NMR-Aided Differentiation of Enantiomers: Signal Enantioresolution

Analytica Chimica Acta xxx (2015) xxx–xxx

G ModelACA 233788 No. of Pages 8

NMR-aided differentiation of enantiomers: Signal enantioresolution

Míriam Pérez-Trujillo a, Teodor Parella a, Lars T. Kuhn b,*a Servei de Ressonància Magnètica Nuclear and Departament de Química, Universitat Autònoma de Barcelona, E-08193 Bellaterra, Barcelona, Spainb European Neuroscience Institute Göttingen (ENI-G) & DFG Research Center Molecular Physiology of the Brain (CMPB)/EXC 171 “Microscopy at theNanometer Range”, Grisebachstrasse 5, 37077 Göttingen, Germany

H I G H L I G H T S G R A P H I C A L A B S T R A C T

� The concept and importance of NMRsignal enantioresolution is described.

� The enantioresolution quotient E forits measurement is introduced.

� E allows more reliable analyses ofenantiodifferentiation data thanprevious methods.

� Using E opens up possibilities foranalyzing data coming from differentNMR methods.

A R T I C L E I N F O

Article history:Received 23 September 2014Received in revised form 21 February 2015Accepted 26 February 2015Available online xxx

Keywords:EnantiodifferentiationEnantioresolutionNMR spectroscopyEnantioresolution quotientChiral complexing agents

A B S T R A C T

NMR-aided enantiodiscrimination using chiral auxiliaries (CAs) is a recognized method for differentiat-ing enantiomers and for measuring enantiomeric ratios (er). Up to the present, the study, optimization,and comparison of such methods have been performed based on the enantiodifferentiation of NMRsignals via analyzing non-equivalent chemical-shift values (DDd) of the diastereoisomeric speciesformed. However, a poor and non-reliable comparison of results is often obtained via the analysis of DDdexclusively. In here, the concept of enantioresolution of an individual NMR signal and its importance forNMR-aided enantiodifferentiation studies is introduced and discussed. In addition, the enantioresolutionquotient, E, is proposed as the parameter to describe its quantification. Complementary to measuringDDd, the experimental determination of E allows a more reliable interpretation of the results and opensup new possibilities for the study of enantiodifferentiation data derived from novel NMR experiments,setup improvements or new CAs. Finally, the different relationships between signal enantiodiffer-entiation, signal enantioresolution, and other main experimental issues of enantiodifferentiationexperiments are addressed.

ã 2015 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Analytica Chimica Acta

journa l home page : www.e l sev ier .com/ loca te /aca

* Corresponding author. Tel.: +49 551 39 12348; fax: +49 551 39 12346.E-mail address: [email protected] (L.T. Kuhn).

http://dx.doi.org/10.1016/j.aca.2015.02.0690003-2670/ã 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: M. Pérez-Trujillo, et al., NMR-aided diffe(2015), http://dx.doi.org/10.1016/j.aca.2015.02.069

1. Introduction

The NMR-aided enantiodifferentiation in the presence of chiralauxiliaries (CAs) is a powerful method for the discrimination of

rentiation of enantiomers: Signal enantioresolution, Anal. Chim. Acta

mailto:[email protected]

http://dx.doi.org/10.1016/j.aca.2015.02.069



http://www.sciencedirect.com/science/journal/00032670

www.elsevier.com/locate/aca

2 M. Pérez-Trujillo et al. / Analytica Chimica Acta xxx (2015) xxx–xxx


individual enantiomers and for the measurement of the enantio-meric ratio (er) – or, alternatively, the enantiomeric excess (ee) – ofmixtures of enantiomers. In addition, it serves to quantify anddetermine the absolute configuration of enantiomeric moleculesunder specific conditions [1–3]. Unlike chiroptical approaches,NMR-aided enantiodifferentiation provides a wider structural viewof molecules in a sample, allowing their simultaneous detection,identification and quantification. In general, the method is based onthe induction of enantiospecificity to the sample via the use of a CA[4–9] orachiral alignmentmedia[10–12] followedby the acquisitionof an NMR spectrum. Among all known CAs, one distinguishesbetween complexing agents (either chiral solvating agents, CSAs[4–6], or chiral metal complexes [7]) and chiral derivatizing agents[8,9] (CDAs). The differentiation of two enantiomeric molecules withCAs is observed via detecting non-equivalences in their respectiveNMR chemical shifts (d) due to the formation of diastereoisomericcomplexes or diastereoisomers between the enantiomeric analyteand the complexing CA or CDA, respectively. In the case ofcomplexing CAs, the formed complexes are commonly in fastexchange on the NMR time scale, and thus, a population-weightedaverage set of chemical shifts results [1–9]. Alternatively, whenworking with chiral alignment media, differences are detectedmainly through other NMR observables, such as dipolar couplingconstants, quadrupolar interactions or residual dipolar couplings[10–12]. When enantiodifferentiation and er measurements are tobe carried out, the working protocol using complexing CAs is fastand straightforward compared to the other methods. Besides, itinvolves minimal manipulation of the sample and excludes kineticresolution effects. The use of complexing CAs has, for a long time,been exercised in pharmaceutical studies [13–16] and theirparticular characteristics make them suitable for developing newapplications in other fields such as metabonomics [17].

Up to the present, the only parameter used to analyze andcompare results obtained from enantiodifferentiation experimentsusing CAs is the d non-equivalence (DDd) of the NMR signals of thediastereoisomeric species formed between the enantiomericanalytes and the CA. This parameter reflects the enantiodiffer-entiation of thesignal,but itdoesnotcommentonthequalityofsuchenantiodifferentiation and often leads to a poor and non-reliablecomparison of results. In this work, the conceptof enantioresolutionof an NMR signal is discussed as a way to overcome this problem.Besides, the enantioresolution quotient (E) is introduced as a newparameter tomeasureNMR signal enantioresolution, thus making ita complementary tool to the use of DDd. We think that theenantioresolution of the signal (measured by E) constitutes a mainexperimental issue to consider in such experiments along with:(i) the signal enantiodifferentiation (measured by DDd); (ii) the ermeasurement; and (iii) the time requiredfor the overall experiment,including the preparation of the sample and the NMR analysis. Therelationship between signal enantioresolution and the aforemen-tioned issues (i–iii) as well as the variables affecting them and theirinterconnection are also addressed.

2. Materials and methods

2.1. Materials

All compounds were purchased from Sigma–Aldrich (TresCantos, Madrid, Spain), except deuterated chloroform (99.80% D)and deuterium oxide (99.96% D) which were obtained fromCortecNet (Paris, France).

2.2. Sample preparation

Samples of spectra shown in Fig. 2 are analogous to the onesdescribed previously [6]. Ibuprofen 1 samples were prepared


mixing 100 mL of a 13.2 mM stock solution of (RS)-1 with 500 mLof D2O or 500 mL of a 12.1 mM b-CD solution in D2O, for sampleswithout and with CSA, respectively; resulting in 2.2 mMsolutions of (RS)-1 without and with 4.6 equiv. of b-CD(Fig. 2E). After that, both solutions were spiked with 70 mLof a 8.0 mM S-1 solution, resulting in a 2.8 mM solution of(R,S)-ibuprofen (35:65) in D2O containing 0 and 3.6 equiv. ofb-CD (Fig. 2A). 1-Aminoindan 2 samples consisted of 600 mLof a 50 mM stock solution of (RS)-2 in CDCl3 without CA and600 mL with 37 mg of R-(�)-1-(9-anthryl)-2,2,2-trifluoroethanol(Pirkle alcohol, PA); resulting in 50 mM solutions of (RS)-2without and with 4.5 equiv. of R-PA (Fig. 2B, D and F–H).Tryptophan 3 solutions were prepared mixing 100 mL of a13.8 mM stock solution of (DL)-3 with 500 mL of D2O or 500 mLof a 14.9 mM (�)-(18-crown-6)-2,3,11,12-tetracarboxylic acid(18C6H4) solution in D2O, for samples without and with CSA,respectively. After that, both solutions were spiked with110 mL of a 10 mM L-3 solution, resulting in a 3.5 mM solutionof (D,L)-3 (28:72) in D2O containing 0 and 3.0 equiv. of(�)-(18C6H4) (Fig. 2C). When D2O was used, a capillary withD2O and 3% of TSP (3-(trimethylsilyl)-[2,2,3,3-2H4]-propionicacid sodium salt) was inserted into the tube and used as anexternal reference.

For the titration shown in Fig. 4, a sample consisting of 600 mLof a fresh stock solution of racemic 2 (55.2 mM) in CDCl3 was used.Consecutive amounts of CA R-PA were added to the sample (0, 6.0,9.0, 11.3, 13.4, 16.2, 20.6 and 34.6 mg, corresponding to 0.0, 0.66,0.98, 1.23, 1.46, 1.77, 2.25 and 3.78 equiv.). Standard 1H, JRES andselective 1H pure shift NMR experiments were recorded after eachaddition of CA.

For the titration shown in Fig. 5, a sample consisting of 600 mLof a fresh solution stock of racemic 3-ethyl-3-(3-hydroxyphenyl)-1-methylazepan-2-one 4 (29 mM) in CDCl3 was used. Consecutiveamounts of CA R-PA were added to the sample (0.0, 2.4, 4.1, 5.3, 7.5,10.2, 12.6, 19.9, 24.7, 27.8, 32.7, 37.9, 41.6 and 46.2 mg, correspond-ing to 0.00, 0.49, 0.84,1.11,1.56, 2.13, 2.63, 4.15, 5.15, 5.78, 6.81, 7.88,8.67 and 9.61 equiv.). Standard 1H NMR experiments wererecorded after each addition of CA.

2.3. NMR experiments and equipment

1H NMR experiments were carried out on a Bruker AVANCEspectrometer (Bruker BioSpin, Rheinstetten, Germany) operatingat 14.1 T (600.13 MHz proton frequency), equipped with a 5 mmtriple resonance inverse probe, a z-axis pulsed field gradient and atemperature control unit. 13C experiments were performed on aBruker AVANCE 500 spectrometer (Bruker Biospin, Rheinstetten,Germany) operating at a field strength of 11.7 T (500.13 MHz and125.76 MHz, 1H and 13C frequencies, respectively) and equippedwith a high-sensitivity cryogenically cooled triple-resonance TCIprobe and a temperature control unit. The probe temperature wasset to 298.0 K for all experiments. Data were collected andprocessed using the software TOPSPIN 3.1 (Bruker BioSpin,Rheinstetten, Germany).

1D 13C NMR standard experiments (Fig. 2A and B) wereperformed using a 90� power gated (pg) proton-decoupled pulsesequence, as described previously [6]. Experiments were acquiredwith an acquisition time of 0.6 s and a relaxation delay (RD) of 2 s.The data were collected into 32 K data points and with a spectralwidth of 27,778 Hz. Each free induction decay (FID) was zero-filledto 64 K data points and multiplied by an exponential apodizationfunction equivalent to a 0.4 Hz line broadening prior to Fouriertransformation. For the enantiodifferentiation study of ibuprofen(2.2 mM) and 1-aminoindan (50 mM), 13C pg experiments wereperformed with 70,000 and 512 number of transients, correspond-ing to experimental times of 34 h and 22 min, respectively.



M. Pérez-Trujillo et al. / Analytica Chimica Acta xxx (2015) xxx–xxx 3


1D 1H NMR standard experiments (Fig. 2C–F and theexperiments performed for titrations in Figs. 4 and 5) wereperformed identically using a standard 90� pulse-acquisitionsequence. The data were collected into 16 K data points during anacquisition time of 0.8 s and using a RD of 2 s. Spectra werecollected in the time domain as FIDs across a spectral width of10,000 Hz as the sum of 8 transients. FIDs were zero-filled to 64 Kdata points and Fourier transformed. The spectra were thenmanually phased and baseline corrected.

2D conventional 1H JRES experiments (Fig. 2G and theexperiments performed for titrations in Fig. 4) were performedas follows: 8 transients were collected for each of the 32 t1increments. Data were recorded into 16 K data points during anacquisition time of 0.8 s and a spectral width of 10,000 Hz. Datawere processed using a sine squared window function followed bya tilting process. E was measured from the internal projectionextracted along the detected dimension after a tilting process.

Frequency-selective 1D 1H pure shift experiments (Fig. 2H andthe experiments performed for titrations in Fig. 4) were acquiredusing the pulse sequence described in a previous work [18]. Thedata were collected into 16 K data points during an acquisition timeof 0.8 s and using a RD of 2 s. The number of concatenated loopswas optimized (n from 20 to 62) and a selective Reburp-shaped180� 1H pulse with a duration of 40 ms was used. FIDs werezero-filled to 64 K data points and multiplied by an exponentialapodization function equivalent to 0.3 Hz line broadening prior toFourier transformation. The spectra were then manually phasedand baseline corrected.

All spectra were manually phased and baseline corrected.Spectra of samples in D2O were referenced to the TSP peakresonance (dH and dC at 0.00 ppm) and spectra of samples in CDCl3,were referenced to the residual solvent signal (dH at 7.26 ppm anddC at 77.0 ppm, respectively).

2.4. Determination of E

The determination of the enantioresolution quotient, E, of anNMR signal, was carried out by the measurement of DDd and W ofthe signal (see Supporting information). These measurementswere done directly from the processed spectra using the “MNova7.0.2” software package (Mestrelab Research S.L., Santiago deCompostela, Spain).

Fig. 1. Schematic representation depicting the main experimental aspects rele


3. Results and discussion

3.1. The importance of NMR signal enantioresolution

The standard 1H NMR experiment is still the most commonexperiment used in NMR-aided enantiodifferentiation using CAsdue to the high abundance of 1H nuclei in organic molecules and itshigh NMR sensitivity. Up to the present, the DDd of a signal is theparameter employed for the analysis of the data coming fromenantiodifferentiation experiments. That is, for instance, the casewhen comparing results coming from experiments (i) usingdifferent CAs [19], (ii) measuring different nuclei by NMR [6]or (iii) employing different NMR pulse sequences [18]. Theenantiodifferentiation value (DDd) is defined as the differencebetween the change in d of one enantiomer (DdR) with respect toother one (DdS), where these changes are relative to the chemicalshift of the enantiomers in the absence of the CA. DDd isdetermined from the NMR spectrum acquired following theaddition of the CA (Eq. (1)) [4–9].

DDd ¼ DdR � DdS (1)

Thus, if DDd 6¼ 0, an NMR signal is enantiodifferentiated. DDd isaffected by the nature of the CA, the concentration of the analyte,the equivalents of complexing CA added to the sample, thetemperature of the sample, and the solvent (Fig. 1A) [20].

Despite of the intrinsic advantages in speed and feasibility ofenantiodifferentiation experiments using complexing CAs, DDdvalues are usually small (0–0.1 ppm) and situations where signalsare not completely split into two well-separated resonances arecommon. DDd indicates the extent of the non-equivalence of dwhich is related to differences in the geometry and/or theassociation constants of the diastereoisomeric complexes formed[2–9]. Nevertheless, DDd does not comment on the quality of theenantiodifferentiation, i.e., a higher DDd value does not necessarilymean a better differentiation of the signals and vice versa. This isexemplified in Fig. 2, showing a set of enantiodifferentiated signalsof a pharmaceutical, ibuprofen 1, a chemical reactant, 1-amino-indan 2, and a biomolecule, tryptophan 3, complexed withdifferent CSAs [6]. The DDd values of the C2 and H3 signals of 2(Fig. 2B and D) are 33.7 and 50.6 ppb, respectively. It is clearly

vant for NMR-aided enantiodifferentiation studies using chiral auxiliaries.



Fig. 2. Examples of different multiplet pattern NMR signals of enantiomeric mixtures of 1, 2 and 3, before and after being enantiodifferentiated by the addition of a CA. NMRresonances originating from standard 13C NMR experiments (A and B); conventional 1H NMR experiments (C–E); the internal projection along the detected dimension of aJ-resolved (JRES) experiment (G), and a 1D, fully homodecoupled, 1H experiment (H) are shown (c.f. Materials and methods section) [18].



visible that the signal of the former, with a lower DDd, is betterseparated into two new resonances than the latter. A similarobservation is made when signals of H2 of 3 and H30 of 2 (Fig. 2Cand F) or, alternatively, signals of H2 of 3 and H3 of 2 (Fig. 2C and D)are compared. In Fig. 2F–H, the enantiodifferentiation of the sameresonance, H30 of 3, is shown. In these examples, the onlyexperimental difference is the NMR experiment employed, i.e., aconventional 1H NMR experiment yielding a multiplet pattern(Fig. 2F) and two different pure-shift NMR experiments [18,21,22]yielding simplified, fully homodecoupled singlet 1H signals (Fig. 2Gand H). As expected, DDd is identical in all cases, but signals fromFig. 2G and H are better separated. Thus, when the quality of theenantiodifferentiation of signals is evaluated, the interpretation ofonly DDd can lead to erroneous conclusions. In this context,the need to refer to the ‘enantioresolution’ of the two newsignals – with respect to the quality of their separation – ratherthan to their ‘enantiodifferentiation’ – with respect to theextension of their separation – is evident.

3.2. NMR signal enantioresolution and the enantioresolutionquotient E

The enantioresolution of an NMR signal is defined as the qualityof the separation of the original signal into two new resonances


due to the induction of enantiospecificity. It depends on theseparation of the two new resonances, DDd, relative to theiraverage width, W, considering the multiplet pattern. Similarconcepts of resolution of two signals are well established in otherfields of chemical analysis, e.g., chromatography [23].

Here, we introduce a new parameter to measure theenantioresolution, i.e., the enantioresolution quotient, E, of anNMR signal, defined as the absolute value of DDd of the splitsignals divided by three times their average W (Eq. (2)).

E ¼ jDDdj3W

(2)

E is a dimensionless parameter providing quantitative informationwith regard to the enantioresolution of an NMR signal. Both DDdand W are measured from the spectrum acquired after the additionof the CA and using the same units (i.e., ppm or Hz). W correspondsto the average width of the split resonances, considering theirpattern multiplicity, measured at half-height of the signal (singletor multiplet; see SI, Fig. S1). Typically, the split signals haveidentical W values given that they correspond to analogous nucleifrom enantiomeric molecules, i.e., they have the same multipletpattern, and due to the fact that the diastereoisomeric complexesformed upon addition of CAs are commonly in fast exchange on theNMR time scale, i.e., no line broadening due to chemical exchange





is observed in the NMR spectrum (see SI, Table S1). In situations ofpartial enantioresolution in which the measurement of W is notstraightforward, fitting procedures are needed.

For a pure Lorentzian lineshape singlet, the denominator, 3W,corresponds to the full width at the tenth-height (Dv1/10) of thepeak (Eq. (3)).

Esinglet ¼jDDdj3Dv1=2

¼ jDDdjDv1=10

(3)

The denominator 3W is chosen taking into account the typicalexperimental range values of DDd when working with complexingCAs [4–7]. For instance, from Eq. (2) follows that an enantiodiffer-entiated singlet of E = 1 and W = 3 Hz is obtained from a DDd of 9 Hz(i.e., 18 ppb in a magnetic field of 500 MHz). For a specific NMRsignal, the DDd and W values will lead to one of the followingscenarios:

ðiÞjDDdj ¼ 0 ! E ¼ 0; ðiiÞjDDdj < 3W ! E < 1;

ðiiiÞjDDdj � 3W ! E � 1

Thus, when E 6¼ 0, there is enantiodifferentiation. Fig. 3exemplifies the different enantioresolution scenarios for a singlet:non (E = 0), partial (0 < E < 1), moderate (E ffi 1) and high (E >> 1)enantioresolution. From Eq. (2) follows that the enantioresolutionof a signal, and therefore, E is increased either by increasing DDd,i.e., adding more CA, reducing the T [20] (Fig. 1A) or by decreasingW. The latter can be favored by increasing the signal dispersionthrough an adequate choice of the experimental conditions (choiceof nuclei, pulse sequence, data processing (Fig. 1B) and/or via usinghigher magnetic fields (B0).

3.3. Real cases of applying the enantioresolution quotient, E

A common way to analyze results in enantiodifferentiationexperiments with complexing CAs is by plotting the DDd of signalsagainst the equivalents of CA added to the sample. Thus, a givensignal yields a characteristic curve under specific conditions. Theplotting of such curves is a convenient visual procedure to compareresults coming from different signals of an analyte enantiodiscri-minated under specific conditions, or, alternatively, from the same

Fig. 3. Evolution of the enantioresolution (E) increas


signal enantiodiscriminated using different experimental condi-tions (e.g., different CAs, different NMR approaches, etc.). Whenthe E of a signal is plotted against equiv. of CA, new curves areobtained which show the enantioresolution achieved for a signal ateach point of the titration with the CA. Plots of E providecomplementary information to that obtained by DDd curves. Thereare cases where the E curves provide a more reliable comparison ofdata than the plots of DDd; two cases are shown to exemplify this:the comparison of (i) the results obtained from different NMRapproaches in the enantiodiscrimination of a specific signal(nucleus) and (ii) the results obtained from different signals(nuclei) of a given analyte using specific enantiodifferentiationconditions.

In the first example, racemic compound 2 was enantiodiffer-entiated adding consecutive amounts of the CSA R-Pirkle alcohol(PA) and the sample was measured at each point of the titrationusing three different 1H NMR approaches: a standard 1D 1H(1H-std), a J-resolved (JRES), and a 1D 1H pure-shift experiment(1H-ps) [18]. Fig. 4 shows the results obtained for the H30 signal atseveral titration points and the results of plotting DDd and Eagainst the equivalents of CA (see SI, Table S2). As expected, theDDd curve plots stemming from the three different NMRapproaches coincide. On the other hand, it is clearly visible thatthe quality of the separation of the split signals is better for thoseexperiments in which W has been reduced, e.g., JRES or 1H-ps (see,for instance, signals after the addition of 1.2 and 1.5 CA equiv.),which can be advantageous with respect to using a standard 1Hexperiment. The plots of E make these differences clear and allow amore consistent comparison of the three experiments.

The second example shows a comparative study of theenantiodiscrimination results of different proton signals of racemic3-ethyl-3-(3-hydroxyphenyl)-1-methylazepan-2-one 4, a precur-sor of diarylether lactams used as cancer chemotherapeutic agents[24], titrated with R-PA as CSA and analyzed through standard 1HNMR experiments. Fig. 5 shows the enantiodiscriminated signalsH11, H12, H15 and H16 of 4 at different titration points. Theseresults are analyzed through DDd and E titration curves (seeTable S3 of the SI). Observing the DDd titration plots, the bestresults are those corresponding to the signals of H12 followed bythose of H11 and H16, and a significantly lower enantiodiffer-entiation is achieved by analyzing the signal of H15. On the otherhand, observing these spectra, it becomes evident that the signals

e of a simulated singlet with a W = Dv1/2 = 2 Hz.



Fig. 4. Results of the H30 signal of racemic 2 (55.2 mM) when titrated with R-PA in CDCl3 and at 298.0 K. (A) H30 NMR signal corresponding to 1D 1H standard (1H-std), JRESprojection (JRES) and frequency selective 1H pure shift (1H-ps) NMR experiments; (B) E and (C) DDd titration curves of H30 signal for each NMR approach. (c.f. SI).



representing H16 signals are well separated (resolved) after theaddition of 5.78 equiv. of CSA; at that point, the other signals arejust partially separated. After the addition of 7.88 equiv. of CSA, theH16 signals are better resolved than signals representing H11,which, in turn, are better separated than the signals of H12 andH15. These observations are clearly visible when E is determinedand the corresponding plots are compared. When analyzingenantioresolution, significantly better results are obtained for thesignals of H16, followed by the results for H11, H12, and H15,respectively. Considering again the enantiodifferentiated signalscompared in Fig. 2 and their corresponding DDd and E values, it isevident that a more reliable comparison of the quality of theenantiodifferentiation achieved for each signal is obtained when Evalues are compared.

Thus, introducing E for studying enantiodifferentiation basedon chemical shift inequivalences opens up new possibilitiesregarding the analyses of the results and the comparison ofenantiodifferentiation methods. DDd remains a relevant parame-ter since it directly reflects the differences in geometry andequilibria of the diastereoisomeric complexes, providing informa-tion of a different kind. For example, methods to determine theassociation constant [25] or the stoichiometry of these complexesare based on Dd measurements [26].

3.4. Relationship between measuring the er and E

In NMR-aided enantiodifferentiation, another important aspectis the measurement of er or ee (Fig. 1C). This is either achieved via


direct integration of the resolved peaks or, alternatively, usingappropriate fitting procedures, always on the condition that theNMR spectrum can be analyzed quantitatively. In all cases, theaccuracy of the measurement will be subject to the signal-to-noiseratio (SNR) [27]. In absence of signal overlap and given thatenantioresolution is complete (E >> 1), the simple integration ofthe resolved signals can provide an accurate er measurement. Onthe other hand, when partial enantioresolution (E < 1) and/orsignal overlap exist, the perpendicular drop method can lead tosignificant errors and the use of fitting procedures is recommended[27,28].

For instance, in the case of two singlets of identical intensityand area, i.e., a racemic mixture, separated by jDDdj ¼ Dv1=10, theoverlapping region corresponds to 10% of each area. A study of twoLorentzian signals separated by Dv1/10 = 3Dv1/2 and havingdifferent relative intensities (scanning a range of intensity ratiosfrom 1 to 100) was carried out in order to explore the values of theoverlapping signal areas. This study simulated a set of enantio-meric mixtures with a range of er going from 1 to 100 and showingan enantioresolved singlet with an E = 1. The results show that,regardless of the er, the measurement of the integration of themajor component has an associated error of <5% using theperpendicular drop method (see SI, Figs. S2 and S3).

Though a complete enantioresolution would imply an accuratemeasurement of the er by direct integration of the resolved signals,an accurate measurement of the er by direct integration does notalways require complete enantioresolution. This is the case, forexample, in partially enantioresolved symmetric multiplets where



Fig. 5. Results of enantiodifferentiated 1H NMR signals of racemic 4 (29 mM) when titrated with R-PA in CDCl3, at 298.0 K and monitored through standard 1D 1H NMRexperiment. (A) H11, H12, H15 and H16 resonances at several titration points; (B) E and (C) DDd titration curves of each proton signal (c.f. SI).



analogous end peaks of both multiplets of diastereoisomericcomplexes are well resolved (see SI, Fig. S4). Finally, when workingwith complexing CAs, the time limiting factor correspondstypically to the duration of the entire NMR experiment since thepreparation of the sample is usually simple and fast. The NMRexperimental time is determined by the type of nucleusbeing observed, the NMR pulse sequence employed, the relativeSNR – depending, for example, on the magnetic field strength andthe specification of the probe being used – and the concentration ofthe analyte (Fig. 1D).

4. Conclusions

In summary, the concept of the enantioresolution of anindividual NMR signal defined as the quality of the separation ofthe original signal into two new resonances due to an induction ofenantiospecificity is presented. In addition, the relationshipbetween the main aspects involved in enantiodifferentiationstudies, like signal enantiodifferentiation, er measurement orexperimental time, and signal enantioresolution has beenaddressed together with the variables affecting them. A newparameter, the enantioresolution quotient, E, is proposed fordescribing the quantification of a signal’s enantioresolution. E is acomplementary tool to using DDd for obtaining more reliable


analyses of enantiodifferentiation data. It depends on theseparation of the two new resonances, DDd, relative to theiraverage width, W, considering the corresponding multiplet patternof a signal. The use of E opens up new possibilities for the analysisand comparison of the results of NMR-aided enantiodifferentiationexperiments. Useful applications involving the application of E arethe analysis of enantiodifferentiated signals derived from(i) different NMR experiments, e.g., 1D 1H NMR standard vs 1HJ-resolved or 1H pure shift experiments [18,21,22], as the use of Ewill consider the differences in the multiplet pattern of anenantiodifferentiated signal; (ii) using different nuclei and NMRexperiments, e.g., 1D 1H vs 1D 13C NMR signals from standardexperiments [6]; (iii) methodological improvements [28,29], e.g.,hyperpolarization methods [30,31]; and (iv) new CSAs as well asprochiral solvating agents (pro-CSAs) [32,33], since the depen-dence on W of E allows a reliable comparison of the resolution ofNMR signals resolved under multiple enantiodifferentitatingconditions.

Supplementary data

Experimental details, figures and tables indicated along the textare provided as Supplementary information (SI). This material isavailable free of charge online.



https://www.researchgate.net/publication/279121491_Hyperpolarization_Methods_in_NMR_Spectroscopy?el=1_x_8&enrichId=rgreq-effbf816-0b4b-4660-8583-7e7b8218342a&enrichSource=Y292ZXJQYWdlOzI3MzQwMDU4MDtBUzoyMTA0NDYxODAxOTYzNjFAMTQyNzE4NTY4NzM1MA==



Acknowledgements

Financial support from the MINECO (project no. CTQ2012-32436) and from the DFG (EXC 171 to LTK) is gratefullyacknowledged. The European Neuroscience Institute Göttingen(ENI-G) is jointly funded by the Göttingen University MedicalSchool and the Max Planck Society. In addition, gratitude isowed to the Servei de Ressonància Magnètica Nuclear (UniversitatAutònoma de Barcelona) for the generous allocation ofspectrometer time.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.aca.2015.02.069.

References

[1] D. Parker, Chem. Soc. Rev. 91 (1991) 1441.[2] T.J. Wenzel, Discrimination of Chiral Compounds Using NMR Spectroscopy,

Wiley-VCH Verlag GmbH & Co. KGaA, 2007.[3] V. Schurig, Top. Curr. Chem. 340 (2013) 21.[4] G. Uccello-Barretta, F. Balzano, Top. Curr. Chem. 341 (2013) 69.[5] A.E. Lovely, T.J. Wenzel, Org. Lett. 8 (2006) 2823.[6] M. Pérez-Trujillo, E. Monteagudo, T. Parella, Anal. Chem. 85 (2013) 10887.[7] T.J. Wenzel, Top. Curr. Chem. 341 (2013) 1.[8] J.M. Seco, E. Quiñoá, R. Riguera, Chem. Rev. 104 (2004) 17.[9] I. Louzao, J.M. Seco, E. Quiñoá, R. Riguera, Chem. Commun. 46 (2010) 7903.

[10] K. Kobzar, H. Kessler, B. Luy, Angew. Chem. 117 (2005) 3205.[11] M. Sarfati, P. Lesot, D. Merlet, J. Courtieu, Chem. Commun. 21 (2000) 2069.[12] P. Lesot, O. Lafon, H. Zimmermann, Z. Luz, J. Am. Chem. Soc. 130 (2008) 8754.


[13] E. Francotte, W. Lindner, Chirality in Drug Research, Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim, Germany, 2006.

[14] T.J. Wenzel, J.D. Wilcox, Chirality 15 (2003) 256.[15] M. Pérez-Trujillo, A. Virgili, Tetrahedron Asymmetry 17 (2006) 2842.[16] J. Redondo, A. Capdevila, I. Latorre, J. Bertran, J. Incl. Phenom. Macrocycl. Chem.

73 (2012) 225.[17] M. Pérez-Trujillo, J.C. Lindon, T. Parella, H.C. Keun, J.K. Nicholson, T.J. Athersuch,

Anal. Chem. 84 (2012) 2868.[18] L. Castañar, M. Pérez-Trujillo, P. Nolis, E. Monteagudo, A. Virgili, T. Parella,

Chem. Phys. Chem. 15 (2014) 854.[19] M. Pomares, F. Sanchez-Ferrando, A. Virgili, A. Alvarez-Larena, J.F. Piniella, J.

Org. Chem. 67 (2002) 753.[20] F.J. Villani, M.J. Costanzo, R.R. Inners, M.S. Mutter, D.E. McClure, J. Org. Chem. 51

(1986) 3715.[21] S.R. Chaudhary, N. Suryaprakash, Chem. Phys. Lett. 555 (2013) 286.[22] R.W. Adams, L. Byrne, P. Király, M. Foroozandeh, L. Paudel, M. Nilsson, J.

Clayden, G.A. Morris, Chem. Commun. 50 (2014) 2512.[23] A. Felinger, Data Analyis and Signal Processing in Chromatography, first ed.,

Elsevier Science, 1998.[24] S.J. deSolms, T.M. Ciccarone, S.C. MacTough, A.W. Shaw, C.A. Buser, M. Ellis-

Hutchings, C. Fernandes, K.A. Hamilton, H.E. Huber, N.E. Kohl, R.B. Lobell, R.G.Robinson, N.N. Tsou, E.S. Walsh, S.L. Graham, L.S. Beese, J.S. Taylor, J. Med.Chem. 46 (2003) 2973.

[25] J. Bouquant, J. Chuche, Tetrahedron Lett. 13 (1972) 2337.[26] P. Job, Ann. Chim. 9 (1928) 113.[27] L. Griffiths, A. Irving, Analyst 123 (1998) 1061.[28] L. Castañar, P. Nolis, A. Virgili, T. Parella, Chem. Eur. J. 19 (2013) 17283.[29] Lokesh, S.R. Chaudhari, N. Suryaprakash, Org. Biomol. Chem. 12 (2014) 993.[30] J.H. Ardenkjaer-Larsen, B.R. Fridlund, A. Gram, G. Hansson, L. Hansson, M.H.

Lerche, R. Servin, M. Thaning, K. Golman, Proc. Natl. Acad. Sci. U. S. A. 100(2003) 10158.

[31] L.T. Kuhn, Hyperpolarization Methods in NMR Spectroscopy, Springer, 2013.[32] A. Shundo, J. Labuta, J.P. Hill, S. Ishihara, K. Ariga, J. Am. Chem. Soc. 131 (2009)

9494.[33] J. Labuta, S. Ishihara, T. Sykorsky, Z. Futera, A. Shundo, L. Hanykova, J.V. Burda,

K. Ariga, J.P. Hill, Nat. Commun. 4 (2013) 2188.



http://refhub.elsevier.com/S0003-2670(15)00297-4/sbref0005


















































https://www.researchgate.net/publication/279121491_Hyperpolarization_Methods_in_NMR_Spectroscopy?el=1_x_8&enrichId=rgreq-effbf816-0b4b-4660-8583-7e7b8218342a&enrichSource=Y292ZXJQYWdlOzI3MzQwMDU4MDtBUzoyMTA0NDYxODAxOTYzNjFAMTQyNzE4NTY4NzM1MA==

NMR-Aided Differentiation of Enantiomers: Signal Enantioresolution

Documents

Transcript of NMR-Aided Differentiation of Enantiomers: Signal Enantioresolution