The Synthesis and Characterization of a Clickable-Photoactive NAADP Analog Active in Human Cells

Abstract

Nicotinic acid adenine dinucleotide phosphate (NAADP) is a potent Ca²⁺ mobilizing second messenger which triggers Ca²⁺ release in both sea urchin egg homogenates and in mammalian cells. The NAADP binding protein has not been identified and the regulation of NAADP mediated Ca²⁺ release remains controversial. To address this issue, we have synthesized an NAADP analog in which 3-azido-5-azidomethylbenzoic acid is attached to the amino group of 5-(3-aminopropyl)-NAADP to produce an NAADP analog which is both a photoaffinity label and clickable. This ‘all-in-one-clickable’ NAADP (AIOC-NAADP) elicited Ca²⁺ release when microinjected into cultured human SKBR3 cells at low concentrations. In contrast, it displayed little activity in sea urchin egg homogenates where very high concentrations were required to elicit Ca²⁺ release. In mammalian cell homogenates, incubation with low concentrations of [³²P]AIOC-NAADP followed by irradiation with UV light resulted in labeling 23 kDa protein(s). Competition between [³²P]AIOC-NAADP and increasing concentrations of NAADP demonstrated that the labeling was selective. We show that this label recognizes and selectively photodervatizes the 23 kDa NAADP binding protein(s) in cultured human cells identified in previous studies using [³²P]5-N₃-NAADP.

Graphical Abstract

1. Introduction

Changes in intracellular calcium regulate diverse processes such as metabolism, secretion, proliferation, muscle contraction, learning and memory and cell death [1–5]. Intracellular Ca²⁺ release can be triggered by diffusible second messengers, inositol-1,4,5-trisphosphate (IP3)¹, cyclic ADP-ribose (cADPR) and nicotinic acid adenine dinucleotide phosphate [6, 7] (NAADP, structure 1 in Figure 1), and these changes are amplified by Ca²⁺ itself through the process of calcium-induced calcium release. NAADP is a very close structural analogue of the more familiar pyridine dinucleotide NADP⁺ in which the nicotinamide moiety is replaced by nicotinic acid. This modest change in structure results in profound changes in biological activity, changing the biological activity of NADP from a redox active co-substrate to the second messenger activity of NAADP [8–10].

Figure 1.

Structure of nicotinic acid adenine dinucleotide phosphate (NAADP, 1; R₁ = H; R₂ = -H), the photoaffinity label 5-azido-NAADP (2; R₁ = N₃; R₂ = -H), 5-(3-(t-Boc-amino)-propyl)-NAADP (3; R₁ = CH₂-CH₂-CH₂-NH-COO-C(CH₃)₃; R₂ = H), and 5-azido-8-ethynyl-NAADP (9; R₁ = N₃; R₂ = H-C≡C-).

The Ca²⁺ mobilizing activity of NAADP has been confirmed in many different cell types. NAADP effectively mobilizes Ca²⁺ at very low concentrations (an EC₅₀ of approximately 10–50 nM when measured using sea urchin egg homogenates [7] and ascidian oocytes [11]). Although mammalian cell homogenates do not respond to NAADP, mammalian cells – for example, mouse pancreatic acinar cells – also release Ca²⁺ in response to low concentrations of microinjected NAADP [12]. NAADP has been shown to be involved in the fertilization process in sea urchin eggs [13, 14], in smooth muscle contraction in several tissues [15], in mediating neuronal functions [16], and in regulating insulin secretion from pancreatic β cells [17, 18]. Its involvement in immunity, where NAADP signaling regulates T cell motility and reactivation, has been observed [19–21].

NAADP was first synthesized using the pyridine base-exchange activity of NAD glycohydrolase (CD38) or Aplysia californica ADP-ribosyl cyclase [22, 23]. In this procedure, NADP is treated with the enzyme in the presence of high (millimolar) concentrations of an added pyridine base (Figure 2). The enzyme catalyst cleaves the pyridinium-ribose bond forming an ADP-ribosyl-enzyme intermediate which subsequently reacts with the added pyridine base, resulting in the formation of a pyridine-substituted NADP analog. In the presence of nicotinic acid and at low pH, formation of useful quantities of NAADP is the result. Aplysia californica ADP-ribosyl cyclase is a member of this family of enzymes that catalyze the pyridine base-exchange reaction, is available as a soluble recombinant protein, and is stable to storage and the conditions of the enzymatic reaction [24]. This enzyme-catalyzed base-exchange reaction has allowed us and others to produce NAADP analogs with substitution on the nicotinic acid moiety [25–28].

Figure 2.

Pyridine base exchange reaction catalyzed by NAD glycohydrolase (CD38) or Aplysia californica ADP-ribosyl cyclase using NADP and nicotinic acid or a substituted nicotinic acid as co-substrates.

The intracellular Ca²⁺ stores targeted by NAADP are different from those regulated by IP3 or cADPR [29, 30]. NAADP-sensitive Ca²⁺ stores reside in acidic compartments termed reserve granules (in sea urchin eggs) or lysosome-like and endosome-like acidic vesicles (in mammalian cells). The identity of the ion channel(s) regulated by NAADP remains controversial. The prime candidates are the two pore channel (TPC) family [9, 31], although other targets have been implicated including the type 1 ryanodine receptor (RyR1) [32] and the transient receptor potential mucolipin 1 channel (TRP-ML1) [33].

NAADP-evoked Ca²⁺ release likely occurs through engagement of NAADP binding proteins (NAADP receptor) that associate with these Ca²⁺ permeable ion channels. These NAADP binding proteins were first implicated using [³²P]-5-N₃–NAADP (2)(Figure 1) as a photoaffinity label to determine the number and the molecular weights of NAADP binding proteins in sea urchin egg homogenates [34] and in mammalian cell extracts [35, 36]. Remarkably, the photoprobe selectively labeled relatively low molecular weight proteins differing from TPCs [34] as well as from the other ion channels proposed to be regulated by NAADP [35]. In sea urchin egg extracts, irradiation of low concentrations of [³²P]-5-N₃-NAADP was shown to specifically label proteins with molecular weights of 45, 40 and 30 kDa, whereas in cultured mammalian cells a ~23 kDa protein doublet was specifically labeled. Labeling of NAADP binding proteins was preserved in pancreatic extracts of TPC1 and TPC2 knock-out mice [35], further indicating that the labeled proteins did not represent either TPC or TPC proteolytic fragments. Lin-Moshier et al. further demonstrated that although NAADP binding was enhanced in HEK 293 cells overexpressing recombinant TPC, enhanced binding was disproportionally low when compared to the expected increase in the number of copies of the TPC [35].

Our photoaffinity labeling studies therefore suggested that NAADP was bound to a novel receptor protein, and that this protein likely controlled one or more Ca²⁺ selective ion-channels through protein-protein interaction [35]. The NAADP binding protein has not been identified, and therefore the mechanism of NAADP-evoked Ca²⁺ mobilization remains unclear. The isolation and characterization of the NAADP binding protein will provide invaluable information on understanding the Ca²⁺ signaling mechanism involved.

To achieve this goal, we sought a bi-functional NAADP analog that contained an aryl azide group which, on irradiation, would covalently crosslink adjacent amino acid residues within a binding site through a photochemical reaction. A second functional group was envisioned which could subsequently react through a “click chemistry” reaction with a fluorescent label or an affinity tag for subsequent isolation of the photo-crosslinked receptor protein. Previous studies on the structure activity relation (SAR) of NAADP analogs using sea urchin egg homogenates suggested that a relatively large substituent could be attached to the nicotinic acid 5-position while still retaining agonist activity and high binding affinity [28]. We had previously synthesized and tested 5-(3-(t-Boc-amino)-propyl)-NAADP (3, Figure 1), and found that it was an agonist with an EC₅₀ in the micromolar range. On this basis we proposed to synthesize substituted nicotinic acid 4, exchange it into NAADP using enzyme catalyzed base-exchange, and produce bifunctional “all-in-one-clickable” NAADP (AIOC-NAADP, 5) (Figure 3). AIOC-NAADP (5) is equipped with both a photoactive aryl azide and a “clickable” aliphatic azide. Similar bi-substituted benzyl groups and bi-substituted benzoic acid groups had been synthesized recently, and used as “clickable” photoaffinity labels in the study of receptors and enzymes [37–40]. In this work have characterized AIOC-NAADP, 5, determined its ability to trigger NAADP mediated Ca⁺² release, and showed in cultured mammalian cells that it photolabels the same proteins that were detected using photoaffinity label 2.

Figure 3.

Synthesis of the AIOC-NAADP analog 5 from NADP and bifunctional nicotinic acid analog 4 using recombinant E98G Aplysia californica ADP-ribosyl cyclase catalyzed base exchange reaction. Reagents and conditions: (a) N-hydroxysuccinimide, N,N′-diisopropylcarbodiimide, DIPEA, THF, RT, 16 h, 63.5% (b)(i) 4 N KOH, 4 h, THF-CH₃OH (1:1); (ii) 1 M aq. HCl, 60.1%.

2. Results

2.1. Synthesis of AIOC-NAADP required nicotinic acid 4 and NADP

The strategy for production of the AIOC-NAADP (5) required the synthesis of the pyridine base 4 from 6 and 7 (Figure 3). The synthesis of 6 [39] and 7 [28] have both been described, and our modifications of the procedures are detailed in Supporting Information Figure S-1. Compounds 6 and 7 were coupled using carbodiimide, resulting in the formation of amide 8, which was hydrolyzed yielding pyridine base 4 (Figure 3).

2.2. Enzyme catalyzed base exchange converted NADP and 4to AIOC-NAADP (5)

Synthesis of NAADP analog 5 was initially attempted using the enzyme catalyzed base exchange reaction with Aplysia californica ADP-ribosyl cyclase, the pyridine base 4, and NADP. A 10–20 eq. of pyridine base was reacted with 1 eq. of NADP in aqueous buffer at pH 4 as illustrated in Figure 3 [23]. The pyridine base exchange with the wild type enzyme requires that the pyridine carboxylic acid be protonated and neutral and this in turn requires that the exchange reaction be carried out at a pH of about 4.5. We found that compound 4 was an inefficient substrate for the base-exchange reaction under these conditions, likely because of its low solubility in buffered water at pH 4.5, limiting the concentration that could be achieved and preventing the base-exchange reaction from occurring in high yield. At higher pH (pH ≥ 5), compound 4 was completely soluble. In an investigation by Graeff et al., X-ray crystallography and site directed mutagenesis studies performed on wild type Aplysia californica ADP-ribosyl cyclase had established the presence of an acidic Glu⁹⁸ residue in the active site [41]. At a high pH, electrostatic repulsion between the Glu⁹⁸ carboxylate and the ionized carboxylate of nicotinic acid prevented the binding and the exchange of the nicotinate anion. The base-exchange reaction using a nicotinic acid derivative therefore could only occur at a low pH at which the nicotinic acid was neutral. We altered the reaction conditions and utilized a mutant homolog of Aplysia cyclase, E98G. E98G has one of the acidic residues, Glu⁹⁸, replaced by a neutral glycine and this eliminates the electrostatic repulsion that could be created in the presence of the pyridine base at a higher pH [42]. As a result, the base-exchange reaction was run at a pH of 7, using 18 eq. of compound 4, 1 eq. of NADP, at 37 °C yielding the AIOC-NAADP (5) in an isolated yield of 35 %.

2.3. Spectroscopic properties of AIOC-NAADP (5)

Following anion exchange purification AIOC-NAADP (5) was shown to be a single component by HPLC and its structure was confirmed by ¹H-NMR, ³¹P-NMR and high resolution mass spectrometry (see 4. Experimental Procedures). The electronic absorption spectrum of 5 at neutral pH exhibits a maximum at 253 nm (Figure 4). As compound 5 is irradiated for 1–3 minutes using a hand-held short wave ultraviolet emitting mineral light, the λ_max was observed to shift to a longer wavelength at 257.5 nm and the intensity of the absorption decreases. A long wavelength tail with absorption between 320–380 nm also appears on irradiation (Figure 4). Irradiation for more than 3 minutes did not elicit any further changes in the UV spectra. The observed changes in the absorption spectra support the presence of an arylazide which upon irradiation is predicted to lose N₂ and produce an electrophilic photoproduct that can react with solvent water.

Figure 4.

UV spectrum of AIOC-NAADP (5)(solid black line) and spectra following the irradiation of 5 with short wavelength UV light for 1 min (dotted red line) and 3 min (dashed and dotted blue line). Additional irradiation produced no further changes. All of the spectra were obtained in 20 mM HEPES at pH 5.

2.4. Recognition of AIOC-NAADP (5) by mammalian and sea urchin egg NAADP receptors

The ability of AIOC-NAADP to elicit calcium release was examined in two biological systems. The first system utilized microinjection of intact SKBR3 cells transfected with TPC2 and the fluorescent calcium indicator protein GCaMP6M [43, 44]. These transfected SKBR3 cells are large and adherent cells which can be microinjected. They respond robustly to microinjected NAADP with increased fluorescence without requiring the addition of a Ca²⁺ sensitive fluorescent dye. Figure 5 shows that microinjection of either NAADP (1) or AIOC-NAADP (5) into single SKBR3 cells resulted in a concentration dependent cytoplasmic fluorescent calcium signal, while buffer injection was without effect. Quantitation of the effects of microinjection are depicted in Figures 5 Panel A and B. Plotting the fold-change in the intensity of the Ca²⁺ signal verses the concentration of 1 or 5 in the micropipette showed saturation of the response characterized by an EC₅₀ of 22 nM for NAADP and an EC₅₀ of 173 nM for AIOC-NAADP (Figure 5 Panel D). Microinjection of either NAADP (1) or AIOC-NAADP (5) into SKBR3 cells with endogenous levels of TPC2 expression also mobilized intracellular calcium stores, an effect which could be blocked by depleting lysosomal calcium content using the vacuolar-ATPase inhibitor, bafilomycin A1 (Figure 5 Panels E and F). The bafilomycin results are similar to those obtained in SKBR3 cells overexpressing TPC1 [43]. The magnitude of the peak calcium responses evoked by NAADP or AIOC-NAADP was approximately 5-fold higher in cells transiently expressing TPC2 compared to non-transfected cells.

Figure 5 -. AIOC-NAADP induces calcium mobilization in SKBR3 cells and sea urchin egg homogenates.

SKBR3 cells transiently overexpressing TPC2-mRuby2 were microinjected with varying concentrations NAADP (Panel A) or AIOC-NAADP (Panel B). Concentrations refer to pipette concentrations. Pseudocolored images of GCaMP6M fluorescence intensity in response to microinjection of the indicated compounds is shown in Panel C, white arrowheads indicate injection sites, scale bars represent 10 μm. Panel D shows the concentration dependence curves for NAADP (solid circles) and AIOC-NAADP (open circles). All data n=3, except the 100 nM data which is an n=4. Panel E shows representative GCaMP6M fluorescence traces from SKBR3 cells with endogenous TPC2 expression after microinjection with NAADP (100 nM, top) or AIOC-NAADP (1μM, bottom) after 30 min incubation with DMSO (0.1%, black) or bafilomycin A1 (100 nM, red). Mock injections (intracellular buffer, grey) responses are shown for comparison. Quantification of the cumulative dataset of injections in SKBR3 cells with endogenous TPC2 expression is shown in panel F, all data are n=3, **p<0.01, ***p<0.001 using the student’s t-test. Panel G shows the concentration dependence of NAADP (closed circles) and AIOC-NAADP on calcium release from sea urchin egg homogenates. Panel H shows the ability of AIOC-NAADP to produce desensitization of the ability of 1 μM NAADP to release calcium homogenate (solid circles) and compete with ³²P-NAADP in a competition-binding assay (open circles). Both sets of data were performed in sea urchin egg homogenates and data represent an n=3).

The second system employed sea urchin egg homogenates, an in vitro system with well characterized NAADP responses [45, 46]. The pharmacological properties of AIOC-NAADP (5) were characterized for Ca²⁺ release and for binding. We found that the NAADP analog 5 was poorly recognized by the sea urchin NAADP receptor. Figure 5 Panel G shows that compound 5 did not elicit release of Ca²⁺ from sea urchin egg homogenates at concentrations less than 100 μM, but did release a small amount of calcium at 1 mM (18% of the amount of calcium released by a 10 μM NAADP (1)). A unique feature of the sea urchin egg extract system is a self-inactivation/desensitization process by which NAADP-induced calcium release can by inactivated by subthreshold concentrations of NAADP in a time- and concentration-dependent manner [47, 48]. Pre-incubation of sea urchin egg extract with varying concentrations of AIOC-NAADP (5) for seven minutes followed by addition of 1 μM NAADP produces a concentration-dependent decrease (desensitization) in the ability of 1 μM NAADP to release calcium (Figure 5H, filled circles). The IC₅₀ for AIOC-NAADP (5) desensitization was 13 μM, whereas NAADP (1) produced desensitization with an IC₅₀ of 0.3 nM (data not shown). Figure 5 Panel H, open circles, also shows that AIOC-NAADP (5) competes with [³²P]-NAADP in a competition-binding assay in sea urchin egg extracts with an IC₅₀ of 6 μM. Overall, the data in Figure 5 suggests that AIOC-NAADP is a potent agonist in the SKBR3 system, but not very potent in the sea urchin egg system.

2.5. Specific photolabeling of NAADP binding proteins using [³²P]-5

Next we determined if [³²P]-labeled AIOC-NAADP ([³²P]-5) would label the same intracellular targets that were detected in our previous studies using [³²P]-5-N₃-NAADP (2) [34–36]. We first showed that [³²P]-AIOC-NAADP (5) would specifically label 23–24 kDa protein(s) in SKBR3 cells (Figure 6 Panel A). NAADP mediated signaling has been extensively characterized in both SKBR3 and in Jurkat cells both in terms of calcium mobilization [20, 35, 43, 49–51] and photolabeling of NAADP binding proteins [20, 36]. Jurkat cells were one of the first mammalian systems shown to be responsive to NAADP [49], and subsequently Jurkat cells were shown to mobilize intracellular NAADP in response to activation of the cell surface T cell receptor [49, 52]. This functional response motivated us to characterize the photolabeling of Jurkat cell extracts with 5, even though we did not possess the technology that would enable us to study Jurkat cells by microinjection.

Figure 6. Comparison of [³²P]-AIOC-NAADP and [³²P]-5-N₃-NAADP photolabeling in SKBR3 cell extracts (Panel A), Jurkat cell extracts (Panel B), and sea urchin egg extracts (Panel C).

SKBR3 (A) or Jurkat (B) cell extracts were photolabeled with either [³²P]-AIOC-NAADP (1.2 nM) or [³²P]-5N₃-NAADP (0.5 nM) in the absence or presence of the indicated concentrations of unlabeled NAADP. Panel C: Sea urchin egg homogenates were photoaffinity labeled with [³²P]-AIOC-NAADP (1 nM) or [³²P]-5N₃-NAADP (0.4 nM). The probe used, irradiation conditions and the presence of unlabeled NAADP are indicated.

The results of this photoaffinity labeling study are shown in Figure 6 Panel B. When Jurkat cell extracts were irradiated in the presence of low concentrations (~1 nM) of [³²P]-5-N₃-NAADP or [³²P]-5, we observed labeling of multiple protein bands (Figure 6B). The labeling patterns with both probes were qualitatively similar, however, ([³²P]-5), produced more intense labeling of most of the bands compared to [³²P]-5-N₃-NAADP (2). No photolabeling of the Jurkat extract was observed without irradiation (see Supporting Information Figure S-2). When unlabeled 1 μM NAADP was included, only a 23 kDa protein labeled by both probes was protected from photoderivatization. This indicates that [³²P]-5-N₃-NAADP (2), [³²P]-5 and NAADP (1) compete for a common binding site, and identifies the 23 kDa protein as a specifically labeled protein. A specifically labeled band at 23 kDa was also observed in SKBR3 cell extracts (Figure 6A), the system we used to demonstrate that AIOC-NAADP (5) was active in mammalian systems in Figure 5. This photolabeling data is consistent with our published data using [³²P]-5-N₃-NAADP (2) in SKBR3 and Jurkat cell extracts [35, 36]. The time course of [³²P]-5 labeling is shown in Supporting Information Figure S-2. The labeling of the 23 kDa protein was detected within 15 seconds and increased rapidly during the first 15 minutes of preincubation prior to irradiation. The labeling then plateaued.

Figure 6 Panel C shows a similar photolabeling system comparing [³²P]-5-N₃-NAADP (2) and [³²P]-5 using sea urchin egg (S. purpuratus) extracts. Neither probe labeled any proteins in the absence of UV irradiation. Irradiation in the presence of [³²P]-5-N₃-NAADP (2) produced labeled proteins with molecular weights of 45, 40 and 30 kDa. The photolabeling of these proteins was protected by the inclusion of 1 μM NAADP, data that is consistent with our previously published data [34]. Irradiation in the presence of [³²P]-5 resulted in the labeling of several additional proteins, however, the labeling of these proteins was non-specific as the inclusion of 1 μM NAADP did not protect the labeling of any of these bands.

The selectivity of photolableing of the 23 kDa protein by [³²P]-5 was evaluated by examining the labeling in the absence and presence of increasing concentrations of NAADP (1), NADP and AIOCNAADP (5). Figures 7A and B show the labeling patterns of Jurkat cell extracts labeled with [³²P]-5 in the presence of increasing concentrations of NAADP and NADP (Figure 7A) or AIOC-NAADP (5) (Figure 7B). Figure 7C compares the intensity of the 23 kDa labeled band, as measured by densitometry of the phosphorimage, as concentrations of the competing ligand were varied. NAADP protected the 23 kDa band from photolabeling most potently (IC₅₀ = 20 nM), whereas 40-fold higher concentrations of NADP (IC₅₀ = 795 nM) were required to achieve similar levels of protection. AIOC-NAADP (5) prevented labeling with an IC₅₀ of 100 nM, a value in reasonable agreement with that measured from micoinjection studies.

Figure 7. Specificity of [³²P]-AIOC-NAADP photolabeling in Jurkat cell extracts.

Panel A. Phosphorimage of a gel examining the effect of NAADP and NADP on [³²P]-AIOC-NAADP (0.3 nM) photolabeling. Panel B. Phosphorimage of a gel examining the effect of AIOC-NAADP on [³²P]-AIOC-NAADP photolabeling. Photolabeling was performed with Jurkat cell extracts in the absence or presence of the indicated concentrations of unlabeled NAADP or NADP or AIOC-NAADP. Panel C. Densitometric analyses of experiments similar to those shown in Panels A and B (n=3)

Since our previous photoaffinity labeling studies relied on [³²P]-5-N₃-NAADP (2), we wanted to carefully demonstrate that the labeling of the 23–24 kDa proteins would show cross competition with 5-N₃-NAADP (2) and AIOC-NAADP (5) as predicted for probes which label the same protein through a common binding site. This is demonstrated in Figure 8, where labeling of the 23 kDa proteins with [³²P]-5-N₃-NAADP (2) was protected by both NAADP and AIOC-NAADP. We also note that in Figure 8 an approximately 45 kDa protein was labeled by 5-N₃-NAADP (2) and protected in the presence of low concentrations of NAADP but poorly protected by AIOC-NAADP. It additionally appears that a 25 kDa protein is labeled by [³²P]-5-N₃-NAADP (2) and partially protected by NAADP and AIOC-NAADP, indicating a further diversity of NAADP binding proteins.

Figure 8 -. Displacement of [³²P]-5-N₃-NAADP with AIOC-NAADP.

Panel A: Phosphorimage of Jurkat cell extracts photolabeled with 6 nM [³²P]-5N₃-NAADP in the absence or presence of the indicated concentrations of NAADP or AIOC-NAADP. Panel B: Densitometry analysis of photolabeling of 23 kDa proteins.

Figure 9 shows the dependence of [³²P]-labeled AIOC-NAADP ([³²P]-5) photolabeling on the concentration of [³²P]-labeled AIOC-NAADP. Figure 9 Panel A shows the intensity of labeling of all bands increases as a function of the concentration of [³²P]-labeled AIOC-NAADP. Only the labeling of the 23 kDa band was labeled specifically as inclusion of unlabeled 2.5 μM NAADP prevented the labeling. A 32 kDa band was partially protected by the 2.5 μM NAADP. The labeling of this band was also partially protected by 1 μM NAADP and NADP (see Figure 7A). Figure 9 PanelB shows the saturation of the 23 kDa labeling by [³²P]-labeled AIOC-NAADP ([³²P]-5). Specific photolabeling of the 23 kDa protein was determined by densitometric analyses of the data shown in Figure 9A (specific labeling is the difference between the labeling in the absence (total labeling) and presence of 2.5 μM NAADP (non-specific labeling). Although we were unable to achieve concentrations of [³²P]-labeled AIOC-NAADP that saturated the labeling, the K_d value of AIOC-NAADP (5) estimated from non-linear regression analysis of the specific labeling curve of Figure 9B was 28 nM. This is consistent with the functional data in Figure 5D and photolabeling data in Figure 7C that AIOC-NAADP (5) is a high affinity probe for NAADP binding sites in mammalian systems.

Figure 9. Photolabeling of Jurkat cell extracts with increasing concentrations [³²P]-AIOC-NAADP.

Panel A. Photolabeling of Jurkat cell extracts with increasing concentrations of [³²P]-AIOC-NAADP in the absence (−) and presence (+) of 2.5 μM NAADP. A phosphorimage of the gel is shown. Panel B. Saturation of the [³²P]-AIOC-NAADP photolabeling of the 23 kDa band. Total labeling (closed circles), non-specific labeling (open circles) and specific labeling (closed square/dashed line) of the 23 kDa band were determined by densitometric analyses of the phosphorimage.

2.6. Proteins photolabeled with [³²P]AIOC-NAADP were biotinylated using “click chemistry” and affinity isolated

We next tested whether the proteins photolabeled by the bifunctional probes could be reacted with reagents containing a biotin affinity tag through “click chemistry”. Figure 10 shows the data for probe 5. Jurkat cell extracts were photoaffinity labeled with [³²P]-labeled probe 5. After removal of the free probe, the sample was subjected to click reaction with four different reagents using either copper (I) catalyzed click chemistry or strain promoted click chemistry using dibenzyl cyclooctyne (DBCO) reagents. The biotin reagents contained either PEG or diazo based linkers between the biotin affinity tag and the clickable moieties (alkyne for Cu (I) catalyzed reactions and DBCO for strain promoted click chemistry). The control for these experiments consisted of a Jurkat cell extract that had been photolabeled with [³²P]-labeled probe 5, but not subjected to click chemistry. After the click reaction, the excess reagents were removed by desalting and the extracts were incubated with streptavidin-agarose beads to absorb the biotinylated proteins. The streptavidin-agarose beads were washed 5-times with phosphate buffered saline to remove proteins bound non-specifically to the beads. Proteins specifically bound to the beads at this point were removed by boiling the beads in 2× Laemmli SDS sample buffer for 10 minutes. The eluted proteins were examined by SDS-PAGE and radiolabeled proteins were detected using a phosphorimager. Figure 10 shows that [³²P]-labeled proteins were eluted from each of the samples subjected to click reaction. No labeled proteins were detected in the control sample that had been photolabeled with 5, but not subjected to click reaction (Figure 10, lane 1). Noteworthy, the 23 kDa protein that we have shown to be specifically labeled with probe 5 is recovered from all the samples subjected to click chemistry (Figure 10, lanes 2–5). This demonstrates that probe 5 is “clickable” after photo-incorporation of the probe into the protein. Based on the radioactivity recovered from the streptavidin-agarose beads, we estimated that the click reagents derivatized between 15 and 30% of the labeled proteins depending on which biotin reagent and click reaction was employed. The best results were with a DBCO PEG-linked biotin. What is not apparent from the radiometric detection is that a high background of biotinylated proteins due to the non-specific reaction of the biotinylated-acetylenes was encountered in every case. This non-specific reactivity that accompanies the click reactions had been observed previously [53, 54]. In the present case, the non-specificity must be overcome before identification of labeled proteins by mass spectrometry or Edman degradation will be possible. Optimizing click-reaction conditions are expected to increase the efficiency of the click chemistry reactions and to decrease non-specific background.

Figure 10.

Proteins photolabeled by bifunctional probe 5 can be biotinylated via click chemistry and recovered using streptavidin-agarose. Human Jurkat cell extract was photoaffinity labeled with 0.3 nM [³²P]-AIOC-NAADP (5) and divided into five equal aliquots. One aliquot served as a control, while the other four were subjected to click chemistry with various biotin containing click chemistry reagents. After click chemistry, all aliquots were incubated with streptavidin-agarose for 12 hours and the beads were subsequently washed with PBS, to remove non-specifically bound proteins. Proteins bound at this point to the beads were eluted by boiling the washed beads in 2 × SDS sample buffer for 10 min. The eluted proteins were run on a SDS-PAGE gel which was dried and exposed to a phosphor image screen for 24 hours. The far left lane labeled Control PAL represents the pattern of photolabeling obtained in the original sample (note: this sample was run on a different gel than the samples eluted from the beads). Lane 1: control sample that was not subjected to click chemistry; Lane 2: click chemistry with dibenzylcyclooctyne(DBCO)-diazo-linked biotin; lane 3: click chemistry with DBCO-PEG-linked biotin; lane 4: Cu(I) click chemistry with alkyne-diazo-linked biotin; lane 5: Cu(I) click chemistry with alkyne-PEG-linked biotin. Note that radiolabeled proteins were only recovered from the streptavidin-agarose beads when the photoaffinity labeled sample was subjected to click chemistry. No labeled proteins were recovered in the control sample (lane 1) that was photoaffinity labeled but not subjected to click chemistry with a biotin reagent. In addition, the 23 kDa protein that we have shown to be specifically labeled with probe 5 is recovered from all the samples subjected to click chemistry (lanes 2–5).

3. Discussion

AIOC-NAADP (5) was designed as a bifunctional NAADP analog that would specifically photolabel and subsequently facilitate the pull-down of NAADP binding proteins. Its synthesis represents an extension of our previously published enzymatic methodology for producing NAADP analogs, in that we have shown that the Aplysia californica ADP-ribosyl cyclase E98G mutant can be used for the production of analogs from hydrophobic nicotinic acid bases that would otherwise fail when using wild type enzyme as the catalyst [28].

Analog 5 was shown to be recognized at low concentration by the human NAADP receptor, but in contrast in sea urchin egg homogenates, 5 was found to be a poor low potency agonist, although it produced sub-threshold desensitization in the low micromolar range. Consistent with these differential potencies at evoking Ca²⁺ release in the human and sea urchin system, 5 showed strong, specific labeling of the 23 kDa NAADP binding protein(s) in different immortalized human cell lines, but no specific labelling of the previously identified NAADP binding proteins in sea urchin homogenate. Differences in the structure activity relationship between human and the sea urchin NAADP receptors have been observed previously [55], and these data reinforce the relationship between the specifically photolabeled binding proteins and the NAADP-evoked Ca²⁺ release mechanism in both systems.

The key finding is that [³²P]-5 specifically photo-labeled the 23 kDa protein(s) in cultured human cells that we had previously specifically labeled using [³²P]-5-N₃–NAADP (2) [35, 36]. Labeling with [³²P]-5 was shown to saturate at low concentration and [³²P]-5 was demonstrated to compete with both NAADP and 5-N₃-NAADP for a common binding site. Our photolabeling studies with [³²P]-5 indicate that despite the significant difference in structure of NAADP (1) and 5-N₃–NAADP (2), versus AIOC-NAADP (5), 5 is capable of specifically recognizing 23 kDa NAADP binding proteins.

Last we demonstrated that the proteins which were photolabeled with [³²P]-5 could be “clicked” to biotinylated acetylene containing labels and the proteins so derivatized isolated by biotin-avidin affinity chromatography. In our brief survey DBCO-biotin was shown to be the most effective, but copper-catalyzed click reaction derivatization was also successful.

This pull-down experiment, was designed to obtain “proof-of-concept”, and was therefore performed radiometrically, to provide extremely high sensitivity to detect photo-labeled target proteins even in the presence of non-specific protein background. Indeed, non-specific labeling of proteins was observed, and in the future conditions must be optimized to reduce the background and permit detection of rare targets such as the NAADP binding protein(s).

We have recently described the characterization of another bi-substituted photoaffinity label: 5-azido-8-ethynyl-NAADP (9; Figure 1) [56]. Dinucleotide 9 contains a 5-azide attached to the nicotinic acid ring and an 8-ethyne attached to the purine moiety. Similar to AIOC-NAADP 5, analog 9 is poorly recognized by sea urchin receptors, but it was active at low concentration when microinjected into cultured human U2OS cells. Our new NAADP analog 5 was shown in this work to be approximately 10-fold more potent than was 9. Furthermore, at low concentration [³²P]-AIOC-NAADP (5) consistently produces a more intense labeling, indicative of a higher photoinsertion yield. AIOC-NAADP (5) will be most useful for click reactions with biotinylated and dye conjugated acetylenes or with methyl ester-phosphines using the Staudinger ligation reaction. Bifunctional NAADP analog 9 will be applied under circumstances where use of an acetylinic photoprobe followed by a click reaction with a biotinylated azide is desired as a strategy for non-specific click reactivity [53]. Neither 5 nor 9 are well recognized by the sea urchin NAADP receptor, and development of probes specific for the sea urchin receptor remains a priority.

The synthesis of AIOC-NAADP 5 is more straightforward than is the synthesis of 9, since 5 can be produced directly from enzyme catalyzed base-exchange between NADP and nicotinic acid analog 4. In contrast, the production of 9 begins with the synthesis of an adenosine nucleoside and requires multiple steps for conversion of the nucleoside into dinucleotide 9.

The results of this study enable us in future work to use 5 to photoderivatize NAADP binding proteins and subsequently attach them to biotinylated acetylene labels by “click chemistry” for isolation by biotin affinity chromatography. The efficacy of click reactions has been shown to be highly condition dependent [57, 58], and future work in applying AIOC-NAADP (5) to the identification of NAADP binding proteins will of necessity focus on optimizing the conditions of photoinsertion and for the click reaction to achieve high specificity and low background. Once this is achieved, AIOC-NAADP (5) should serve as an effective tool for the labeling and the isolation of NAADP binding proteins in mammalian cells.

4. Experimental Procedures

4.1. Biological Methods

Plasmids

Plasmid encoding TPC2-mRuby2 was produced by subcloning the TPC2 ORF from previously described plasmid encoding TPC2-myc [59] into pcDNA3-mRuby2 at the NheI restriction site, using standard molecular biology techniques. Plasmids encoding pcDNA3-mRuby2 and GCaMP6M were gifts from Michael Lin and Douglas Kim, respectively (Addgene plasmids, #40260, #40754) [60, 61]

Cell culture and transfection

In this study, Jurkat cell extracts are used for most [³²P]-AIOC-NAADP photoaffinity labeling experiments (Figures 6B, 7, 8, 9, and Supplementary Figure S-2) due to the low amount of non-specific labeling present in this cell line - however, photoaffinity labeling of SKBR3 cell extracts is also reported (Figure 6A). Microinjection of AIOC-NAADP to monitor changes in cytosolic calcium concentrations was conducted in SKBR3 cells as our group is equipped with an injection system that is only suited for adherent cell microinjections, and Jurkat cells are cultured in suspension. Jurkat cells were maintained in RPMI medium, and SKBR3 cells were maintained in McCoy’s 5A medium. Cell culture media was supplemented with 10% fetal bovine serum, 100 units/ml penicillin and streptomycin and 292 μg/ml L-glutamine, and cells were cultured at 5% CO₂ and 37 °C. For transfection, SKBR3 cells were seeded in collagen-coated 35 mm glass-bottom dishes (MatTek) at a density of 2 × 10⁵ cells/ml. The following day, cells were transfected with Lipofectamine 2000 according to the vendors protocol. All cell lines were purchased from ATCC. All cell culture and transfection reagents were purchased from Thermo Fisher.

Microinjection and confocal Ca²⁺ imaging

SKBR3 cells were transfected with plasmids encoding TPC2-RFP and GCaMP6M. Dishes were mounted on an Olympus IX81 microscope for imaging. Cells were perfused with Ca²⁺-free Hank’s Balanced Salt Solution (Thermo Fisher) at a rate of 0.5 ml/min. Isolated cells expressing TPC2-RFP and GCaMP6M were identified by fluorescence. Cell morphology was assessed by acquiring z-stack images and reconstructing three-dimensional models of each cell to be injected. Regions that were not relatively close to either the nucleus or the cell periphery were targeted for injection sites. Femtotip (Eppendorf) injection pipettes were backfilled with intracellular buffer (110 mM KCl, 10 mM NaCl, 20 mM HEPES, pH 7.2) containing either vehicle or drug, and were positioned using an Injectman-4 (Eppendorf) micromanipulation system. Cells were injected at a z-position that was approximately 70% of the cell thickness at the site of injection using a Femtojet4i (Eppendorf). Injection parameters were 85 hPa injection pressure, 40 hPa compensation pressure, 0.5 s injection duration, 45° injection angle, and 600 μm/s injection speed. Cells to be injected were imaged (λ_ex 488nm, λ_em 514±15nm bandpass) using a Plan-Apochromat 60x/1.42 oil differential interference contrast objective, and fluorescence changes were monitored using a Yokogawa spinning disk confocal (CSU-X-M1N), and an Andor iXon Ultra 888 EMCCD camera. Image acquisition and data collection was done using Metamorph version 7.10.

Ca²⁺ release, competition ligand binding and sub-threshold desensitization

Studies using the Strongylocentrotus purpuratus sea urchin egg homogenate system were performed as described previously [28].

4.2. Chemical Methods

4.2.1. General Methods

TLC sheets consisting of silica gel layer (200 μm) with a fluorescent indicator coated on a plastic backing were purchased from Baker-flex (Radnor, PA), and were visualized by examination under a 254 nm UV light or sprayed with a 0.5% of ninhydrin (2,2-dihydroxyindane-1,3-dione) dissolved in ethanol solution.

Preparative normal phase and reverse phase purifications were performed using an automated chromatography system equipped with an in-line UV detector (200 – 360 nm) purchased from Teledyne-Isco (Combiflash R_f system, Lincoln, NE). Pre-packed silica gel columns containing 4 g, 12 g, 40 g, and 80 g of absorbent and reverse phase C-18 columns containing 15.5 g of absorbent were purchased from Teledyne-Isco.

Analytical and large scale anion exchange purifications were performed using a Bio-Rad BioLogic Duoflow HPLC system (Bio-Rad Laboratories, Herculus, CA). Macroporous AG MP-1 anion exchange resin (200 – 400 mesh, Cl^- form) was purchased from Bio-Rad Laboratories and converted into the trifluoroacetate form according to the manufacturer’s instructions. Preparative scale columns used with this resin were packed in an OMNI FIT glass column (1.5 cm × 11.5 cm). A prepacked Bio-Rad Uno-Q anion exchange column (7 × 35 mm) was used for analytical separations. The mobile phase consisted of water and trifluoroacetic acid solution (0 – 0.1 M). The flow rate, gradient and injection volume was controlled by the unit.

Open column anion exchange purification was performed using diethylaminoethyl (DEAE) cellulose, (DE 53, Whatman Laboratory Products #7153093), packed in a glass Bio-Rad Econo-Column (2.5 cm × 50 cm). The DEAE cellulose resin was conditioned according to the manufacturer’s instructions. The flow was controlled using a peristaltic pump (Tris, Teledyne-ISCO, Lincoln, NE) and an automated fraction collector. The mobile phase was ammonium bicarbonate (0 – 600 mM).

Storage and Dilution of Aplysia californica ADP-ribosyl cyclase (WT and E98G) from concentrated aliquots

Recombinant Aplysia californica ADP-ribosyl cyclase was produced by the procedure of H. C. Lee et al. [24]. Concentrated solutions of purified WT enzyme were stored at −80 °C at a concentration of 10 mg/ml in 100 μl aliquots. A single 100 μl aliquot of the wild type enzyme was removed from the freezer and thawed in an ice bucket to maintain a temperature of 0 °C. The concentrated aliquot was diluted to 5 ml using 20 mM HEPES buffer (pH = 7.5) to make a final concentration of 0.2 mg/ml. The diluted enzyme solution was mixed thoroughly and transferred into polypropylene plastic vials in 60 μl (12 μg) portions and stored at −80 °C until future use. For the synthesis of NAADP (1), three 60 μl portions (36 μg) of the wild type enzyme were utilized for the base exchange of 0.012 mmol of NADP.

Concentrated solutions of purified the mutant E98G Aplysia californica ADP-ribosyl cyclase [42] were acquired and stored at −80 °C at concentrations of 0.62 mg/ml and 0.11 mg/ml in 100 μl aliquots. The concentrated E98G enzyme was removed from the freezer and thawed in an ice bucket to maintain a temperature of 0 °C and then diluted using 20 mM HEPES (pH = 7.5) to make a final concentration of 0.1 mg/ml. The diluted aliquots were mixed thoroughly and transferred into polypropylene plastic vials in 60 μl (6 μg) portions and stored at −80 °C. For the synthesis of NAADP analog 5, three or more, 60 μl aliquots of E98G enzyme (18 μg) were used for the exchange reaction of 0.012 mmol of NADP.

HPLC Method 1: analytical anion exchange chromatography

A sample volume of 50 μl of diluted solution (2 mg/ml in nucleotide) was injected onto the 7 mm × 35 mm (1.3 ml) analytical column packed with AG MP-1 resin in trifluoroacetate form. The chromatography was developed at a flow rate of 5 ml/min by applying 10 ml water followed by a linear gradient of 0–45 mM aqueous TFA over a volume of 85 ml. This step was followed by column cleaning and re-equilibration wherein the column was washed with 15 ml of 100 mM aqueous TFA followed by 15 ml deionized water for re-equilibration prior to the next injection.

HPLC Method 2: preparative anion-exchange using a shallow gradient

Diluted reaction solution (2 mg/ml in dinucleotide) was loaded into a 5 ml injection loop and loaded onto the 1.5 cm × 11.5 cm column packed with AG MP-1 resin in trifluoroacetate form. The chromatography was developed at a flow rate of 5 ml/min by applying 25 ml of water, followed by a linear gradient of 0–45 mM TFA/water formed over a total volume of 280 ml. This step was followed by column cleaning and re-equilibration wherein the column was washed with 30 ml of 100 mM aqueous TFA followed by 40 ml deionized water for re-equilibration prior to the next injection. Effluent samples that showed UV absorbance at 254 nm were collected manually.

Open-Column Chromatography, Method 3—preparative anion-exchange chromatography on columns of DEAE-cellulose

The DEAE- cellulose (DE-53) anion exchange chromatographic system was used for desalting, polishing, and final purification of the dinucleotides that had already been purified by the AG MP-1 column. DE-53 cellulose which was equilibrated with 100 mM NH₄HCO₃ was packed into an open glass column 2.5 cm × 30 cm (column volume = 147 ml). 200 ml of deionized water was applied to the column until the pH of the effluent, which was previously at pH 8, was neutral. The dry sample of dinucleotide and any residual salts was diluted to 1 mg/15 ml and loaded onto the column. The separation was developed by the application of a linear gradient formed between 250 ml of water and 250 ml of 800 mM NH₄HCO₃ at a flow rate of 2 ml/min using a slight positive pressure generated by a peristaltic pump. Fractions were collected and monitored for absorbance at 254 nm. Fractions that showed absorbance at this wavelength were combined, frozen and lyophilized.

4.2.2. Synthesis and Characterization of New Compounds 4, 5 and 8

5-(3-(3-Azido-5-(azidomethyl) benzamido)propyl) nicotinic acid (4)

A 10 ml vial was wrapped in an aluminum foil and equipped with a stirring bar. Compound 8 (500 mg, 1.22 mmol) was dissolved in 5 ml of 1:1 CH₃OH-tetrahydrofuran (THF) solution and stirred at room temperature. 4 N KOH (1 ml) was added dropwise to the reaction mixture and stirred for 4 h. At this time TLC (CH₃OH-CH₂Cl₂-acetic acid 9:90:1) indicated the consumption of the starting material. HCl (1 M, 4 ml) was added to the reaction mixture and stirring continued for another 10 min. The solvents in the reaction mixture were distilled in vacuo and the resulting residue was purified by column chromatography on C-18 column using Combiflash R_f system (0–60% CH₃OH-H₂O) and the product collected and dried in vacuo to afford an off-white oil (279.3 mg, 60%). TLC C-18 alumina backed TLC plates (RP18, Analtech (Newark, DE) cat# 350016) R_f 0.38 in 50% CH₃OH-H₂O, ¹H NMR (600 MHz, D₂O with water suppression) δ 8.52 (s, 1H), 8.35 (s, 1H), 8.02 (s, 1H), 7.17(s, 1H), 7.14(s, 1H), 7.06 (s, 1H), 4.36 (s, 2H), 3.38 (t, 2H, J = 6.18 Hz), 2.74 (t, 2H, J = 7.08 Hz), 1.99 (p, 2H, J = 6.60 Hz); ¹³C NMR (150.9 MHz, CD₃OD) δ 167.3, 167.2, 152.2, 147.5, 141.2, 138.6, 137.5, 137.0, 136.5, 127.2, 123.0, 121.0, 117.1, 53.3, 39.2, 30.1, 29.7.

5-[3-(3-Azido-5-(azidomethyl)benzamido)propyl]-NAADP (5)

A 10 ml vial was wrapped in an aluminum foil and equipped with a stirring bar. 5-[3-(3-Azido-5-(azidomethyl)benzamido)propyl]-nicotinic acid (4) (68.4 mg, 0.18 mmol) was dissolved in distilled water (5 ml) and transferred to the 10 ml vial. The reaction mixture was continuously stirred for 30 min in an incubator (37 °C). After this time the pH was adjusted to 7 with 0.1–0.5 M NaOH and stirred for another 10 min. This procedure was repeated until the pH was stabilized to 7. After the pH was stabilized, NADP (10 mg, 0.01 mmol) was added to the vial and stirred at 37 °C for 10 min. Next, mutant E98G, Aplysia californica ADP-ribosyl cyclase (180 μl, 18 μg) was added to the reaction solution and stirred at 37 °C for 3.5 h. Analysis by HPLC (method 1) indicated the consumption of NADP after 3.5 h. The sample was then purified by preparative HPLC according to method 2. 5-[3-(3-Azido-5-(azidomethyl)-benzamido)propyl]-NAADP (5) eluted between 37 and 45 min (40–45 % B). The fractions collected from the HPLC were combined and pH of solution was adjusted to 7 using 0.1–1 M NaOH. The neutral solution was frozen at −80 °C protected from light by a covering of aluminum foil and lyophilized to give a white amorphous residue. Next, the dry residue was dissolved in 150 ml water and then loaded onto the DEAE-cellulose anion exchange column for further purification according to method 3. The fractions that displayed a UV absorbance at 254 nm were collected, frozen at −80 °C and lyophilized to afford a white powder: (4.2 mg, 34.9%). ¹H NMR (600 MHz, D₂O with water suppression) δ 8.79 (s, 1H), 8.72 (s, 1H), 8.54 (s, 1H), 8.20 (s, 1H), 7.90 (s, 1H), 7.13 (s, 1H), 6.99 (m, 2H), 5.86 (d, 1H, J = 5.16 Hz), 5.82 (d, 1H, J = 5.16 Hz), 4.37 (t, 1H, J = 5.04 Hz), 4.35 (m, 1H), 4.24–3.99 (m, 8H), 3.28 (tt, 2H, J = 6.60 Hz), 2.80 (t, 2H, J = 7.44 Hz), 1.90 (p, 2H, J = 6.84 Hz); ³¹P NMR (161.9 MHz, D₂O) δ 0.52, −10.74 (dd). HRMS calcd for C₃₂H₃₉N₁₃O₁₉P₃⁺: 1002.1698, Found m/z: 1002.1694 (M⁺). HPLC Method 1: t = 25–32 min, 32 mM TFA; UV absorption spectrum: λ_max 253 nm, ε₂₅₃ = 2.78 × 10⁺⁴ L mol⁻¹cm⁻¹; ε₂₆₀ = 2.57 × 10⁺⁴ L mol⁻¹cm⁻¹; ε₂₈₀ = 7.16 × 10⁺³ L mol⁻¹cm⁻¹.

Determination of the extinction coefficients of AIOC-NAADP (5)

Solutions of 5 were prepared, the absorption spectrum measured, and the concentrations of the dinucleotide determined by measurement of the quantity of organic phosphate in a 100 μl aliquot according to the method of Ames, 1966 [62].

Ethyl-5-[3-(3-azido-5-(azidomethyl)benzamido)propyl]nicotinate (8)

A 10 ml vial was wrapped in an aluminum foil and equipped with a stirring bar. Compound 6, 3-azido-5-azidomethylbenzoic acid (190.6 mg, 0.87 mmol) was dissolved in THF (5 ml) and N-hydroxysuccinimide (109.6 mg, 0.95 mmol) added and stirred for 10 min at room temperature. Compound 7 (ethyl 5-(3-aminopropyl)nicotinate) [28] (199.8 mg, 0.96 mmol), diisopropylcarbodiimide (163.8 mg, 1.29 mmol) and diisopropylethylamine (589.4 mg, 4.56 mmol) were added consecutively and the reaction mixture was stirred continuously at room temperature for 16 h. At this time, TLC showed the total consumption of the starting material and a distinct new spot at an R_f of 0.33 (silica gel, ethyl acetate) The reaction mixture was concentrated in vacuo, and the concentrated residue was purified by column chromatography on silica gel (0 −10% CH₃OH-DCM) affording a yellow oil. For analytical purposes, compound 8 was crystallized from ethyl acetate to afford needle like colorless crystals. (226.4 mg, 63.5%). TLC R_f 0.33 in ethyl acetate (100%), ¹H NMR (600 MHz, CDCl₃) δ 9.08 (d, 1H, J = 1.98 Hz), 8.64 (d, 1H, J = 2.16 Hz), 8.15 (t, 1H, J = 2.10 Hz), 7.44 (s, 1H), 7.41 (t, 1H, J = 1.86 Hz), 7.11 (q, 1H, J = 1.56 Hz), 6.36 (t, 1H, -NH, J = 5.22 Hz), 4.43 (q, 2H, J = 7.14 Hz), 4.42 (s, 2H), 3.55 (q, 2H, J = 6.30 Hz), 2.80 (dd, 2H, J = 7.62 Hz), 2.02 (quintuplet, 2H, J = 7.74 Hz), 1.43 (t, 3H, J = 7.14 Hz); ¹³C NMR (150.9 MHz, CDCl₃) δ 166.3, 165.4, 153.4, 148.8, 141.5, 138.2, 136.9, 136.7, 136.5, 126.1, 122.5, 121.2, 117.5, 72.5, 53.5, 39.5, 20.5, 20.0, 17.0. HRMS calcd for C₁₉H₂₀N₈O₃: 409.1737 (M + H). Found 409.1739 (M + H). Anal. calcd for C₁₉H₂₀N₈O₃∙ ¼ EtOAc:  C, 55.81; H, 5.15; N, 26.03. Found:  C, 55.49; H, 4.98; N, 26.18.

4.2.3. Radiochemical Synthesis

[³²P]-NAADP (1), [³²P]-5-N₃-NAADP (2) and [³²P]-AIOC-NAADP (5) — The [³²P]-labeled probes were synthesized using a two-step strategy used previously [34]. In the first step, [³²P]-NAD is converted to [³²P]-NADP using human recombinant NAD kinase (Enzo Life Sciences, Farmingdale, NY; ALX-201–236). The resulting [³²P]-NADP is converted to the appropriate probe through a base-exchange reaction catalyzed by wild type [24] or a mutant E98G Aplysia ADP-ribosyl cyclase [42]. The conditions for each base-exchange reaction for the production of each probe are detailed below. The [³²P]-NADP was dried in a Savant SpeedVac concentrator prior to the base-exchange reactions.

[³²P]-NAADP was produced by adding 200 μl of 50 mM nicotinic acid in 100 mM sodium acetate, pH 4 to a vial of dried [³²P]-NADP. The reaction was started by adding 0.5 μl of 0.1 mg/ml wild type Aplysia ADP-ribosyl cyclase and allowed to incubate for 2 h at room temperature in the dark.

[³²P]-5-N₃-NAADP was produced by adding 100 μl of 50 mM 5-N₃-nicotinic acid, pH 7 to a vial of dried [³²P]-NADP. The reaction was started by adding 0.5 μl of 0.1 mg/ml E98G Aplysia ADP-ribosyl cyclase and allowed to incubate for 5 h at room temperature in the dark. The E98G ADP-ribosyl cyclase mutant is able to catalyze the base-exchange reaction at neutral pH, which is useful in situations where the nicotinic acid analog being used in not soluble at pH 4 [42].

[³²P]-AIOC-NAADP was produced by adding 50 μl of 50 mM AIOC-nicotinic acid, pH 7.5 to a dried aliquot of [³²P]-NADP. The reaction was started by adding 1 μl of 0.1 mg/ml E98G Aplysia ADP-ribosyl cyclase and allowed to incubate at room temperature in the dark for 5 h. All reactions were purified by chromatography on a AG MP-1 column as previously described [34].

4.2.4. Photoaffinity Labeling Studies

The conditions for the preparation of sea urchin egg homogenates and their photoaffinity labeling has been previously described [34–36]. Jurkat and SKBR3 cell extracts were prepared as previously described [35, 36]. Briefly, whole cell lysates (WCL) were prepared by resuspension of trypsinized or pelleted cells in HEPES buffer (20 mM HEPES, protease inhibitors (Roche), pH 7.3) prior to sonication on ice. The sonicated sample was then centrifuged (1,000 × g for 10 min, 4 °C) and the pellet discarded. The cytosolic fraction was prepared from the WCL by dual centrifugation steps. Supernatant from a second spin (10,000 g for 20 min, 4 °C) was collected and re-spun (100,000 g for 1 hr, 4 °C) to yield the cytosolic cell extract that was used for photoaffinity labeling in this study.

The standard labeling condition utilized 0.3 – 3 nM ³²P- probe (5-N₃-NAADP or AIOC-NAADP). Samples were incubated with [³²P] -photoprobe for 5–90 min on ice prior to photoactivation effected by exposure of samples to UV light for 2 min. PAL samples were then incubated (<15 min) with SDS sample buffer supplemented with 2- mercaptoethanol (10%) to reduce free label. The photolabeled samples were analyzed by SDS-PAGE on 12% TGX gels from Bio-Rad. The resulting gels were stained with Simply Blue Safestain and air-dried. The photolabeling was analyzed by exposing the dried gels to MP storage phosphor screens (Packard Instruments). The screens were developed using a Typhoon storage phosphor system. Densitometric analysis was accomplished using Image J software.

4.2.5. Click chemistry conditions

Jurkat cell extracts photoaffinity labeled with AIOC-NAADP was subjected to click chemistry in order to determine if proteins photo-crosslinked were able to be biotinylated (results shown in Figure 10). Jurkat cytosol was first photoaffinity labeled with either unlabeled AIOC-NAADP or [³²P]-AIOC-NAADP in separate reactions. The photoaffinity labeling reactions were done in a total volume of 480 μl containing 1.5 mg/ml Jurkat cytosol and either 100 nM unlabeled AIOC-NAADP or 1 nM [³²P]-AIOC-NAADP (5). Following the UV irradiation under the standard photolabeling protocol, the two samples were mixed and desalted using Roche Quick Spin Columns (Sephadex G-25) to remove free probe. The desalted sample (2.3 ml) was divided into five equal (460 μl) aliquots and subjected to either copper(I) catalyzed click chemistry or strain promoted click chemistry. Diazo-biotin-alkyne, acetylene-PEG-biotin, Diazo-biotin-DBCO and DBCO-PEG4-biotin and THPTA were obtained from Click Chemistry Tools, Phoenix, AZ. The copper(I) catalyzed click chemistry conditions were performed under conditions optimized by Finn, Banyon, McKart and colleagues [63, 64]. Each reaction contained 72 μM diazo-biotin-alkyne or acetylene-PEG-biotin, 5 mM aminoguanidine, 0.1 mM CuSO₄, 0.5 mM THPTA and 5 mM ascorbic acid in a total volume of 530 μl (460 μl of desalted photolyzed sample). The CuSO₄ and THTPA were mixed in a 1:5 ratio just prior to addition. The reactions were allowed to proceed 2 h at room temperature. The strain promoted click chemistry reactions were performed as follows. Diazo-biotin-DBCO or DBCO-PEG4-biotin were added to 460 μl of the desalted photolyzed Jurkat sample to a final concentration of 36 μM and incubated at room temperature for 16 h. The fifth aliquot of desalted photolyzed Jurkat cytosol sample served as a control. The click reagents were removed after the reaction by desalting as described above. The control and desalted samples from the four click reactions were then allowed to incubate with 200 μl of streptavidin-agarose (SAA) beads for 16 h at 4 °C on a rotator. The samples were then centrifuged at 4000 rpm for 2 min and the supernatant was removed and the SAA pellets were washed 5 times with 5 ml of phosphate-buffered saline (PBS), pH 7.4. The final washed SAA pellets were suspended in 2× SDS sample buffer and boiled for 20 min to release proteins bound to the SAA beads. The SAA beads were transferred to a Costar SpinS centrifugal filter unit and centrifuged to remove the SAA beads. The proteins released were analyzed by SDS-PAGE on 12% TGX gels from Bio-Rad.

Highlights to accompany: The Synthesis and Characterization of a Clickable-Photoactive NAADP Analog Active in Human Cells.

• A clickable-photoactive NAADP analog was synthesized from simple materials

• This analog was recognized at low concentration by human NAADP receptors

• Clickable-photoactive NAADP was validated as a probe for NAADP binding proteins

Abbreviations

cADPR

cyclic adenosine diphosphate ribose

DBU

1,8-diazabicyclo[5.4.0]undec-7-ene

DBCO

dibenzocyclooctyne

DCM

dichloromethane

DEAE

diethylaminoethyl

HRMS

high resolution mass spectrometry

IP3

inositol 1,4,5-trisphosphate

NAADP

nicotinic acid adenine dinucleotide phosphate

PEG

polyethyleneglycol

RyR

ryanodine receptor

SAR

structure activity relationship

SAA

streptavadin-agarose

t-BuONO

tert-butyl nitrite

TFA

trifluoroacetic acid

THF

tetrahydrofuran

THPTA

tris(benzyltriazolylmethyl)amine

TMS-N₃

azidotrimethylsilane

TPC

two pore channel

TRP-ML1

transient receptor potential cation channel, mucolipin subfamily 1

WCL

whole cell lysate