Received: 22 October 2019 Revised: 30 December 2019 Accepted: 2 January 2020

DOI: 10.1002/mrc.4990

SPECIAL ISSUE RESEARCH ARTICLE

What is the form of muscimol from fly agaric mushroom

(Amanita muscaria) in water? An insight from NMR
experiment supported by molecular modeling

Teobald Kupka | Małgorzata A. Broda | Piotr P. Wieczorek

Faculty of Chemistry, University of Opole,

Opole, Poland
Abstract
The biologically active alkaloid muscimol is present in fly agaric mushroom
Correspondence (Amanita muscaria), and its structure and action is related to human neuro-
Teobald Kupka and Małgorzata A. Broda,
Faculty of Chemistry, University of Opole, transmitter γ-aminobutyric acid (GABA). The current study reports on deter-
48,Oleska Street, 45-052 Opole, Poland. mination of muscimol form present in water solution using multinuclear 1H
Email: [email protected], broda@uni.
and 13C nuclear magnetic resonance (NMR) experiments supported by density
opole
functional theory molecular modeling. The structures of three forms of free
muscimol molecule both in the gas phase and in the presence of water solvent,
modeled by polarized continuous model, and nuclear magnetic isotropic
shieldings, the corresponding chemical shifts, and indirect spin–spin coupling
constants were calculated. Several J-couplings observed in proton and carbon
NMR spectra, not available before, are reported. The obtained experimental
spectra, supported by theoretical calculations, favor the zwitterion form of
muscimol in water. This structure differs from NH isomer, previously deter-
mined in dimethyl sulfoxide (DMSO) solution. In addition, positions of signals
C3 and C5 are reversed in both solvents.

KEYWORDS
GIAO NMR, indirect spin-spin coupling constants, molecular modeling, muscimol, NMR in
water

1 | INTRODUCTION In Scheme 1, three potential forms of muscimol

together with the corresponding atom labeling are
Muscimol is an interesting small heterocyclic compound shown.
originating from the decomposition of ibotenic acid[1] In analogy to amino acids, muscimol contains two
which is present in widely spread in Europe fly agaric possible groups able to donate/accept proton: –OH and –
mushroom.[1–3] Our recent studies concentrated on NH2 with pKa values of 4.8 and 8.4.[7] These values are
ibotenic acid and muscimol, as well as detection and deter- somehow larger in thiomuscimol[7] (6.1 and 8.9). Taking
mination of trace amounts of these alkaloids.[4,5] Its close into an account the equilibrium present in acid/base
relation to GABA and significant impact on human central solutions,[8] in physiological conditions in water at pH
nervous system[6] indicates its importance in our life for near 7, one could expect a situation closer to pKa = 8.4,
ages.[1] Therefore, it is worth reminding the readers its with protonated –NH3+ and deprotonated hydroxyl
unique and imposing appearance in the forest (Figure 1). group –O−.

Magn Reson Chem. 2020;1–10. wileyonlinelibrary.com/journal/mrc © 2020 John Wiley & Sons, Ltd. 1
2 KUPKA ET AL.

independent atomic orbital approach[14–16] allows rea-

sonable prediction of isotropic nuclear magnetic
shieldings and the corresponding theoretical chemical
shifts.[17–20] Somehow more difficult calculations pro-
vide theoretical indirect spin–spin coupling constants
(SSCCs).[21–23] In case of real-size molecules density
functional theory[24,25] is widely accepted due to its
good accuracy-to-computational expense ratio. Among
a huge variety of density functionals, the popular
hybrid Becke three-parameter Lee, Yang, and Parr
one[26–28] (B3LYP) is still in use.[29] In addition, B3LYP
functional has been shown fairly accurate in predicting
various molecular parameters,[30,31] including nuclear
magnetic shieldings and chemical shifts.[17,32] Thus, in
a benchmark study, a fairly good performance of
B3LYP among over 40 other density functionals in
predicting gas-phase shieldings of nine small molecules
in the complete basis set limit was observed.[17]
Recently, we also published several studies using
B3LYP functional for prediction of NMR parameters in
FIGURE 1 Amanita muscaria in its natural environment
some small and middle-size molecules.[33–38]
However, one should be aware that in case of some
Its crystal structure[9] proves the presence of a zwitter- “exotic” molecular systems, B3LYP-predicted chemical
ion form (a) of muscimol in the solid state. This structure shifts and coupling constants could be less
was also supported by low level ab initio calculations.[9,10] accurate.[38,39]
The possibility of the existence of this GABA agonist as Thus, in the current study, we are interested in find-
neutral (hydroxylated) compound (b) and in the NH form ing an answer to the question what is the form of
(c) in the gas phase and DMSO was studied theoretically in muscimol in water? Is it the same as in polar DMSO solu-
our recent work.[11] In addition, our detailed experimental tion[11] or in crystal[9]? Besides, is NMR spectroscopy
1
H and 13C nuclear magnetic resonance (NMR) studies in alone (e.g., without hints from other techniques) suffi-
DMSO also favored its NH form.[11] However, no experi- cient to provide these answers, or do we need to support
mental NMR studies in water (D2O) were completed arbi- experiment with molecular modeling?
trarily, assuming similar interaction of both polar solvents These questions are not yet answered in the literature
with solute molecules.[11] Besides, no special attention was but seem to be important for understanding its mecha-
paid to routine shimming and a possibility of determina- nism of action at atomistic level. The gained knowledge
tion of very small H–H coupling constants.[11] could help in the design and preparation of new and effi-
NMR spectroscopy[12] is considered the most power- cient acetylcholinesterase inhibitors.[1,9]
ful technique of structure elucidation for organic com- In these combined experimental NMR and theoretical
pounds. On the other hand, in some difficult cases, studies, we will attempt to answer the above mentioned
interpretation of structure in the gas phase and solu- questions by measuring 1H and 13C NMR chemical shifts,
tion, as well as NMR spectral assignment, could be proton–proton, and carbon–proton coupling constants in
supported by reliable molecular modeling studies. Such water and critically compare the measured and calcu-
works start with full structure optimization, followed lated parameters with the previously reported data in
by vibrational analysis.[13] In many cases, gauge- DMSO.[11]

S C H E M E 1 Atom numbering in
zwitterion (a), neutral (b), and NH (c) forms
of muscimol
KUPKA ET AL. 3

2 | E X P E R I M E T A L AN D Grimme D3 dispersion term[42] and pc-3 basis set.[43,44]

C O M P U T A T I O N A L DE T A I L S Polarization-consistent basis sets of Jensen[43,44] are care-
fully designed and considered as very efficient and accu-
2.1 | NMR experiment rate for modeling the structure of organic molecules. In
addition, the use of dispersion correction[42] could help in
Very small amounts of muscimol, from 1 to 3 mg (Abcam more accurate positioning of both side substituents with
Biochemicals, Cambridge) were dissolved in 0.7-ml 99.8% respect to main ring. Polarized continuum model[45]
D2O in 5 mm outer diameter NMR tubes. All high-resolu- (PCM) was used—to approximately mimics the presence
tion 1-D 1H and 13C NMR measurements were conducted of water solvent. We are aware that PCM is a crude
at room temperature (20
C) using Bruker Ultrashield approximation of solvent–solute interaction but is in
Avance 400 MHz NMR spectrometer and standard TOP- common use as a trade-off between accuracy and compu-
SPIN 1.3 software for both data acquisition and tational cost.[35] Very large grid (INT [GRID = 150590])
processing. The corresponding proton and carbon fre- and very tight optimization criteria were used. The subse-
quencies were 400.1316 and 100.6218 MHz, respectively. quent frequency calculations produced all positive vibra-
Typical convention for proton decoupled {−1H} and tions assuring the presence of minimum energy
coupled {+1H} carbon-13 NMR spectra notation was used structures.
(13C{+1H} and 13C{−1H} NMR). NMR parameters were modeled using a fairly com-
Because tetramethylsilane (TMS) is not soluble in plete and flexible Pople-type basis set: 6-311++G
water, all chemical shifts were reported with respect to (3df,2pd). Gauge-independent atomic orbital
sodium trimethylsilylpropanesulfonate (DSS). In addi- approach[14–16] was used to calculate isotropic nuclear
tion, a small drop of p-dioxane was used as the secondary magnetic shieldings of muscimol in the gas phase and
reference in 13C{−1H} NMR spectra.[40] First, larger water at B3LYP/6-311++G(3df,2pd) level of theory
amount of DSS, producing a visible signal, should be using previously optimized B3LYPD3/pc-3 structures in
used as reference in carbon spectra. However, its –SO3− the gas phase and water. Benzene was used as refer-
group could somehow influence the presence of different ence of theoretical chemical shift (the same level of
muscimol forms. Second, dioxane is a small and neutral theory was applied to obtain the corresponding ben-
organic molecule showing a strong single line. Thus, it's zene proton and carbon nuclear isotropic shieldings).
very small 1H NMR linewidth, achieved during careful In case of benzene, the corresponding theoretical 13C
shimming, was helpful for verifying and measuring very and 1H nuclear shieldings, σ(C)o and σ(H)o, in the gas
small proton–proton SSCCs. phase are 49.732 and 24.112 ppm, respectively. These
For better digital resolution, long acquisition times values in water are negligibly smaller (49.192 and
were used for recording proton and carbon spectra 23.937 ppm). The experimental carbon and proton
(16 and 4 s, respectively). The magnetic field homogene- chemical shifts δ(C)o and δ(H)o of benzene in solution
ity tuning (shimming) in the sample was performed with are 128.5 and 7.21 ppm, respectively. This way of
great care and proton linewidth of 0.17 Hz (!) was chemical shift calculation should somehow cancel the
observed for dioxane in water. No line broadening systematic errors. The theoretical carbon and proton
(LB) was applied for accurate reading of proton–proton chemical shifts (in ppm) for nuclei “i” were calculated
coupling constants (LB = 0). However, LB = 1.0 Hz was as follows:
used for better accuracy of carbon–proton SSCCs reading
from noisy 13C{+1H} spectra. In addition, in the pres- δðCÞi = σðCÞo − σ ðCÞi + 128:5: ð1Þ
ented spectra peaks were labeled as “s” (singlet), “d”
(doublet), “t” (triplet), and “m” (multiplet), and the δðHÞi = σðHÞo −σ ðHÞi + 7:21: ð2Þ
corresponding composite patterns were marked as “dt”
(doublet of triplets) or “td” (triplet of doublets).
Final B3LYP/6-311++G(3df,2pd) calculations were
carried out with option NMR = mixed, allowing basis set
2.2 | Theoretical approach uncontraction for more accurately derived indirect
SSCCs. Here we will only analyze the proton–proton and
All calculations were performed with Gaussian 16 B.01 carbon–proton SSCC parameters visible in the experi-
program package[41] running in parallel on 12 processors. mental spectra of muscimol in D2O.
Initial structures of free molecule in its three forms (see To assess the suitability of the muscimol model, devi-
Scheme 1) were created with GaussView and then opti- ations between all predicted and observed “n” (Theor.–
mized using B3LYP hybrid density functional with Exp.) values were calculated, and the final choice was
4 KUPKA ET AL.

made on the basis of the lowest root-mean-square (RMS) is longer by about 0.03 Å than the corresponding double
deviation calculated as follows: bond in NH form. On the other hand, a typical single
bond C3–N2 in NH form of muscimol in water is about
vhffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffii
u 
u x − x 2 + …x −x 2 1.40 Å and is about 0.09 Å longer than the double
t 1 1exp n nexp C3═N2 bond, calculated for a neutral form. Thus, in case
RMS = : ð3Þ
n of zwitterion, an intermediate value between single and
double C–N bond is calculated (1.35 Å). This indicates a
possibility of partial delocalization of negative charge in
3 | R ES U L T S A N D D I S C U S S I O N zwitterion form of muscimol between O6 and N2 atoms
(it resembles a carboxylate group).
3.1 | Structural and energetic aspects of A brief comparison of relative energies of zwitterion
zwitterion form of muscimol and NH forms of muscimol molecule, optimized at
B3LYP/pc3 level of theory, indicates that its first form is
At the very beginning, we decided to consider some geo- significantly destabilized with respect to neutral molecule
metrical aspects of muscimol in its zwitterion form (by 52.16 kcal/mol). The NH form is only slightly des-
related to a potential balance between its two resonance tabilized in the gas phase (by 1.73 kcal/mol). However,
structures by reanalyzing some of previously published the presence of water within PCM model results in a sig-
data[11]: (–)O–C═N– and O═C–N(–). The possible (and nificant stabilization of zwitterion (higher in energy by
predominant) resonance structures of the zwitterion are 11.03 kcal/mol) and NH form (higher energy by
deduced from C4 to O6 and from C3 to N2 distances, 0.26 kcal/mol). Obviously, we are aware that the PCM
reported in our previous study.[11] Thus, for brevity, in method introduces only a crude approximation of sol-
Table S1 in the supporting information, the B3LYP/6-311 vent–solute interaction, and more elaborated approaches,
++G(3df,2dp)-calculated C4–O6 and C3–N2 distances for containing implicit solvent molecules, could produce
the three studied forms in the gas phase, DMSO, and more accurate results. However, in the current short
water are gathered. It is apparent from Table S1 that neu- study, we will not attempt to calculate such time-con-
tral and NH forms of muscimol contain single and double suming results.
bonds C–O and C═O, respectively. These values in the
gas phase are about 0.02 Å shorter than in polar solvents.
Moreover, the C3–O6 distance in the gas phase is about
0.1 Å shorter in case of zwitterion with respect to typical 3.2 | NMR experiment
single C–O bond, calculated for the neutral form of
muscimol. However, the presence of water decreases this A typical 1H NMR spectrum of very small amount of
difference to 0.08 Å. Thus, the C3–O6 bond in zwitterion about 2 mg of muscimol in D2O is shown in Figure 2.

1
FIGURE 2 H NMR spectrum of diluted
solution of muscimol in D2O with clearly visible
peaks originating from dioxane (D) and
remaining nondeuterated solvent (HOD)
KUPKA ET AL. 5

Dioxane (D) at 3.749 ppm and the remaining HOD TABLE 1 Experimental 1H and 13C NMR chemical shiftsa of
solvent peak at 4.827 ppm are also clearly shown in Fig- muscimol in D2O and DMSO-d6
ure 2. Two insets present enlarged fragments of H(4) trip- Signal Chemical Shift (ppm)
let at about 5.808 ppm and CH2 doublet at about 1
H NMR
4.138 ppm, both originating from dissolved muscimol. In
addition, no NH protons at lower magnetic field (in the DMSOb Water Lit.c (water)
range of 10 to 15 ppm, not shown) were noticed due to C(4)H 5.823 (s) 5.808 (t) 5.85
fast exchange with D2O. Positions of multiplet peaks are C(7)H2 3.629 (s) 4.138 (d) 4.18
also given in Hz to show the approximate magnitude of HOD — 4.827 —
the spin couplings. The multiplets are partly overlapped Dioxane — 3.749 —
due to fairly small 4JHH coupling constant (0.68 Hz !).
DSS — 0.000 —
Besides, the relative intensity ratio (1:2) of proton H(4) to 13
C NMR
CH2 is apparent from the corresponding signal integral
intensities (1.00:1.97). Overall, this spectrum seems to be C3 170.38 (d) 180.230 (d) 178.1
very easy to understand but unfortunately is not suffi- C5 175.34 (m) 166.436 (dt) 164.4
cient to assign to only one of the structures, shown in C(4)H 92.38 (dt) 102.025 (dt) 100.1
Scheme 1. C(7)H2 38.16 (t) 37.758 (td) 35.8
The corresponding 13C{−1H} spectrum of muscimol in
Dioxane — 69.287 —
D2O with strong signal of dioxane, used as secondary ref-
DSS — 0.000 —
erence, is presented in Figure 3. In addition, Figure S1
shows multiplets resulting from carbons coupled with Abbreviation: DSS, sodium trimethylsilylpropanesulfonate.
a
The remaining water (HOD) and dioxane signals are also included (no
protons, recorded in the subsequent 13C{+1H}
correction of −2.01 ppm applied to the raw 13C NMR data yet).
experiment. b
From Kupka and Wieczorek.[11]
In Table 1, the 1H and 13C NMR chemical shifts of c
Old experimental data from Jäger and Frey.[46]
muscimol in water with respect to DSS (dioxane was used
as the secondary reference) are gathered. It is interesting
to notice a change of C3 and C5 chemical shifts in differ- chemical shifts indicate a significant change of structure
ent solvents. Thus, in DMSO, the two most deshielded (and possible a charge distribution) near these carbon
carbon signals are C5 at 175.54 ppm and C3 at atoms. Thus, this observation alone suggests a presence
170.38 ppm, respectively. A reverse order is observed in of muscimol in different forms in two strongly polar sol-
water: C3 at 180.230 ppm and C5 at 166.436 ppm, respec- vents, so often used in studies of biological systems.
tively. This assignment is based on 2D heteronuclear TMS is not soluble in water, and referencing NMR
multiple bond correlation spectra (not discussed, see samples in aqueous media is not so simple as in organic
Figure S2) and carbon–proton couplings discussed later. solvents. In a separate experiment, dioxane was
However, one should realize that such huge change of referenced to DSS signal set at 0.000 ppm. Next, the

13
FIGURE 3 C{−1H} NMR spectrum of
diluted solution of muscimol in D2O with clearly
visible peak originating from dioxane (D).
Position of dioxane is referenced to sodium
trimethylsilylpropanesulfonate at 0 ppm
6 KUPKA ET AL.

observed 13C chemical shifts of muscimol in D2O were T A B L E 2 Experimental proton–proton and carbon–proton
referenced to dioxane. In this case, dioxane was used as coupling constants of muscimol in D2O compared with earlier
secondary chemical shift standards, respectively. How- literature dataa, recorded in DMSO-d6
ever, experimental chemical shifts of TMS and DSS could Type SSCC (Hz)
differ. According to Zhang,[47] the difference between
TMS and DSS in water is not negligible (−2.647 ppm). On DMSO-d6 D 2O
the other hand, Hoffman[48] suggests the difference of 1
H NMR spectrum
−2.74 ppm. However, according to Harris et al,[49] the 4
JH(4)H(7) 0.83 0.68
DSS signal is shifted by −2.01 ppm with respect to exter- 13 1
C{+ H} NMR spectrum
nal TMS. This nicely explains a difference of about 2 ppm 1
JC(4)H 182.19 180.84
in carbon chemical shifts of muscimol, reported earlier 1
JC(7)H 136.53 144.43
and observed in the current study (see third and fourth
2
columns of Table 1). In fact, the chemical shift of dioxane JC(5)H(4) 4.93 9.73
2
observed in our 13C NMR spectra is 69.287 ppm and in JC(5)H(7) 3.03 5.78
the corrected spectra (using a value of −2.01 ppm) is 2
JC(3)H(4) 2.25 3.12
67.277 ppm. The latter number is very close to the value 3
JC(4)H(7) 2.80 2.60
of 67.19 ppm, reported in a very popular collection of 3
JC(7)H(4) — 0.78
residual solvent peaks.[50] Therefore, the subsequent
Abbreviations: NMR, nuclear magnetic resonance; SSCC, spin–spin coupling
experimental carbon chemical shifts of muscimol in
constant.
water includes the correction of −2.01 ppm. It is also a
Earlier data in DMSO-d6 are from Kupka and Wieczorek[11]; however, the
important to notice a large difference (about 0.5 ppm) for previously reported spectra were processed again to make possible
proton chemical shift of methylene group in both sol- determination of all small coupling constants, and no line broadening was
applied.
vents (position of H4 proton is essentially unchanged).
For brevity, in Figure S1, enlarged fragments of
muscimol carbon signals splitted by protons are shown. sensitive to the presence of water are oxygen and nitro-
Due to one-bond coupling, the signal originating from gen shieldings of zwitterions (not measured in our experi-
CH2 fragment at 37.758 ppm is splitted into a triplet of ment). For example, the shielding of O(6) changes from
doublets, and a doublet of triplets is visible at about 69.774 ppm in vacuum to 138.908 ppm in water. Simi-
100 ppm. A significantly weaker signal, splitted into a larly, isotropic shielding of N(2) increases from −167.699
doublet with a small coupling constant, is visible at about to −115.098 ppm. Carbon and proton signals are signifi-
160 ppm. The most unshielded carbon signal at about cantly less sensitive to the presence of water.
180 ppm is composed of triplet of doublets. For brevity, a comparison of theoretically predicted
As in case of proton spectra, assignment of these reso- proton and carbon chemical shifts of the three forms of
nances is not difficult. The analysis of proton–carbon free muscimol in vacuum and in water with experimental
connectivity was additionally supported by 2-D heter- data in D2O is shown in Table S2. However, a suitability
onuclear multiple bond correlation experiment (not of a selected muscimol model can be better assessed by
shown and not discussed here). checking the size of deviation between theoretically
All observed H–H and C–H coupling constants are predicted proton and carbon chemical shifts and the
shown in Table 2. It is surprising to see significant experimental values for muscimol in D2O, as well as the
changes in two-bond carbon–proton couplings, measured corresponding RMS deviations (see Table 4).
in two, highly polar solvents. In fact, 2JC(5)H(4) and 2JC(5)H Interestingly, it is apparent from Table 4 that a fairly
(7) couplings recorded in DMSO are about two times large RMS deviation (8.46–9.92 ppm) is calculated for all
smaller than in water. This could indicate the presence of three models of isolated molecule of muscimol in the gas
muscimol in different forms in both solvents. phase. Moreover, the RMS values are nearly the same for
neutral and NH forms in water (8.65 to 8.83 ppm). How-
ever, moving from free zwitterion molecule in the gas
3.3 | Molecular modeling of NMR phase to its solution in water, the corresponding RMS
parameters of free muscimol molecule in drops significantly (from 9.92 to 3.15 ppm). Thus, calcu-
the gas phase and in water lated NMR chemical shieldings of three potential struc-
tures favor the presence of zwitterion form of muscimol
All calculated isotropic nuclear magnetic shieldings of in water that resembles its crystal structure.
isolated muscimol molecule and those in water are gath- In Table 5, deviations of theoretical coupling con-
ered in Table 3. It is apparent from Table 3 that the most stants of muscimol in the gas phase and water from the
KUPKA ET AL. 7

T A B L E 3 Isotropic nuclear magnetic shieldings of muscimol in its zwitterion, neutral, and NH forms calculated using B3LYP-D3/6-311
++G(3df,2pd) method in the gas phase and watera

Zwitterion (ppm) Neutral (ppm) NH form (ppm)

Atom Vacuum Water Vacuum Water Vacuum Water

O(1) 40.409 19.001 −31.144 −25.110 77.754 66.135
O(6) 69.774 138.908 241.103 240.124 −33.838 20.559
N(2) −167.699 −115.098 −109.760 −100.167 30.202 30.388
N(8) 206.030 207.693 219.442 221.212 222.258 224.090
C(3) 0.914 −4.956 5.626 4.202 4.584 0.779
C(4) 65.768 73.688 90.114 88.324 83.177 82.893
C(5) 32.244 17.017 −2.462 −4.280 −5.060 −7.476
C(7) 132.779 138.876 139.349 139.997 138.609 139.367
C(4)H 25.677 25.870 26.073 25.783 26.364 26.130
CH2 27.288 27.438 28.074 27.982 28.245 28.097
+
NH3 26.412 26.772
NH2 30.7518 30.5035 30.760 30.487
OH 26.674 26.060
N(2)H 23.881 23.154
a
At B3LYP-D3/pc-3 calculated equilibrium geometries in vacuum and in water, respectively.

TABLE 4 Deviations of the three forms of muscimol-predicted carbona and proton chemical shifts in the gas phase and in water from
experimental values in D2O

Atom Exp.a (ppm) Deviation (Theor.–Exp.) (ppm)

Zwitterion Neutral NH form

Vacuum Water Vacuum Water Vacuum Water
C3 178.220 −0.902 4.428 −4.572 −4.730 −5.614 −1.307
C4 100.015 12.451 3.990 −11.897 −10.647 −4.959 −5.215
C5 164.426 −18.438 −3.750 16.269 17.547 18.867 20.743
C7 35.748 9.705 3.068 3.135 1.946 3.876 2.577
(C4)H 5.808 −0.162 −0.531 −0.559 −0.443 −0.849 −0.791
CH2 4.138 −0.103 −0.429 −0.890 −0.972 −1.061 −1.088
RMS 9.916 3.148 8.544 8.646 8.455 8.828

Abbreviation: RMS, root-mean square.

a
Experimental carbon data are corrected by −2.01 ppm.

currently measured values in water are gathered. Because spectral resolution. In addition, two couplings seem to be
only positive values of SSCC parameters are observed about two times smaller (compare 2JC(5)H(4) and 2JC(5)H(7)
experimentally, the RMS values in Table 5 were calcu- in DMSO and in water in Table 2). Thus, in this study,
lated using absolute values of deviations. In the case of we observed eight individual couplings for muscimol in
coupling constants, the smaller RMS value is also consis- water (only five SSCCs were reported for DMSO solu-
tently predicted for zwitterion form of muscimol in tion[11]). It is apparent from Table 5 that RMS values for
water. the three forms of muscimol in the gas phase are about
It was possible to observe one coupling constants 4–5 Hz. On the contrary, the RMS of zwitterion in water
from the proton spectrum and seven from the 13C{+1H} drops to 3.7 Hz. We are aware that the difference
NMR spectrum. Interestingly, in recently reported NMR between the lowest RMS deviation and the remaining
spectra of muscimol in DMSO,[11] no 4JH(4)H(7) and 3JC(7)H structures is only about 0.5 to 1.0 Hz. However, it is
(4) couplings were observed, possibly due to lower worth reminding that the accuracy of theoretically
8 KUPKA ET AL.

T A B L E 5 Deviations of muscimol indirect spin–spin coupling constants calculated for the three forms of free molecule and of those in
water solution from available experimental data in D2O (this work)

Atom Exp. (Hz) Deviation (Theor.–Exp.) (Hz)

Zwitterion Neutral NH form

Vacuum Water Vacuum Water Vacuum Water
1
JC(4)H 179.60 3.165 4.316 8.241 11.688 10.017 12.396
1
JC(7)H 144.43 9.871 9.042 −3.624 −1.9565 −2.882 −0.948
3
JC(4)H(7) 2.60 −0.625 −0.152 0.194 0.262 0.29 0.31
3
JC(7)H(4) 0.78 0.27 0.367 0.063 0.099 0.644 0.54
4
JH(4)H(7) 0.68 0.19 −0.112 −0.255 −0.298 −0.278 −0.316
2
JC(3)H(4) 2.83 −0.258 2.318 −0.296 −0.028 2.571 2.971
2
JC(5)H(4) 9.73 3.852 1.431 0.97 0.71 2.304 1.49
2
JC(5)H(7) 5.78 −9.613 −1.68 −7.918 −3.349 −7.561 −3.712
RMS 5.187 3.722 4.256 4.364 4.721 4.742

Abbreviation: RMS, root-mean square.

predicted nuclear shieldings and chemical shifts is sig- 4 | CONCLUSIONS

nificantly higher than for coupling constants.[17,21]
Moreover, it is difficult to select one, optimized density Detailed analysis of muscimol 1H, 13C{−1H}, and 13C
functional to predict couplings acting via one, two, or {+1H} NMR spectra in D2O is reported. Three potential
more chemical bonds.[21] In fact, the B3LYP hybrid forms of muscimol—neutral, NH, and zwitterion—
functional works relatively well for these parameters, were considered in molecular modeling of structure,
and there is no need to check the performance of isotropic nuclear shieldings, chemical shifts, and indi-
more or less exotic functionals and much larger basis rect SSCCs using hybrid B3LYP density functional and
sets.[11] In fact, comparing the reported performance of fairly large basis sets. The best matching (or the
dedicated basis sets for SSCC (for example pcJ-n[51]) in smallest RMS deviation) of theoretical and experimen-
some benchmark calculations, one could accept the tal proton and carbon chemical shifts and proton–pro-
results presented in Table 5 as fairly accurate. At this ton and carbon–proton coupling constants was
point, we want to thank and respond to the anony- observed for the zwitterion form of muscimol in water.
mous reviewer suggesting checking the performance of This differs from previously reported NH form of
one or two density functionals in prediction of muscimol in DMSO solution and could be of interest
muscimol nuclear shieldings and chemical shifts. In for future studies of GABA analogs in water. The
Tables S4A and S4B, we include muscimol shieldings B3LYP results were additionally verified using PBE0
calculated with arbitrary-selected two additional func- and BHandH density functionals.
tionals—PBE0 and BHandH in combination with the
same Pople basis set. The corresponding chemical shift A C KN O WL ED G EME N T S
deviations from experimental values and the calculated All calculations were performed using hardware and
RMS values are included in Tables S5A and S5B. As software of Wrocław Supercomputer Center (WCSS).
before for B3LYP results, the lowest RMS values are Support from the Institute of Chemistry and the Uni-
calculated for the zwitterion form of muscimol in versity of Opole is also greatly acknowledged. In addi-
water. Finally, deviations between theoretical and tion, we want to thank Dorota Wieczorek for NMR
experimental SSCC values in water calculated with measurements of our samples with exceptionally great
PBE0 and BHandH density functionals also favor the care.
presence of zwitterion in water (Tables S6A and S6B).
In fact, it seems that in case of muscimol in water, ORCID
the latter density functional produces the lowest RMS Teobald Kupka https://orcid.org/0000-0002-6252-3822
values for chemical shifts and coupling constants. We Małgorzata A. Broda https://orcid.org/0000-0002-4092-
should mention that the BHandH density functional 3593
was fairly accurate[52,53] in predicting NMR parameters Piotr P. Wieczorek https://orcid.org/0000-0002-0016-
of selected small molecules. 0114
KUPKA ET AL. 9

R EF E RE N C E S [28] B. Miehlich, A. Savin, H. Stoll, H. Preuss, Chem. Phys. Lett.

[1] P. Krogsgaard-Larsen, L. Brehm, K. Schaumburg, Acta Chem. 1989, 157, 200.
Scand. 1981, 35, 311. https://doi.org/10.3891/acta.chem.scand. [29] I. Alkorta, J. Elguero, C. Dardonville, F. Reviriego, D. Santa
35b-0311 María, R. M. Claramunt, M. Marín-Luna, J. Phys. Org. Chem.
[2] D. Michelot, L. M. Melendez-Howell, Mycol. Res. 2003, 2019, in press, e4043. https://doi.org/10.1002/poc.4043
107, 131. [30] S. F. Sousa, P. A. Fernandes, M. J. Ramos, J. Phys. Chem. A
[3] K. Stebelska, Ther. Drug Monit. 2013, 35, 420. 2007, 111, 10439.
[4] S. Deja, E. Jawie n, I. Jasicka-Misiak, M. Halama, [31] B. A. Saeed, R. S. Elias, F. S. Kamounah, P. E. Hansen, Magn.
P. Wieczorek, P. Kafarski, P. Młynarz, Magn. Reson. Chem. Reson. Chem. 2018, 56, 172. https://doi.org/10.1002/mrc.4677
2014, 52, 711. https://doi.org/10.1002/mrc.4104 [32] F. Blanco, I. Alkorta, J. Elguero, Magn. Reson. Chem. 2007, 45,
[5] P. P. Wieczorek, D. Witkowska, I. Jasicka-Misiak, 797. https://doi.org/10.1002/mrc.2053
A. Poliwoda, M. Oterman, K. Zieli nska, in Studies in Natural [33] A. Buczek, D. Siodłak, M. Bujak, M. Makowski, T. Kupka,
Products Chemistry, (Ed: Atta-ur-Rahman) Vol. 46, Elsevier M. A. Broda, inStruct. Chem. 2019, 30, 1685.
2015, 133. [34] Ł. Gajda, T. Kupka, M. A. Broda, M. Leszczy nska, K. Ejsmont,
[6] D. Zhang, Z.-H. Pan, M. Awobuluyi, S. A. Lipton, Trends Magn. Reson. Chem. 2018, 56, 265.
Pharmacol. Sci. 2001. https://doi.org/10.1016/S0165-6147(00) [35] R. Walesa, T. Kupka, M. A. Broda, Struct. Chem. 2015, 26,
01625-4. 1083. https://doi.org/10.1007/s11224-015-0573-0
[7] P. Krogsgaard-Larsen, B. Frolung, U. Kristiansen, B. Ebert, [36] R. Faber, J. Kaminský, S. P. A. Sauer, in Gas Phase NMR,
Glutamate and GABA Receptors and Transporters: Structure, Chapter: Rovibrational and Temperature Effects in Theoretical
Function and Pharmacology, Taylor and Francis, New York Studies of NMR Parameters, (Eds: K. Jackowski, M. Jaszu nski)
2002 236. The Royal Society of Chemistry 2016, 218.
[8] N. Tro, Chemistry: A Molecular Approach, 5 edition ed., Pear- [37] T. Kupka, M. Stachow, L. Stobinski, J. Kaminsky, J. Mol.
son 2019. Graph. Model. 2016, 67, 14. https://doi.org/10.1016/j.jmgm.
[9] L. Brehm, K. Frydenvang, L. M. Hansen, P. O. Norrby, 2016.04.008
P. Krogsgaard-Larsen, T. Liljefors, Struct. Chem. 1997, 8, 443. [38] T. Kupka, M. Leszczy nska, K. Ejsmont, A. Mnich,
https://doi.org/10.1007/BF02311703 M. A. Broda, K. Thangavel, J. Kaminský, Int. J. Quant. Chem.
[10] G. Serdaroglu, Int. J. Quantum. Chem. 2011, 111, 3938. https:// 2019, 119, e26032.
doi.org/10.1002/qua.22809 [39] J. Adamson, R. B. Nazarski, J. Jarvet, T. Pehk, R. Aav,
[11] T. Kupka, P. P. Wieczorek, Spectroch. Acta Part A 2016, 153, ChemPhysChem 2018, 19, 631. https://doi.org/10.1002/cphc.
216. https://doi.org/10.1016/j.saa.2015.08.026 201701125
[12] J. A. Pople, W. G. Schneider, H. J. Bernstein, High-resolution [40] K. Jackowski, Chem. Anal. (Warsaw) 1987, 32, 947.
Nuclear Magnetic Resonance, New York, McGraw-Hill 1959. [41] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
[13] J. B. Foresman, A. Frisch, Exploring Chemistry with Electronic M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
Structure Methods; Ed, Second ed., Gaussian Inc, Pittsburg, PA G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato,
1996. A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts,
[14] R. Ditchfield, Mol. Phys. 1974, 27, 789. B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov,
[15] F. London, J. Phys, Radium (Paris) 1937, 8, 397. J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini,
[16] K. Wolinski, J. F. Hinton, P. Pulay, J. Am. Chem. Soc. 1990, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson,
112, 8251. D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng,
[17] T. Kupka, M. Stachow, M. Nieradka, J. Kaminsky, T. Pluta, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Chem, Theory Comput. 2010, 6, 1580. https://doi.org/10. J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
1021/ct100109j H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery Jr.,
[18] T. Helgaker, W. Klopper, H. Koch, J. Noga, J. Chem. Phys. J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd,
1997, 106, 9639. E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith,
[19] V. Semenov, D. Samultsev, L. Krivdin, Magn. Reson. Chem. R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell,
2019, 57. https://doi.org/10.1002/mrc.4851 J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam,
[20] R. B. Nazarski, P. Wałejko, S. Witkowski, Org. Biomol. Chem. M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin,
2016, 11. K. Morokuma, O. Farkas, J. B. Foresman, D. J. Fox. Gaussian
[21] T. Kupka, M. Nieradka, M. Stachow, T. Pluta, P. Nowak, 16 Rev. B.01, Wallingford, CT, 2016.
H. Kjaer, J. Kongsted, J. Kaminsky, J. Phys. Chem. A 2012, 116, [42] S. Ehrlich, J. Moellmann, W. Reckien, T. Bredow, S. Grimme,
3728. https://doi.org/10.1021/jp212588h ChemPhysChem 2011, 12, 3414. https://doi.org/10.1002/cphc.
[22] T. Helgaker, M. Jaszunski, K. Ruud, Chem. Rev. 1999, 99, 293. 201100521
[23] Y. Y. Rusakov, L. B. Krivdin, S. P. A. Sauer, E. P. Levanova, [43] F. Jensen, J. Chem. Phys. 2001, 115, 9113.
G. G. Levkovskaya, Magn. Reson. Chem. 2010, 48, 44. [44] F. Jensen, J. Chem. Phys. 2003, 118, 2459.
[24] J. K. Labanowski, J. W. Anzelm, Density Functional Methods [45] S. Miertus, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117.
in Chemistry, Springer-Verlag, New York 1991. [46] V. Jäger, M. Frey, Liebigs Ann. Chem 1982, 1982, 817. https://
[25] P. Hohenberg, W. Kohn, Phys. Rev. 1964, 136, B864. doi.org/10.1002/jlac.198219820423
[26] A. D. Becke, Phys. Rev. A 1988, 38, 3098. [47] S. Zhang, J. Biophys. Chem. 2011, 2, 208.
[27] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785. [48] R. E. Hoffman, Magn. Reson. Chem. 2006, 44, 606.
10 KUPKA ET AL.

[49] R. K. Harris, E. D. Becker, S. M. Cabral de Menezes,

P. Granger, R. E. Hoffman, K. W. Zilm, Pure Appl. Chem 2008, How to cite this article: Kupka T, Broda MA,
80, 59.
Wieczorek PP. What is the form of muscimol from
[50] H. E. Gottlieb, V. Kotlyar, A. Nudelman, J. Org. Chem. 1997,
fly agaric mushroom (Amanita muscaria) in water?
62, 7512.
[51] F. Jensen, J. Chem. Theory Comput. 2006, 2, 1360. An insight from NMR experiment supported by
[52] T. Kupka, Magn. Reson. Chem. 2009, 47, 674. molecular modeling. Magn Reson Chem. 2020;1–10.
[53] T. Kupka, Magn. Reson. Chem. 2009, 47, 959. https://doi.org/10.1002/mrc.4990

S UP PO RT ING IN FOR MAT ION

Additional supporting information may be found online
in the Supporting Information section at the end of this
article.