This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights
21
Embed
Synthesis and stereochemistry of new 1,3-thiazolidine systems based on 2-amino-2-(mercaptomethyl)propane-1,3-diol: 4,4-bis(hydroxymethyl)-1,3-thiazolidines and...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies areencouraged to visit:
Synthesis and stereochemistry of new 1,3-thiazolidine systemsbased on 2-amino-2-(mercaptomethyl)propane-1,3-diol: 4,4-bis(hydroxymethyl)-1,3-thiazolidines and c-5-hydroxymethyl-3-oxa-7-thia-r-1-azabicyclo[3.3.0]octanesq
Cristina Morar a, Carmen Sacalis a, Pedro Lameiras b, Albert Soran a, Hassan Khartabil b,Cyril Antheaume c, Ioan Bratu d, Oana Moldovan a, Mircea Darabantu a,*
a “Babes-Bolyai” University, Department of Chemistry, 11 Arany J�anos St., 400028 Cluj-Napoca, RomaniabUniversity of Reims Champagne-Ardenne, ICMR, UMR 7312, BP 1039, 51687 Reims, FrancecUniversity of Strasbourg, Faculty of Pharmacy, 74 Rhin St., BP 60024, 67401 Illkirch Cedex, FrancedNational Institute for Research and Development of Isotopic and Molecular Technologies (INCDTIM), 65-103, Donath St., PO Box 700,400293 Cluj-Napoca 5, Romania
a r t i c l e i n f o
Article history:Received 11 July 2013Received in revised form 9 September 2013Accepted 24 September 2013Available online 1 October 2013
The thiaminalisation of 2-amino-2-(mercaptomethyl)propane-1,3-diol [‘2-(hydroxymethyl)cysteinol’]with aryl(di)aldehydes is reported. The resulting new class of 2-aryl-4,4-bis(hydroxymethyl)-1,3-thiazolidines is investigated by NMR and IR spectroscopy in tandem with DFT calculations, permittingstructural assignments that are discussed in terms of conformational analysis, anomeric effects and ring-chain tautomerism. These acquired data are subsequently exploited. After treatment with formaldehyde,the subsequent (double) regio- and diastereoselective oxaminalisation of the 1,3-thiazolidine building-block affords the first non-symmetric series of a thiazolidin-oxazolidine fused system singly function-alised at the C-5 position. An unexpected rearrangement, which consists of the partial relocation of theAr ligand from the 1,3-thiazolidine to the 1,3-oxazolidine ring, is observed as a major influence on thesubstitution of the Ar ring. The first single crystal X-ray analysis of the title bicyclic system, which dis-closes the homo- and/or heterochiral non-bonding interactions, is also presented.
� 2013 Elsevier Ltd. All rights reserved.
1. Introduction
The heterocyclic saturated cis-fused system 3-oxa-7-thia-r-1-azabicyclo[3.3.0]-c-5-octane 1 (Scheme 1) has been known sincethe 1970s.1 This compound can be viewed as a chiral (1R*,5S*)3-oxa-7-thia analogue of the core alkaloid pyrrolizidine,
r-1-azabicyclo[3.3.0]-c-5-octane 2, as well an S-analogue of themuch better documented 3,7-dioxa-r-1-azabicyclo[3.3.0]-c-5-octane 3.2
Currently, there are two known routes by which the thiazolidin-oxazolidine skeleton of type 1 can be built. In the method wenamed the ‘Cysteine based approach’ (Scheme 2),2 one of its enan-tiomers (optionally as ethyl cysteinate) is sequentially cyclo-condensed with aldehydes, proceeding via 1,3-thiazolidines Ia, toyield thiazolidin-azalactones of type IIa.1,3
Their reduction provides series 1a, including the parent system:R1¼R2¼H.1b In the 1980s, the above chemistry was developed bySeebach et al.3c,d as an R-cysteine diastereoselective cyclisation(IIa, R1¼t-Bu, R2¼H) towards the benefit of subsequent asymmetricfunctionalisation via metallation at the C-5 position of IIa.4
The same methodology was applied to obtain 2,4-disubstituted-1,3-thiazolidines Ia with impact upon the stereochemistry of sat-urated five-membered rings5 and notable bioactivity.6
Recently, Saiz et al.7a reported an alternative pathway directedtowards optically active derivatives of 1a, specifically via 1,3-thiazolidines Ib. This strategy also uses a two-step manipulation
NOS
N7 3 7 35 5 5
1 1 1
H H
1
NOO
H
2 3
Scheme 1.
q Tribute to Prof. Ioan SILAGHI-DUMITRESCU. In memoriam.* Corresponding author. Tel.: þ40 264 59 38 33; fax: þ40 264 59 08 18; e-mail
0040-4020/$ e see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.tet.2013.09.070
Tetrahedron 69 (2013) 9966e9985
Author's personal copy
with different aryl-aldehydes, Ar1(2)eCH]O. The successful prep-aration of series (4S)-Ia, (4S)-Ib and (5S)-1a was claimed to accessa potential DCL (dynamic combinatorial library) because, depend-ing on the electrophilicity of Ar1(2)eCH]O, thia- versus oxamin-alisation equilibrium distributions were observed.
The synthesis of the first thiazolidin-oxazolidine condensedsystem 1 (Scheme 1), which was singly functionalised at theC-5 position with a hydroxymethyl group, was reported by usbased on the ‘Cysteinolic approach’ (Scheme 3, compound 1b,R3¼H).2
The key intermediate in this approach is the thioamino-1,3-diolIVb, which is otherwise known as a ‘2-(hydroxymethyl)cysteinol’(stable as its hydrochloride IVa only)2a,2b and has either the 1,3-oxazoline IIIa or the 1,3-thiazolidin-2-thione IIIb as precursors;the latter was proposed by Saiz et al.7b recently.
One must examine the essential differences between the twoseries of bicycles known so far, which are 1a and 1b. While com-pounds 1a (Scheme 2) are optically active with no further de-velopments regarding the chemistry on the C-(4)eC-(5)eC-(6)sequence, series 1b (Scheme 3), although a racemate, still containsan exploitable functionality at C-5.
The increased interest in thiazolidin-oxazolidines 1 impelled usto enlarge the ‘Cysteinolic approach’ by targeting new compounds inseries 1b, which arise from the consecutive cyclocondensation of‘2-(hydroxymethyl) cysteinol’ IVb with two different aldehydes. Inthis regard, the previously unreported tandem reaction using firstAreCH]O and then H2C]O was interesting to us because of thevery distinct reactivity of these electrophiles against the triplenucleophile IVb. No similar synthetic or structural approaches havebeen reported to date.
2. Results and discussion
2.1. Synthesis and structure of novel 2-aryl-1,3-thiazolidineseries
Initially, we examined the thiaminalisation of the free base IVb,which is derived from IVa (Scheme 3), via reaction with variousaryl-(di)aldehydes to enact a classically disfavoured Baldwin’s5-endo-trig cyclisation.8 We chose to explore a larger diversity ofcarbonyl electrophiles than were investigated in the synthesis of
1,3-thiazolidines Ia (Scheme 2). In this context, the use of thionated(pseudo)ephedrines9 or the much simpler 2-aminoethanethiol(cysteamine) were reported.10 The resulting cyclocondensateswere investigated via conformational analysis,10a as bio-molecules10b,c and ring-chain tautomers.11
The aim of this section is to report the synthesis of a new familyof 2-aryl-4,4-bis(hydroxymethyl)-1,3-thiazolidines and examinethe structure of these molecules in terms of configurational anal-ysis, anomeric effects and ring-chain tautomerism. These proper-ties would be used as the basis for the next ring-closure to generatenew thiazolidin-oxazolidine derivatives 1b (Scheme 3).
2.1.1. Synthesis. Scheme 4 depicts our first step, which was thia-minalisation, and the isolated yields from reactions using a range ofaryl-(di)aldehydes under three different sets of conditions.
In the presence of a stoichiometric amount of aryl-(di)aldehyde(aei), the free base IVb was generated in situ from IVa via acid-ebase interchange. Because IVb exhibited high redox instability,2
three types of reaction conditions, AeC, were tested. All threeconditions employed mild conditions and an inert atmosphere. We
NH2
COORHS
NOS
OH
NOS
H
5
R = H, Et
R1 R2
R1 R2S NH
COOR
R1 S NH
CH2OH
Ar1
"Cysteine based approach"
1a
R1 = R2 = HR1 R2
H, t-BuAr
1, Ar2
5
red.
R1-CH=O
red.(R1 = Ar
1)
R2-CH=O
Ar2-CH=OIa
Ib
IIa 6 4
2
4
4
Scheme 2.
NH2OHHO
OHO N
Me
CH2OHCH2OH
i) Ph-CO-SHii) HCl
Ac2O
NH3+Cl-OHHS
OH
CS2
S NH
S
CH2OHCH2OH HCl
"Cysteinolic approach"
TRIS"2-(hydroxymethyl)serinol"
IIIa
IIIb
IVa IVb 1b
2 R3-CH=ONH2
OHHS
OH
NOS
CH2OH
R3 R3
5
R3 = H, Ar"2-(hydroxymethyl)cysteinol"
Scheme 3.
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9967
Author's personal copy
previously reported2a,b the isolation of a type 3 dimeric S,S-cis-fused oxazolidine (Scheme 1) as a side product of the reactionbetween IVb (6% partial conversion) and two molar equivalent offormaldehyde, yielding derivative 1b (R3¼H, 94% conversion,Scheme 3). Therefore, we have been cautious ever since becauseformaldehyde acted not only as an electrophile, but also as an ox-idant acting on IVb. Therefore, it was clear to us that the thiamin-alisations depicted in Scheme 4 ‘trapped’ IVb using our carbonylelectrophiles with moderate success.y The lowest yield was in thecase of p-nitrobenzaldehyde, which contained a nitro group thatcan act as another oxidant for the thiol group (compound 4a,method A),7c and the results confirmed the above suspicion. Even ifmilder conditions were applied, e.g., C, the result was almost thesame. All of the other investigated aryl-(di)aldehydes, bei, dem-onstrated that our protocols were viable for generating the productin medium to satisfactory yields. Because the dimeric compound 4jexhibited a low solubility in common solvents, it was isolated asa material contaminated with 5% of its precursor, the mono-thiazolidine 4i.
The reactions were monitored using TLC (UV, 254 nm), whichdisplayed the normal evolution of our reactions towards the de-sired structures. In contrast, additional TLC visualisation in an I2bath allowed the detection of non-aromatic side products, whichresulted from the oxidative degradation of IVb. To eliminate theseproducts in three cases, which included 4b, 4c and 4jþ4i, directcrystallisation was successful. The other members of series 4
required purification by column chromatography, which wasmonitored by double TLC control (vide supra), followed by crys-tallisation. Once isolated, 4aej were stable indefinitely under am-bient conditions.
2.1.2. Structural assignments. Our structural investigations werebased on NMR (in solution) and IR spectroscopy (in solid state) intandem with DFT calculations.
The NMR spectra of condensates 4aej established their identity(Scheme 4) as exclusively (4aef, 4h, 4i) or largely (4g, 4j) the 1,3-thiazolidine forms (Fig. I-SD and Table I-SD in Supplementarydata). The 2D-1H,1H-NOESY chart for compound 4a (Fig. 1, dashedred lines) revealed that the thiazolidine ring faces were stericallynon-equivalent (cis vs trans) (Scheme 5) because only one nOe in-teraction H-2/CH2OH was observed.
Therefore, the saturated heterocyclic skeleton was hetero-facial,12 with two stereogenic centres located at C-2 and N-3. Incontrast to the other 1,3-thiazolidines, e.g., of type Ia (Scheme 2)6b
or those built on thionated (pseudo)ephedrines,9 no epimerisationoccurred at C-2. Therefore, we could adopt the Ar ligand at C-2position as a fiducial substituent.12 The remaining exocyclic li-gands, NeH, C(5)eH and CeCH2OH are henceforth referred to withthe use of descriptors: -c (cis) and -t (trans). However, the spatialproximities, which were described by the 2D-1H,1H-NOESY chartof compound 4a (Fig. 1), were consistent with the usual rapidpseudorotation equilibria between three dominant Envelopeconformers, 4a (EC-2), 4a (EN) and 4a (EC-4). The conformation of ENwas also assigned by X-ray analysis of a dimeric N,N0-methyl-enebis-1,3-thiazolidine,5a while the EC-4 conformer was elucidatedby 1H NMR spectroscopy as early as 1974, which occurred in thecase of 2-tert-butyl-1,3-thiazolidine.10a The 1H NMR data of
direct crystallisation. cSide product in the synthesis of 4j + 5j
dAs total conversion of IVa into 4i, 4j and 5j
HSNH2
OH
OH
IVa IVb
HS N
CH2OHCH2OH
R1 R1
= R2 = R3A - C
4a-i4j (R1 = p-R2)
5g (R1 = o-OH)5j (R1 = p-R3)
Scheme 4.
y Saiz et al. (Ref. 7b) indirectly supported this observation by reacting the1,3-thiazolidin-2-thione IIIb (Scheme 3), rather than IVb, with 2 mol. equiv. of thesame R-substituting aryl-aldehyde, R¼p-Cl, p-Br, m-Br, p-F, p-F3C and p-, m-diCl.
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859968
Author's personal copy
compound 4a were extrapolated to the entire series 4aej, focus-sing our investigations on four major items (Scheme 5, Fig. 1 andTable I-SD):
(a) Preferred arrangement of the Ar C-2 ligand
The Ar C-2 ligand was placed in a pseudo-equatorial position, asin conformers 4a (EC-2)$4a (EN), with a dominant bisectionalorientation. This stereochemistry promoted the lower field ab-sorption of the CH2OH-c protons relative to the CH2OH-t. Mutatis-mutandis, d(OH-c) was higher than d(OH-t) by approximately0.2e0.3 ppm. In contrast, the C-5 methylene was, magnetically,more influenced by the proximity of the sulfur, which was similarto other (condensed)-1,3-thiazolidine systems.3b Therefore, thedata indicate that d(H-5-c)<d(H-5-t).
(b) Assignment of the relative configuration at the N-3 position
Except for compound 4b, a noteworthy vicinal 3JH,H (NHeCH)coupling pattern, which was approximately 11.5 Hz, was observed(e.g., thiazolidine 4a, Fig. 1, yellow arrows). Due to the fast pseu-dorotation of the pentaheterocycle, this was an averaged J value,mediating the environments illustrated in Scheme 5. However, the
magnitude of the 3JH,H (NHeCH) coupling pattern allowed us toestimate the NH proton life time of its spin state, s(s), where s>1/3J[Hz, s�1]w0.09 s.13a The size of the s value precluded the N-pyra-midal inversion,9,13b as well as an acidebase interchange, implyingthe existence of an NH group.13c The same shape of the NH reso-nance (in CDCl3) in epimeric 2-aryl-1,3-thiazolidines Ia and Ib(Scheme 2, R1¼Ar) was listed, with no comments, by Saiz et al.7a Toour knowledge, to the above unusual splitting of an amine proton(Fig. 1),13a,b no attention was paid by other authors,3e,5a,b,6a,9,10a,11b,c
including those investigating 2-aryl-1,3-thiazolidines.3e,6a,11b Basedon the 3JH,H (NHeCH) value and the NOESY chart for compound 4a,we ascertained a pseudo-trans (axeax$axeeq) relationship be-tween H-2-t/NH-c (Scheme 5), in spite of the inadequacy of theKarplus equation10a,14 for this fluctuation of the HeNeCeH di-hedral angle magnitude in a five-membered saturated hetero-cycle15aec z. In addition to the C-2 (R*), an N(R*) relativeconfiguration, which is stable for at least 0.09 s, can be assigned.
Fig. 1. Relevant nOe interactions (dashed red lines) in the 2D-1H,1H-NOESY experiment of compound 4a (on 500 MHz timescale, DMSO-d6, 298 K) and the scalar 3JH,H (NHeCH)coupling (yellow arrows).
SN
SN
H*-t
H
CH2OHCH2OH-c
HH
HH
OH-t OH
H-c
H-t
H
H
O2N O2NH-c H
24
5
SCH
2OH
CH2OH
H
HN
H
H
O2N
4a (EC-2) 4a (EN) 4a (EC-4)*bold face: nOe interactions between protons in a given conformer as disclosed by the 2D-1H,1H-NOESY Chart of compound 4a (Figure 1)
cis face
trans face
Scheme 5.
z A similar 3JH,H (NHeCH) coupling in DMSO-d6 was previously detected, with nocomments, even on 60 MHz 1H NMR timescale in 2-aryl-1,3-oxazolidines, i.e.,O-analogues of 4 (Scheme 3).15a We also described this splitting in various2-aryl-1,3-oxazolidine related systems.15b,c It does not depend on the ability of theNMR solvent to act as hydrogen bond donor (DMSO-d6) or acceptor (CDCl3).15d
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9969
Author's personal copy
(c) The anomeric effect occurring in the pNeCeSe sequence
We mention the related configurational analysis of S�elambaronet al.5a concerning the molecular structure of a dimeric N,N0-methylenebis-1,3-thiazolidine of type Ia (Scheme 2). The pseudo-axial orientation of its N,N0-methylene exocyclic ligand againstthe pseudo-equatorial orientation of the nitrogen lone pair (lpN-eq)were put in conjunction with the following: (i) the decrease in thebond length of NeC(2) (�0.019 �A) versus NeC(4) and (ii) the in-crease of the SeC(2) bond length (þ0.035 �A) against SeC(5). Thesefindings provide evidence for an endo anomeric effect arising fromawell-positioned donor, lpN-eq, overlapping with an appropriatelyoriented acceptor, which is the antiperiplanar antibonding s*(CeS)orbital. According to the literature,16a this is a general behaviouroccurring in saturated 1,3-X,N pentaheterocycles (X¼O,2b,15c,16,17S2b,5,10c), i.e., in those containing the pNeCeXe sequence. The1,3-thiazolidine system is, however, much less explored.
Because we were unable to obtain single crystals for a similarX-ray analysis of series 4, we performed DFT calculations on com-pounds 4a, 4d, 4h and 4g; these compounds were chosen as rep-resentative due to the different electronic and steric influence ofthe Ar-ring R1-substituents (Charts 1 and 2, Table 1).
From several molecular arrangements, the optimised geome-tries, which were predicted by these calculations, agreed with theconformations of the envelope EC-2, EN and EC-4 types described by
1H NMR data analysis (Scheme 5, Chart 1). The two most stableconformers were differentiated by the steric relationship betweenligands H-2 and NH. The pseudo-trans arrangement of the NH-c/H-2-t in 2-aryl-1,3-thiazolidines 4a, 4d and 4h, conferred more sta-bility than the alternative pseudo-cis arrangement in the gas phase,as well as in solution (Table 1). The differences in stability D(E, H orG298)<0, pseudo-trans against pseudo-cis, increased with the in-creasing electron-donating character of the p-R1 substituentNO2<H<NMe2 (Table 1).
The noticeable ‘reversed’ stabilisation in the case of compound4g, D(E, H or G298)>0, pseudo-trans against pseudo-cis, was notsurprising because of the expected intramolecular hydrogen-bonding, bisectional-phenol o-(O)eH/lpN-ax (thiazolidine).
In the ‘thiaminalic zone’ (pNeCeSe) of the selected compounds,the shortening of the NeC(2) bond with respect to NeC(4)[w(�0.029 �A) in gas phase, w(�0.025 �A) in DMSO] alongside theincreased length of the C(2)eS bond relative to SeC(5)[w(þ0.065 �A) in gas phase, w(þ0.077 �A) in DMSO] were morepronounced in conformers 4 (pseudo-trans).
To examine these geometric fluctuations with respect to themanifestation of an anomeric effect we performed NBO and3JH,H(NHeCH) coupling constant analyses for 1,3-thiazolidine 4d atthe B3LYP/6-311þþG** level of theory (Chart 2).
The Edel values for the hyperconjugation of lpN-eq/s* C(2)-S in4d (pseudo-trans) were the highest encountered in our investigation
Chart 1. Optimised geometries of 2-aryl-1,3-thiazolidines 4a, 4d, 4h and 4g in the gas phase and in DMSOa at the B3LYP/6-311þþG** level of theory.
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859970
Author's personal copy
Chart 2. Relevant energies (Edel.) of the NBO hyperconjugative interactions and 3JH,H(NHeCH) coupling constant in the gas phase and in DMSOa (at the B3LYP/6-311þþG** level oftheory) in 1,3-thiazolidine 4d.
Table 1Computed energies of 2-aryl-1,3-thiazolidines 4a, 4d, 4h and 4g in the gas phase and in DMSOa (at the B3LYP/6-311þþG** level of theory)
Stability, (kJ/mol) as 4(pseudo-trans)b�4 (pseudo-cis)b
Selected bond distances, d (�A), contractions Dd (�A)as [NeC(2)]�[NeC(4)]<0, elongations Dd (�A) as [SeC(2)]�[SeC(5)]>0
Selected hydrogen bond distances, d (�A)c,d
DE DH DG298 NeC(4) NeC(2) SeC(5) SeC(2) (N)H/O (O)H/N o-OH/Xe
Gas phase�6.44 a �5.40 a �6.01 a 4a, 4d, 4h, 4g (pseudo-trans)�8.55 d �7.69 d �9.21 d 1.482 a 1.449 a 1.829 a 1.885 a 2.369 a 2.441 a d
�9.45 h �8.49 h �9.51 h 1.480 d 1.451 d 1.829 d 1.887 d 2.409 d 2.381 d d
þ 0.10 g �0.12 g �1.04 g 1.479 h 1.452 h 1.828 h 1.891 h 2.437 h 2.371 h d
1.468 g 1.443 g 1.836 g 1.919 g 2.797 g 2.308 g 2.259 g�0.033 a þ0.056 a�0.029 d þ0.058 d�0.027 h þ0.063 h�0.025 g þ0.083 g4a, 4d, 4h, 4g (pseudo-cis)1.488 a 1.464 a 1.832 a 1.850 a 2.372 a 2.402 a d
1.486 d 1.467 d 1.832 d 1.851 d 2.395 d 2.351 d d
1.485 h 1.470 h 1.831 h 1.855 h 2.411 h 2.321 h d
1.489 g 1.478 g 1.833 g 1.854 g 2.328 g 2.718 g 1.838 g�0.024 a þ0.018 a�0.019 d þ0.019 d�0.015 h þ0.024 h�0.011 g þ0.021 g
DMSO�6.89 a �5.97 a �7.64 a 4a, 4d, 4h, 4g (pseudo-trans)�9.86 d �8.85 d �10.89 d 1.471 a 1.445 a 1.838 a 1.908 a 2.938 a 2.430 a d
11.22h �10.33h �11.80h 1.471 d 1.446 d 1.837 d 1.910 d 2.936 d 2.381 d d
þ7.39 g þ6.92 g þ4.53 g 1.470 h 1.448 h 1.836 h 1.917 h 2.932 h 2.372 h d
1.470 g 1.444 g 1.838 g 1.923 g 2.901 g 2.427 g 2.236 g�0.026 a þ0.070 a�0.025 d þ0.073 d�0.022 h þ0.081 h�0.026 g þ0.085 g4a, 4d, 4h, 4g (pseudo-cis)1.489 a 1.464 a 1.832 a 1.852 a 2.440 a 2.382 a d
1.488 d 1.467 d 1.832 d 1.853 d 2.450 d 2.339 d d
1.487 h 1.471 h 1.832 h 1.858 h 2.456 h 2.314 h d
1.480 g 1.474 g 1.844 g 1.874 g 2.464 g 3.002 g 1.798 g�0.025 a þ0.020 a�0.021 d þ0.021 d�0.016 h þ0.026 h�0.006 g þ0.030 g
a PCM calculation using the CPCM polarisable conductor continuum model.b The lpN and the proton H-2 are the references for the descriptors pseudo-cis and pseudo-trans.c Intramolecular hydrogen bonds involving the geminal hydroxymethyl groups only are not listed for reason of simplicity.d SrvdW (O, H) (sum of the van der Waals radii)¼2.60 �A.e X¼S in 4g (pseudo-trans), X¼O in 4g (pseudo-cis).
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9971
Author's personal copy
[50.12 kJ/mol (gas phase), 55.35 kJ/mol (DMSO)] and are in line withthe modifications of the bond distances in the same region.Wewereable to assign the structures of our 1,3-thiazolidines using ‘bonding/non-bonding’ formulations (Scheme 6).17
Examination of the hydrogen bond ‘network’ betweenNH/CH2OH-c/CH2OH-t in the two types of conformers (Chart 1) insolution exposed normal interactions, except for the NH-c (pseudo-axial) hydrogen in 4 (pseudo-trans); the distances for this exception[(N)H/O (w2.927 �A)] were the only ones greater than the sum ofthe van der Waals radii (O, H), 2.60 �A. These exceptions we con-sidered as indirect evidence for the magnitude of the endoanomeric effect. This anomeric effect caused the NH-c hydrogen toadopt a pseudo-axial orientation, therefore precluding its partici-pation in an (N)H/O(H) association. To this end, the 4d (pseudo-trans) calculated value of the 3JH,H(NHeCH) coupling in solution(11.7 Hz, Table 1) validated our experimental data, which was11.5 Hz (Table I-SD).
Additionally, these 3JH,H (NHeCH) constants are typically foundin motifs such as �CHeNHe(C]X)e4�CHeNHþ](CeX�)e,(X¼O, S, NH, Co, etc.),13a,b i.e., in those containing the nitrogen in-volved in partial double bonds formed because of a stabilisinglpN/p conjugation. According to our data, a lpN/s* hyper-conjugative relationship in pNeCeXe (X¼S, O) sequences, sym-bolised as eNHeCHeXe4eNHþ]CH] [X�dcan be deduced aswell; this deduction requires a 3JH,H (NHeCH) value. This diagnosisalso applies for compound 4g, in which the (pseudo-trans) con-former, which displays identical 3JH,H (NHeCH)¼11.1 Hz values(calculation vs experiment, Table I-SD) was less stable than the(pseudo-cis) one.
(d) The ring-chain tautomerism
The ring-chain tautomerism of Schiff bases derived from (het-ero)aryl carbonyl compounds and amino(poly)(thi)ols involves thereversible nucleophilic addition of an XH group to an imine doublebond and is well established (Scheme 7).11 The situations in whichX¼S and n¼0, have received little attention.6a,10c,11b,c To examinethe condensates of IVbwith aryl-(di)aldehydes aei (Scheme 4), ourstudies utilized a combination of IR (in solid state) with NMR (insolution) data.
In the solid state (Chart 3), the IR spectra of the mono-condensates 4aeh displayed a typical band located in the aro-matic region at approximately 1600 cm�1; we assigned this reso-nance to a conjugated nC]C
sk absorption and not to a nC]Nsk one.13b
The dimeric condensate 4j exhibited the same band at 1606 cm�1
(Chart 3), but it was very weak (almost flat) and therefore consis-tent with the presence of amajor Ci symmetricmeso-4j s-trans formin the solid state. The signal observed at 1703 cm�1 ðnC]O
skÞ in-dicated the previously mentioned contamination of compound 4jwith 4i (Scheme 4, see also Fig. II-SD).
In solution (Scheme 4, Fig. 1, Fig. I-SD and Table I-SD), the NMRspectra identified the ‘cysteinolic’ mono-condensates of aryl-aldehydes aef, h and i as 1,3-thiazolidine ring forms only. Mean-while, the mono-condensate of salicylaldehyde g, was a mixturewith a constant composition: 4g (1,3-thiazolidine) 72% versus 5g(Schiff Base) 28% (Figs. III-SDeV-SD). Therefore, the ring-chaintautomerism in series 4aei existed only in the case of 4g andmanifested as a spontaneous equilibration attained when passingfrom the solid state to solution, i.e., a partial (28%) ring-opening of4g towards 5g. In line with the literature data,11bed this generalbehaviour of series 4aei is not surprising considering the muchhigher nucleophilicity of the SH group relative to the OH group.
We compared the ‘cysteinolic’ mono-condensates of aryl-aldehydes g and h with their ‘serinolic’ O-analogues, based onTRIS (Chart 4), to determine how this difference applied to oursystem.15cWhile the salicylaldehyde derivative of TRIS existed 100%as 7g (Schiff base, E-diastereomer), the ‘cysteinolic’ mono-condensate consisted of a mixture of 4gþ5g (Table I-SD, Fig. III-SD). As early as 1990,18 the stabilisation of the condensates ofa plethora of aminoalcohols with salicylaldehyde as the E-di-astereomeric Schiff bases was explained by an intramolecular hy-drogen bond generated in six-membered chelates (Chart4);2c,13a,15b,c,16a this ‘rule’ applied, but only partially (14%) in thecase of the Schiff base 5g. The 13C NMR spectrum indicated that 5gwas a 1:1 E/Z mixture (Figs. IV-SD and V-SD). Next, even thestrongest electron-donor p-substituent, Me2N in aryl-aldehyde h,could not stabilise any ‘cysteinolic’ Schiff base (Table I-SD) in so-lution. Only the 1,3-thiazolidine 4h was detected.
The ring-chain tautomeric behaviour of the dimeric thiazoli-dine 4j (Chart 5) in solution was similar to the behaviour of 4g;a spontaneous but minor ring-opening (1%) occurred towards thecorresponding double Schiff base 5j (Figs. VI-SDeVIII-SD). Thetautomeric composition remained constant over time. Theamount of the double Schiff base 5j was too small to ascertainany stereochemistry; therefore it was considered to be an un-determined distribution of (E,E)-, (Z,Z)-, and (E,Z)-diastereomericspecies. In contrast to the corresponding TRIS derivative of ter-ephthaldialdehyde,15c no ‘mixed tautomer’ of type Schiffbased1,3-thiazolidine could be detected. Therefore, only thestereochemistry of the major double ring form 4j in solutionmeso- and/or rac- was left to be assigned. The 1H and 13C NMRspectra, when collected with increased sensitivity, exposeda unique and complete set of signals fully consistent with both ofthe configurational diastereomers. Historically, we have found 4jto be poorly soluble in almost all suitable solvents for elucidatingthis type of structural mystery. Separation by HPLC (appropriateeluent EtOH abs) was unsuccessful. The use of a CSR (chiral shiftreagent) based on EuIII, such as Eu(hfc)3,13a,16b was ruled outbecause only DMSO-d6 was practical for NMR investigations.Therefore, we concluded that the meso-4j form, which was pre-viously determined to be the major constituent in the solid state,was also the major species in solution due to the lack of addi-tional evidence otherwise.
2.2. Synthesis and structure of novel c-5-hydroxymethyl-3-oxa-7-thia-r-1-azabicyclo[3.3.0]octanes
2.2.1. Synthesis. When we attempted the one-pot double ring-closure of the in situ-generated thioamino-1,3-diol IVb with twomolar equivalents of benzaldehyde, our results were very
(Het)Ar: (Hetero)Aryl groupRG: Releasing Group (Me, Et), optionally H
n = 0, 1; m = 0 - 2; X = O, NH, S
Scheme 7.
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859972
Author's personal copy
disappointing (Chart 6). In addition to the oxidative alteration ofIVb, its partial conversion also indicated intermediate 4d was sur-prisingly unreactive towards benzaldehyde. Diastereomers 8 werepreviously investigated via DFT by Saiz et al.,7b but there was nosynthesis to support the calculations. In our hands, the experi-mental difference in stability, DG298¼G298[8(c,c)]�G298[8(2-t,8-c)],was about �1.31 kJ/mol, which was much smaller than the valueprovided by the computations (�7.10 kJ/mol). The mixture of 4dþ8could only be isolated by column chromatography as a single mixedfraction. If we increased the electrophilicity of the aryl-aldehydepartner, i.e., by using two molar equivalents of p-nitro-benzaldehyde, under the same conditions (Chart 6), then only thethiaminalisation process took place to furnish thiazolidine 4a
(Scheme 4). We deduced that the nucleophilicity of the NH groupon the 2-aryl-1,3-thiazolidines 4a and 4d might have been di-minished by the anomeric effect (Section 2.1.2(c), Scheme 6).5b
Therefore, we changed our strategy; a stoichiometric amount ofp-formaldehyde was used as the oxaminalisation electrophile withthe isolated members of the series: 4aeh, 4j. In a test experiment,no reaction occurred between 4d and formaldehyde in ethanol atroom temperature within 24 h. With heating, monitoring by TLCdetected a near complete trans-thiaminalisation, producing theknown derivative 1b (Scheme 3, R3¼H).2a,b In contrast, the syn-thesis performed in benzene, which utilized a DeaneStark trap forthe continuous removal of water, provided both positive and sur-prising results (Chart 7).
Chart 3. Details form IR spectra (KBr) of condensates 4aeh and 4j indicating their thiazolidine nature.
X NH
OH
CH2OHCH2OHC(CH2OH)2HXH2C
N
OH
H-im
X NH
CH2OHCH2OH
NMe2
C(CH2OH)2
HXH2C
NH-im
NMe2
C(CH2OH)2N
CH2XHHO
H-im
X = SX = O
(E)-5g
(E)-7g
4g
6g
(Z)-5g
(Z)-7g
4h
6h
(E)-5h
(E)-7h
2
2
Chart 4. Comparative ring-chain tautomerism data of the mono-condensates of ‘2-(hydroxymethyl)cysteinol’ (IVb) and ‘2-(hydroxymethyl)serinol’ (TRIS) with salicylaldehyde (g)and p-dimethylaminobenzaldehyde (h).
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9973
Author's personal copy
No acid catalyst, such as p-toluene sulfonic acid (PTSA), wasneeded, most likely because of the traces of formic acid usuallypresent in p-formaldehyde. The quantitative conversion of 4dproduced a mixture of the regioisomers, 9d (c) and 10d (c), whichwere isolated by column chromatography as a mixture (singlefraction). The starting material 4d did not act as an a priori N-,S-protected form of IVb because the reaction also resulted in analmost complete ‘migration’ of the benzaldehyde, eSe(Ph)CHeNo(1,3-thiazolidine)/pNe(Ph)CHeOe (1,3-oxazolidine).
This behaviour has not yet been reported in the literature.Our results prompted us to screen, the rest of the 2-aryl-1,3-
thiazolidines, 4aec, 4eeh and 4j (Chart 8), under identical condi-tions (Chart 7).
According to our data, the above reactivity of the benzaldehyde-based 1,3-thiazolidine 4d was the exception rather than the rule.The repeated failures to obtain thiazolidin-oxazolidine condensedsystems based on the tandem p-chlorobenzaldehyde (b) or m-hydroxybenzaldehyde (e)/formaldehyde systems should be viewedexceptions, as well; the explanation for these systems is still ob-scure to us. These observations led us to consider the influence of
the aryl-ring substituent R1 over not only the regioselectivity of thecyclocondensations, Ar-8 (series 9) versus Ar-2 (series 10) but alsothe diastereoselectivities, Ar-8(2)-c versus Ar-8(2)-t. In this context,when using a long reaction time (13 h), we observed a thermody-namic distribution of the products.
First, the regioselectivity was elucidated using 13C as Jmod, DEPTand 2D-1H,13C-HSQS NMR experiments [Figs. IX-SD and X-SD, com-pounds 9d (c),10d (c)]. The spectral data for the 1,3-thiazolidine ringdisplayed dC (eSeCH2eNo)w56 ppm but w75 ppm in aneSeCH(Ar)eNomotif; meanwhile, in the 1,3-oxazolidine condensedcounterpart, dC (pNeCH2eOe) was located at w86 ppm but atw95 ppm appeared the pNeCH(Ar)eOe sequence. Next, the 2JH,Hcoupling patterns at C-4 displayed values typical for cis-fused 1,3-oxazolidine systems 3 (Scheme 1), 8.5e9.0 Hz15b,c,17b,c,19 (8.5 Hz inthe parent compound 1b, R3¼H, Scheme 2).2a,b Finally, the 2JH,Hvalues at C-6, 12 Hz were identical with those in the reference 1b(R3¼H, Scheme 2).2a,b
We determined the stereochemistry of the diastereomers startingfrom the 2D-1H,1H-NOESY data for compounds 9d (c)þ10d (c) [Chart7, Fig. XI-SD, 9j (c, c0) and 10j (c, c0)] (Figs. XII-SD and XIII-SD). Overall,the configurational assignments listed in Chart 8 agreed with ourprevious findings for the cis-fused 1,3-oxazolidine systems 3 type
S NH
CH2OHCH2OH C(CH2OH)2HSH2C
N
S NH
CH2OHCH2OH
SHN
CH2OHHOH2C
S NH
CH2OHCH2OHCH2SH(HOH2C)2C
N
H-im
H-im
meso-4j 94%
R
S
S (R)
S (R)
rac-4j(E,E)-, (Z,Z)-, (E,Z)-5j 1%
2 2
2' 2'
Chart 5. 1H and 13C NMR double ring-chain tautomerism data concerning the double-condensate of ‘2-(hydroxymethyl)cysteinol’ (IVb) with terephthaldialdehyde.
NOS
NH3+Cl-OHHS
OH 2.20 eq. Ph-CH=O0.46 eq. K2CO30.06 eq. H2OBenzene - Dean Stark trapreflux / 13 h / N2
S NH
CH2OH-c*
CH2OH-t
Ph H-t* Ph Ph NOS
Ph Ph
IVa 4d (49%)**8 (c, c) (5%) 8 (2-t, 8-c) (3%)
2
OH-c OH-cH-t
5H-c
*Stereochemical descriptors c (cis) and t (trans) are referred with respect to the fiducial substituents (r) : Ph (in 4d) and the lone pair at N-1 in 8.**Partial conversions of IVa
H-t
H-t H-t H-t H-c
H-t
4 6 48 2 8 2
Chart 6. Results of the cyclocondensation of ‘2-(hydroxymethyl)cysteinol’ IVbwith 2 mol. eq. of benzaldehyde and the NMR assignment of the product nature and stereochemistry.
NOSS NH
CH2OHCH2OH
PhPh
H
NOS
HPh
OHOHH1 eq.
H2C=OBenzenereflux13 h
4d 9d (c) 10d (c) (8%)* (92%)
8 2 8 2
H
HH
*Partial conversions of 4d
6 4
Chart 7. Results of the cyclocondensation of 1,3-thiazolidine 4d with 1 mol equivalentof formaldehyde and the NMR assignment of the product nature and stereochemistry.
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859974
Author's personal copy
derivatives,2c,16b,c,19 (Scheme 1); the ‘all-cis’ spatial arrangements ofthe ligands lpN(1), C(5)eCH2OH and Ar-2(8) predominated.
2.2.2. Rationalisation of the synthetic results. With the above as-signments in mind, we dealt with providing plausible mechanisticexplanations, which will be referred to as routes ieiv, for the regio-and diastereoselectivities listed in Chart 8.
(i) Route towards thiazolidin-oxazolidines 9 as the main direction(Scheme 8)
The major regioisomers 9a, 9c, 9feh and 9j were obtained viathis route. This route required the preliminary formation of theexpected N-substituted methyleneiminium cations 11 with pNþ]
CH2 diastereotopic faces. The Si face was more exposed to the nu-cleophile addition of CH2OH-t, and this addition afforded mostlythe 9 (c) series. The reaction with CH2OH-c on the Re face wouldencounter the Ar ligand as a steric obstruction; this geometry was
reinforced by the change of the N-hybridisation, from sp3 in 4 to sp2
in 11. Formation of the minor diastereomers 9 (t) involved thepartial inversion of the thiazolidine 4 (2R*,3R*) relative configura-tion to 1S*,5R*,8R* in series 9 (t).
The diastereoselectivity for 9 (c)/9 (t) was approximately 96:4,which we could also justify by considering a more appropriatecationic intermediate, i.e., 1-thionia-4-azabicyclo[2.2.1]hexane 12.With this intermediate, the bridged C-5 methylene group wasplacedwithin the exclusive vicinity of the CH2OH-t nucleophile. Weviewed the generation of species 12 to be another consequence ofthe diminished nucleophilicity of the NH group in the 1,3-thiazolidines 4, as well as the increase in nucleophilicity fur-nished by using sulfur, due to the anomeric effect (Section 2.1.2(c),Scheme 6).5b The main 9 (c) diastereomers preserved the initialrelative configuration of the thiazolidine 4 (2R*,3R*), but the con-figuration of the bicyclic skeleton (1R*,5S*) was inverted with re-spect to 9 (t). The most illustrative example of the proposedmechanistic model in Scheme 8 was provided by compound 9h (c),
NOS
CH2OH
S NH
CH2OHCH2OH
1.0 eq. H2C=OBenzenereflux13 h
R1
NOS
CH2OH
R1
NOS
CH2OH
R1 R1
n = 1.0 eq. 4a-c, 4e-h 9a-c, 9e-h; 10a-c, 10e-h; R1= a (p-O2N); b (p-Cl); c (p-Br); e (m-HO); f (p-HO); g (o-HO); h (p-Me2N)n = 0.5 eq. 4j (R1 = p-R2) 9j (R1 = p-R4, R1 = p-R4'); 10j (R1 = p-R5, R1 = p-R5')
Ar-8(')-c Ar-8(')-t Ar-2(')-c Ar-2(')-t
series 9* series 10*
NOS
CH2OH
R1
158
51
2
*Represented as 1R*,5S* relative configurations of the unsubstituted cis-fused thiazolidin-oxazolidine skeleton; for the inversion of these configurations1R*,5S* 1S*,5R* see discussion and Schemes 8-12.
= R2R4 = R4' = = R5 = R5'
n eq.
Chart 8. Results of the cyclocondensation of 1,3-thiazolidine 4aec, 4eeh and 4j with formaldehyde and the 1H NMR assignment of the product nature and stereochemistry.
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9975
Author's personal copy
which was isolated as a single stereoisomer (Figs. XIV-SD andXV-SD, see later discussion about the molecular structure of thiscompound in Figs. 2e4).
(ii) The blocked route of the Ar-group relocation (Scheme 9)
As shown in Chart 8, when the R1 ring substituent was a (strong)electron-donor Br (c) and Me2N (h), but not HO (f, g, see later dis-cussion), we did not observe the translocation of the Ar-group. Areasonable mechanistic model justifying this versatile ‘immobility’began from intermediates 12, 12c and 12h, and proceeds via anSN1-like mechanism of the 2-aryl-1,3-thiazolidine ring-opening.The C(6)eSþe bond cleavage was assisted by the high lp/p sta-bilisation provided by the p-R1 substituent. In the resulted benzyl-type carbocations and the free rotation about the C(benzyl)eN(3)bond ensured that the appropriate Z-geometry was adopted for thefuture iminium double bond. The generation of (Z)-13c, -13h, whichare ‘pushepull’ cationic Schiff bases, was the final step of this route.In summary, this mechanism proceeds as follows: isomerisation ofthe methyleneiminium cations 11c, 11h (Scheme 8), via 12c, 12h,into the more stable (Z)-13c, 13h (Scheme 9). This mechanismexplains why the 1,3-oxazolidine ring-closure did not occur; the‘pushepull’ cationic Schiff bases (suggested by the symbol ‘]þ’),which are weak, soft electrophiles, encountered two strong as hardnucleophiles, which were the hydroxymethyl groups. Overall, inagreement with the classical CurtineHammett principle (withWinsteineHolness additions),20 if R1 was Br (c) or Me2N (h), theequilibria were shifted back, to the ‘good direction’ i, with forma-tion of derivatives 9 (Scheme 8)
(iii) Route towards thiazolidin-oxazolidines 10 as the minordirection (Scheme 10)
In contrast, if the R1 substituent was not only a strong electron-donor but also easily deprotonated (p-HO f and o-OH g), the above
route (ii) (Scheme 9) could continue. For example, in the prochiral(Z)-13f species with pNþ]C(H)e enantiotopic faces, a prototropictautomerism involving the p-HO group and the thiazolidinium sp2
hybridised nitrogen might create the chiral 1,3-thiazolidiniumcation 14f that bears a p-quinoid like ligand with >C]C(H)e dia-
stereotopic faces (o-quinoid if R1¼o-OH, g) at N(3) position (inScheme 10, an S* configuration at N(3) position was adopted arbi-trarily). The acquired pyramidal sp3 configuration at N(3) in tandemwith the formation of free in (disfavoured)$out (favoured)rotamers about the p-quinoid ligand facilitated the diaster-eoselective interactions between the CH2OH (pro-S) nucleophile,resulting in either (pro-S)/Si-attack/10f (c) (major, 85%, Chart 8)or (pro-S)/Re-attack/10f (t) (minor, 15%). A similar model shouldalso apply to the case involving the o-HO substituent (g), but
S+
CH2OH
CH2OH
N
R1
N+SCH2OH
CH2OH
R1
X
12c, 12h
R1: Br (c), Me2N (h) (Z)-13c, -13h
1
6S
CH2OH
CH2OH
N
+R1
H
S
CH2OH
CH2OH
N
R1
H H
H1
3
4
+
+NS
CH2OHCH2OH
R1
H
••
••
Scheme 9.
SN
H
Ar CH2OH-c
Hcis face
trans face
SAr
N+
t-HOH2C
S
O
N
CH2
CH2OHAr-8-c N
S
CH2OH
ArS+Ar CH2OH
N
t-HO
Si
c-HORe
O
SN
CH2OH
NSO
CH2OH
Ar
H2C=O / H+
(-H2O)
Ar-8-t1
5
88
1
5
8 5
1 1
5
8
CH2OH-t
Re attack 4%)
- H+
Si attack( 96%)
- H+
1
4
2
3
5
6
11: Methyleneiminium type cation12: 1-Thionia-4-azabicyclo[2.2.1]hexane
type cation
4
(2R*,3R*)
9 (c) (1R*,5S*,8R*)
O
23
9 (t) (1S*,5R*,8R*)3
2
11
12
Scheme 8.
NS
CH2OH (pro-S)CH
2OH
HSi - attack
- H+
NOS
CH2OH
Ar-2-c
1
5
2
NS
CH2OH (pro-S)CH
2OH
H
O
Re - attack NOS
CH2OH
Ar-2-t
1
5
2- H+
14f "out"
14f "in"
10f (c)(1R*,2S*,5S*)
10f (t)(1R*,2R*,5S*)
3
3
H
+
H
+
S*
S*
ON+
SCH2OH
CH2OH
O
(Z)-13f
H
H(- H+)
(+ H+)
Scheme 10.
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859976
Author's personal copy
obviously not for m-HO (e) because the qualitative regio- and dia-stereoselectivities between (f) and (g) were comparable.
(iv) Route towards thiazolidin-oxazolidines 10 as the main direction(Schemes 11 and 12)
This route was dominant in the absence of an electron-donatingsubstituent on the phenyl ring (d) that was able to stabilise thecorresponding benzyl carbocations. However, the generation of thesame type of intermediate as 12 (Scheme 8), in this case 12d(Scheme 11), allowed an extension of the mechanistic approachesalready proposed. Therefore, the 12d ring opened in a concertedprocess to form the iminium pNþ]C(6) double bond with C(6)eSþ
bond cleavage. Consequently, the resultant cationic Schiff base 13dhad an E-configuration for the pNþ]Co double bond. The majordifference between (E)-13d and (Z)-13c, -13h, -13g, -13h was itsstabilisation; with (E)-13d, the benzyl carbon was stabilised by thethiazolidine nitrogen rather than via resonance with the phenylring. Consequently, the enantiotopic faces of the iminium doublebond were equally exposed to the addition from the enantiotopichydroxymethyl nuclophiles to determine the same (Ph-2-c) finalorientation of the Ph ligand.
In particular, the same mechanism (Scheme 11) should be evenmore compelling if the R1 substituent was a strong electron-acceptor, such as p-O2N (a), to explain formation of the minorregioisomer 10a (c) (6%) (Chart 8, Scheme 12).
The 10a (t) (4%) diastereomer was formed in a subsequent epi-merisation step assisted by the electron donation from oxygenmoieties, such as prochiral E-oxonium cations: Ar-2-c$Ar-2-t(oxazolidine ring-opening$ring-closure16b). During the equili-bration, the relative configuration of the cis-fused thiazolidin-oxazolidine skeleton was inverted (1R*,5S*)$(1S*,5R).
2.2.3. Additional structural evidence. Finally, the proposed mecha-nistic models also could still operate with the 1,4-terephthaldialdehyde derivatives to yield the regioisomeric di-mers 9j (c, c0) and 10j (c, c0) in a 47:13 ratio, respectively. As2D-1H,1H-NOESY experiments revealed (Figs. XII-SD and XIII-SD),all of the 1,4-phenylene connections between the bicyclic units
were either 8(0)-c in 9j or 2(0)-c in 10j regardless the regioisomerism.Therefore, the remaining problem was another attempt to differ-entiate the configurational diastereomers, whichweremeso againstrac, for the dimeric condensates 9j and 10j (Chart 9).
Towards this purpose, we studied 9j and 10j as the para-phe-nylene derivatives bearing two thiazolidin-oxazolidine dipolar
units, which were homomorphic in rac-9j, -10j but heteromorphicin meso-9j, -10j. Subsequently, to ensure an equal steric compres-sion at ortho-phenylene positions, only the s-trans conformers wereadopted (Chart 9). These structures are Ci symmetric in the meso-form, but C2 symmetric in the rac-forms. Consequently, the twobicyclic dipoles were remote in meso-9j, -10j but close in rac-9j,-10j. Therefore, from a magnetic point of view, the deshieldinginfluence promoted by the bicyclic dipoles was expected to behigher in rac-environments than in the meso-environments. Thisinfluence was detected, but was very weak: � 0.01e0.02 ppm, on13QC NMR 125 MHz timescale, in the case of compound 9j and�0.01 ppm, on 1H NMR 700 MHz timescale with respect to OHprotons in 10j. The diastereoselectivity (meso- against rac-) wasslightly higher for the oxazolidineeoxazolidine coupling of 10j thanfor thiazolidineethiazolidine of 9j.
The IR spectra of dimers 9j and 10j (Fig. XVI-SD) did not differ;when compared to the IR spectra of the 4j precursor (Chart 3), theconjugated nC]C
sk absorption at approximately 1600 cm�1 wasalmost absent, suggesting that these structures primarily adopt themeso-forms in solid state.
Compound 9h (c) produced diffractable crystals suitable forX-ray analysis. This derivative crystallises as a racemate (Figs. 2 and3) with four molecules in the unit cell (Fig. 4).
As in cis-fused 1,3-oxazolidine systems 3 (Scheme 1),16 a contrac-tionof theN(1)eC(2)bond (�0.043�A) relative to theN(1)eC-(5)bondwas observed and the bond length of O(3)eC(2) increased (þ0.024�A)relative to O(3)eC(4). These fluctuations are related to the manifes-tation of an endo anomeric effect, which is a hyperconjugation in-volving the lpN(1) (donor)/s*O(3)eC-(2) (acceptor) orbitals.According to the crystallographic data, the a priori antiperiplanarorientation required by this stereoelectronic relationship occurs be-cause the O(3)-anti envelope geometry of the oxazolidine ring hasa relevant dihedral angle C(4)eC(5)eN(1)eC(2)/C(4)eO(3)eN(1) at
S+
CH2OH
CH2OH
N N+S
Ph
CH2OH (pro-R)
CH2OH (pro-S)N
OS
CH2OH
Ph
NOS
CH2OH
Ph
15
2
25
1
HH
1
46
12d
pro-S Si-face
pro-R Re-face
10d (c)(1R*,2S*,5S*)
10d (c)(1S*,2R*,5R*)(E)-13d
- H+
- H+
Scheme 11.
NOS
CH2OH
C6H4-NO2-pNH
S
CH2OH
O+
H C6H4-NO2-pH
2 NHS
CH2OH
O+
H C6H4-NO2-p
NOS
CH2OH
C6H4-NO2-p
+ H+
- H+
- H+
+ H+2
H
10a (c) 10a (t)(1R*,2S*,5S*) (1S*,2S*,5R*)
E
Re
E
Re
1
5 5
1
Scheme 12.
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9977
Author's personal copy
Chart 9. Configurational diastereomers of dimeric compounds 9j and 10j and their assignment by 13QC NMR (on 125 MHz timescale, DMSO-d6, 298 K) and 1H NMR (on 700 MHztimescale, DMSO-d6, 298 K).
Fig. 2. Molecular structure of compound 9h as the (1R,5S,8R) enantiomer, with 40% probability ellipsoids.
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859978
Author's personal copy
143.8�. In conclusion, the stereochemistry of the 1,3-oxazolidine ringin the solid state was dictated by themanifestation of the above endoanomeric effect.
In contrast, in the thiazolidine counterpart, only a small elon-gation of the S(7)eC(8) bond (þ0.013 �A) versus S(7)eC(6) wasobserved; meanwhile, the bond length N(1)eC(8) was identical toN(1)eC(5). Therefore, the opposite S(7)-syn orientation of thethiazolidine envelope should be viewed as a requirement for theequatorial-bisectional linkage of the aromatic ring, which was al-ready predicted by the 1H NMR data (Scheme 5) and DFT compu-tations (Chart 1).
Some interesting intermolecular associations deserve attention(Figs. 3 and 4). Intermolecular hydrogen-bonding between O(3)/H-(OCH2) was homochiral and led to the formation of supramo-lecular chains (Fig. 4). Further associations differentiated themethyl groups of the NMe2 framework as follows (Fig. 4):
(i) Through (N)CH3/p (1,4-phenylene) contacts between thechains of the enantiomers with the same configuration
(ii) Through weak interaction (N)CH3/(CH2)O(H) between chainsof enantiomers with an opposite configuration.
Weak heterochiral non-bonding interactions doubly associatedwith the thiazolidine rings as H(6-c)/S.
3. Conclusions
The thiaminalisation of 2-amino-2-(mercaptomethyl)propane-1,3-diol with aryl(di)aldehydes afforded a new series of chiral 4,4-bis(hydroxymethyl)-1,3-thiazolidines with medium to satisfactory
yields. The overall results are caused by the redox instability ofthe thioamino-1,3-diol. On the 1H NMR timescale, the above2-aryl(bis)-1,3-thiazolidines exhibited no detectable epimerisationat the C-2 position, no N-pyramidal inversion and no NH acidebaseinterchange. In the thiaminalic pNeCeS- part of these compounds,the 1H NMR data, when combined with DFTcalculations, revealed anendo anomeric effect assigned as lpN/s* CeS hyperconjugation.The IR spectroscopy in the solid state indicated that these structuresexist in their ring forms. However, the condensates of salicylalde-hyde and terephthaldialdehyde displayed weak ring (1,3-thiazolidine)/chain (Schiff base) tautomerism, due to their spon-taneous partial ring-opening in solution. After treating our 4,4-bis(hydroxymethyl)-1,3-thiazolidines with a stoichiometric amountof formaldehyde, a new class of C-2,-8-substituted-3-oxa-7-thia-r-1-azabicyclo[3.3.0]-c-5-octanes, which are singly functionalised at theC-5 position with an exploitable hydroxymethyl group, were ob-tained in satisfactory to good yields, regio- and diastereoselectivities.Depending on the nature of the p-,m- or o-substituent of the phenylring in the initial 2-aryl-1,3-thiazolidines, a C-8/C-2 migration ofthe aromatic group occurred during the oxaminalisation. Severalmechanistic models were proposed relating to this unexpected re-activity; the proposals consisted of the formation of (C-substituted)-methyleneiminium and 1-thionia-4-azabicyclo[2.2.1]hexanes cat-ions as non-isolable intermediates. The first molecular structure ofa 3-oxa-7-thia-r-1-azabicyclo[3.3.0]-c-5-octane derivative, whichbears a p-dimethylaminophenyl group in position C-8, exhibited ananomeric effect that manifested as hyperconjugation between thelpN(1) (donor)/s*O(3)eC-(2) (acceptor) orbitals, i.e., occurring inthe 1,3-oxazolidine ring only. In contrast, entire thiazolidin-oxazolidine units were involved in four types of (supramolecular)
Fig. 3. View of the homochiral associations through hydrogen bonds between the (1R,5S,8R) enantiomers of compound 9h.
Fig. 4. View along ‘b’ axis of the unit cell showing hydrogen-bonding and weak interactions between enantiomers in the crystal of compound 9h. Hydrogen atoms not involved ininteractions are omitted for clarity.
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9979
Author's personal copy
interactions, homochiral pO/H-(OCH2) and (NC)H3/p but hetero-chiral (NC)H3/Oo and (CH)H/So.
4. Experimental section
4.1. General
Melting points were carried out on ELECTROTHERMAL� in-strument. Conventional NMR spectra were recorded on a Bruker�
AM 300 instrument operating at 300 and 75 MHz for 1H and 13Cnuclei, respectively. 2D-1H,1H-NOESY, 2D-1H,13C-HSQC, 13C-HMBC,13QC and 13C DEPT experiments were recorded either on Bruker�
AM 500, 600 or 700 instruments operating at 500, 600 or 700 MHzfor 1H and at 125, 150 or 175 MHz for 13C nuclei. All chemical shifts(d values) are given in parts per million (ppm); all homocouplingpatterns (nJH,H values) are given in hertz (Hz). In the NMR de-scriptions, some specific abbreviations were used: ‘br t’ (broadtriplet), ‘br d’ (broad doublet), ‘br m’ (broad multiplet). TLC wasperformed by using aluminium sheets with silica gel 60 F254(Merck�); column chromatography was conducted on Silica gel Si60 (40e63 mm, Merck�). IR spectra were performed on a JASCO�
FT-IR 6100 spectrometer. Only relevant absorption maxima arelisted, throughout, in cm�1: s (strong), m (medium) and w (weak).Microanalyses were performed on a Carlo Erba� CHNOS 1160 ap-paratus. Mass spectra (MS) ESI were recorded on Bruker� EsquireInstrument. All reagents and solvents were of commercial qualityand used as such with no supplementary purification. Synthesisand data of compounds IIIa, IVa,1b (R3¼H), 6g, 7g, 6h and 7hwerereported elsewhere.2a,b,15c
4.2. Computational details
Full geometry optimization and frequency calculation hasbeen carried out for conformers 4d (pseudo-trans) and (pseudo-cis) in the gas phase at the B3LYP21/6-311þþG** level of theory,22
which represents a good compromise between accuracy andcomputational cost as shown in many related studies. Solventeffects have been included using the CPCM23 (polarisable con-ductor) model. We have considered several molecular confor-mations in each case, though only results for the most stable arereported here. Natural bond orbital analysis24 (at the B3LYP/6-311þþG** level of theory) has been done to determine the con-tribution of specific donoreacceptor orbital interaction to con-formational stability. Coupling constants were calculated usingthe GIAO25 method in the gas phase and in DMSO ( 3¼46.826).The Gaussian 09 program,26 together with Gabedit graphicalinterface,27 was used.
4.3. X-ray analysis
Crystallographic data for compound 9h (c) were collected atroom temperature on a Bruker SMART APEX diffractometer, usinggraphite monochromated Mo Ka radiation (l¼0.71073 �A). For thispurpose, the crystal was mounted on a cryo-loop with Paratone-Noil. The structure was solved by direct methods (SHELXS-97)28 andrefined by full matrix least-squares procedures based on F2 withall measured reflections (SHELXL-97).29 All non-hydrogen atomswere refined anisotropically. H atoms were introduced in theiridealized positions and refined as riding. Further details on thedata collection and refinement methods can be found in Table II-SD. The drawings were created with the Diamond program.30
The supplementary crystallographic data for this paper can bedownloaded free of charge from The Cambridge CrystallographicData Centre via www.ccdc.cam.ac.uk/data_request/cif (see Table II-SD for details).
4.4. Typical procedures for the preparation of compounds4aej
4.4.1. Method A
4.4.1.1. Preparation of compound 4d. Under a dry nitrogen at-mosphere, to a benzene (40 mL) solution containing benzaldehyde(0.613 g, 0.587 mL, 5.780 mmol), 2-amino-2-(mercaptomethyl)propane-1,3-diol hydrochloride IVa (1.000 g, 5.780 mmol) wasadded with vigorous stirring. In the resulting suspension, a solutionobtained by dissolving anhyd K2CO3 (0.400 g, 2.890 mmol) inwater(1.5 mL) was injected. The reaction mixture was refluxed, under N2,for about 8 h (until no more water separated in a DeaneStark trap).At room temperature, the suspension was filtered off and mineralswerewell washed with anhyd THF (50 mL). The organic filtrate wasevaporated under reduced pressure to dryness and the solid resi-due was purified by column chromatography on silica gel (eluentligroin/acetone 1.5:1 v/v) to give the desired compound 4d (0.665 g,51% yield).
4.4.2. Method B
4.4.2.1. Preparation of compound 4g. Under a dry nitrogenatmosphere, to an EtOH (25 mL) solution containing2-hydroxybenzaldehyde (0.247 g, 0.210 mL, 2.022mmol), 2-amino-2-(mercaptomethyl)propane-1,3-diol hydrochloride IVa (0.350 g,2.022 mmol) was added with vigorous stirring. To the resultingsuspension, Et3N (0.204 g, 0.280mL, 2.022mmol) was injected. Thereaction mixture was refluxed for 8 h and then evaporated underreduced pressure to dryness. The residue was redissolved inanhyd THF on heating (40 �C, 3�15 mL) and filtered off. The com-bined THF solution was evaporated under reduced pressure todryness. The residue was purified by column chromatography onsilica gel, the desired compound being isolated as a single fraction.This was additionally triturated with CH2Cl2/ligroin 1:4 v/v at�18 �C to yield compound 4g (0.242 g, 50% yield).
4.4.3. Method C
4.4.3.1. Preparation of compound 4b. Under a dry nitrogenatmosphere, to an EtOH (15 mL) solution containing4-chlorobenzaldehyde (0.284 g, 2.022 mmol), 2-amino-2-(mer-captomethyl)propane-1,3-diol hydrochloride IVa (0.350 g,2.022 mmol) was added with vigorous stirring. To the resultedsuspension, Et3N (0.204 g, 0.280mL, 2.022mmol) was injected. Thereactionmixturewas stirred at room temperature for 24 h and thenevaporated under reduced pressure to dryness. The residue redis-solved in anhyd THF on heating (40 �C, 3�15 mL) and filtered off.The combined THF solution was evaporated under reduced pres-sure to dryness. The residue was first triturated with THF/Et2O 1:4v/v then crystallised from EtOH/CH2Cl2/Et2O 0.5:5:3 v/v/v at�18 �Cto yield the desired compound 4b (0.189 g, 36% yield).
4.5. Typical procedure for the preparation of compounds 9and 10. Preparation of compounds 9f and 10f
1,3-Thiazolidine 4f (0.241 g, 1.0 mmol) and p-formaldehyde(0.030 g, 1.0 mmol) in benzene (25 mL) were refluxed for 13 h withcontinuous removal of water in a DeaneStark trap. The resultingsuspensionwas evaporated under reduced pressure to dryness. Thecrude solid was triturated from THF/ligroin 1:4 v/v at �18 �C toafford compounds 9f (c), 9f (t) as a single crop (0.200 g, 79% partialconversion of 4f). Themother liquor was evaporated under reducedpressure to dryness and the residue was purified by column chro-matography on silica gel (eluent CH2Cl2/Et2O 5:3 v/v) to yieldcompounds 10f (c), 10f (t) as a single fraction (0.040 g, 16% partialconversion of 4f).
4.5.1. c-5-Hydroxymethyl-c-8-(4-nitrophenyl)-3-oxa-7-thia-r-1-azabicyclo[3.3.0]octane 9a (c); c-5-hydroxymethyl-t-8-(4-nitrophenyl)-3-oxa-7-thia-r-1-azabicyclo[3.3.0]octane 9a (t); c-5-hydroxymethyl-c-2-(4-nitrophenyl)-3-oxa-7-thia-r-1-azabicyclo[3.3.0]octane 10a (c); and c-5-hydroxymethyl-t-2-(4-nitrophenyl)-3-oxa-7-thia-r-1-azabicyclo[3.3.0]octane 10a (t). These compoundswere isolated by column chromatography as a single fraction (el-uent CH2Cl2/Et2O 0.75:1 v/v) with a 82:8:6:4 9a (c)/9a (t)/10a (c)/10a (t) composition of the mixture. Total conversion of 4a 83%(0.164 g mixture starting from 0.190 g 4a); yellow oil. Found: C,51.15; H, 4.66; N, 10.11%. C12H14N2O4S (282.07) requires: C, 51.05; H,
C. Morar et al. / Tetrahedron 69 (2013) 9966e99859982
The financial support from Grant provided by the ResearchCouncil Romania (Project PN-II-ID-PCE-2011-3-0128) is gratefullyacknowledged. The computational calculations reported in thiswork were performed using resources from the Centre de Calculde Champagne-Ardenne (ROMEO, University of ReimsChampagne-Ardenne, France) and the Centre de RessourcesInformatiques de Haute Normandie (CRIHAN, Rouen, France). Also,we would like to thank the Plateforme de Mod�elisation Mol�eculaireMulti-�echelle (P3M, University of Reims Champagne-Ardenne,France) for various supports. H.K. thanks Prof. E. H�enon for reg-ular and lasting scientific discussions and for his work on relatedsystems.
Supplementary data
Supplementary data associated with this article can be found inthe online version, at http://dx.doi.org/10.1016/j.tet.2013.09.070.
References and notes
1. (a) Broadbent, H. S.; Burnham, W. S.; Sheely, R. M.; Olsen, R. K. J. Heterocycl.Chem. 1976, 13, 337e348; (b) Habernel, G. G.; Ecsy, W. Heterocycles 1977, 7,1027e1032.
2. (a) But, A.; Batiu, C.; Ramondenc, Y.; Darabantu, M. Rev. Roum. Chim. 2009, 54,127e133; (b) But, A.; Lameiras, P.; Silaghi-Dumitrescu, I.; Batiu, C.; Guillard, S.;Ramondenc, Y.; Darabantu, M. Lett. Org. Chem. 2010, 7, 283e290; (c) for a re-view of the (condensed)-1,3-oxazolidine systems 3 (Scheme 1), see: Darabantu,M. Curr. Org. Synth. 2010, 7, 235e275.
3. (a) Enz, W.; Cecchinato, M. Helv. Chim. Acta 1961, 44, 706e709; (b) Crabb, T. A.;Hall, M. J.; Williams, R. O. Tetrahedron 1973, 29, 3389e3398; (c) Seebach, D.;Weber, T. Tetrahedron Lett. 1983, 24, 3315e3318; (d) Seebach, D.; Weber, T. Helv.Chim. Acta 1984, 67, 1650e1661; (e) Luhowy, R.; Meneghini, F. J. Am. Chem. Soc.1979, 101, 420e426; (f) Hamril, A.; P�era1, M.-H.; Fillion, H. J. Heterocycl. Chem.1987, 24, 1629e1633; (g) Gonzales, A.; Lavilla, R.; Piniella, J. F.; Alvarez-Larena,A. Tetrahedron 1995, 51, 3015e3024; (h) Skvortsov, A. N.; Uvarov, V. M.; deVekki, D. S.; Studentsov, E. P.; Skvortsov, N. K. Russ. J. Gen. Chem. 2010, 80,2007e2021; (cf. Zh. Obsch. Khim. 2010, 80, 1697e1711 ).
4. For this chemistry highlighting the importance derivatives of type IIa (Scheme2) as scaffolds: (a) total synthesis of Biotin (vitamin H, coenzyme R): Deroose, F.D.; De Clercq, P. J. J. Org. Chem. 1995, 60, 321e330; (b) total synthesis of theantimicrobial Micacocidin: Ino, A.; Hasegawa, Y.; Murabayashi, A. TetrahedronLett. 1998, 39, 3509e3516; (c) synthesis of segments of Farnesyl Transferase:Yang, H.; Sheng, X. C.; Harrington, E. M.; Ackerman, K.; Garcia, A. M.; Lewis, M.D. J. Org. Chem. 1999, 64, 242e251.
5. (a) S�elambaron, J.; Carr�e, F.; Fruchier, A.; Roque, J.-P.; Pavia, A. A. Tetrahedron2002, 58, 4439e4444; (b) for a similar diagnosis in 1,3-oxazolidine systemssee: S�elambaron, J.; Monge, S.; Carr�e, F.; Roque, J.-P.; Pavia, A. A. Tetrahedron2002, 58, 9559e9566.
6. (a) as masked sulfhydryl agents for inhibition of melanogenesis, see: Napoli-tano, A.; d’Ischia, M.; Prota, G.; Havens, M.; Tramposch, K. Biochim. Biophys. Acta1991, 1073, 416e422; (b) as anti-inflammatory agents, see: Chiarino, D.; Fer-rario, F.; Pellacini, F.; Sala, A. J. Heterocycl. Chem. 1989, 26, 589e593.
7. (a) Saiz, C.; Wipf, P.; Manta, E.; Mahler, G. Org. Lett. 2009, 11, 3170e3173; (b)Saiz, C.; Wipf, P.; Mahler, G. J. Org. Chem. 2011, 76, 5738e5746; (c) as a typicalexample see: Tocher, J. H.; Edwards, D. I. Biochem. Pharmacol. 1995, 50,1367e1371.
8. (a) Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976, 734e736; (b) Baldwin, J. E.;Cutting, J.; Dupont, W.; Kruse, L.; Silberman, L.; Thomas, R. C. J. Chem. Soc.,Chem. Commun. 1976, 736e738.
10. (a) Wilson, G. E., Jr.; Bazzone, T. J. J. Am. Chem. Soc. 1974, 96, 1465e1470; (b) asa new class of non-narcotic antitussive agents (N-acylated forms), see: Gan-dolfi, C. A.; Di Domenico, R.; Spinelli, S.; Gallico, L.; Fiocchi, L.; Lotto, A.; Menta,E.; Borghi, A.; Rossa, C. D.; Tognella, S. J. Med. Chem. 1995, 38, 508e525; (c) asnew ligands for HgII complexation (N-acyl derivatives, analogous of the cor-responding penicillin moiety), see: Chandrasekhar, S.; Chopra, D.; Gopalaiah,K.; Guru, R.; Tayur, N. J. Mol. Struct. 2006, 837, 118e131.
11. (a) Valters, R.; Flitsch, W. Ring-chain Tautomerism; Plenum: New York, NY, 1985;(b) F€ul€op, F.; Mattinen, J.; Pihlaja, K. Tetrahedron 1990, 46, 6545e6552; (c) Al-kseyev, V. V.; Zelenin, K. N. Khim. Geterosikl. Soedin. 1998, 1068e1071; (d) L�az�ar,L.; F€ul€op, F. Eur. J. Org. Chem. 2003, 16, 3025e3042; (e) Darabantu, M. Curr. Org.Synth. 2010, 7, 120e152.
12. Eliel, E. L.; Wilen, H. S. Stereochemistry of the Organic Compounds; John Wiley &Sons: New York, NY, Chichester, UK, Brisbane, Toronto, Singapore, 1994; p 758;1199, 1200.
13. (a) Friebolin, H. Basic One- and Two Dimensional NMR Spectroscopy; VCH Ver-lagsgesellschaft mbH: Weinheim, Germany/New York, NY, 1991; p 93;263e291; (b) Silverstein, R. M.; Webster, F. X.; Kiemle, D. J. SpectrometricIdentification of Organic Compounds; John Wiley and Sons: Somerset, New Jer-sey, U.S.A, 2005; p106; 153e155; (c) for example, 10�10 s for the acidebaseinterchange and 10�8 s for the N-pyramidal inversion in the case of N-substituted piperidines, see: Crowley, P. J.; Robinson, M. J. T.; Ward, M. G. Tet-rahedron 1977, 33, 915e925.
14. (a) Karplus, M. J. Chem. Phys. 1959, 30, 11e15; (b) Karplus, M. J. Am. Chem. Soc.1963, 85, 2870e2871.
15. (a) Vanelle, P.; De Meo, M. P.; Maldonado, J.; Nouguier, R.; Crozet, M. P.; Laget,M.; Dumenil, G. Eur. J. Med. Chem. 1990, 25, 241e250; (b) Darabantu, M.; Pl�e, G.;Mager, S.; Cotora, E.; Gaina, L.; Costas, L.; Mates, A. Tetrahedron 1997, 53,1873e1890; (c) Maiereanu, C.; Darabantu, M.; Pl�e, G.; Berghian, C.; Condamine,E.; Ramondenc, Y.; Silaghi-Dumitrescu, I.; Mager, S. Tetrahedron 2002, 58,2681e2693; (d) Ghiviriga, I.; Oniciu, D. C. Chem. Commun. 2002, 2718e2719and the literature cited therein.
20. Seeman, J. I. Chem. Rev. 1983, 83, 83e134.21. (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648e5652; (b) Stephens, P. J.; Devlin, F.
J.; Chablowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623e11627; (c)Hertwig, R. H.; Koch, W. Chem. Phys. Lett. 1997, 268, 345e351.
22. Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72,650e654.
23. (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995e2001; (b) Cossi, M.;Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem. 2003, 24, 669e681.
24. Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. NBO, Version 3.1Madison, WI, 1988.
25. (a) McWeeny, R. Phys. Rev. 1962, 126, 1028e1034; (b) Ditchfield, R. Mol. Phys.1974, 27, 789e807; (c) Dodds, J. L.; McWeeny, R.; Sadlej, A. J. Mol. Phys. 1980, 41,1419e1430; (d) Wolinski, K.; Hilton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112,8251e8260.
26. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman,J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Car-icato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg,J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Naka-jima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J.E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.;Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S.S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.;Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.;Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels,A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,Revision A.02; Gaussian: Wallingford, CT, 2009.
27. Allouche, A.-R. J. Comput. Chem. 2011, 32, 174e182.28. Emsley, J. Die Elemente; Walter de Gruyter: Berlin, Germany, 1994.29. Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2008, A64, 112e122.30. Brandenburg, K. DIAMOND; Crystal Impact GbR: Bonn, Germany, 2006.
C. Morar et al. / Tetrahedron 69 (2013) 9966e9985 9985