET probe, we next tested it along with the other compounds to determine the potential cell toxicity due to introduction of the coumaryl and DABCYL groups. To examine cell toxicity, K562 cells were treated with the test compounds and examined after 72 hours of exposure. Little to no direct toxicity was observed at doses of 100 μM or lower (Figure S3, Table 3), with the exception of compounds 14 and 18 which were toxic at 100 μM. These concentrations were well above the effective concentrations which were in the nanomolar to picomolar range. This indicates that the release of coumaryl and DABCYL substituents is relatively non-toxic and supports their cellular use as components of various probes.

Table 3.

Activity of test compounds for stimulation of human γδ T cells and toxicity towards K562 leukemia cells.^a

	Proliferation EC50 μM (95% CI)	ELISA EC50 μM (95% CI)	Toxicity IC50 μM (95% CI)
14	0.099 (0.021 to 0.47)	0.11 (0.011 to 1.1)	36 (11 to 120)
16	0.79 (0.57 to 1.1)	18 (16 to 20)	>100
17	0.21 (0.076 to 0.56)	2.9 (2.3 to 3.6)	>100
18	0.00050 (0.00034 to 0.00074)	0.0075 (0.0056 to 0.010)	27 (17 to 43)
22	5.8 (0.031 to 1100)	>10	>100
23	0.00069 (0.00018 to 0.0026)	0.12 (0.089 to 0.17)	>100

^aData shown is from three independent experiments (n=3).

3. Conclusions

Aryl esters of phosphonates are significantly more resistant to cleavage by nucleophiles than the corresponding methyl esters.⁷ Through selective cleavage of phosphonate methyl esters, it has proven possible to construct mixed phosphonate diesters that include both a fluorescent coumarin and a DABCYL-derived dark quencher. For example, a phosphonate dimethyl ester can be selectively converted to the mono ester, mono acid form and then esterified with 7-hydroxycoumarin. The resulting mixed aryl methyl diester can be selectively cleaved at the methyl group and then allowed to react with the new chloromethyl ester of DABCYL (10) to generate an aryl acyloxymethyl ester (i.e. compound 18). The sub-nanomolar activity of this compound in our T cell bioassays indicates that this complex form does indeed release the phosphonate payload after cellular uptake. A similar synthetic strategy was used to prepare the mixed phosphonate diester 23. In this case, after a tyramine derivative of a fluorescent coumarin was incorporated as one phosphonate ester, nucleophilic cleavage of the methyl group followed by reaction with compound 10 gave the mixed aryl acyloxy ester 23. Again, observation of sub-nanomolar biological activity indicates effective payload release despite the substantial size of these phosphonate substituents.

The biological activity of these new mixed phosphonate diesters was assessed in cellular models of γδ T cell function, including the ability to stimulate proliferation of γδ T cells and the ability of loaded K562 cells to stimulate T cell cytokine production. The results of these assays demonstrate that complex aryl acyloxymethyl phosphonate diesters readily undergo cellular activation, while mixed phosphonate diesters that include a simple methyl ester show little or no activity, presumably because they are more stable to metabolism. These results clearly indicate that the large size of phosphonate esters is not a detriment to payload release, and also reaffirms that simple alkyl esters do not readily undergo metabolic cleavage.

The photophysical properties of various compounds were examined by UV-visible and fluorescence spectroscopy. The coumarin 20 displayed pH dependent fluorescence that followed a Stern-Volmer relationship when titrated with compound 9, a known dark quencher. This encouraged design and synthesis of the novel phosphonate 23, which contains both a fluorophore and a dark quencher within the same phosphonate. The lowered emission of this compound relative to a control without the dark quencher component (i.e. compound 22) serves as a proof of concept that intramolecular FRET can be achieved by proper pairing of a fluorescent coumarin and a dark quencher based on the DABCYL moiety. Further research will be needed to optimize the utility of such probes, for example in construction of a phosphonate form that becomes fluorescent only upon metabolism.

Two of these compounds displayed sub-nanomolar potency for γδ T cell activation, suggesting they were metabolized as expected, but significant fluorescence could not be observed from the metabolites within the live cell. This lack of metabolite fluorescence might be due to the diffusion of the neutral FRET donor out of the cell while the acceptor remains within, or this might be due to photoinduced electron transfer by the liberated tyramine-linked coumarin.³⁹ Nonetheless, we were able to establish biochemical activation of the aryl acyloxy phosphonates, which supports the hypothesis that very complex FRET components can be incorporated in a single phosphonate and still serve as substrates for metabolic activation with minimal cellular toxicity of the byproducts. Further investigation will be needed to determine what structural modification concerning fluorophore choice and placement will be necessary for compounds like these to be utilized to probe delivery of phosphonate payloads.

4. Experimental

4.1. Chemical Synthesis

4.1.1. General Experimental Methods.

Acetonitrile and dichloromethane were distilled from calcium hydride prior to use. Tetrahydrofuran was distilled from sodium metal immediately prior to use. Triethylamine and dimethylformamide were dried over molecular sieves. The sodium iodide was dried overnight in an oven. All other reagents and solvents were purchased from commercial sources and used without further purification. Flame-dried glassware under a positive pressure of nitrogen or argon was utilized for all reactions in non-aqueous solvents and these reactions were conducted with a magnetic stir bar. For TLC analyses, pre-coated silica polyester-backed plates (200 μM thickness, UV254 indicator) were visualized under both shortwave ultraviolet light (254 nm) and by p-anisaldehyde stain (93% absolute ethanol, 3.5 % 18 M sulfuric acid, 1% acetic acid, 2.5% p-anisaldehyde). Silica gel (60Å, 40 – 60 μM) was used for flash column chromatography. The NMR spectra were obtained at 400 MHz for ¹H, 100 MHz for ¹³C, and 161 MHz for ³¹P in CDCl₃ with (CH₃)₄Si (¹H, 0.00 ppm) or (CD₃)₂SO (¹H, 2.45 ppm; ¹³C, 40.5) CDCl₃ (¹H, 7.26; ¹³C, 77.0 ppm), as the internal standards. High-resolution mass spectrometry data were obtained by GC-TOF.

4.1.2. 4-(Dimethylaminoazo)benzene-4-carboxylic acid (9).

To a solution of 4-aminobenzoic acid (5.0 g, 36.0 mmol, 1.0 eq) in 12 M hydrochloric acid (15 mL, 157 mmol, 4.4 eq), 1 M NaNO₂ (36.5 mL, 36.5 mmol) was added dropwise and the solution was stirred at 0 °C for 30 min. A chilled 0.11 M solution of N,N’-dimethylaniline (365 mL, 40.1 mmol, 1.1 eq) was added and the reaction mixture was stirred for 2 h. The resulting precipitate was collected by filtration and recrystallized from 50% EtOH in H₂O to provide compound 9 as a red solid (5.71 g, 59%). Both ¹H and ¹³C NMR data matched reported literature values.³¹

4.1.3. (E)-Chloromethyl 4-((4-(dimethylamino)phenyl)diazenyl)benzoate (10).

Compound 9 (1.0 g, 3.7 mmol, 1.0 eq) was added to a flask with MeOH (20 mL) and NaOH (148 mg, 3.7 mmol, 1.0 eq) at rt and the resulting solution was left to stir for 10 min before removal of solvent via evaporative distillation. To the resulting solid, tetrabutylammonium hydrogen sulfate (TBAH, 0.25 g, 0.77 mmol, 0.21 eq), and sodium bicarbonate (2.6 g, 31.0 mmol, 8.3 eq) were added and the solids were dissolved with CH₂Cl₂ (20 mL) and H₂O (20 mL) to provide a biphasic mixture. After it was cooled to 0 °C, chloromethyl chlorosulfate (0.94 mL, 9.3 mmol, 2.5 eq) was added. The mixture was vigorously stirred for 24 h before extraction with CH₂Cl₂ (3 x 10mL). The deep red extract was dried (Na₂SO₄) and concentrated in vacou, and the resulting solid was purification via recrystallized from warm MeOH to provide compound 10 as a red solid (1.1 g, 92%); ¹H NMR (400 MHz, CDCl₃) δ 8.22 (d, J = 8.7 Hz, 2H), 7.92 (t, 4H), 6.79 (d, J = 9.4 Hz, 2H), 6.01 (s, 2H), 3.15 (s, 6H); ¹³C NMR (100 MHz, CDCl₃) δ 164.8, 156.8, 153.1, 143.7, 131.1 (2C), 128.5, 125.8 (2C), 122.2 (2C), 111.5 (2C), 69.4, 40.3 (2C); (TOF MS EI) m/z calcd for C₁₆H₁₆ClN₃O₂ (M+H)⁺ 318.1009, found 318.1015.

4.1.4. ((((E)-5-Hydroxy-4-methylpent-3-en-1-yl) ((2-oxo-2H-chromen-7-yl)oxy)phosphoryl)oxy)methyl 4-((E)-(4-(dimethylamino)phenyl)diazenyl)benzoate (18).

Compound 15 (1.0 g, 3.1 mmol, 1.0 eq) and DABCO (359 mg, 3.2 mmol, 1.03 eq) were dissolved in acetonitrile (6.2 mL) and the solution was heated at reflux for 13 h. After the reaction mixture was concentrated in vacuo, the resulting residue, compound 10 (984 mg, 3.1 mmol, 1.0 eq) and NaI (474 mg, 3.2 mmol, 1.03 eq) were dissolved in acetonitrile (12 mL) and the solution was heated at reflux for 20 h. The solvent was removed via evaporative distillation and the residue was dissolved in a mixture of CH₂Cl₂ and H₂O. The aqueous layer was washed with CH₂Cl₂ (3x), the combined organic layers were dried (Na₂SO₄) and filtered, and the filtrate was concentrated in vacuo. The resulting residue (84 mg, 0.142 mmol, 1.0 eq) was dissolved in CH₂Cl₂ (10 mL) followed by the addition of SeO₂ (13 mg, 0.136 mmol, 0.8 eq), 4-hydroxybenzoic acid (3.1 mg, 0.023 mmol, 0.2 eq), and ^tBuOOH (0.06 mL, 0.568 mmol, 4.0 eq) and the reaction mixture was stirred vigorously for 21 h. Saturated NaHCO₃ was added and the aqueous layer was washed with CH₂Cl₂ (3x). The combined organic layer was dried (MgSO₄) and filtered, and the filtrate was concentrated. Final purification with flash column chromatography (4% CH₃OH in ether ether) provided compound 18 as a red solid (25 mg, 29%): ¹H NMR (400 MHz, CDCl₃) δ 7.99 (d, J = 8.7 Hz, 2H), 7.94 (d, J = 9.2 Hz, 2H), 7.78 (d, J = 8.4 Hz, 2H), 7.46 (d, J = 9.6 Hz, 1H), 7.26 (d, J = 8.7 Hz, 1H) 7.20 (s, 1H), 7.16 (d, J = 8.3 Hz, 1H), 6.80 (d, J = 9.7 Hz, 2H), 6.24 ( d, J = 9.5 Hz, 1H) 5.98 (dd, J_HP = 17.2 Hz, J = 5.4 Hz, 1H), 5.84 (dd, J_HP = 11.4 Hz, J = 5.6 Hz, 1H), 5.46 (td, J = 7.1, 1.2 Hz, 1H), 4.02 (s, 2H), 3.15 (s, 6H), 2.51 (m, 2H), 2.15 (m, 2H), 1.68 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 164.5, 160.1, 156.6 154.8, 153.1, 152.3 (d, J_PC = 11.3 Hz), 143.7, 142.5, 136.9, 130.9 (2C), 128.9, 128.0, 125.8 (2C), 122.9 (d, J_PC = 15.1 Hz), 122.0 (2C), 117.4 (d, J_PC = 6.3 Hz), 116.1, 115.8, 111.5 (2C), 109.5 (d, J_PC = 6.3 Hz), 81.9 (d, J_PC = 7.3 Hz) 68.2, 40.3 (2C), 28.4 (d, J_PC = 139.8 Hz,), 27.1 20.4 (d, J_PC = 5.9 Hz); ³¹P NMR (160 MHz, CDCl₃) δ 30.4; HRMS (ES+, m/z) calcd. for (M+H)⁺ C₃₁H₃₃N₃O₈P: 606.2005; found: 606.2006

4.1.5. ((((E)-5-Hydroxy-4-methylpent-3-en-1-yl)(methoxy)phosphoryl)oxy)methyl 4-((E)-(4-(dimethylamino)phenyl)diazenyl)benzoate (14).

Compound 13 (124 mg, 0.597 mmol, 1.0 eq) and 1,4-diazabicyclo[2.2. 2]octane (74 mg, 0.658 mmol, 1.0 eq) were dissolved in acetonitrile (10 mL) and the solution was heated at reflux for 24 h. The solution was concentrated in vacuo, to provide a brown residue. To the residue, (E)-chloromethyl 4-((4-(dimethylamino) phenyl) diazenyl) benzoate (0.19 g, 0.597 mmol, 1.0 eq) and NaI (0.89 g, 0.597 mmol, 1.0 eq) were added to the flask, then dissolved in acetonitrile (10 mL) and the solution was heated at reflux for 24 h. The reaction mixture was quenched by addition of brine, the aqueous layer was washed with CH₂Cl₂ (3x), the combined organic fractions were dried (Na_sSO₄), and the filtrate was concentrated in vacuo. Final purification via flash column chromatography (15% acetone in CH₂Cl₂) gave compound 14 as a red solid (63 mg, 22%): ¹H NMR (400 MHz, CDCl₃) δ 8.21 (d, J = 8.7 Hz, 2H), 7.92 (d, J = 9.2 Hz, 2H), 7.90 (d, J = 8.5 Hz, 2H), 6.78 (d, J = 9.4 Hz, 2H), 5.95 (dd, J = 13.7, 2.0 Hz, 2H), 5.37 (td, J = 6.2, 1.2 Hz, 1H), 3.97 (s, 2H), 3.76 (d, J_HP = 11.2 Hz, 3H), 3.13 (s, 6H), 2.36 (m, 2H), 1.91 (m, 2H), 1.62 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 164.9, 156.7, 153.0, 143.8, 136.5, 131.0 (2C), 128.5, 125.7(2C), 123.4 (J_PC= 15.8 Hz), 122.2 (2C), 111.5 (2C), 82.0 (d, J_PC= 6.0 Hz), 68.3, 52.0 (d, J_PC= 7.3 Hz), 40.3 (2C), 25.8 (d, J_P= 140.0 Hz), 20.5 (d, J_PC= 4.9 Hz), 13.6; ³¹P NMR (160 MHz, CDCl₃) δ 33.7; HRMS (ES+, m/z) calcd. for (M+H)⁺ C₂₃H₃₀N₃O₆P: 476.1950; found: 476.1959.

4.1.6. Methyl (2-oxo-2H-chromen-7-yl)(4-methylpent-3-en-1-yl)phosphonate (15).

Dimethyl (4-methylpent-3-en-1-yl) phosphonate (1.5 g, 7.8 mmol, 1.0 eq) was dissolved in CH₂Cl₂ (70 mL) and was cooled to 0 °C before addition of DMF (3 drops) and oxalyl chloride (2.1 mL, 23.4 mmol, 3.0 eq). After 24 h, the volatiles were removed in vacuo and the resulting residue was dissolved in THF (70 mL). After the subsequent addition of 7-hydroxycoumarin (1.26 g, 7.8 mmol, 1.0 eq) and triethylamine (1.1 mL, 7.8 mmol, 1.0 eq), the reaction mixture was left to stir for 1 h at rt. The reaction was quenched by addition of 1.0 M NaOH and the aqueous layer was extracted with ethyl acetate (3x). The combined organic layers were dried (Na₂SO₄) and filtered, and the filtrate was concentrated in vacuo. Final purification via flash chromatography (100% ethyl acetate) provided compound 15 as a brown residue (1.93 g, 92%); ¹H NMR (300 MHz, CDCl₃) δ 7.65 (d, J = 9.6 Hz, 1H), 7.42 (d, J = 9.3 Hz, 1H), 7.16 (s, 1H), 7.13 (d, J = 1.3 Hz, 1H), 6.31 (d, J = 9.6 Hz, 1H), 5.07 (td, J = 8.5, 2.7 Hz, 1H), 3.82 (d, J_PH = 11.0 Hz, 3H), 2.32 (m, 2H), 1.92 (m, 2H), 1.63 (s, 3H), 1.57 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 160.3, 155.0, 153.3 (d, J_PC = 8.2 Hz), 142.9, 133.5 (d, J_PC = 1.4 Hz), 129.0, 122.3 (d, J_PC = 16.7 Hz), 117.1 (d, J_PC = 4.1 Hz), 115.9, 115.5, 108.9 (d, J_PC = 5.0 Hz), 53.0 (d, J_PC = 7.2 Hz), 25.7 (d, J_PC = 137.9 Hz), 25.6, 20.9 (d, J_PC = 4.7 Hz), 17.7; ³¹P NMR (120 MHz, CDCl₃) δ 30.8; HRMS (ES+, m/z) calcd. for (M+H)⁺ C₁₆H₂₀O₅P: 323.1048; found: 323.1041.

4.1.7. (E)-Methyl (2-oxo-2H-chromen-7-yl)(5-hydroxy-4-methylpent-3-en-1-yl)phosphonate (16).

Compound 15 (0.976 g, 3.03 mmol, 1.0 eq) was dissolved in CH₂Cl₂ (16 mL), followed first by the addition of selenium (IV) oxide (0.302 g, 2.72 mmol, 0.8 eq), then by p-hydroxybenzoic acid (0.070 g, 0.51 mmol, 0.2 eq), and finally by t-BuOOH (1.42 mL, 14.5 mmol, 4.0 eq). After 24 h at room temperature the reaction was quenched by addition of a saturated sodium bicarbonate solution, and the aqueous layer was washed with CH₂Cl₂ (3x). The combined organic extracts were dried (MgSO₄) and filtered, and the filtrate was concentrated in vacuo. The resulting residue was dissolved in MeOH (16 mL) and then cooled to 0 °C. Sodium borohydride (275 mg, 7.26 mmol, 2.4 eq) was added in portions and allowed to react for 5 h. The reaction then was quenched by addition of ammonium chloride and extracted with CH₂Cl₂ (3x). The combined organic extracts were dried over MgSO₄ and filtered, and the filtrate was concentrated in vacuo. Final purification via column chromatography (0 to 50% acetone in ethyl ether) gave compound 16 as a brown oil (197 mg, 19%): ¹H NMR (300 MHz, CDCl₃) δ 7.65 (d, J = 9.5 Hz, 1H), 7.43 (d, J = 9.1 Hz, 1H), 7.15 (m, 2H), 6.32 (d, J = 9.5 Hz, 1H), 5.39 (td, J = 7.2, 2.8 Hz, 1H), 3.93 (s, 2H), 3.78 (d, J = 11.7 Hz, 3H), 2.50 (br, 1H), 2.38 (m, 2H), 1.96 (m, 2H), 1.61 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 160.3, 154.8, 153.1 (d, J_PC = 8.2 Hz), 142.8, 136.7, 129.0, 122.6 (d, J_PC = 16.3 Hz), 117.1 (d, J_PC = 4.1 Hz), 115.9, 115.5, 108.8 (d, J_PC = 4.6 Hz), 53.0 (d, J_PC = 7.0 Hz), 25.3 (d, J_PC = 138.9 Hz), 20.4. (d, J_PC = 4.7 Hz), 13.6; ³¹P NMR (120 MHz, CDCl₃) δ 30.6; HRMS (ES+, m/z) calcd. for (M+Na)⁺ C₁₆H₁₉O₆P: 361.0817; found: 361. 0815.

4.1.8. Sodium (E)-2-oxo-2H-chromen-7-yl(5-hydroxy-4-methylpent-3-en-1-yl)phosphonate (17).

Compound 16 (215 mg, 0.634 mmol, 1.0 eq) was dissolved in acetonitrile (1 mL), followed by addition of sodium iodide (93 mg, 0.622 mmol, 0.98 eq) and the solution was heated at reflux for 12 h. After the solvent was removed via evaporative distillation, the resulting oil was dissolved in CH₂Cl₂ and H₂O, and the organic layer washed with H₂O (3x). The aqueous layer was freeze-dried to give compound 17 as a yellow solid (167 mg, 76%): ¹H NMR (400 MHz, D₂O) δ 7.90 (d, J = 9.5 Hz, 1H), 7.54 (d, J = 8.3 Hz, 1H), 7.10 (s, 2H), 6.30 (d, J = 9.5 Hz, 1H), 5.37 (td, J = 8.5, 1.2 Hz, 1H), 3.83 (s, 2H), 2.23 (m, 2H), 1.71 (m, 2H), 1.52 (s, 3H); ¹³C NMR (100 MHz, D₂O) δ 164.4, 155.0 (d, J_PC = 7.6 Hz), 154.2, 145.7, 134.9, 129.6, 126.1 (d, J_PC = 16.2 Hz), 118.2 (d, J_PC = 4.0 Hz), 115.3, 113.3, 108.5 (d, J_PC = 4.4 Hz), 67.5, 26.7 (d, J_PC = 134.9 Hz), 21.2 (d, J_PC = 4.4 Hz), 12.9; ³¹P NMR (120 MHz, D₂O) δ 26.1; HRMS (ES+, m/z) calcd. for (M+Na)⁺ C₁₅H₁₆NaO₆P: 323.0708; found: 323.0699.

4.1.9. N-(4-Hydroxyphenethyl)-7-methoxy-2-oxo-2H-chromene-3-carboxamide (20).

Thionyl chloride (10 mL, 0.92 M) was added to a flask that contained 7-methoxycoumarin-3-carboxylic acid (2.0 g, 9.1 mmol, 1.0 eq) and the neat solution was heated to reflux for 3 h. Afterwards, the thionyl chloride was removed via evaporative distillation and the resulting bright yellow solid was dissolved in warm toluene (150 mL) and placed in an addition funnel. Concurrently in a large reaction vessel, tyramine (1.25 g, 9.1 mmol, 1.0 eq) was dissolved in toluene (350 mL) and the coumarin solution was introduced slowly to the stirring tyramine solution. After full conversion, as determined by TLC, aq. NH₄Cl was added to the reaction mixture, the organic layer was washed with a saturated solution of NH₄Cl (3x) and the organic fractions were concentrated via evaporative distillation. The resulting light brown solid then was fully dissolved in minimal boiling methanol and left to cool until a white precipitate formed. The precipitate was collected by filtration to provide compound 20 (1.09 g, 35%): ¹H NMR (400 MHz, (CD₃)₂SO) δ 9.20 (b, 1H), 8.83 (s, 1H), 8.68 (t, J = 5.5 Hz 1H), 7.91 (d, J = 8.8 Hz, 1H), 7.10 (d, J = 2.3 Hz, 2H), 7.05 (dd, J = 8.6, 2.3 Hz, 2H), 6.69 (d, J = 8.5 Hz, 2H), 3.90 (s, 3H), 3.50 (m, 2H), 2.72 (t, J = 7.2 Hz, 2H); ¹³C NMR (100 MHz (CD₃)₂SO)δ 164.9, 161.7, 161.3, 156.6, 156.2, 148.3, 132.02, 130.0, 129.7 (2 C), 115.7 (2 C), 115.2, 114.1, 112.6, 100.7, 56.7, 41.4, 34.7; HRMS (TOF EI MS) m/z calcd for C₁₉H₁₇O₅N (M+Na)⁺ 362.1004, found 362.0999.

4.1.10. (E)-5-(Methoxy(4-(2-(7-methoxy-2-oxo-2H-chromene-3-carboxamido) ethyl) phenoxy) phosphoryl) -2-methylpent-2-en-1-yl acetate (22).

Compound 21 (260 mg, 0.97 mmol, 1.0 eq) was dissolved in CH₂Cl₂ (2.4 mL, 0.4 M) and the solution was cooled to 0 °C. After 10 min, DMF (4 drops) was added followed by the dropwise addition of oxalyl chloride (0.34 mL, 3.9 mmol, 4.0 eq) and the reaction mixture was left to stir and warm to room temperature. After the volatiles were removed via evaporative distillation, the residue was dissolved in THF (2.4 mL). To the solution, triethylamine (0.11 mL, 0.97 mmol, 1.0 eq) followed by N-(4-hydroxyphenethyl)-7-methoxy-2-oxo-2H-chromene-3-carboxamide (329 mg, 0.97 mmol, 1.0 eq) were added, and the solution was left to stir at rt for 45 min. The volatiles then were removed and the resulting residue was dissolved in 1.0 M NaOH and ethyl acetate, and the aqueous layer was washed with ethyl acetate (3x). The organic layers were combined and dried (Na₂SO₄), and then concentrated. Purification via column chromatography (10-50% acetone in ethyl ether) afforded compound 22 as a white solid (630 mg, 22%): ¹H NMR (400 MHz, CDCl₃) δ 8.83 (s, 2H), 7.58 (d, J = 8.7 Hz, 1H), 7.23 (d, J = 8.5 Hz, 2H), 7.15 (d, J = 8.5 Hz, 2H), 6.94 (dd, J = 11.0, 2.3 Hz, 1H), 6.85 (d, J = 2.3 Hz, 1H), 5.47 (t, J = 6.7 Hz, 1H), 4.45 (s, 2H), 3.91 (s, 3H), 3.79 (d, J = 11.0 Hz, 3H), 3.68, (q, J = 6.8 Hz, 2H), 2.91 (t, J = 6.8 Hz, 2H), 2.43 (m, 2H), 2.07 (s, 3H), 1.95 (m, 2H), 1.67 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 170.9, 164.9, 162.0, 161.8, 156.6, 149.1 (d, J_PC = 8.4 Hz), 148.3, 135.7, 131.7 (d, J_PC = 1.4 Hz), 130.9, 130.1 (2 C), 127.2 (d, J_PC = 17.0 Hz), 120.5 (d, J_PC = 4.1 Hz, 2 C), 114.7, 114.0, 112.4, 100.3, 69.6, 56.0, 52.8 (d, J_PC = 6.6 Hz), 41.2, 35.0, 25.0 (d, J_PC = 139.9 Hz), 21.0, 20.8 (d, J_PC = 4.8 Hz), 13.9; ³¹P NMR (120 MHz, CDCl₃) δ 29.48. HRMS (TOF EI MS) m/z calcd for C₂₈H₃₃NO₉P (M+H)⁺ 558.1893, found 558.1898.

4.1.12. ((((E)-5-Acetoxy-4-methylpent-3-en-1-yl) (4-(2-(7-methoxy-2-oxo-2H-chromene-3-carboxamido) ethyl) phenoxy) phosphoryl)oxy) methyl4-((E)-(4-(dimethylamino) phenyl) diazenyl) benzoate (23).

Compound 22 (820 mg, 1.8 mmol, 1.0 eq) and DABCO (201 mg, 1.8 mmol, 1.0 eq) were dissolved in acetonitrile and the solution was heated at reflux for 12 h. After sodium iodide (270 mg, 1.8 mmol, 1.0 eq) and (E)-chloromethyl 4-((4-(dimethylamino) phenyl)diazenyl) benzoate (571 mg, 1.8 mmol, 1.0 eq) were added to the reaction mixture, it was heated again at reflux for 12 h. The volatiles were removed in vacuo and the residue was dissolved in CHCl₃ and saturated aqueous Na₂S₂O₂. After the aqueous layer was washed with CHCl₃ (3x), the combined organic fractions were washed with brine and dried (MgSO₄) and concentrated. Final purification of the residue via column chromatography on silica (3-8% methanol in ethyl ether) provided compound 23 as a red solid (369 mg, 38%): ¹H NMR (400 MHz, CDCl₃) δ 8.78 – 8.76 (m, 2H), 8.09 (d, J = 8.7 Hz, 2H), 7.90 (d, J = 9.1 Hz, 2H), 7.86 (d, J = 8.7 Hz, 2H), 7.57 (d, J = 8.7, 1H), 7.15 (m, 4H), 6.94 (d, J = 8.6 Hz, 1H), 6.83 (s, 1H), 6.73 (d, J = 9.2 Hz, 2H) 5.99 (dd, J_PH= 14.9 Hz, J = 5.2 Hz, 1H), 5.88 (dd, J_PH = 12.3 Hz, J = 5.2 Hz, 1H), 5.48 (t, J = 5.5 Hz, 1H) 4.43 (s, 2H), 3.90 (s, 3H), 3.59 (m, 2H), 3.07 (s, 6H), 2.81 (d, J = 7.2 Hz, 2H), 2.48 (m, 2H), 2.06 (m, 5H), 1.65 (s, 3H); ¹³C NMR (100 MHz, CDCl₃) δ 170.9, 164.8, 164.7, 161.9, 161.7, 156.6, 156.5, 152.9, 148.4 (d, J_PC = 9.5 Hz), 148.1, 143.7, 136.0, 131.9 (2C), 131.0, 130.9 (2C), 130.0 (2C), 128.41, 126.86 (d, J_PC = 17.6 Hz),125.7 (2C), 120.7 (2C), 114.7, 113.9, 112.4, 111.4 (2C), 100.3, 82.0 (d, J_PC = 5.9 Hz), 69.5, 56.0, 41.1, 40.21 (2C), 34.9, 26.0 (d, J_PC = 140.1 Hz), 21.0, 20.9, 20.6 (d, J_PC = 4.8 Hz), 13.9; ³¹P NMR (120 MHz, CDCL₃) δ 29.12. HRMS (TOF EI MS) m/z for C₄₃H₄₆N₄O₁₁P (M + H)⁺ 825.2901, found 825.2900.

4.2. Biological Assays

4.2.1. Reagents and supplies.

The following reagents and supplies were used for this study. Buffy coat obtained from Research Blood Components (Boston, MA) was the source of the peripheral blood mononuclear cells (PBMCs). K562 cells were from Sigma Aldrich (St. Louis, MO). The FITC-conjugated anti-γδ-TCR (5A6.E91) antibody, pooled human plasma, and BCA assay kit were from Fisher (Waltham, MA). The phycoerythrin-conjugated anti-CD3 (UCHT1) antibody and interferon γ enzyme-linked immunosorbent assay kit were from Biolegend (San Diego, CA). The CellQuanti-Blue Cell Viability Assay Kit was from BioAssay Systems (Hayward, CA). The interleukin 2 and the TCR γ/δ+ T Cell Isolation Kit were from Miltenyi (Bergisch Gladbach, Germany). HMBPP was from Echelon (Salt Lake City, UT). POM₂-C-HMBP was synthesized.¹¹

4.2.2. T cell proliferation.

T cell proliferation in response to the test compounds was assessed by flow cytometry staining of T cell populations following compound exposure. PBMCs were stimulated for 3 days with test compounds at various doses (10-fold serial dilutions with maximum concentration determined in a pilot assay). Cells were cultured for another 11 days after compound removal as described previously.^11,20 Briefly, cells were suspended in 100 μL of FACS buffer (2% BSA in PBS). Cells were co-stained with TCR and CD3 antibodies at 4 °C for 30 min, washed twice, and then fixed in 3% paraformaldehyde. Data were collected using a BD Fortessa and analyzed using FlowJo and GraphPad Prism. Data were analyzed by non-linear regression using the log (agonist) versus response – variable slope (four parameter) model in which the bottom was defined as the value of the non-stimulated controls and the top was defined as the highest value of the two positive controls.

4.2.3. Enzyme-linked immunosorbent assay (ELISA).

The ability of γδ T cells to produce interferon γ in response to cells loaded with test compounds was determined by ELISA as follows. PBMCs were expanded in T cell media and purified by magnetic bead isolation.²⁰ K562 cells were treated with compounds at different concentrations for 240 min, washed twice, then mixed with γδ T cells. Each well contained 200 μL with 12,000 T cells and 4,000 K562 cells. The co-culture was incubated for 20 hours and the concentration of interferon γ in the media was determined by ELISA. In each experiment, compounds were evaluated in comparison to negative controls that contained cells without compound. Data was analyzed with GraphPad Prism. ELISA data were analyzed by non-linear regression using the log (agonist) versus response – variable slope (four parameter) model.

4.2.4. K562 viability.

The impact of test compounds on K562 cell viability in the absence of T cells was determined using the Cell QB assay as previously described.²⁰ Briefly, cells were cultured for 72 hours in the presence of test compounds, exposed to the QB reagent during the final 2 hours, and the color change was detected by spectrophotometry using a plate reader. Proliferation data were analyzed by non-linear regression using the log (inhibitor) versus response – variable slope (four parameter) model in which the top was defined as the value of the untreated control cells and the bottom was defined as 0.

4.2.5. Plasma stability.

The novel compounds were assessed for stability in human plasma. Pooled human plasma was diluted to 50% with phosphate buffered saline at pH 7.5. Test compounds were added at a final concentration of 100 μM in a volume of 250 μL. Samples were incubated for various times, then a 50 μL aliquot was removed and proteins crashed with 200 μL of LCMS grade acetonitrile and vigorous mixing. Debris was pelleted by centrifugation at 10,000 rcf for 2 min.

4.2.5. LCMS.

After extraction from plasma, the amount of remaining test compounds was quantified by LCMS. The plasma extracts were evaluated by LCMS with a Waters Synapt G2-Si in positive mode. For most compounds, a gradient starting at 25% acetonitrile then increasing to 80% acetonitrile over 8 min was used. For compound 23 a gradient starting at 50% acetonitrile then increasing to 95% acetonitrile over 8 min was used. Masses corresponding to the molecular ion [M+H]⁺, the sodium adduct [M + Na]⁺, and the dehydration product [M - OH]⁺ or deacetylation product [M – OC=OCH₃]⁺ were generally observed, though this data varied by compound and condition. The integrated peak values for all three ions when applicable were summed and compared to those of t = 0 min for each compound.

4.3. Fluorescence experiments

4.3.1. Fluorescence Spectra.

The relevant compounds were analyzed by spectrophotometry to characterize their fluorescent properties. Stock solutions were prepared by dissolving all test solutions in DMSO. A tris(hydroxymethyl)aminomethane (Tris) buffer was used for all experiments. All buffers were brought to the correct pH by the addition of NaOH (aq) or HCl (aq). Eppendorf and Rainin micropipettes were used to transfer reagents and volumetric flasks were used for all dilutions. The cuvette used was from Hellma Analytics (semi-micro cell type 114F-QS with PTFE stopper): light path (10 mm x 4 mm), quartz (200 nm - 2500 nm), volume (1600 μL). Either a Cary UV-Vis NIR Spectrophotometer or the HORIBA Scientific FluroMax-4 (1 nm slit widths, excitation at 340 nm) was used to obtain the spectra. All stock solutions were prepared at 1.7E⁻³ M and serial dilutions with buffer were used to prepare the following stock solutions for all compounds: 6.8E⁻⁵ M, 1.4E⁻⁵ M, 2.7E⁻⁶ M, 5.4E⁻⁷ M, 1.1E⁻⁷ M, 2.2E⁻⁸ M, 4.0E⁻⁹ M, & 8.7E⁻¹⁰ M. In all experiments, the coumarin containing compounds serve as the donors and DABCYL is the acceptor. Decomposition was observed with the aryl acyloxy compounds after 24 h in all serial dilutions, but not in stock solution. All calibration curves were determined by measurements at the following serial dilutions: 6.8E⁻⁵ M, 1.4E⁻⁵ M, 2.7E⁻⁶ M, 5.4E⁻⁷ M, 1.1E⁻⁷ M, 2.2E⁻⁸ M, 4.0E⁻⁹ M, & 8.7E⁻¹⁰ M using both the Cary and the fluorometer instruments. The probes were measured at 23 μM and also compared to the calibration curve data. A relative quantum yield of 0.03 was determined at a concentration of 3 x 10⁻³ M in ethanol by comparison to anthracene according to standard procedures.³⁸

4.3.2. Intermolecular fluorescence quenching experiment.

A spectrophotometric titration experiment was performed to assess the ability of the DABCYL group to effectively quench the coumaryl linker. All experiment runs were maintained at the same donor concentration (23 μM) in the cuvette with the general equation: [Donor] + [Acceptor] + [buffer] = X volume in cuvette (1.5 mL).

0.50 mL + 0.00 mL + 1.00 mL

0.50 mL + 0.10 mL + 0.90 mL

0.50 mL + 0.20 mL + 0.80 mL

0.50 mL + 0.30 mL + 0.70 mL

0.50 mL + 0.40 mL + 0.60 mL

0.50 mL + 0.50mL + 0.50 mL

0.50 mL + 0.70 mL + 0.30 mL

0.50 mL + 0.85 mL + 0.15 mL

0.50 mL + 1.00 mL + 0.00 mL

Each sample was prepared immediately prior to use and stored in a glass vial. For each experiment, both the Cary and fluorometer were used.

Highlights for Incorporation of a FRET Pair within a Phosphonate Diester.

• Phosphonate acyloxymethyl derivatives of the quencher DABCYL have been synthesized.

• Complex aryl acyloxymethyl phosphonates have sub-nanomolar activity as BTN3 agonists.

• A tyramine linker to the phosphonate enables coumarin fluorescence and FRET.