Ondřej Baszczyňski1, Joanna M Watt2,1, Monika D Rozewitz3, Andreas H Guse3, Ralf Fliegert3, Barry V L Potter2,1. 1. Wolfson Laboratory of Medicinal Chemistry, Department of Pharmacy and Pharmacology , University of Bath , Bath BA2 7AY , U.K. 2. Medicinal Chemistry & Drug Discovery, Department of Pharmacology , University of Oxford , Mansfield Road , Oxford OX1 3QT , U.K. 3. The Calcium Signalling Group, Department of Biochemistry and Molecular Cell Biology , University Medical Center Hamburg-Eppendorf , Martinistrasse 52 , 20246 Hamburg , Germany.
Abstract
TRPM2 (transient receptor potential cation channel, subfamily M, member 2) is a nonselective cation channel involved in the response to oxidative stress and in inflammation. Its role in autoimmune and neurodegenerative diseases makes it an attractive pharmacological target. Binding of the nucleotide adenosine 5'-diphosphate ribose (ADPR) to the cytosolic NUDT9 homology (NUDT9 H) domain activates the channel. A detailed understanding of how ADPR interacts with the TRPM2 ligand binding domain is lacking, hampering the rational design of modulators, but the terminal ribose of ADPR is known to be essential for activation. To study its role in more detail, we designed synthetic routes to novel analogues of ADPR and 2'-deoxy-ADPR that were modified only by removal of a single hydroxyl group from the terminal ribose. The ADPR analogues were obtained by coupling nucleoside phosphorimidazolides to deoxysugar phosphates. The corresponding C2″-based analogues proved to be unstable. The C1″- and C3″-ADPR analogues were evaluated electrophysiologically by patch-clamp in TRPM2-expressing HEK293 cells. In addition, a compound with all hydroxyl groups of the terminal ribose blocked as its 1″-β- O-methyl-2″,3″- O-isopropylidene derivative was evaluated. Removal of either C1″ or C3″ hydroxyl groups from ADPR resulted in loss of agonist activity. Both these modifications and blocking all three hydroxyl groups resulted in TRPM2 antagonists. Our results demonstrate the critical role of these hydroxyl groups in channel activation.
TRPM2 (transient receptor potential cation channel, subfamily M, member 2) is a nonselective cation channel involved in the response to oxidative stress and in inflammation. Its role in autoimmune and neurodegenerative diseases makes it an attractive pharmacological target. Binding of the nucleotide adenosine 5'-diphosphate ribose (ADPR) to the cytosolic NUDT9 homology (NUDT9 H) domain activates the channel. A detailed understanding of how ADPR interacts with the TRPM2 ligand binding domain is lacking, hampering the rational design of modulators, but the terminal ribose of ADPR is known to be essential for activation. To study its role in more detail, we designed synthetic routes to novel analogues of ADPR and 2'-deoxy-ADPR that were modified only by removal of a single hydroxyl group from the terminal ribose. The ADPR analogues were obtained by coupling nucleoside phosphorimidazolides to deoxysugar phosphates. The corresponding C2″-based analogues proved to be unstable. The C1″- and C3″-ADPR analogues were evaluated electrophysiologically by patch-clamp in TRPM2-expressing HEK293 cells. In addition, a compound with all hydroxyl groups of the terminal ribose blocked as its 1″-β- O-methyl-2″,3″- O-isopropylidene derivative was evaluated. Removal of either C1″ or C3″ hydroxyl groups from ADPR resulted in loss of agonist activity. Both these modifications and blocking all three hydroxyl groups resulted in TRPM2 antagonists. Our results demonstrate the critical role of these hydroxyl groups in channel activation.
The nonselective cation
channel TRPM2 (transient receptor potential
cation channel, subfamily M, member 2) is activated in a Ca2+-dependent manner after binding of the nucleotide ADP-ribose (ADPR)
to its cytosolic C-terminal NUDT9H domain that shares
homology with a mitochondrial nucleotide pyrophosphatase NUDT9.[1] While earlier studies indicated that the NUDT9H domain also has a low pyrophosphatase activity, hydrolyzing
ADPR to AMP and ribose 5-phosphate (R5P),[2] a recent study showed that the Nudix box motif in TRPM2 does not
support catalysis and the production of AMP might have been due to
spontaneous hydrolysis of ADPR at alkaline pH.[3]Reactive oxygen species (ROS) and genotoxic stress can result
in
release of ADPR from the nucleus due to the activation of the poly(ADP-ribose)
polymerase-1 (PARP-1) and poly(ADP-ribose) glycohydrolase pathways.[4,5] The ADPR so generated can then activate TRPM2, resulting in prolonged
Ca2+-entry, mitochondrial Ca2+-overload, and
apoptosis,[6] thereby contributing to cell
damage in post ischemic reperfusion injury during myocardial infarction[7] and stroke.[8]Besides this role in cell death, TRPM2 also participates in physiological
processes like inflammation.[9] In neutrophil
granulocytes and dendritic cells, TRPM2 contributes to chemotaxis.[10−12] The chemotaxis of murine neutrophils in response to fMLP is independent
of PARP-1 but can be inhibited by 8-Br-ADPR, a compound that inhibits
activation of TRPM2 by ADPR, and by knock-out of CD38, a glycohydrolase
that producesADPR from NAD.[10] In macrophages
and monocytes, TRPM2 is involved in secretion of chemokines and cytokines
in response to ROS and pro-inflammatory cytokines,[13,14] whereas in effector T cells, it plays a role in proliferation and
secretion of pro-inflammatory cytokines.[15] Inhibition of TRPM2 has been shown to reduce tissue damage after
stroke by preventing invasion of neutrophils,[16] and knock-out of TRPM2 ameliorates the symptoms of experimentally
induced autoimmune encephalomyelitis, a model for multiple sclerosis.[15] TRPM2 is thus an attractive pharmacological
target for the treatment of neurodegenerative and autoimmune diseases.[17,18]Further investigations into the role of TRPM2 in physiology
and
pathophysiology require specific modulators of channel function, and
their rational design will benefit from a better understanding of
its structure–activity relationship (SAR). Recently, TRPM2
structures from three different species were elucidated by cryo-EM.
The structure of RPM2 from the sea anemone Nematostella
vectensis nvTRPM2 lacks the NUDT9H domain, probably because of the flexibility of this part of the
molecule.[19] TRPM2 from zebrafish (Danio rerio) drTRPM2 has been solved in the apo state
and in an ADPR-bound state.[20] While ADPR
could not be localized in the NUDT9H domain, electron
density corresponding to ADPR was found in the N-terminal
MHR1/2 domain. Comparison of the conformations of the apo and the
ADPR-bound state and mutational analysis indicate that in drTRPM2,
gating of the channel occurs via binding of ADPR to the MHR1/2 domain
instead of NUDT9H.[20] Structures
of humanTRPM2 (hTRPM2) have been resolved for the apo state, an ADPR
bound (primed) state and an open conformation bound to ADPR and Ca2+.[21] Mutagenesis of the MHR1/2
domain and removal of the NUDT9H domain show that
in the humanTRPM2 channel, gating occurs after binding of ADPR to
the NUDT9H domain, but again poor resolution prevents
the placement of ADPR in the presumed binding pocket. Structure-based
drug design therefore still awaits a high-resolution structure of
either the full-length channel or the isolated NUDT9H domain of TRPM2.Recently, we synthesized ADPR analogues to
explore the role of
the adenosine, pyrophosphate, and terminal ribose motifs in activation
of TRPM2.[22,23] The pyrophosphate-forming couplings used
employed morpholidate- or CDI-mediated methodologies. Yields were
generally low and reaction times were generally long and improvements
are warranted. To our surprise, the majority of structural modifications
of ADPR led to compounds that do not activate TRPM2, indicating that
all three moieties are required for channel opening.[23] These studies also revealed that hydroxyl group removal
at C2′ resulted, in the case of ADPR, in a compound that is
a TRPM2 agonist with significantly higher efficacy than ADPR itself[23] and also improved the antagonist 8-Ph-ADPR (IC50 11 μmol/L compared to 8-Ph-2′-deoxy-ADPR, IC50 3 μmol/L).[22] We also highlighted
the importance of the terminal ribose for TRPM2 activation by evaluating
a series of modified ADPR analogues.[29] While
simple ADP neither activated nor antagonized the channel, introduction
of substituents at the β-phosphate that increasingly resembled
the terminal ribose (Figure ) returned antagonist properties, but none of the analogues
exhibited agonist activity. The fact that β-(tetrahydrofuran-2-yl)methyl-ADP,
an analogue with the ribofuranose backbone, but lacking the hydroxyl
groups, did not activate the channel but instead antagonized TRPM2
indicated that one or more of the hydroxyl groups might be important
for the gating of TRPM2. Previous studies have shown hydroxyl group
deletion to be a valuable tool in SAR elucidation for complex bioactive
molecules.[24−26] While such endeavors may be synthetically protracted
in order to effect a desired precision edit to the parent molecule,
they can divulge key mechanistic information.[27,28]
Figure 1
Structure
of ADPR and known terminal ribose analogues.[29]
Structure
of ADPR and known terminal ribose analogues.[29]We report here the synthesis and
electrophysiological evaluation
of novel terminal-ribose-modified analogues of the TRPM2 agonists
ADPR and 2′-deoxy-ADPR to study the role of the individual
hydroxyl groups of the terminal ribose in TRPM2 activation. The critical
step of analogue formation was achieved by combining a sugar phosphate
with a P-activated nucleotide to form a single molecule linked by
a pyrophosphate bond. A recently reported improved procedure for the
preparation of NDP sugars from nucleoside phosphorimidazolides[36] uses 2–4 equiv of magnesium chloride
to achieve high reaction yields and short reactions times for nucleoside
5-phosphorimidazolide coupling to sugar phosphates.
Results and Discussion
The terminal ribose of ADPR is essential for the ligand-driven
activation of the cation channel TRPM2.[29] To investigate the SAR of this part of ADPR in more detail, we focused
upon the synthesis of all three possible 1″-deoxy, 2″-deoxy,
and 3″-deoxy terminal riboseADPR derivatives 2, 3, and 4 by selective deletion of the
appropriate hydroxyl group (Figure ). The similar reactivity of the ribosehydroxyl groups
and potential for ring opening of the cyclic hemi-acetal required
development of multistep synthetic routes with selective masking and
deprotection. The modified ribose building blocks were then coupled
to either AMP or 2′-deoxy-AMP via pyrophosphate coupling reaction
(Figure ). We also
synthesized a corresponding analogue 1 with all hydroxyl
groups present, but blocked with small alkyl groups to interrogate
the H-bond donating capability of the terminal ribose (Figure ). The chemically stable analogues
were evaluated regarding their agonist and antagonist activity by
patch-clamp experiments in whole cell configuration using HEK cells
with stable expression of humanTRPM2.
Figure 2
Pyrophosphate coupling
reaction and ADPR analogues prepared in
this study.
Pyrophosphate coupling
reaction and ADPR analogues prepared in
this study.At first, routes to the
required modified terminal ribose 5-phosphates
were designed. It was envisaged that such phosphates would be ideal
coupling partners for activated nucleotide imidazolides. Protected
ribose monophosphate 9 was chosen to exploit the ligand
space around the terminal ribose, H-bond donating capability, and
its good susceptibility for coupling with activated nucleosides. The
precursor of 9, compound 8, was obtained
from protected ribose[30]7 by
phosphitylation with dibenzyl N,N-diisopropylphosphoramidite and subsequent oxidation using hydrogen
peroxide[31] to afford 8 (79%)
(Scheme A). Compound 8 was then hydrogenated using Pd/C to afford 9 (82%). Ribose-5-monophosphate 9 was converted to its
tributylammonium salt and freeze-dried. Synthesis of 1-deoxyribose-5-phosphate 10, to explore the role of the anomeric OH group of the terminal
ribose, started from 2,3-O-isopropylidene ribose 13 that was protected at position 5 by a trityl group, to
afford 14 (65%) (Scheme B). Reduction of 14 by sodium borohydride
led to compound 15 (85%). Reaction of 15 with tosyl chloride in pyridine gave no significant result at room
temperature, but heating the reaction mixture up to 60 °C helped
to afford the protected 1-deoxyribose 16 (80%). However,
deprotection of 16 with acetic acid failed after multiple
attempts, as both the 5-O-trityl and 2,3-O-isopropylidene protecting groups were cleaved, leading
to the undesired fully deprotected 1-deoxyribose or to complex mixtures.
Using the alternative deprotection of 16 with formic
acid in diethylether[32] afforded compound 17 (32%).[33] Phosphitylation and
subsequent oxidation of 17 with di-tert-butyl N,N-diisopropylphosphoramidite
gave the phosphate derivative 18 (41%). Careful deprotection
of 18 with aqueous trifluoroacetic acid at low temperature
afforded the target 1-deoxyribose-5-phosphate 10 (49%)
that was subsequently converted to its tributylammonium salt.
Scheme 1
Synthesis of Modified Terminal Riboses 9 and 10
Reagents and conditions: (a)
5-Ph-1-H-tetrazole, dibenzyl N,N-diisopropylphosphoramidite, DCM, 20 °C, 1 h; (b)
triethylamine, H2O2, 0–20 °C, 1
h; (c) hydrogen (balloon), Pd/C, 5 h, TEAB (1 M), 20 °C; (d)
Dowex D50 (H+), tributylamine; (e) tritylchloride, pyridine,
20 °C, 16 h; (f) sodium borohydride, ethanol, 0–20 °C,
2 h; (g) tosyl chloride, pyridine, 60 °C, 16 h; (h) HCOOH/diethylether,
20 °C, 16 h; (i) 5-Ph-1-H-tetrazole, di-tert-butyl N,N-diisopropylphosphoramidite,
DCM, 20 °C, 1 h; (j) triethylamine, H2O2, 0–20 °C, 1 h; (k) aqueous TFA, 0–20 °C,
4 h; (l) Dowex D50 (H+), tributylamine.
Synthesis of Modified Terminal Riboses 9 and 10
Reagents and conditions: (a)
5-Ph-1-H-tetrazole, dibenzyl N,N-diisopropylphosphoramidite, DCM, 20 °C, 1 h; (b)
triethylamine, H2O2, 0–20 °C, 1
h; (c) hydrogen (balloon), Pd/C, 5 h, TEAB (1 M), 20 °C; (d)
Dowex D50 (H+), tributylamine; (e) tritylchloride, pyridine,
20 °C, 16 h; (f) sodium borohydride, ethanol, 0–20 °C,
2 h; (g) tosyl chloride, pyridine, 60 °C, 16 h; (h) HCOOH/diethylether,
20 °C, 16 h; (i) 5-Ph-1-H-tetrazole, di-tert-butyl N,N-diisopropylphosphoramidite,
DCM, 20 °C, 1 h; (j) triethylamine, H2O2, 0–20 °C, 1 h; (k) aqueous TFA, 0–20 °C,
4 h; (l) Dowex D50 (H+), tributylamine.Synthesis of the 2-deoxy version of the terminal ribose (Scheme ) started from commercially
available 2-deoxyribose-5-phosphate sodium salt 19 (Sigma-Aldrich)
that was converted to its corresponding mono-tributylammonium salt 11 by using Dowex resin (H+ form) followed by titration
with tributylamine to pH ≈ 7 (Scheme A). Partial decomposition (20–30%)
of 19 was observed during the transformation to 11 via 31P NMR, suggesting that 2-deoxyribose-5-monophosphate
is less stable toward changes in pH (see Supporting Information, S14–16) compared to its 1-deoxy- and 3-deoxy-counterparts.
Synthesis of 3-deoxyribose-5-monophosphate 12 started
from 1,2-O-isopropylidene xylose 20,[34] which was protected at the 5-hydroxyl group
as the TBDPS ether to give 21 (83%) (Scheme B). Reaction of 21 with 1,1′-thiocarbonyldiimidazole in DCM afforded compound 22 (63%). Reductive deoxygenation of 22 with
tributyltin hydride led to the compound 23 (76%) which
was subsequently deprotected in buffered TBAF to give 1,2-O-isopropylidene-3-deoxyribose 24 (82%). Phosphitylation
and subsequent oxidation of 24 gave the desired di-tert-butyl phosphate derivative 25 (50%). However,
several attempts to deprotect compound 25 with aqueous
TFA led to a complex mixture, suggesting that di-tert-butyl protected phosphate 25 was not the optimal precursor
for 12. Therefore, 24 was phosphitylated
and subsequently oxidized to its dibenzylphosphate derivative 26 (72%) that was successfully deprotected to give 27 (68%). Hydrogenation of 27 followed by neutralization
of the free phosphate using tributylamine afforded the target analogue
3-deoxyribose-5-phosphate as its mono-tributylammonium salt 12 (67%).
Scheme 2
Synthesis of Modified Terminal Riboses 11 and 12
Reagents and conditions:
(a)
Dowex D50 (H+), tributylamine; (b) TBDPS chloride, DMAP,
pyridine, 20 °C, 16 h; (c) 1,1′-thiocarbonyldiimidazole,
DCM, reflux, 2 h; (d) tributyltinhydride, AIBN, toluene, 116 °C,
3 h; (e) TBAF·3H2O, acetic acid, 20 °C, 3 h;
(f) 5-Ph-1-H-tetrazole, di-tert-butyl N,N-diisopropylphosphoramidite, DCM, 20
°C, 1 h; (g) triethylamine, H2O2, 0–20
°C, 1 h; (h) various conditions with aqueous TFA; (i) 5-Ph-1-H-tetrazole, dibenzyl N,N-diisopropylphosphoramidite, DCM, 20 °C, 1 h; (j) triethylamine,
H2O2, 0–20 °C, 1 h; (k) aqueous
TFA, 0 °C, 1.5 h; (l) hydrogen (balloon), Pd/C, 4 h, 20 °C;
(m) tributylamine.
Synthesis of Modified Terminal Riboses 11 and 12
Reagents and conditions:
(a)
Dowex D50 (H+), tributylamine; (b) TBDPS chloride, DMAP,
pyridine, 20 °C, 16 h; (c) 1,1′-thiocarbonyldiimidazole,
DCM, reflux, 2 h; (d) tributyltinhydride, AIBN, toluene, 116 °C,
3 h; (e) TBAF·3H2O, acetic acid, 20 °C, 3 h;
(f) 5-Ph-1-H-tetrazole, di-tert-butyl N,N-diisopropylphosphoramidite, DCM, 20
°C, 1 h; (g) triethylamine, H2O2, 0–20
°C, 1 h; (h) various conditions with aqueous TFA; (i) 5-Ph-1-H-tetrazole, dibenzyl N,N-diisopropylphosphoramidite, DCM, 20 °C, 1 h; (j) triethylamine,
H2O2, 0–20 °C, 1 h; (k) aqueous
TFA, 0 °C, 1.5 h; (l) hydrogen (balloon), Pd/C, 4 h, 20 °C;
(m) tributylamine.In previous work, we used
morpholidate chemistry or 1,1′-carbonyldiimidazole-based
methodology in coupling reactions, but these had been less than satisfactory
during ADPR analogue preparation. To find the most suitable conditions
here for pyrophosphate bond formation between the corresponding ribose-5-monophosphate
and either AMP or 2′-dAMP, two different methods were tried.
Activation of AMP tributylammonium salt with 1,1′-carbonyldiimidazole[35] followed by addition of the triethylammonium
salt of ribose 9 did show formation of the desired pyrophosphate
product 1 by high-performance liquid chromatography (HPLC).
However, several byproducts were also observed (see Supporting Information Figure S1a, S2). In contrast, following
the Dabrowski-Tumanski procedure[36] activation
of the AMP tributylammonium salt using imidazole, 2,2′-dithiodipyridine
(Aldrithiol) and triphenyl phosphine cleanly afforded the AMP-imidazolide
that was isolated by precipitation with a cold solution of NaI in
acetone. When stirred with the triethylammonium salt of 9 and magnesium chloride in dimethylformamide (DMF), the AMP-imidazolide
gave almost exclusively the desired compound 1 (see Supporting Information Figure 1b, S2–S3).
Finally, purified target ribose derivatives 9, 10, 11, and 12 were individually
coupled to imidazolide-activated AMP or 2′-deoxy-AMP (Scheme ). Adenosine-5′-monophosphate 28 and 2′-deoxy-adenosine-5′-monophosphate 29 were transformed to their imidazolides 30 and 31, by reaction of the corresponding mono-tributylammonium
salt of 28 and 29 and imidazole using the
Aldrithiol and triphenylphosphine condensation protocol.[36] Imidazolides 30 and 31 were precipitated from the reaction mixture by addition of a 0.1
M solution of sodium iodide in cold, anhydrous acetone and were directly
used for coupling. The target modified ADPR analogues 1–6 were then prepared by coupling[36] the corresponding imidazolide 30 or 31 and mono-tributylammonium salt of the modified
ribose-5-phosphate 9, 10, 11, or 12 to give variable yields of the desired ADPR
analogues 1 (24%), 2 (66%), 3 (8%), 4 (17%), 5 (42%), and 6 (27%). Final compounds 1–6 were purified by
semi-preparative HPLC using triethylammonium bicarbonate (TEAB) buffer
and isolated as the corresponding triethylammonium salts.
Scheme 3
Synthesis
of Target Terminal-Modified ADPR Analogues
Reagents
and conditions: (a)
Aldrithiol, imidazole, triethylamine, triphenylphosphine, 20 °C,
16 h; (b) NaI, acetone (precipitation), 0 °C; (c) MgCl2, dimethylformamide, 20 °C, 3–16 h.
Synthesis
of Target Terminal-Modified ADPR Analogues
Reagents
and conditions: (a)
Aldrithiol, imidazole, triethylamine, triphenylphosphine, 20 °C,
16 h; (b) NaI, acetone (precipitation), 0 °C; (c) MgCl2, dimethylformamide, 20 °C, 3–16 h.
Stability
of Terminal-Modified ADPR Analogues 1–6
The stability of ADPR analogues 1–6 was evaluated using analytical HPLC. Aqueous solutions of 1–6 stored at 5 °C were compared
with freshly prepared standard solutions of 1–6 (from solid samples stored at −20 °C). All compounds
except 3 were stable in aqueous solution for several
days when stored in the fridge (5 °C). Compound 3, 2″-deoxy-ADPR, was found to be unstable under these conditions.
We also observed decomposition of 3 during transport
and sample preparation, so we could not do electrophysiological experiments
with 3 (Supporting Information Figure 2, S4). The instability of 3 may correspond
with the lower stability of 2-deoxyribose-5-monophosphate 11 that was sensitive toward changes in pH during conversion of the
commercially available sodium salt to the tributylammonium salt (conversion
of 19 to 11). This decomposition led to
hydrolysis of the 5′-phosphate (approx. 20–30% de-phosphorylation
of 11 was observed via 31P NMR). The corresponding
decomposition by de-phosphorylation in 2″-deoxy-ADPR 3 might cleave the pyrophosphate bond and would be predicted
to lead to inactive fragments.
Evaluation of Novel Analogues
Against TRPM2
Recently,
we investigated the importance of the ADPR terminal ribose for the
activation of TRPM2.[29] This can be investigated
by whole cell patch-clamp experiments. During these experiments, a
glass pipette with a diameter much smaller than a cell and filled
with a solution mimicking the ionic composition of the cytosol is
attached to the plasma membrane. By applying suction to the pipette,
the membrane patch underneath the pipette is ruptured. In voltage-clamp
experiments, a voltage is applied between an electrode in the pipette
solution and a bath electrode in the bath solution surrounding the
cells. This voltage results in ions moving through channels in the
membrane resulting in a current that can be recorded.[37,38] In HEK293 cells expressing TRPM2, addition of ADPR to the pipette
solution results in a current that is absent in wild type HEK293 cells.
ADPR analogues can be tested for their ability to activate the channel,
and they can also be tested for antagonist activity by adding them
in addition to ADPR. We found that adenosine 5′-monophosphate
(AMP) and adenosine 5′-diphosphate (ADP) neither activated
or inhibited the channel even when applied in large excess over ADPR.[29] Interestingly, replacement of the terminal ribose
of ADPR with small substituents led to analogues that did not activate
the channel, but inhibited activation by ADPR, indicating that they
compete with ADPR for the ligand binding site of TRPM2. Thus, activation
of TRPM2 may be attributed to specific interactions between the hydroxyl
groups of the terminal ribose of ADPR and the ligand binding NUDT9H domain of TRPM2. We evaluated the deoxy analogues 2 and 4 and an ADPR analogue 1 where
all hydroxyl groups of the terminal ribose are masked and could no
longer act as hydrogen bond donors (Figure ). It was not possible to evaluate 3 biologically because of instability issues (see above and SI). During the experiments, we included 8-Br-ADPR
as a control. 8-Br-ADPR has previously been shown to inhibit activation
of TRPM2 by ADPR most likely by competing with the agonist.[10]
Figure 3
Effect of ADPR analogues on whole cell currents in TRPM2
expressing
HEK293 cells. Conditions: (a) Effect of ADPR analogues on whole cell
currents in TRPM2 expressing HEK293 cells. Experiments were done as
outlined in the Experimental Section. ADPR
or the indicated ADPR analogue was added to the pipette solution at
a concentration of 100 μmol/L. (b) Effect of ADPR analogues
on TRPM2 whole cell currents elicited by ADPR. In this case, the pipette
solution contained either no nucleotide (buffer), 100 μM ADPR
or a combination of 100 μM ADPR with 900 μM of the indicated
ADPR analogue. 8-Br-ADPR was included as the inhibitor control. Points
indicate the maximum current from individual cells. The number of
cells for each condition is indicated at the bottom. Bars represent
the mean of the log-transformed currents (ns = non-significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). Control
conditions on both panels include overlapping sets of data because
on some days, both agonist and antagonist experiments were performed.
Effect of ADPR analogues on whole cell currents in TRPM2
expressing
HEK293 cells. Conditions: (a) Effect of ADPR analogues on whole cell
currents in TRPM2 expressing HEK293 cells. Experiments were done as
outlined in the Experimental Section. ADPR
or the indicated ADPR analogue was added to the pipette solution at
a concentration of 100 μmol/L. (b) Effect of ADPR analogues
on TRPM2 whole cell currents elicited by ADPR. In this case, the pipette
solution contained either no nucleotide (buffer), 100 μM ADPR
or a combination of 100 μM ADPR with 900 μM of the indicated
ADPR analogue. 8-Br-ADPR was included as the inhibitor control. Points
indicate the maximum current from individual cells. The number of
cells for each condition is indicated at the bottom. Bars represent
the mean of the log-transformed currents (ns = non-significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). Control
conditions on both panels include overlapping sets of data because
on some days, both agonist and antagonist experiments were performed.As expected, the masked analogue
1″-β-O-Me-2″,3″-O-iPr-ADPR 1 did not activate
TRPM2 on its own (Figure a). Because it is more space-filling than
ADPR, we were unsure whether it would compete with ADPR for binding
to the NUDT9H domain. When applied in excess over
ADPR, however, it indeed inhibited activation of TRPM2 (Figure b), showing that further steric
constraints than those simply at the 1″-position of the terminal
ribose are tolerated and may potentially be exploitable for TRPM2
antagonist optimization.Interestingly, an analogue 32 (Figure ) of the
low-affinity partial TRPM2 agonist
α-β-methylene ADPR (AMPCPR)[39] masked in the same way as 1 was recently described
to be a high affinity TRPM2 antagonist that inhibits ADPR-elicited
TRPM2 currents with an IC50 of 5.7 μM.[40] Luo et al. here did not explicitly mention whether
they tested this compound for TRPM2 activation but, because it fully
inhibits the channel at 100 μM, it is unlikely to exhibit agonist
activity. Methylene analogues may show increased stability toward
cellular pyrophosphatases like NUDT9 and NUDT5 which may be retained
in the cell during whole cell patch-clamp experiments. Earlier reports
that indicated an ADPR pyrophosphatase activity of the NUDT9H domain of TRPM2,[1] that might
contribute to degradation of agonists with pyrophosphate bridge have
recently been disputed.[3] The observed difference
could also result from different experimental conditions as we buffered
the intracellular Ca2+ concentration to 200 nmol/L with
10 mmol/L of EGTA, whereas Luo et al. left the Ca2+ concentration
largely unbuffered (50 μmol/L EGTA) which would result in a
much steeper dependence of the current on the agonist concentration
and a more pronounced inhibition by the antagonist.
Figure 4
Structures of masked
ADPR 1, α-β-methylene
ADPR (AMPCR) 32, 2″-O-acetyl-ADPR 33, and 3″-O-acetyl-ADPR 34.
Structures of masked
ADPR 1, α-β-methylene
ADPR (AMPCR) 32, 2″-O-acetyl-ADPR 33, and 3″-O-acetyl-ADPR 34.Contrary to our expectations,
both ADPR analogues 2 and 4 lacking hydroxyl
groups at the terminal ribose
did not induce significant whole cell currents in TRPM2-expressing
HEK cells at 100 μmol/L (Figure a). This was unexpected as O-acetyl-ADPR,
a product of NAD-dependent histone deacetylases of the Sir2 family,
is an effective activator of TRPM2 that binds to the NUDTH domain with similar affinity as ADPR, as shown by UV cross-linking
experiments and activates TRPM2 with a comparable concentration dependence
as ADPR.[41] 2″-O-acetyl-ADPR 33, the product of the deacetylases, undergoes
in neutral solution rapid intramolecular transesterification resulting
in roughly equal amounts of 2″-O-acetyl-ADPR 33 and 3″-O-acetyl-ADPR 34.[42] Either the nucleotide binding site
of TRPM2 is indifferent to the position of the acetyl group, which
would indicate that neither the 2″- nor the 3″-hydroxyl
group is essential as hydrogen bond donor, or there is a specificity
for one of the isomers which would mean that only half of the O-acetyl-ADPR molecules in solution are effective agonists.
Because the latter seems unlikely, this may suggest that the 2″-OH
and 3″-OH groups are hydrogen-bond acceptors, a feature that
could be retained even when acetylated. We have previously shown that
α-1″-O-methyl-ADPR and β-1″-O-methyl-ADPR (Figure ) do not activate TRPM2, but both inhibit channel activation
by ADPR.[29] This indicates that the 1″-hydroxyl
group in ADPR could not be masked or that there is no space in this
region to accommodate even a relatively small methyl group, without
losing agonist activity in contrast to the 2″- and 3″-hydroxyl
groups. So why does 3″-deoxy-ADPR 4 not activate
the channel? Possible explanations may be that the conformation of
the furanose ring or the ratio between the two anomeric forms of the
ribose is affected by the absence of the 3″-hydroxyl group
in a way that prevents the proper hydrogen bonding necessary for channel
activation or that this hydroxyl is required as a hydrogen bond acceptor.The five-membered ribosesugar ring is puckered because of nonbonded
interactions, and the energetically most stable conformation has all
substituents as far apart as possible.[43] Thus, the different substitution patterns in the terminal ribose
of ADPR, 1″-deoxy-ADPR 2, 2″-deoxy-ADPR 3, and 3″-deoxy-ADPR 4, might be expected
to produce differing types of puckering. Extensive analysis of the
ribose and 2′-deoxy-ribose rings in RNA and DNA nucleotides
has shown that while ribose adopts a primarily C3′-endo configuration,
2′-deoxy-ribose adopts a primarily C2′-endo form.[44] Such changes alter the orientation of the ribose
ring substituents, and this may affect the orientation of the substituents
and the way they are presented to the binding site, as well as the
stability of a given ADPR analogue by bringing nucleophilic hydroxyl
groups into proximity for intramolecular attack on the pyrophosphate.
Unfortunately, the broad nature of the peaks in the 1H
NMR spectra of the ADPR analogues means that it is not possible fully
to determine the coupling constants and hence to analyze fully each
terminal ribose ring conformation and suggests that rapid interconversion
between configurations may be taking place in solution.ADPR
is a mixture of terminal ribose α- and β-anomers
and it is not known whether one or both anomers activate TRPM2. Analysis
of 1H NMR spectra demonstrates that both 2″-deoxy-ADPR 3 and 3″-deoxy-ADPR 4 are also mixtures
of the α- and β-forms in the following ratios (α/β,
2″-deoxy-ADPR 3, 1:1.17; 3″-deoxy-ADPR 4; 1:4). Both analogues still adopt cyclic (as opposed to
open chain form) structures for the terminal ribose. Given that both
anomers (α- and β-forms) are present in both cases, it
would seem unlikely that this could be the sole cause of their observed
inactivity.Surprisingly, di-deoxy ADPR analogues 5 and 6 did not have significant antagonist activity
(data not shown).
This was unexpected because the potency of our previous antagonist
8-Ph-ADPR (IC50 = 11 μM) was improved by the 2′-deoxy
modification (8-Ph-2′-deoxy-ADPR IC50 = 3 μM).The recent cryo-EM structures of humanTRPM2 show that in the absence
of ADPR interactions between the NUDT9H domain and
MHR1/2 domains of the same and adjacent subunits lock the channel
in a closed conformation.[21] Binding of
ADPR impacts the conformation of the NUDT9H domain
of TRPM2 in a way that leads to disengagement of the inter-subunit
interactions, priming the channel for full activation by binding of
Ca2+ ions to a cytosolic site near the pore region.[21] The NUDT9H domain of TRPM2
has a bi-lobed structure with the two lobes forming a cleft that is
supposed to bind ADPR. While comparison of the apo and the ADPR bound
state shows a narrowing of this cleft upon binding of ADPR,[21] the low local resolution of the structure in
the NUDT9H domain does not currently allow for localization
of ADPR in this binding pocket or for identification of the interactions,
which would trigger the conformational changes leading to the primed
state. In the absence of a high-resolution structure, SAR data therefore
remain essential for ligand based drug design. Our current data further
support the role of interactions involving the terminal ribose of
ADPR in the gating of TRPM2.Recently, key differences between
hTRPM2 and the invertebrate N. vectensis variant nvTRPM2 were identified. Thus,
two reported synthetic ADPR analogue hTRPM2 antagonists were nvTRPM2
agonists.[45] Moreover, a regulatory function
of NUDT9H in nvTRPM2 opposed to that in hTRPM2 was
uncovered through the action of another synthetic analogue, inosine
5′-diphosphate ribose IDPR. Thus, ADPR analogues, such as those
reported here, are at the cutting edge of progress to unravel the
mechanism of TRPM2 function.
Conclusions
Building
on our previous study that showed the essential role of
the terminal ribose of ADPR in gating of the Ca2+-permeable,
nonselective cation channel TRPM2, we synthesized ribose and 2′-deoxyribose
analogues of ADPR 1–6, each lacking
one of the hydroxyl groups of the terminal ribose. The convergent
synthesis of the targets 1–6 consisted
of two separate steps: (a) synthesis of particular deoxyribose monophosphates 9–12 and (b) coupling analogues 9–12 to the AMP and 2′-deoxy-AMPimidazolides 30–31, following the
Dabrowski-Tumanski procedure.[36] Such an
approach seems to be highly efficient in preparation of such complex
molecules as modified ADPR analogues, offering improved yields in
our hands compared to pyrophosphate formation using AMP-morpholidates
or activation using CDI. While 2″-deoxy-ADPR 3 proved to be unstable, analogues 1, 2, 4, 5, and 6 showed acceptable stability
and were tested electrophysiologically in patch-clamp cell experiments.
Neither 1″-deoxy-ADPR 2 nor 3″-deoxy-ADPR 4 was able to activate TRPM2 significantly; instead, both 2 and 4 were weak antagonists of ADPR-mediated
TRPM2 activation in whole cell experiments, further highlighting the
sensitivity of channel activation to structural changes of the terminal
ribose. Thus, 1″-β-O-Me-2″,3″-O-iPr-ADPR, 1″-deoxy-, and 3″-deoxy-ADPR
(1, 2 and 4) were all antagonists
of ADPR-mediated Ca2+-release in whole cell patch-clamp
experiments. Unlike in our previous observations, where including
an additional 2′-deoxy-modification generated a more potent
TRPM2 antagonist, the corresponding 1″- and 3″, 2′-dideoxy-analogues 5 and 6 were less potent antagonists of TRPM2.
Results further highlight the significance of the ADPR terminal ribose
for activation of TRPM2 and the use of synthetic ADPR analogues in
general. The synthesis of such ADPR analogues as chemical biology
tools using the methods outlined here is proving invaluable in the
wider TRPM2 field.[45] Further analogues
to interrogate individual hydroxyl group stereochemistry or individually
mask each of the three terminal hydroxyl groups to probe hydrogen
bonding interactions at the TRPM2 binding site will undoubtedly shed
more light on the role of the terminal ribose.
Experimental
Section
General
All reagents and solvents were of commercial
quality and were used without further purification, unless described
otherwise. Triethylamine was dried over potassium hydroxide, distilled,
and then stored over potassium hydroxide pellets. H2O was
of MilliQ quality. Unless otherwise stated, all reactions were carried
out under an inert atmosphere of nitrogen. All 1H, 13C, and 31P NMR spectra of the final compounds
were collected, either on a Varian Mercury 400 MHz or Bruker AVANCE
III 500 MHz Spectrometer. Chemical shifts (δ) are reported in
parts per million (ppm) and all 1H and 13C NMR
assignments are based on COSY, HSQC, HMBC, and DEPT experiments. Abbreviations
for splitting patterns are as follows: br, broad; s, singlet; d, doublet;
t, triplet; m, multiplet etc. High-resolution time-of-flight mass
spectra were obtained on a Bruker Daltonics micrOTOF mass spectrometer
using electrospray ionisation. Analytical HPLC analyses were carried
out on a Waters 2695 Alliance module equipped with a Waters 2996 Photodiode
Array Detector (210–350 nm). The chromatographic system consisted
of a Hichrom Guard Column for HPLC and a Phenomenex Synergi 4u MAX-RP
80A column (150 × 4.60 mm), eluted at 1 mL/min with a gradient
of MeCN in 0.05 M TEAB. Semi-preparative HPLC was carried out on a
Waters 2525 pump with manual FlexInject. The chromatographic system
consisted of a Phenomenex Gemini, 5u, C18, 110A column (250 ×
10.00 mm), eluted at 5 mL/min.
2,3-O-Isopropylidene-D-ribofuranose 13 (4 g, 21.1
mmol) was dissolved in pyridine (10 mL). Trityl chloride (7.05 g,
25.3 mmol) was added to the stirred solution, and the mixture was
stirred at 20 °C for 16 h. Water (30 mL) was added, and the mixture
was stirred additional 10 min. The aqueous phase was extracted with
dichloromethane (3×). The combined organic phase was dried with
MgSO4, solids were removed by filtration, and the solvent
was evaporated in vacuo. The crude product was purified by silica
gel chromatography petroleum ether/EtOAc (1:0 → 0:1, v/v).
This procedure afforded the title compound 14 as a white
amorphous solid (5.9 g, 65%). 1H NMR[25] (400 MHz, DMSO-d6): δ
7.42–7.32 and 7.30–7.24 (2 × m, 12 + 3H, o, m, p-Tr), 6.41 (d,
1H, J = 4.0 Hz, OH), 5.15 (d, 1H, J = 4.0 Hz, H-1), 4.56 (dd, 1H, J = 6.0 Hz, J = 1.0 Hz, H-2), 4.36 (d, 1H, J = 6.0
Hz, H-3), 4.11–4.05 (m, 1H, H-4), 3.16–3.05 (m, 2H,
H-5), 1.37 and 1.23 (2 × s, 2 × 3H, CH3iPr). HRMS (ES+) calcd for C27H28NaO5, 455.1829 [M + Na]+; found, 455.1842.
2,3-O-Isopropylidene-5-O-trityl-d-ribitol (15)
2,3-O-Isopropylidene-5-O-trityl-d-ribofuranose 14 (5.2 g,
11.98 mmol) was dissolved in EtOH and cooled to 0 °C. Sodium
borohydride (460 mg, 12.1 mmol) was added to the solution in three
portions over the period of 30 min. The mixture was stirred at 20
°C for 2 h. Water (80 mL) was added, and the solution was carefully
acidified with acetic acid to pH 6. Then, the aqueous phase was extracted
with dichloromethane (3×). The combined organic phase was dried
(Na2SO4), solids were removed by filtration,
and the solvent was evaporated in vacuo. The crude product was purified
by Isco-Flash chromatography using CH2Cl2/acetone
(1:0 → 1:1, v/v). This procedure afforded the title compound 15 as an amorphous white solid (4.35 g, 83%). 1H NMR (400 MHz, DMSO-d6): δ 7.48–7.41
(m, 6H, o-Tr), 7.35–7.28 (m, 6H, m-Tr), 7.26–7.20 (m, 3H, p-Tr), 5.15 (d, J = 5.6 Hz, OH-4), 4.80 (t, J = 5.6 Hz,
OH-1), 4.17–4.10 (m, 1H, H-2), 4.08–4.01 (m, 1H, H-3),
3.78–3.66 (m, 2H, H-4 and H-1a), 3.51–3.43
(m, 1H, H-1b), 3.15–3.11 (m, 1H, H-5a), 3.06–3.01 (m, 1H, H-5b), 1.20 (s, 6H, CH3iPr). 13C{1H}NMR (100
MHz, DMSO-d6): δ 144.0 (3 ×
C-ipso-Tr), 128.4 (6 × C-o-Tr),
127.7 (6 × C-m-Tr), 126.8 (3 × C-p-Tr), 107.4 (C(CH3)), 85.7 (C(Tr)), 78.2 (C-2), 76.4 (C-3), 68.2 (C-4), 66.1 (C-1), 60.0
(C-5), 27.8 and 25.4 (CH3iPr). HRMS (ES+) calcd for C27H30NaO5, 457.1985
[M + Na]+; found, 457.2004.
1-Deoxy-2,3-O-isopropylidene-d-ribofuranose 17 (94 mg, 0.49 mmol) and 5-phenyl-1H-tetrazole (145
mg, 0.99 mmol) were co-evaporated with
dry toluene (2×). Then, the dry mixture was dissolved in dichloromethane
(2 mL), and di-tert-butyl N,N-diisopropylphosphoramidite (234 μL, 0.74 mmol) was
added dropwise. The mixture was stirred at 20 °C for 2 h after
which TLC analysis (petroleum ether/EtOAc, 2:1, v/v) showed the reaction
was complete. It was cooled to 0 °C and triethylamine (411 μL,
2.97 mmol) was added, followed by hydrogen peroxide (35% aq, 109 μL,
1.24 mmol). The resulting mixture was stirred at 20 °C for 1
h. The mixture was diluted with EtOAc (20 mL) and extracted with aqueous
Na2SO3 (10%, w/v). The organic phase was dried
(Na2SO4), solids were filtered off, and the
solvent was evaporated in vacuo. The crude product was purified by
Isco-Flash chromatography using petroleum ether/EtOAc (1:0 →
0:1, v/v), to afford the title compound 18 as a colorless
liquid (74 mg, 41%). 1H NMR (400 MHz, CDCl3):
δ 4.85–4.81 (m, 1H, H-2), 4.76 (dd, 1H, J = 6.4 Hz, J = 1.2 Hz, H-3), 4.23–4.18 (m,
1H, H-4), 4.2–3.94 (m, 4H, H-1, H-5), 1.51 (s, 3H, CH3iPr), 1.48 (s, 18H, CH3tBu), 1.33 (s, 3H, CH3iPr). 31P NMR (160 MHz, CDCl3): δ −10.1 (s, P). 13C{1H}NMR (100 MHz, CDCl3): δ
112.6 (C(CH3)), 83.1 (d, J = 8.0 Hz, C(tBu)), 82.6 (d, J = 7.0 Hz, C-4), 82.3 (C-3), 81.2 (C-2), 73.9 (C-1), 66.6
(d, J = 7.0 Hz, C-5), 29.8–29.7 (m, CH3tBu), 26.5 and 24.9 (CH3iPr)). HRMS (ES+) calcd for C16H31NaO7P, 389.1700 [M + Na]+; found, 389.1699.
1-Deoxy-2,3-O-isopropylidene-D-ribosfuranose-5-O-di-tert-butyl phosphate 18 (36 mg,
0.098 mmol) was dissolved in aqueous methanol (1:1, v/v, 2 mL), and
the solution was cooled to 0 °C. Trifluoroacetic acid (2 mL)
was added dropwise and the solution was allowed to warm up to 20 °C
and stirred for an additional 4 h. The solution was evaporated to
dryness and the residue co-evaporated with water (3×) and with
methanol (3×). The procedure afforded the pure phosphate derivative
as a colorless gum (21 mg, quantitative) which was directly dissolved
in water (2 mL), neutralized with tributylamine (33 μL, 0.14
mmol), the solvent evaporated, and the residue co-evaporated with
EtOH (2×). The title compound 10 was obtained as
the mono-tributylammonium salt (18.7 mg, 49%). 1H NMR (400
MHz, MeOH): δ 4.21–4.16 (m, 1H, H-2), 4.12–4.07
(m, 1H, H-3), 4.03–3.87 (m, 4H, H-1a, 4, 5a,b), 3.69 (dd, 1H, J = 9.6 Hz, J =
4.0 Hz, H-1b), 3.10–3.03 (m, 6H, NCH2), 1.73–1.64 (m, 6H, NCH2CH), 1.41 (m, 6H, N(CH2)2CH), 1.00 (t, 9H, J = 7.6 Hz, CH3Bu). 31P NMR (160 MHz, MeOH):
δ 1.09 (s, P). 13C{1H}NMR (100 MHz, MeOH):
δ 83.3 (d, J = 8.0 Hz, C-4), 73.7 (C-1), 73.6
(C-3), 72.5(C-2), 66.3 (d, J = 5.0 Hz, C-5), 54.0
(NCH2), 27.0 (NCH2CH2), 21.0 (N(CH2)2CH2), 14.0 (CH3Bu). HRMS (ES–) calcd for C5H11O7P, 213.0170 [M
– H]−; found, 213.0162.
2-Deoxy-d-ribofuranose-5-phosphate sodium Salt (19)
Compound 19 (sodium salt) was purchased
from Sigma-Aldrich (UK) and found to be a mixture of α and β
anomers in the ratio 1/1.17 α/β). 1H NMR (500
MHz, D2O): δ 5.52–5.47 (m, 1.6H, H-1α,
β, deuterium exchange), 4.51–4.46 (m, 1H, H-3β),
4.38–4.34 (m, 1H, H-3α), 4.25–4.20 (m, 1H, 4α),
4.04–3.99 (m, 1H, H-4β), 3.85–3.81 (m, 2H, H-5β),
3.81–3.75 (m, 2H, H-5α), 2.51–2.42 (m, 1H, H-2α),
2.22–2.14 (m, 2H, H-2β), 1.91–1.83 (m, 1H, H-2α). 31P NMR (202 MHz, D2O): δ 3.65 (P-β)
and 3.68 (P-α). 13C{1H}NMR (126 MHz, D2O): δ 98.4 (C-1β), 98.0 (C-1α), 84.9 (d, J = 10.08 Hz, C-4β), 84.6 (d, J =
8.82 Hz, C-4α), 71.6 (C-3β), 71.1 (C-3α), 64.6 (d, J = 5.0 Hz, C-5β), 63.7 (d, J = 3.8
Hz, C-5α), 40.8 (C-2β), 40.6 (C-2α). Compound 19 was transformed to its tributylammonium salt.
2-Deoxy-D-ribofuranose-5-phosphate Mono-tributylammonium
Salt (11)
Commercially available 2-deoxy-D-ribofuranose-5-phosphate sodium salt 19 (Sigma-Aldrich)
(8 mg, 0.029 mmol) was dissolved in water (2 mL) and acidified to
pH 1–2 by using the ion-exchange resin Dowex (D50 H+). The resin was filtered off and the free phosphate was neutralized
by addition of one equivalent of tributylamine (6.9 μL, 0.029
mmol). The aqueous solution was freeze-dried (quant.). (NMR analysis
showed partial decomposition (approx. 20–30%), of 2′-deoxyribose.
The compound itself is sensitive to basic conditions (pH 9)). 1H NMR (500 MHz, D2O): δ 5.50–5.47
(m, 0.54H, H-1β), 5.43–5.41 (dd, 0.46H, J = 5.5 Hz, J = 2.5 Hz, H-1α), 4.43–4.39
(m, 0.53H, H-3β), 4.29–4.25 (m, 0.47H, H-3α), 4.15–4.1
(m, 0.46H, H-4α), 4.02–3.97 (m, 0.7H, H-5β), 3.94–3.85
(m, 1.32H, H-4β, 5α, 5β), 3.84–3.77 (m, 0.52H,
H-5α), 3.09–3.01 (m, 10H, NCH(CH2)2CH3), 2.38–2.30
(m, 0.5H, H-2α), 2.12–1.99 (m, 1H, H-2β), 1.86–1.79
(m, 0.5H, H-2α), 1.72–1.60 (m, 10H, NCH2CHCH2CH3),
1.47–1.35 (m, 10H, NCH2CH2CHCH3), 1.04–0.94 (m, 15H,
NCH2(CH2)2CH). 31P NMR (202 MHz, D2O): δ
1.10 (P-β) and 1.03 (P-α). 13C{1H}NMR (126 MHz, D2O): δ 100.0 (C-1β), 99.5
(C-1α), 86.3 (d, J = 8.4 Hz, C-4β), 85.6
(d, J = 8.7 Hz, C-4α), 73.5 (C-3β), 73.0
(C-3α), 67.3 (d, J = 5.4 Hz, C-5β), 66.1
(d, J = 5.3 Hz, C-5α), 53.8 (NCH(CH2)2CH3), 43.0 (C-2β), 42.6 (C-2α), 26.9 (NCH2CHCH2CH3),
21.1 (NCH2CH2CHCH3), 14.0 (NCH2(CH2)2CH).
1,2-O-Isopropylidene-d-xylofuranose
(20)[34]
Finely powdered d-xylose (10 g, 67 mmol) was suspended in acetone (260 mL) containing
sulphuric acid (10 mL) and stirred for 30 min until it was dissolved.
The solution was cooled to 0 °C, and a solution of Na2CO3 (13 g in 112 mL H2O) was added. The mixture
was stirred at 20 °C for 1 h, and then solid Na2CO3 (7 g) was added. The slurry was stirred at 20 °C for
an further 30 min. Solids were removed by filtration and washed with
acetone. Acetone was evaporated, and the water phase was extracted
with EtOAc (3×) and dried with Na2SO4.
Isco-Flash chromatography using petroleum ether/EtOAc (1:0 →
1:1, v/v) afforded the title compound 20 as a colorless
viscous oil (4.35 g, 34%) and doubly protected 1, 2-O-isopropylidene-3,5-O-isopropylidene-d-xylofuranose
(3.29 g, 21%).[34]1H NMR (400
MHz, CDCl3): δ 5.98 (d, 1H, J =
3.6 Hz, H-1), 4.52 (d, 1H, J = 3.6 Hz, H-2), 4.32
(d, 1H, J = 2.4 Hz, H-4), 4.19–4.11 (m, 2H,
H-3, 5a), 4.05 (dd, 1H, J = 12.4 Hz, J = 2.4 Hz, H-5b), 1.48 and 1.32 (CH3iPr). 13C{1H}NMR (100 MHz,
CDCl3): δ 112.0 (C(CH3)2), 105.1 (C-1), 85.9 (C-2), 78.7 (C-3), 77.3 (C-4), 61.5 (C-5),
26.9 and 26.3 (CH3iPr). HRMS (ES+) calcd for C8H14NaO5, 213.0733
[M + Na]+; found, 213.0787.
AMP sodium salt (50 mg, 0.128 mmol, Sigma-Aldrich)
was dissolved in water (2 mL) and treated with Dowex D50 (H+) until all the phosphate was transformed to free acid (pH 1–2).
Then, the resin was filtered off and washed with water, and the solution
of AMP (free acid) was titrated with tributylamine (30 μL, 0.13
mmol, 1 equiv, pH 7). The solution of adenosine-5′-phosphatemono-tributylammonium salt (28) was freeze-dried to obtain
a light white powder (quantitative).AMP tributylammonium salt
(15 mg, 0.028 mmol) and imidazole (19 mg, 0.28 mmol) were co-evaporated
with EtOH (2×) and with toluene (2×). Aldrithiol (18.6 mg,
0.085 mmol) was added followed by DMF (220 μL), and the suspension
stirred until all solid matter was dissolved. Triethylamine (16 μL,
0.113 mmol, distilled-dry-stored over KOH in the dark) and triphenylphosphine
(22.2 mg, 0.085 mmol) were added, and the solution was stirred at
20 °C for 16 h. Product precipitation: The reaction mixture was
cooled to 5 °C and a cold solution of NaI (33.8 mg) in dry acetone
(2.2 mL) was added at 5 °C. The white precipitate was filtered
off and washed with cold acetone. The product was dried under high
vacuum for 16 h and stored under argon (crude compound, hygroscopic,
yield was not quantified). 1H NMR (400 MHz, CD3OD): δ 8.38 (s, H-8), 8.19 (s, H-2), 7.88–7.87 (m, 1H,
imidazole), 7.28–7.26 (m, 1H, imidazole), 6.99–6.97
(m, 1H, imidiazole), 6.04 (d, 1H, J = 6.0 Hz, H-1′),
4.65 (t, 1H, J = 5.6 Hz, H-2′), 4.25 (dd,
1H, J = 5.2 Hz, J = 3.6 Hz, H-3′),
4.18–4.13 (m, 1H, H-4′), 4.06–3.96 (m, 2H, H-5a,b′). 31P NMR (160 MHz, CD3OD): δ −9.03 (s,
P). 13C{1H}NMR (100 MHz, CD3OD):
δ 157.4 (C-6), 153.9 (C-2), 150.9 (C-4), 140.9 (C-8), 140.8
(d, J = 5.5 Hz, imidazole), 129.7 (d, J = 10.4 Hz, imidazole), 121.3 (d, J = 5.7 Hz, imidazole),
120.3 (C-5), 89.2 (C-1′), 85.3 (d, J = 8.8
Hz, C-4′), 75.8 (C-2′), 72.19 (C-3′), 66.8 (d, J = 5.9 Hz, C-5′). HRMS (ES–) calcd
for C13H15N7O6P, 396.0827
[M – H]−; found, 396.0853.
2′-Deoxy-AMP was treated
in the same way as AMP to generate the mono Bu3N salt (29).2′-Deoxy-AMPmono-tributylammonium salt 29 (145 mg, 0.28 mmol) and imidazole (191 mg, 2.8 mmol) were
put into the flask and co-evaporated with EtOH (2×) and with
toluene (2×). Aldrithiol (186 mg, 0.84 mmol) was added to the
flask followed by dry DMF (1.5 mL) and stirred until all solid matter
was dissolved. Triethylamine (156 μL, 1.13 mmol, distilled-dry-stored
over KOH in the dark) and triphenylphosphine (221 mg, 0.85 mmol) were
added to the mixture, and the mixture was stirred at 20 °C for
16 h. The reaction mixture was cooled to 5 °C, and a cold solution
of NaI (337 mg, 2.25 mmol) in dry acetone (15 mL) was added at 5 °C.
The white precipitate was filtered off and washed with cold acetone.
The product was dried under high vacuum and stored under argon (crude
compound, hygroscopic, yield was not quantified). 1H NMR
(400 MHz, CD3OD): δ 8.36 (s, H-8), 8.18 (s, H-2),
7.87–7.86 (m, 1H, imidazole), 7.26–7.24 (m, 1H, imidazole),
6.98–6.96 (m, 1H, imidazole), 6.44 (dd, 1H, J = 7.6 Hz, J = 1.2 Hz, H-1′), 4.49 (q, 1H, J = 2.8 Hz, H-3′), 4.08–4.04 (m, 1H, H-4′),
3.99–3.95 (m, 2H, H-5′), 2.75 (ddd, 1H, J = 13.6 Hz, J = 7.6 Hz, J = 1.6
Hz, H-2a′), 2.42 (dq, 1H, J = 13.6 Hz, J = 2.8 Hz, H-2b′). 31P NMR (160 MHz, CD3OD): δ −9.00
(s, P). 13C{1H}NMR (100 MHz, CD3OD):
δ 157.3 (C-6), 153.8 (C-2), 150.5 (C-4), 140.9 (C-8), 140.7
(d, J = 5.5 Hz, imidazole), 129.6 (d, J = 10.5 Hz, imidazole), 121.3 (d, J = 5.7 Hz, imidazole),
120.3 (C-5), 87.5 (d, J = 8.7 Hz, C-4′), 85.5
(C-1′), 72.9 (C-3′), 67.0 (d, J = 5.9
Hz, C-5′), 41.2 (C-2′). HRMS (ES–)
calcd for C13H15N7O5P,
380.0878 [M – H]−; found, 380.0883.
3-Deoxy-d-ribofuranose-5-phosphate tributylammonium
salt 12 (12 mg, 0.03 mmol) and MgCl2 (5.9
mg, 0.06 mmol) were dissolved in DMF (1 mL), co-evaporated with dry
toluene (2×), and evaporated to dryness. Dry DMF (600 μL)
was added to the solid residue, and the mixture was stirred at 20
°C until all solids were dissolved (5–10 min). Then, 2′-deoxy-AMP-imidazolide 31 (10.8 mg, 0.027 mmol) was added dropwise in DMF (400 μL).
The mixture was stirred at 20 °C for 16 h. HPLC showed consumption
of all starting material. The solvent was evaporated in vacuo and
the crude product was dissolved in TEAB buffer (10 mL, 0.1 M) and
purified by semi-preparative, reverse-phase HPLC using a gradient
of TEAB (0.1 M)–acetonitrile (95/5 → 35/65, v/v). The
title compound 6 was obtained as a colorless amorphous
solid (7.5 μmol, 5.32 mg, 27%, 1.8 × Et3N salt). 1H NMR (500 MHz, D2O): δ 8.48 (s, 1H, H-8),
8.24 (s, 1H, H-2), 6.51 (t, 1H, J = 7.0 Hz, H-1′),
5.26 (d, 0.3H, J = 4.0 Hz, H-1″α), 5.21
(s, 0.7H, H-1″β), 4.78–4.75 (m, 1H, H-3′,
obstructed by water signal), 4.49–4.39 (m, 1H, H-4″α,
β), 4.31–4.23 (m, 1.3H, H-2″α, 4′),
4.20–4.11 (m, 2.7H, H-2″β, 5a,b′), 4.05–3.98 (m,
0.7H, H-5″βa), 3.94–3.84 (m, 1H, H-5″αa, βb), 3.82–3.76 (m, 0.3H, H-5″αb), 3.19 (q, 11H, J = 7.5 Hz, NCH), 2.84 (ddd, 1H, J = 7.5
Hz, J = 6.5 Hz, J = 1.0 Hz, H-2a′), 2.63–2.57
(m, 1H, H-2b′), 2.10–2.01 (m, 0.3H, H-3″αa), 2.00–1.92
(m, 1.7H, H-3″αb, βa,b),
1.27 (t, 16.5H, J = 7.5 Hz, NCH2CH). 31P NMR (202 MHz,
D2O): δ −11.12 (m, α-pyrophosphate),
−11.38 (m, β-pyrophosphate). 13C{1H}NMR (126 MHz, D2O): δ 155.6 (C-6), 152.7 (C-2),
148.7 (C-4), 139.9 (C-8), 118.6 (C-5), 102.1 (C-1″β),
97.6 (C-1″α), 85.8 (d, J = 8.8 Hz, C-4′),
83.6 (C-1′), 78.3 (d, J = 8.2 Hz, C-4″β),
75.5 (d, J = 8.4, C-4″α), 75.4 (C-2″β),
71.3 (C-3′), 70.4 (C-2″α), 68.8 (d, J = 5.4 Hz, C-5″β), 67.9 (d, J = 5.5
Hz, C-5″α), 65.4 (d, J = 4.9 Hz, C-5′),
46.6 (NCH), 39.0 (C-2′),
32.4 (C-3′β), 31.7 (C-3′α), 8.2 (NCH2CH). HRMS (ES–) calcd for C15H22N5O12P2, 526.0746 [M – H]−; 526.0761.
HPLC tR = 3.68 min.
Pharmacology
Materials
ADPR was obtained from Sigma-Aldrich.
Cell Culture
HEK293
cells were kept at 37 °C and
5% CO2 in complete medium (DMEM with 4.5 g/L glucose, Glutamax-I,
10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin)
at 37 °C. For the maintenance of HEK293 clones with stable expression
of TRPM2 400 μg/mL G418 sulfate was added to the medium.
Transfection
and Generation of Cell Lines
Generation
of the clonal HEK293 cell line with stable expression of humanTRPM2
has been described previously.[16] In brief,
HEK293 were transfected with an expression vector encoding the full-length
of humanTRPM2 and EGFP (pIRES2-EGFP-TRPM2). Cells that successfully
integrated the expression vector were then enriched by selection with
400 μg/mL G418 sulfate (Biochrom). Clonal cell lines were established
from these cells by limiting dilution. Expression of TRPM2 was confirmed
by Ca2+ measurement and whole cell patch clamp.
Patch-Clamp
Measurements
The day before the experiments,
cells from a clonal HEK293 cell line with stable expression of humanTRPM2 were seeded to 35 mm tissue culture dishes at a low density.
For the whole cell patch-clamp experiments, the culture medium was
replaced by a bath solution containing 1 mM CaCl2, 140
mM NMDG, 5 mM KCl, 3.3 mM MgCl2, 1 mM CaCl2,
5 mM d-glucose, and 10 mM HEPES, pH 7.4. Patch pipettes were
pulled from borosilicate glass capillaries with an outer diameter
of 1.5 mm and an inner diameter of 1.05 mm using a Sutter P-97 and
filled with a pipette solution containing 120 mM KCl, 8 mM NaCl, 1
mM MgCl2, 10 mM HEPES, 10 mM EGTA, and 5.6 mM CaCl2 (resulting in 200 nM free [Ca2+] as calculated
by CaBuf software (G. Droogmans, formerly available from ftp://ftp.cc.kuleuven.ac.be/pub/droogmans/cabuf.zip). These pipettes had a resistance between 1.5 and 3.5 MΩ.
Data were recorded using an EPC-10 amplifier and PatchMaster software
(HEKA Elektronik, Germany). Fast and slow capacity transients were
compensated, series resistance was compensated by 70%. After establishing
whole cell configuration, cells were held at −50 mV and channel
activation was followed by applying voltage ramps from −85
to +20 mV over 140 ms every 5 s for a total of 450 s. For further
analysis, the maximum outward current at +15 mV during the course
of the recording was taken as a measure of channel activity. To test
for agonist activity, ADPR analogues were included in the pipette
solution at 100 μM. To test for antagonist activity, the pipette
solution contained 100 μM ADPR and 900 μM of the ADPR
analogue under test. All experiments were performed at room temperature.
Statistical analysis
The data from patch-clamp experiments
were analyzed using GraphPad Prism (version 7.04, GraphPad Software
Inc.) Because the distribution of currents is skewed toward higher
values, data were log-transformed. Log-transformed data were tested
for significant differences using one-way-Anova followed by post hoc
testing against the respective control (buffer for agonist experiments,
ADPR for antagonist experiments) using Bonferroni correction for multiple
testing (α = 0.05). In the charts, the horizontal bar indicates
the mean of the log-transformed data.
Authors: Wenyi Zhang; Iwona Hirschler-Laszkiewicz; Qin Tong; Kathleen Conrad; Shao-Cong Sun; Linda Penn; Dwayne L Barber; Richard Stahl; David J Carey; Joseph Y Cheung; Barbara A Miller Journal: Am J Physiol Cell Physiol Date: 2005-11-23 Impact factor: 4.249
Authors: Elena Fonfria; Cesar Mattei; Kerstin Hill; Jon T Brown; Andrew Randall; Christopher D Benham; Stephen D Skaper; Colin A Campbell; Barry Crook; Paul R Murdock; Jennifer M Wilson; Frank P Maurio; Davina E Owen; Paula L Tilling; Shaun McNulty Journal: J Recept Signal Transduct Res Date: 2006 Impact factor: 2.092
Authors: Olivera Grubisha; Louise A Rafty; Christina L Takanishi; Xiaojie Xu; Lei Tong; Anne-Laure Perraud; Andrew M Scharenberg; John M Denu Journal: J Biol Chem Date: 2006-03-24 Impact factor: 5.157
Authors: Anne-Laure Perraud; Betty Shen; Christopher A Dunn; Karsten Rippe; Megan K Smith; Maurice J Bessman; Barry L Stoddard; Andrew M Scharenberg Journal: J Biol Chem Date: 2002-11-08 Impact factor: 5.157
Authors: A L Perraud; A Fleig; C A Dunn; L A Bagley; P Launay; C Schmitz; A J Stokes; Q Zhu; M J Bessman; R Penner; J P Kinet; A M Scharenberg Journal: Nature Date: 2001-05-31 Impact factor: 49.962
Authors: Elena Fonfria; Ian C B Marshall; Christopher D Benham; Izzy Boyfield; Jason D Brown; Kerstin Hill; Jane P Hughes; Stephen D Skaper; Shaun McNulty Journal: Br J Pharmacol Date: 2004-08-09 Impact factor: 8.739
Authors: Ondřej Baszczyňski; Joanna M Watt; Monika D Rozewitz; Ralf Fliegert; Andreas H Guse; Barry V L Potter Journal: RSC Adv Date: 2020-01-09 Impact factor: 4.036
Figure 1
Figure 2
Scheme 1
Scheme 2
Scheme 3
Figure 3
Figure 4