(3: S,4 R)-3,4-Dihydroxy-N-Alkyl-l-homoprolines: Synthesis and computational mechanistic studies

Article abstract of DOI:10.1039/c9ob02141h

This is the first synthetic report of (3S,4R)-dihydroxy-N-Alkyl-l-homoprolines described so far. 2,4-O-Benzylidene-d-erythrose was obtained from d-glucose with an improved yield, and then transformed into the title (3S,4R)-dihydroxy-N-Alkyl-l-homoprolines, in a two-step strategy, with excellent overall yields. Hydrogenolysis of the benzyl group led to the NH congener. The synthesis of final products from 1,4-lactone intermediates was studied by computational means either under acidic or basic conditions. The theoretical mechanism studies fully explain the experimental results: (a) an equilibrium between l-homoprolines and their bicyclic counterparts is established in acids; (b) the equilibrium suffers a complete displacement towards the l-homoproline side in a basic medium.

Full text of DOI:10.1039/c9ob02141h

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
(3S,4R)-3,4-Dihydroxy-N-alkyl-L-homoprolines:
synthesis and computational mechanistic studies†
Cite this: Org. Biomol. Chem., 2019,
17, 10052
a
David S. Freitas, ‡a Cristina E. A. Sousa, ‡a Joana Parente,a Artem Drogalin,
a
António Gil Fortes, a Nuno M. F. S. A. Cerqueira *b and Maria J. Alves
*
This is the first synthetic report of (3S,4R)-dihydroxy-N-alkyl-L-homoprolines described so far. 2,4-O-
Benzylidene-D-erythrose was obtained from D-glucose with an improved yield, and then transformed into
the title (3S,4R)-dihydroxy-N-alkyl-L-homoprolines, in a two-step strategy, with excellent overall yields.
Hydrogenolysis of the benzyl group led to the NH congener. The synthesis of final products from 1,4-
lactone intermediates was studied by computational means either under acidic or basic conditions. The
theoretical mechanism studies fully explain the experimental results: (a) an equilibrium between
L-homoprolines and their bicyclic counterparts is established in acids; (b) the equilibrium suffers a com-
plete displacement towards the L-homoproline side in a basic medium.
Received 3rd October 2019,
Accepted 5th November 2019
DOI: 10.1039/c9ob02141h
rsc.li/obc
attained through the introduction of amino acids other than
L-proline; β,γ-amino acids, amino acids of the D-series, and
1. Introduction
Carbohydrates have served as chiral starting materials for the various types of cyclic monomers.10,15 β-Amino acids are the
synthesis of several amino acids.1–3 Some of these syntheses best-studied among those compounds.16,17 In this context, the
are multi-stage, involving considerable manipulation of pro- development of asymmetric synthetic routes for the synthesis
tecting groups. However, some elegant protocols are known for of homoprolyl residues is highly desirable as new templates in
3,4-dihydroxyprolines.4 Three members of the L-series have the design of specific turn-mimics of biologically active mole-
been isolated from natural sources: 1, from Navicula pelicul- cules. The present work intends to be a contribution to the
losa;5 2 from Amanita virosa;6 and 3, the sixth residue of the required pool of homoproline molecules. Besides, the two
decapeptide sequence of Mefp1 (Fig. 1), an adhesive protein, hydroxyl groups at 3,4-positions of the proline ring being polar
from the marine mussel Mytilus edulis.7 Though, the syntheses and capable of hydrogen-bonding can be easily tunable into
of homoproline congeners are very scarce, only two methods nonpolar groups, e.g. by acetal protection. Additional impor-
have been published for these compounds: one, by the Davis tance of these compounds comes from the recent synthesis of
group,8 and another from our laboratory.9 This is, as far as we their D-congeners, obtained by a divergent synthesis from 2,4-
know, the first report for the synthesis of L-3,4-dihydroxypro- O-benzilidene-D-erythrose.9 Knowing that a matching configur-
lines. A major feature of proline amino acids in Nature is ation of amino acids is crucial in shaping the outcome for the
related to the rigidity imposed into the peptide sequence in peptide secondary structure,15 the choice between D- and
which proline is incorporated, leading to protein folding L-structures is preeminent in the design of new peptide
(α-,β-,γ-turns), which is known to play important roles in: (i) foldamers.
the tertiary structure of proteins, (ii) molecular/cellular reco-
2,4-O-Benzilidene-D-erythrosylimines generated in situ from
gnition processes, and (iii) interactions between substrates D-erythrose display a complete facial selectivity towards 1-(tert-
and receptors.10–14 However, in the modified peptide back- butyldimethylsilyloxy)-1-methoxyethene nucleophilic addition
bones the conformational torsions have been successfully giving the R configuration at the new stereogenic centre. The
acid treatment of adducts followed by basic treatment triggers
aDepartamento de Química, Universidade do Minho, Campus de Gualtar,
4710-057 Braga, Portugal. E-mail: mja@quimica.uminho.pt
bUCIBIO@REQUIMTE, Departamento de Biomedicina da Faculdade de Medicina da
Universidade do Porto, Alameda Professor Hernâni Monteiro, 4200-319 Porto,
Portugal. E-mail: nunoscerqueira@med.up.pt
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c9ob02141h
‡The first authors.
Fig. 1 Dihydroxyprolines isolated from natural sources.
10052 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
a reaction cascade leading to an L-homoproline in excellent lactones 7a–d during the course of chromatographic purifi-
1
overall yields. The computational results are consistent with cation. The H NMR spectra of adducts 6a–d showed the pres-
the formation of (3S,4R)-dihydroxy L-homoprolines in two ence of lactones 7a–d as contaminants. In case a, an enriched
stages from the imine adducts.
sample of 7a was obtained during chromatography.
In a previous work, the acid treatment of either compound
9 or 10 led to 1,4-lactone 11. Scheme 2 shows how the spatial
proximity between the attacking oxygen (O-1) and the carbonyl
can easily lead to 1,4-lactone 11 in both cases. Computational
mechanistic studies were developed for some of these cases.21
Accordingly, compounds 6a–d (together with their contami-
2. Experimental results and
discussion
2,4-O-Benzylidene-D-erythrose18 (4) is obtained on a 10 g scale
from D-glucose, by adapting a known procedure for the syn- nants 7a–d) were subjected to strong acid conditions with the
thesis of 2,4-O-isopropylidene-D-erythrose.19 2,4-O-Benzylidene- aim to generate lactones 12a–d (Scheme 3).
D-erythrose reacted with alkyl amines, as they did with aro-
It is likely that the matching configuration of the amino
matic amines before20 to give the respective N-alkylimines 5. group and the 1,2-dihydroxyethyl arm in lactone 12 makes the
Imines 5 were found to be excellent chiral templates able to aminocyclization a much easier process than the aminocycliza-
induce complete stereoselectivity in nucleophilic additions tion of its epimer 11.9 Lactone 11 had to undergo bromination
with 1-(tert-butyldimethylsilyloxy)-1-methoxyethene, furnishing at the primary hydroxyl group in order to make cyclization
single adducts 6a–d, in high yields (Scheme 1). Compounds occur. Theoretical calculations have confirmed that compound
6a–d were purified by flash column chromatography in silica. 12a can be converted into 13a under other acids besides HBr
It was found that silica promotes cyclization to six-membered (HCl, TFA, see the third section) by a relatively easy process.
Scheme 1 Synthesis of D-erythrosyl alkylimines 5 and its reactions with 1-(tert-butyldimethylsilyloxy)-1-methoxyethene – (i) alkylamines 0.7 equiv.,
THF (dry), 1 h, 35 °C, 4 Å MS, N2 (g); (ii) BF3·OEt2 0.7 equiv., 1-(tert-butyldimethylsilyloxy)-1-methoxyethene 0.7 equiv., THF (dry), 4 Å MS, N2 (g),
−84 °C to rt, 2 h; (iii) ethyl acetate : pet. ether (1 : 1), r.t., prolonging contact with silica (in a chromatographic column).
Scheme 2 Synthesis of 1,4-lactone 11 obtained by the acid treatment of D-erythrose 6-carbon chain derivatives 9 and 10.
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem., 2019, 17, 10052–10064 | 10053

View Article Online
Paper
Organic & Biomolecular Chemistry
genation in the presence of Pd/C at room temperature
(Scheme 4).
When a D2O solution of the obtained compound 8a con-
tained in a NMR tube was treated with TFA-d8 8a it was found
that an equilibrium between 8a and 13a is re-established.
Computational mechanistic studies clearly show how ener-
getics favors the equilibrium back in the acidic medium.
3. Computational studies
Scheme 3 Synthesis of lactones 12a–d obtained by the acid treatment
of compounds 6a–d.
Computational studies were carried out, using density func-
tional theory, to understand the mechanism of the reactions
involving the reaction of 1,4-lactones 12a under different
acidic conditions, eventually followed by basic conditions, and
from which the N-benzyl (3S,4R)-dihydroxy L-homoproline 8a is
obtained.
The computational results have shown that although the
reaction is done experimentally in one-pot synthesis, the
mechanism goes through several sequential stages and
involves the formation of the intermediate 13a, as described in
the next sections.
The treatment of compound 6a with TFA led to an evanescent
compound, tentatively assigned as compound 12a. It was
observed over the reaction period in aliquots studied by 1H
NMR. After a 5 h reaction course at room temperature, this
compound was the main species in the reaction mixture. Over
the time (5–8 days), a dynamic equilibrium had been reached
leading to a 1 : 3 : 1 mixture of three components: 13a (major)
together with 8a and 12a (Scheme 4). Any attempt to isolate
compound 12a by chromatography was found to be not poss-
ible; the mixture systematically evolves to compounds 13 and
8. The obtained products were found to be in a 2 (13a) : 1 (8a)
3.1. Stage 1: formation of the reaction intermediate 13a from
1,4-lactone 12a
ratio in reactions run in standard HCl–dioxane, a 1 : 1 mixture An intramolecular cyclisation of lactone 12a occurs under
with HBr in acetic acid, and a 3 (13a) : 1 (8a) with 6 M trifluor- acidic conditions at the first stage (Fig. 2). In compound 12a,
oacetic acid in dioxane. In the last case, the ratio did not nitrogen N1 and carbon C2 are at a 3.13 Å distance from each
change by prolonging the reaction time for a week. The full other in the reactants. The proton from HCl is stabilized by a
conversion of the bicyclic compounds 13 into L-homoprolines hydrogen bond network provided by one water molecule W1
8 occurs by the basification of the reaction medium with KOH and the hydroxyl group (O2) of the lactone.
or triethylamine followed by acidic resin treatment, to adjust
The transition state of the reaction is characterized by an ima-
the pH to 7. In some spectra of homoprolines 8, vestigial ginary frequency at 386.33i cm−1. This stationary point reveals
amounts of the respective bicyclic compounds are always that nitrogen N1 and carbon C2 become closer to each other
present. Compounds 8a–d are generally formed in high yields. (2.26 Å) at the same time that the bond lengths between carbon
The debenzylation of homoproline 8a was achieved by hydro- C2 and oxygen O2 become extended (1.87 Å versus 1.44 Å in the
Scheme 4 Synthesis of L-homoprolines 8a–d – (i) HCl 6 M, dioxane : water 1 : 1, rt, 2 h; (ii) (a) KOH 6 M; rt, overnight; (b) Amberlite IR-120 (H+), pH
7, rt; (iii) (a) Pd/C, 10 mol%, H2, 1 atm, MeOH, 1,1,2-trichloroethane (1.1 equiv.), 2 days; (b) KOH (0.01 M) until pH 12.
10054 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 2 Aminocyclization mechanism for the synthesis of bicyclic compound 13a from 12a.
reactants), enabling the simultaneous proton transfer among 3.2. Stage 2: formation of L-homoproline 8a from bicyclic
them, through the HCl molecule and the water molecule W1.
At the end of this stage, the intra-molecular cyclisation is
compound 13a
complete. Carbon C2 and nitrogen N1 become covalently The next stage of the mechanism involves the formation of the
bonded to each other by a single bond (1.52 Å versus 3.13 Å in homoproline 8a from the intermediate 13a, obtained in the
the reactants). At the same time, one proton migrates from end of the first stage. Experimentally, this reaction was found
nitrogen N1 to the water molecule W1 and from the latter one to diverge whenever acidic or basic conditions were employed.
to oxygen O2 of the lactone, through HCl. This process results Under acidic conditions, an equilibrium is observed between
in the cleavage of the bond between oxygen O2 and carbon C2 compounds 13a and 8a. However, the yields of the reactions
(3.17 Å versus 1.44 Å in the reactants) and the formation of one change significantly depending on the type of acid that is
water molecule (W2).
This reaction requires a free activation energy of 19.2 kcal reaction is very efficient, giving solely compound 8a.
mol−1 and is exergonic with −21.6 kcal mol−1
The computational results explain the mechanistic differ-
employed (HCl, HBr and TFA). Under basic conditions, the
.
The intra-molecular network of hydrogen bonds plays an ences, both under acid and basic conditions as described in
important role in this step since: (i) it enables the correct detail below.
alignment of nitrogen N1 and carbon C2 within the lactone,
3.2.1. Acid catalysis. Under acid conditions, the reaction
making them ready for the nucleophilic attack, and (ii) it involves the cleavage of the lactone ring in intermediate 13a.
enables the proton transfer from nitrogen N1 to the hydroxyl The reaction requires the direct participation of one water
group bonded to carbon C2, through the HCl molecule. The molecule and an acid species molecule, and it is completed in
HCl molecule behaves as a catalyst in the reaction facilitating two sequential steps (Fig. 3). The mechanism is very similar
the proton transfer through the water molecules that is deter- for the different acids that were employed (HCl, HBr or TFA).
minant for the cyclisation process.
In the first step, the adduct formation occurs. The proton
Experimentally, compound 12, once formed, easily evolves from the acid species approaches the oxygen O1 of the reactant
to 13. The computational results showed a free activation and establishes one hydrogen bond with it (in average 1.68 Å),
energy (ΔG‡ = 19.2 kcal mol−1) compatible with the reaction’s and a weaker interaction with the neighbouring water mole-
room temperature. As this conversion is very exergonic (ΔGr = cule (in average 2.43 Å).
−21.6 kcal mol−1) the reverse step is not possible, which is in
The transition states of these additions are characterized by
accordance with experimental results.
one imaginary frequency (see the first step in Fig. 3). The bond
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem., 2019, 17, 10052–10064 | 10055

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 3 Mechanism involved in lactone cleavage in compound 13a to form compound 8a under HCl, HBr and TFA.
length between carbon C1 and oxygen O1 increases to around group from the water molecule becomes covalently bonded to
1.33 Å (1.22 Å in the reactants), at the same time the oxygen carbon C1 (around 1.44 Å). At the same time, the proton from
from the water molecule (W1) becomes aligned with carbon C1 the water molecule migrates to the acid.
and very close to it (in average 1.69 Å versus 3.30 Å in the
reactants), and the acid proton establishes a bond with HBr is used and higher when TFA and HCl are employed
oxygen O1. (Table 1). No significant differences are observed in the free
The free activation energy of the first step is lower when
The adduct is formed when the carbon C1–oxygen O1 bond reaction energy. In all the studied cases, the reactions are
is 1.38 Å (versus 1.22 Å in compound 13a), and the hydroxyl endergonic and with similar values.
10056 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
The second step involves the ring cleavage with the for- with HCl. The intermediate 13a evolved to compound 8a in
mation of compound 8a.
equilibrium with 13a in 2 (13a) : 1 (8a).
In the adduct, carbon C1 and oxygen O2 are bonded to
In a separated experiment, it was found that compound 8a
each other by a covalent bond (1.50 Å in HBr and HCl, 1.47 Å quickly sets an equilibrium back with 13a. The experiment
in TFA) and the proton from the acid species is bonded to a consisted of adding TFA-d8 to a solution of 8a in D2O con-
water molecule W1 (1.07 Å, except for TFA (1.54 Å)). tained in a NMR tube. After 30 minutes, a 3 (13a) : 1 (8a) ratio
The transition states of the second step are characterized by was verified.
one imaginary frequency (see the second step in Fig. 3). The
The energetic profile obtained from the theoretical calcu-
bond between carbon C1 and oxygen O2 starts to elongate (in lations for this transformation is only 18 kcal mol−1 for the
average 1.65 Å), and one hydrogen molecule from the water limiting step, which is likely to take place at room temperature.
molecule W1 becomes very closer to oxygen O2 (in average From the thermodynamic point of view compound 13 is
1.31 Å versus 1.53 Å in HBr and HCl, and 1.78 Å in TFA in the favored over 8a in ca. 5 kcal mol−1. Interestingly, the inter-
adducts).
mediate adduct for the direct/inverse processes requires
In the product, the bond between carbon C1 and oxygen O2 similar free activation energies for both sides of the reaction
is cleaved (2.45 Å in HBr/HCl, 2.72 Å in TFA), and the protic (direct: 10.6 kcal mol−1; inverse: 9.3 kcal mol−1). The reaction
acid is on the way to be regenerated from the hydroxyl group point of equilibrium was found to be dependent on the
converted into carbonyl.
ground energy of 8a and 13a.
Overall, the addition–elimination reaction mechanisms do
3.2.2. Basic catalysis. Under basic conditions, only one
not show significant differences within the acids employed product is obtained in the end of the reaction, compounds 8a–
(HBr, HCl and TFA). The calculated free energies show that the d. The mechanism requires the presence of a base, KOH, and
first step is the rate-limiting one in all studied reactions. The two water molecules. The reactions are complete in two
reaction with the best full energetic profile (faster kinetics) is sequential steps (Fig. 4).
the one where HBr is employed. This happens due to the
The computational results have shown that the reaction
largest atomic radius of bromine that makes the proton requires two steps; the first step involves the formation of a
shuttle more efficiently in HBr, especially in the first step (see zwitterionic reaction intermediate.
Fig. 3). The slowest reaction is the one where TFA is employed,
due to the heist activation free energies that are observed in carbon C1 and the oxygen O of KOH is 3.05 Å. Its position is
both steps (Table 1). stabilized by two water molecules that are aligned with oxygen
The energetic profile of the theoretical calculations of O2, through a network of hydrogen bonds (2.65 Å and 1.89 Å).
hydrolysis in HBr in acetic acid showed a free activation energy The transition state of this reaction is characterized by one
In the reagent, compound 13a, the bond length between
of 12.9 kcal mol−1 in the first step, and 3.5 kcal mol−1 in the imaginary frequency at 78.47i cm−1. In these structures, the
second step. As so, an open equilibrium with very low free acti- central role played by the potassium cation on the reaction
vation energies in both sides should operate. The ratio becomes clear. It favours: (i) the alignment of the hydroxide
between compounds 13a/8a is nevertheless difficult to predict with carbon C1 (2.55 Å), (ii) the hydrogen bond between water
from the differences in activation and reaction energies.
molecule 2 (W2) and the carbonyl group (1.81 Å) and, (iii) the
The energetic profile of the theoretical calculations of hydro- hydrogen bond between water molecule 2 (W2) with the pot-
lysis in HCl showed that the first step (Δ‡ = 21.9 kcal mol−1) con- assium (2.81 Å).
trols the lactone hydrolysis process. At room temperature the
In the end of this reaction, a stable zwitterionic reaction
energetic barrier is predicted to be surpassed, counting with an intermediate is obtained. Carbon C1 acquires a tetrahedral
energy supplement released in the formation of compound 13a. configuration, and oxygen O1 becomes negatively charged
The second step showed a very low free activation energy (Δ‡ = (−1.00 a.u. versus −0.78 a.u. in the reactants). The negative
1.3 kcal mol−1), and it is exergonic (Δr = −5.9 kcal mol−1). Being charge is stabilized by one water molecule, W2, that is now
so, the reversal process needs only 16 kcal mol−1 to be sur- very close to the positively charge potassium. The water mole-
mounted. These findings go in line with the experimental equi- cule W1 is also close to the potassium cation, but also inter-
librium observed in the hydrolysis of the bicyclic lactone 13a acts with oxygen O2 through a short hydrogen bond (1.99 Å).
This reaction is almost spontaneous requiring a very low
free activation energy (1.9 kcal mol−1) and it is exergonic with
−9.4 kcal mol−1
.
Table 1 Activation free energies (Δ‡) and reaction free energies (Δr) for
the transformation of compound 13a into 8a in TFA, HCl, and HBr
The second step involves the cleavage of the zwitterionic
intermediate into compound 8a.
First step
Second step
The reactant of this step is very similar to the product of
the last step. Carbon C1 is firmly attached to oxygen O2 by a
covalent bond (1.59 Å) and the potassium cation (0.94 a.u.)
continues to be stabilized by the oxygen atoms provided by the
two water molecules and the hydroxyl group that is now
attached to carbon C1.
Δ‡
Δr
Δ‡
Δr
(kcal mol−1
)
(kcal mol−1
)
(kcal mol−1
)
(kcal mol−1
)
13a/8a
TFA 22.9
HCl 21.9
HBr 12.9
13.6
11.8
13.6
10.6
1.2
3.5
−8.3
−5.9
−5.2
3 : 1
2 : 1
1 : 1
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem., 2019, 17, 10052–10064 | 10057

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 4 Mechanism involved in lactone unit cleavage in compound 13a to form compound 8a under basic conditions.
The transition state of the studied reaction is characterized
In the product of this reaction, the formation of compound
by one imaginary frequency at 87.83i cm−1. The bond length 8a is complete. The bond between carbon C1 and oxygen O2 is
between carbon C1 and oxygen O2 starts to elongate (2.01 Å cleaved (2.41 Å versus 1.59 Å in the reactant), and KOH is
versus 1.59 Å in the reactant) and one of the hydrogens from regenerated after the proton transfer from water molecule W2
water molecule W1 becomes very close to oxygen O2.
to water molecule W1.
Fig. 5 Energetic profile of all the studied reactions to yield the formation of N-benzyl-(3S,4R)-dihydroxy L-homoprolines 8a under acidic and basic
conditions.
10058 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
Similar to the previous step, this reaction requires a low Hz, 1H, H-4), 4.38 (dd, J = 10.8, 5.2 Hz, 1H, H-6), 4.77 (s, 2H, H-
free activation energy (6.3 kcal mol−1), being slightly endergo- 1″), 5.69 (s, 1H, H-2), 7.36–7.65 (m, 10H, Ph–CH), 7.98 (d, J =
nic with 4.0 kcal mol−1
.
1.6 Hz, 1H, H-1′) ppm.
From the experimental work, it is known that the equili-
4.3.3. Synthesis of (2R,4S,5R)-4-((E)-((benzo[d][1,3]dioxol-4-
brium is completely shifted towards to the right giving 8a. ylmethyl)imino)methyl)-2-phenyl-1,3-dioxan-5-ol
(5b).
This is in accordance with the theoretical results obtained Aldehyde 4 (150 mg; 0.72 mmol); THF (5 mL); piperonylamine
1
which display an energetically favoured profile (Fig. 5).
(63 μL; 76 mg; 0.50 mmol). H NMR (400 MHz, CDCl3) δ 3.73
(t, J = 10.4 Hz, 1H, H-6), 4.02 (ddd, J = 10.4, 8.8, 5.2 Hz, 1H, H-
5), 4.20 (dd, J = 8.8, 1.6 Hz, 1H, H-4), 4.37 (dd, J = 11.2, 5.2 Hz,
1H, H-6), 4.55 (s, 2H, H-1″), 5.58 (s, 1H, H-2), 6.69–6.83 (m, 3H,
H–Ph), 7.37–7.42 (m, 3H, H–Ph), 7.51–7.53 (m, 2H, H–Ph), 7.92
(d, J = 1.2 Hz, 1H, H-1′) ppm.
4. Experimental
4.1. General
All reagents were purchased from Sigma-Aldrich, Acros, TCI or
Alfa Aesar and used without further purification, except for
p-toluenesulfonic acid monohydrate which was dried under a
vacuum pistol at 120 °C, THF was dried by reflux under Na(s),
and BF3·OEt2 by distillation. Aldehyde 4 18,19 and imine 6a 9
were synthesized following the procedures described in the lit-
erature. An optimized synthesis of 4 was obtained from a syn-
thetic description of an analog described in the literature.19
4.3.4. Synthesis of (2R,4S,5R)-4-((E)-((4-fluorobenzyl)imino)
methyl)-2-phenyl-1,3-dioxan-5-ol (5c). Aldehyde 4 (150 mg,
0.72 mmol); THF (5 mL); 4-(fluoro)benzylamine (57 μL; 63 mg;
1
0.50 mmol). H NMR (400 MHz, CDCl3) δ 3.73 (t, J = 10.4 Hz,
1H, H-6), 4.03 (ddd, J = 10.4, 8.8, 5.2 Hz, 1H, H-5), 4.22 (dd, J =
8.8, 1.6 Hz, 1H, H-4), 4.49 (dd, J = 10.8, 5.2 Hz, 1H, H-6), 4.62
(s, 2H, H-1″), 5.59 (s, 1H, H-2), 7.04 (t, J = 8.8 Hz, 2H, H–Ph),
7.23 (dd, J = 5.2, 3.6 Hz, 2H, H–Ph), 7.35–7.42 (m, 3H, H–Ph),
7.52 (dd, J = 7.2, 2 Hz, 2H, H–Ph), 7.96 (d, J = 1.2 Hz, 1H, H-1′).
4.2. Optimized synthesis of 2,4-benzylidene-D-erythrose
4.3.5. Synthesis
of
(2R,4S,5R)-4-((E)-((4-chlorobenzyl)
(i) Acetal protection of D-glucose to yield (2R,4aR,6S,7S,8S,8aS)-
2-phenylhexahydropyrano[3,2-d][1,3]dioxine-6,7,8-triol.18
(ii) To a solution of NaIO4 (8.2 g; 0.038 mol; 2.2 equiv.) in
water (62 mL), refrigerated at 0 °C, and kept at pH = 4, was
added dropwise a suspension of (2R,4aR,6S,7S,8S,8aS)-2-phe-
nylhexahydropyrano[3,2-d][1,3]dioxine-6,7,8-triol (5 g), in water
(21 mL), with permanent pH control (pH ≥ 4), by the addition
of aqueous sodium hydroxide 8 M (∼8 mL). The temperature
was kept under 10 °C until total dissolution (∼3 h). After the
3 h reaction time, the pH was adjusted to 6.5, and the stirring
was continued for another 2 h, at room temperature. The solu-
tion was evaporated and the residue was dried under vacuum.
Ethyl acetate (40 mL) is added to the residue and the flask was
heated for 2 min at 80 °C under stirring. The suspension
formed was filtered and the solid residue was washed with
ethyl acetate (3 × 30 mL). The combined liquid fractions were
dried over anhydrous sodium sulphate for 1 h, and concen-
trated in the rotary evaporator to yield 3.69 g (95%) of 2,4-O-
benzylidene D-erythrose.18,19
imino)methyl)-2-phenyl-1,3-dioxan-5-ol (5d). Aldehyde
4
(200 mg, 0.96 mmol); THF (10 mL); 4-(chloro)benzylamine
1
(82 μL, 95 mg, 0.67 mmol). H NMR (400 MHz, CDCl3) δ 3.72
(t, J = 10.4 Hz, 1H, H-6), 4.03 (ddd, J = 10, 8.8, 5.2 Hz, 1H, H-5),
4.22 (dd, J = 8.8, 1.2 Hz, 1H, H-4), 4.38 (dd, J = 10.8, 5.2 Hz,
1H, H-6), 4.62 (s, 1H, H-1″), 5.59 (s, 1H, H-2), 7.19 (d, J = 6.8
Hz, 2H, H–Ph), 7.23–7.43 (m, 5H, H–Ph), 7.52 (dd, J = 7.2, 2.0
Hz, 2H, H–Ph), 7.97 (d, J = 1.2 Hz, 1H, H-1′).
4.4. Reaction of D-erythrosylbenzylimine 5a–d with 1-(tert-
butyldimethylsilyloxy)-1-methoxyethene
4.4.1. General procedure. The imine reaction mixture
obtained in the precedent step (0.34–0.67 mmol) was refriger-
ated at −84 °C for 15 min, and BF3·OEt2 (42–83 μL; 47–95 mg;
0.34–0.67 mmol) was added under magnetic stirring followed
by 1-(tert-butyldimethylsilyloxy)-1-methoxyethene (79–150 μL;
63–130 mg; 0.34–0.67 mmol). The mixture was kept stirring
under the conditions described for 1–2 h, and then allowed to
recover at room temperature. The reaction mixture was passed
through a pad of Celite®, and the filtrate was evaporated in
the rotary evaporator to give a brownish oil, which was re-dis-
4.3. Synthesis of D-erythrosyl benzylimine 5a–c
4.3.1. General procedure. To a solution of aldehyde 4 solved in chloroform (40 mL) and successively washed with
(150–200 mg; 0.72–0.96 mmol) in dry THF (4–10 mL), contain- water (3 × 20 mL), aqueous sat. NaHCO3 (3 × 20 mL) followed
ing activated 4 Å molecular sieves (1.0 g), was added the amine by saturated NaCl solution (2 × 20 mL). The organic layers
(57–84 µL; 63–110 mg; 0.50–0.67 mmol) under magnetic stir- were combined and dried with anhydrous magnesium sulfate,
ring and an N2 atmosphere. The reaction mixture was kept stir- filtered, and the solvent was evaporated to give a yellow oil
1
ring for 1 h at 35 °C. According to the H NMR, products are residue. The crude product was subjected to column chrom-
clear yellow oils, virtually pure, and used in the next step atography (silica, petroleum ether : ethyl acetate).
without purification.
4.4.2. Synthesis
of
(R)-methyl-3-(benzylamino)-3-
4.3.2. Synthesis of (2R,4S,5R)-4-((E)-(benzylimino)methyl)- ((2R,4S,5R)-5-hydroxy-2-phenyl-1,3-dioxan-4-yl)propanoate (6a).
2-phenyl-1,3-dioxan-5-ol (5a). Aldehyde 4 (150 mg, 0.72 mmol); Imine 5a (100 mg, 0.34 mmol) obtained previously; THF
THF (4 mL); benzylamine (55 μL, 54 mg, 0.50 mmol).
(4 mL); BF3·OEt2 (42 μL; 47 mg; 0.34 mmol); 1-(tert-butyldi-
1H NMR (400 MHz, CDCl3) δ 3.73 (t, J = 10.8 Hz, 1H, H-6), methylsilyloxy)-1-methoxyethene (79 μL; 63 mg; 0.34 mmol);
4.04 (ddd, J = 10.4, 8.8, 5.2 Hz, 1H, H-5), 4.22 (dd, J = 8.8, 1.6 2 h. Product: yellow oil (112 mg; 0.30 mmol) η = 88%.(a)
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem., 2019, 17, 10052–10064 | 10059

View Article Online
Paper
Organic & Biomolecular Chemistry
[α]D17 = −13.3 (c 0.6, MeOH); IR νmax 3323, 1724 cm−1; 1H NMR 1″), 3.94 (dd, J = 9.2, 3.6 Hz, 1H, H-4), 4.00 (ddd, J = 10.4, 9.2,
(400 MHz, CDCl3) δ 2.57 (dd, J = 16.4, 10.4 Hz, 1H, H-2′), 2.94 4.8 Hz, 1H, H-5), 4.03 (d, J = 12.8 Hz, 1H, H-1″), 4.33 (dd, J =
(dd, J = 16.4, 3.2 Hz, 1H, H-2′), 3.55 (ddd, J = 10.4, 4.4, 3.2 Hz, 10.8, 5.2 Hz, 1H, H-6), 5.51 (s, 1H, H-2), 7.01–7.06 (m, 2H, H–
1H, H-1′), 3.65 (s, 3H, OMe), 3.65 (t, J = 10.8 Hz, 1H, H-6), 3.90 Ph), 7.27–7.47 (m, 7H, H–Ph) ppm; 13C NMR (100 MHz, CDCl3)
(d, J = 12.4 Hz, 1H, H-1″), 3.93 (dd, J = 9.6, 4.4 Hz, 1H, H-4), δ 33.4 (C-2′), 51.3 (OMe), 52.0 (C-1″), 53.7 (C-1′), 62.7 (C-5), 70.9
4.00 (ddd, J = 10.4, 9.6, 5.2 Hz, 1H, H-5), 4.03 (d, J = 12.4 Hz, (C-6), 78.3 (C-4), 101.6 (C-2), 115.4 (d, J = 21 Hz, C-4″ or C-6″),
1H, H-1″), 4.35 (dd, J = 10.8, 5.2 Hz, 1H, H-6), 5.54 (s, 1H, H-2), 115.7 (d, J = 21 Hz, C-4″ or C-6″), 126.1, 128.3, 128.4, 129.1,
7.30–7.48 (m, 10H, H–Ph) ppm; 13C NMR (100 MHz, CDCl3) 129.5 (CH–Ph), 129.8 (d, J = 8.0 Hz, C-3″ or C-7″), 130.5 (d, J =
δ 33.9 (C-2′), 51.9 (OMe), 52.6 (C-1″), 56.4 (C-1′), 63.0 (C-5), 70.9 8.0 Hz, C-3″ or C-7″), 132.7, 137.3 (C-q), 162.5 (d, J = 246 Hz,
(C-6), 77.9 (C-4), 101.6 (C-2), 126.1, 127.7, 128.2, 128.5, 128.7, C-5″), 172.5 (CvO). HRMS (ESI): calcd for C21H26FNO5:
129.1 (CH–Ph), 137.5, 138.4 (C-q), 172.8 (CvO) ppm; HRMS 390.1717 (M + H+); found: 390.1717.
(ESI): calcd for [C21H26NO5 ]: 372.1766 (M + H+); found:
372.1803.
4.4.5. Synthesis of (R)-methyl 3-((4-chlorobenzyl)amino)-3-
((2R,4R,5R)-5-hydroxy-2-phenyl-1,3-dioxan-4-yl)propanoate (6d).
+
(a) A small fraction enriched in product 7a (3 mg) was Imine 5d (222 mg; 0.67 mmol); 1-(tert-butyldimethylsilyloxy)-1-
obtained from the chromatographic purification of compound 6a. methoxyethene (150 µL, 130 mg; 0.67 mmol); BF3·OEt2 (83 µL,
Compound 7a: 1H NMR (400 MHz, CDCl3) δ 2.61 (dd, J = 95 mg, 0.67 mol). Product: yellow oil (280 mg, 0.67 mmol); η =
24
18.0, 7.6 Hz, 1H, H-7), 3.12 (dd, J = 18.0, 7.6 Hz, 1H, H-7), 3.38 quant. [α]D = −42 (c 0.6, DCM); IR (nujol) νmax 3327,
(dt, J = 8.8, 7.6 Hz, 1H, H-8), 3.77 (t, J = 9.2 Hz, 1H, H-4), 3.77 1728 cm−1; 1H NMR (400 MHz, CDCl3) δ 2.57 (dd, J = 16.4, 10.0
(d, J = 13.2 Hz, 1H, CH2–Ph), 3.83 (t, J = 10.4 Hz, 1H, H-6), 3.87 Hz, 1H, H-2′), 2.93 (dd, J = 16.0, 3.2 Hz, 1H, H-2′), 3.50 (dt, J =
(d, J = 13.2 Hz, 1H, CH2–Ph), 4.21 (ddd, J = 10.0, 9.6, 5.2 Hz, 10.0, 3.6 Hz, 1H, H-1′), 3.64 (dd, J = 10.4, 9.6 Hz, 1H, H-6), 3.65
1H, H-5), 4.45 (dd, J = 10.4, 4.8 Hz, 1H, H-6), 5.60 (s, 1H, H-2), (s, 3H, OMe), 3.87 (d, J = 12.4 Hz, 1H, H-1″), 3.91 (dd, J = 10.0,
7.28–7.49 (m, 10H, H–Ph) ppm; 13C NMR (100 MHz, CDCl3) δ 4.0 Hz, 1H, H-4), 3.98 (d, J = 12.8 Hz, 1H, H-1″), 3.99 (ddd, J =
36.2 (C-2′), 50.9 (C-7), 53.7 (C-8), 68.3 (C-6), 69.3 (C-5), 79.8 9.6, 5.2, 4.4 Hz, 1H, H-5), 4.33 (dd, J = 10.8, 5.2 Hz, 1H, H-6),
(C-4), 101.9 (C-2), 126.1, 127.3, 128.1, 128.4, 128.6, 129.4 (CH– 5.53 (s, 1H, H-2), 7.26–7.39 (m, 7H, H–Ph), 7.45–7.48 (m, 2H,
Ph), 136.6, 139.3 (C-q), 168.4 (CvO) ppm.
H–Ph); 13C NMR (100 MHz, CDCl3) δ 33.9 (C-2′), 51.8 (OMe),
4.4.3. Synthesis of (R)-methyl 3-((benzo[d][1,3]dioxol-4- 51.9 (C-1″), 56.8 (C-1′), 63.1 (C-5), 70.9 (C-6), 78.0 (C-4), 101.6
ylmethyl)amino)-3-((2R,4R,5R)-5-hydroxy-2-phenyl-1,3-dioxan- (C-2), 126.1, 128.3, 128.8, 129.1, 129.7 (CH–Ph), 133.4 (C-q),
4-yl)propanoate (6b). Imine 5b (171 mg; 0.50 mmol); 1-(tert- 136.9 (C-q), 137.4 (C-q) 172.8 (CvO). HRMS (ESI): calcd for
butyldimethylsilyloxy)-1-methoxyethene (110 µL, 95 mg, C21H26ClNO5: 406.1421/408.1392 (M + H+); found: 406.1420/
0.50 mmol); BF3·OEt2 (62 µL, 71 mg, 0.50 mmol). Product: 408.1392.
24
yellow oil (150 mg, 0.36 mmol); η = 72%; [α]D = −16.0 (c 0.9,
5. Treatment of the adducts 6a–d with acids
5.1. Hydrochloric acid
DCM); IR (nujol) νmax 3330, 1731 cm−1 1H NMR (400 MHz,
;
CDCl3) δ 2.56 (dd, J = 16.4, 10.4 Hz, 1H, H-2′), 2.93 (dd, J =
16.0, 3.2 Hz, 1H, H-2′), 3.52 (dt, J = 10.4, 3.6 Hz, 1H, H-1′), 3.65
5.1.1. General procedure. (i) To
a solution of the
(dd, J = 10.4, 9.5 Hz, 1H, H-6), 3.66 (s, 3H, OMe), 3.79 (d, J = 12 β-aminoester 6a–d (60–111 mg; 0.15–0.27 mmol) in dioxane
Hz, 1H, H-1″), 3.93 (dd, J = 9.2, 4.0 Hz, 1H, H-4), 3.94 (d, J = (2–4 mL) was added hydrochloric acid 37% (1.5–2.8 mL) to
12.4 Hz, 1H, H-1″), 3.99 (ddd, J = 9.2, 5.2, 4.8 Hz, 1H, H-5), 4.33 form a 6 M solution. The reaction mixture was stirred for 2 h
(dd, J = 10.8, 5.2 Hz, 1H, H-6), 5.53 (s, 1H, H-2), 5.94 (d, J = 2 at room temperature, evaporated in the rotary evaporator until
Hz, 2H, H-8″), 6.75–6.83 (m, 3H, H–Ph), 7.27–7.41 (m, 5H, H– a solid foam or a brown oil was formed (a mixture of com-
Ph); 13C NMR (100 MHz, CDCl3) δ 33.9 (C-2′), 51.9 (OMe), 52.3 pounds 13 and 8 in a 2 : 1 ratio). The reaction mixture was
(C-1″), 56.7 (C-1′), 63.1 (C-5), 70.9 (C-6), 77.9 (C-4), 101.0 (C-8″), used in the next stage.
101.6 (C-2), 108.3, 108.8, 121.7 (C-, C-4″, C-7″), 126.1, 126.2,
(ii) To the reaction mixture obtained in (i) (60–111 mg,
128.2, 128.4. 129.1 (CH–Ph), 132.3 (C-2″), 137.5 (C-q), 147.1 0.15–0.27 mmol) was added water (2.5–3.3 mL) and solid KOH
(C-q), 147.9 (C-q), 172.8 (CvO). Elemental analysis, calcd for (23–33 mg, 0.14–0.20 mmol). The reaction mixture was stirred
C22H25NO7: C, 63.6; H, 6.07; N, 3.37. Found C, 63.2; H, 6.10; N, overnight. An aqueous suspension of acid resin IR-120 (H+)
3.10.
was added until pH 7, and kept under hand stirring. The resin
4.4.4. Synthesis of (R)-methyl 3-((4-fluorobenzyl)amino)-3- was filtered off, and the solvent was removed in the rotary
((2R,4R,5R)-5-hydroxy-2-phenyl-1,3-dioxan-4-yl)propanoate evaporator to give a pasty residue. This was dissolved in
(6c). Imine 5c (158 mg; 0.50 mmol); 1-(tert-butyldimethyl- ethanol (2 mL), filtered and evaporated to give light brownish
silyloxy)-1-methoxyethene (110 µL, 95 mg, 0.50 mmol), oils (20–70 mg, 0.074–0.26 mmol, 48–96.3% yield), as the
BF3·OEt2 (62 µL, 71 mg, 0.50 mmol). Product: yellow oil respective L-homoprolines 8a–d.
24
(96 mg, 0.25 mmol); η = 50%; [α]D = −27.0 (c 0.6, DCM); IR
5.1.2. Synthesis of 2-((2S,3S,4R)-N-benzyl-3,4-dihydroxypyr-
(nujol) νmax 3331, 1732 cm−1 1H NMR (400 MHz, CDCl3) rolidin-2-yl)acetic acid (8a). (i) β-Aminoester 6a (100 mg,
;
δ 2.67 (dd, J = 16.0, 9.2 Hz, 1H, H-2′), 2.95 (dd, J = 16.8, 4.1 Hz, 0.27 mmol); 1,4-dioxane (3 mL); HCl 37% (2.5 mL).
1H, H-2′), 3.60 (dt, J = 9.6, 4.0 Hz, 1H, H-1′), 3.64 (dd, J = 10.8, Quantitative mixture of compounds 13a : 8a in a 2 : 1 ratio, by
9.6 Hz, 1H, H-6), 3.66 (s, 3H, OMe), 3.93 (d, J = 13.2 Hz, 1H, H- 1H NMR.
10060 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
1
Compound 13a(a): H NMR (400 MHz, D2O) δ 3.10 (dd, J = (1.5 mL). A quantitative mixture of compounds 13c : 8c in a
19.2, 2.0 Hz, 1H, H-3), 3.30 (dd, J = 19.2, 8.4 Hz, 1H, H-3), 3.78 2 : 1 ratio. Compound 13c was identified by 1H NMR
(dd, J = 12.4, 2.0 Hz, 1H, H-7), 3.90 (dd, J = 12.4, 3.6 Hz, 1H, H- (400 MHz, D2O) δ 2.91 (dd, J = 19.2, 2.0 Hz, 1H, H-3), 3.12 (dd,
7), 4.17 (d, J = 13.2 Hz, 1H, H-1′), 4.32–4.36 (m, 2H, H-3a + H- J = 19.2, 8.0 Hz, 1H, H-3), 3.59 (dd, J = 12.4, 2.0 Hz, 1H, H-7),
6), 4.44 (d, J = 13.2 Hz, 1H, H-1′), 4.87 (dd J = 5.6, 4.4 Hz, 1H, 3.70 (dd, J = 12.4, 3.2 Hz, 1H, H-7), 3.99 (d, J = 13.6 Hz, 1H, H-
H-6a), 7.44–7.51 (m, 10H, H–Ph) ppm; 13C NMR (100 MHz, 1′), 4.10 (d, J = 13.2 Hz, 1H, H-1′), 4.23–4.27 (m, 2H, H-3a + H-
D2O) δ 31.5 (C-3), 49.2 (C-1′), 55.6 (C-3a), 61.1 (C-7), 68.1 (C-6), 6), 4.77 (dd, J = 6.0, 4.4 Hz, 1H, H-6a), 7.01–7.05 (t, 2H, H-4′ +
80.8 (C-6a), 128.9 (C-2), 129.0 (C-3′ or C-4′ or C-5′), 129.3 (C-3′ H-6′), 7.28–7.31 (m, 2H, H-3′ + H-7′).
or C-4′ or C-5′), 129.5 (C-3′ or C-4′ or C-5′), 175.3 (CvO) ppm.
(ii) To a mixture obtained in step (i) (43 mg) was added
(a) The 1H NMR and 13C NMR descriptions were taken from water (2.5 mL) and solid KOH (23 mg, 0.14 mmol). Product 8c:
24
the crude material of the reaction of compound 6a with TFA, (20 mg, 0.074 mmol); η = 48%; [α]D = −23.0 (c 0.6, MeOH); IR
where a 3 : 1 ratio of compounds 13a (3) : 8a (1) was isolated.
1
(nujol) νmax 3347, 2926, 2856, 1660 cm−1; H NMR (400 MHz,
(ii) To the mixture obtained in step (i) (72.6 mg) was added D2O) δ 2.55 (dd, J = 14.8 Hz, 6.8 Hz, 1H, CH2CO2H), 2.60 (dd, J =
water (3.0 mL) and solid KOH (30 mg, 0.18 mmol). Product 8a: 14.8 Hz, 6.4 Hz, 1H, CH2CO2H), 3.27 (td, J = 6.8 Hz, 2.8 Hz, 1H,
24
(70 mg, 0.26 mmol); η = 96.3%; [α]D = −20.5 (c 0.4, MeOH); H-2), 3.50 (dd, J = 12.0 Hz, 6.0 Hz, 1H, H-5), 3.67 (dd, J = 12.0 Hz,
IR (nujol) νmax 3187, 1680 cm−1 1H NMR (400 MHz, D2O) δ 4.8 Hz, 1H, H-5), 3.73 (dd, J = 6.0, 2.8 Hz, 1H, H-3), 3.76–3.80 (m,
;
2.89 (dd, J = 18.0, 6.0 Hz, 1H, CH2CO2H), 3.95 (dd, J = 18.0, 6.4 1H, H-4), 3.79 (d, J = 12.8 Hz, 1H, H-1′), 3.96 (d, J = 13.2 Hz, 1H,
Hz, 1H, CH2CO2H), 3.51 (dd, J = 12.0, 6.4 Hz, 1H, H-5), 3.65 H-1′), 7.13–7.18 (m, 2H, H–Ph), 7.39–7.43 (m, 2H, H–Ph); 13C
(dd, J = 12.0, 4.0 Hz, 1H, H-5), 3.80 (td, J = 6.8, 3.2 Hz, 1H, H- NMR (100 MHz, D2O) δ 38.5 (CH2CO2H), 49.5 (C-1′), 55.7 (C-2),
2), 3.86 (td, J = 5.6, 4.0 Hz, 1H, H-4), 3.91 (dd, J = 5.6, 3.2 Hz, 62.3 (C-5), 71.5 (C-3), 73.4 (C-4), 115.2 (d, J = 22.0 Hz, C-4′ or
1H, H-3), 4.32 (d, J = 13.2 Hz, 1H, H-1′), 4.46 (d, J = 13.2 Hz, C-6′), 115.4 (d, J = 22.0 Hz, C-4′ or C-6′), 130.5 (d, J = 8.0 Hz, C-3′
1H, H-1′), 7.54 (s, 5H, H–Ph) ppm; 13C NMR (100 MHz, D2O) δ or C-7′), 130.6 (d, J = 8.0 Hz, C-3′ or C-7′), 134.2 (C-2′), 161.9 (d, J
30.2 (CH2CO2H), 46.4 (C-1′), 52.9 (C-2), 59.3 (C-5), 66.0 (C-3),
70.3 (C-4), 126.6 (C-3′ or C-4′ or C-5′), 127.1 (C-3′ or C-4′ or C13H19FNO5: 288.1242 (M + H2O + H+); found: 288.1242.
C-5′), 127.2 (C-3′ or C-4′ or C-5′), 127.8 (C-2′), 172.2 (CvO) 5.1.5. Synthesis 2-((2S,3S,4R)-N-(4-chlorobenzyl)-3,4-dihy-
ppm; elemental analysis, calcd for C13H17NO4: C, 62.14; H, droxypyrrolidin-2-yl)acetic acid (8d). (i) β-Aminoester 6d
6.82; N, 5.57. Found C, 62.20; H, 6.65; N, 5.76. (111 mg, 0.27 mmol); solvent: 1,4-dioxane (4 mL); HCl 37%
= 241.0 Hz, C-5′), 180.3 (CvO). HRMS (ESI): calcd for
5.1.3. Synthesis of 2-((2S,3S,4R)-N-piperonyl-3,4-dihydroxy- (2.8 mL). A quantitative mixture of compounds 13d : 8d in a
pyrrolidin-2-yl)acetic acid (8b). (i) β-Aminoester 6b (100 mg, 2 : 1 ratio. Compound 13d was identified by 1H NMR
0.24 mmol); 1,4-dioxane (3 mL); HCl 37% (2.5 mL). A quanti- (400 MHz, D2O) δ 3.02 (dd, J = 19.2, 2.0 Hz, 1H, H-3), 3.23 (dd, J
tative mixture of compounds 13b : 8b in a 2 : 1 ratio.
= 19.2, 8 Hz, 1H, H-3), 3.70 (dd, J = 12.4, 2.0 Hz, 1H, H-7), 3.83
1
Compound 13b was identified by H NMR (400 MHz, D2O) (dd, J = 12.4, 3.6 Hz, 1H, H-7), 4.09 (d, J = 13.2 Hz, 1H, H-1′),
δ 3.00 (dd, J = 18.8, 1.6 Hz, 1H, H-3), 3.23 (dd, J = 18.8, 8.0 Hz, 4.22–4.29 (m, 2H, H-3a + H-6), 4.35 (d, J = 13.2 Hz, 1H, H-1′),
1H, H-3), 3.72 (dd, J = 12.4, 2.0 Hz, 1H, H-7), 3.84 (dd, J = 12.0, 4.80 (dd, J = 6.0, 4.8 Hz, 1H, H-6a), 7.34–7.45 (m, 4H, H–Ph).
3.2 Hz, 1H, H-7), 4.03 (d, J = 13.2 Hz, 1H, H-1′), 4.14 (d, J = 13.6
(ii) To the mixture obtained in step (i) (83 mg) in water
Hz, 1H, H-1′), 4.25–4.31 (m, 2H, H-3a + H-6), 4.67 (dd, J = 5.6 (3.3 mL) was added solid KOH (33 mg, 0.20 mmol). Product
24
Hz, 4.8 Hz, 1H, H-6a), 5.95 (s, 2H, H-8′)(a), 6.88–6.94 (m, 3H, 8d: (66 mg; 0.23 mmol); η = 85.0%; [α]D = + 6.0 (c 0.7,
1
H–Ph)(a)
.
MeOH); IR (nujol) νmax 3346, 1690 cm−1; H NMR (400 MHz,
(a) These signals are coincident in the 1H NMR spectrum of D2O) δ 2.64 (dd, J = 16.4 Hz, 6.8 Hz, 1H, CH2CO2H), 2.70 (dd, J
the mixture of the two compounds 13b and 8b. = 16.4 Hz, 6.4 Hz, 1H, CH2CO2H), 3.52 (dd, J = 11.6 Hz, 5.6 Hz,
(ii) To a mixture obtained in step (i) (78 mg) was added 1H, H-5), 3.60 (td, J = 6.4 Hz, 3.2 Hz, 1H, H-2), 3.67 (dd, J =
water (3.0 mL) and solid KOH (30 mg, 0.18 mmol). Product 8b: 12.8 Hz, 4.4 Hz, 1H, H-5), 3.80–3.87 (m, 2H, H-3 + H-4), 4.16 (d,
24
(56 mg, 0.19 mmol); η = 79%; [α]D = −25.0 (c 0.6, MeOH); IR J = 13.2 Hz, 1H, H-1′), 4.28 (d, J = 13.2 Hz, 1H, H-1′), 7.44 (d, J =
(nujol) νmax 3350, 1586 cm−1 1H NMR (400 MHz, D2O) 8.4 Hz, 2H, H–Ph), 7.48 (d, J = 8.4 Hz, 2H, H–Ph); 13C NMR
;
δ 2.40–2.70 (m, 2H, CH2CO2H), 3.47–3.58 (m, 2H, H-4 + H-5), (100 MHz, D2O) δ 34.4 (CH2CO2H), 47.8 (C-1′), 55.9 (C-2), 61.6
3.68 (br d, J = 12.0 Hz, 1H, H-5), 3.82 (br s, 2H, H-3 + H-4), 4.12 (C-5), 69.1 (C-3), 72.6 (C-4), 128.7, 130.7 (CH–Ph), 130.5, 134.1
(d, J = 13.2 Hz, 1H, H-1′), 4.24 (d, J = 13.2 Hz, 1H, H-1′), 5.93 (s, (C-q), 177.4 (CvO). HRMS (ESI): calcd for C13H17ClNO4:
1H, H-8′), 5.94 (s, 1H, H-8′), 6.85–6.94 (m, 3H, H–Ph); 13C NMR 286.0846 (M + H+); found: 286.0851.
(100 MHz, D2O) δ 35.5 (CH2CO2H), 48.8 (C-1′), 55.5 (C-2), 61.7
5.2. Trifluoroacetic acid
(C-5), 69.6 (C-3), 72.7 (C-4), 100.9 (C-8′), 108.3, 109.1, 122.9
(CH–Ph), 127.2, 146.9, 147.1 (C-q), 178.1 (CvO). HRMS (ESI):
calcd for C14H20NO7: 314.1234 (M + H2O + H+); found: acid: (3aR,6R,6aS)-4-benzyl-6-hydroxyhexahydro-2H-furo[3,2-b]
314.1233. pyrrol-2-one (13a) and 2-((2R,3S,4R)-1-benzyl-3,4-dihydroxypyr-
5.2.1. Formation of a mixture of salts of trifluoroacetic
5.1.4. Synthesis of 2-((2S,3S,4R)-N-(4-fluorobenzyl)-3,4- rolidin-2-yl) acetic acid (8a). To a solution of β-aminoester 6a
dihydroxypyrrolidin-2-yl)acetic acid (8c). (i) β-Aminoester 6c (30 mg; 0.08 mmol) in 1,4-dioxane (1.62 mL) was added tri-
(60 mg, 0.15 mmol); solvent: 1,4-dioxane (2 mL); HCl 37% fluoroacetic acid until a 6 M acid solution was formed
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem., 2019, 17, 10052–10064 | 10061

View Article Online
Paper
Organic & Biomolecular Chemistry
(1.38 mL). The reaction mixture was run at room temperature, tures were determined through intrinsic reaction coordinate
under magnetic stirring for one week. The solvents were (IRC) calculations that were followed by a final geometry
removed in the rotary evaporator, leaving a brownish oil con- optimization.33 In all cases, the geometry optimizations and
sisting of a virtually pure mixture of compounds 13a and 8a, in the stationary points were obtained with standard Gaussian
a 3 : 1 ratio and quantitative yield.
convergence criteria. The transition-state structures, reactants
and products were all verified by vibrational frequency calcu-
lations. All the TS structures have only one imaginary fre-
5.3. Bromic acid in acetic acid
5.3.1. Formation of a mixture of salts of hydrobromic acid: quency with the correct transition vector, whereas the reac-
(3aR,6R,6aS)-4-benzyl-6-hydroxyhexahydro-2H-furo[3,2-b]pyrrol- tants and products have all the frequencies positive.
2-one (13a) and 2-((2R,3S,4R)-1-benzyl-3,4-dihydroxypyrrolidin-
The free final energies were calculated with density func-
2-yl) acetic acid (8a). A solution of β-aminoester 6a (50 mg; tional theory, as a sum of the electronic energies, using the all-
0.13 mmol) in 33% bromic acid in acetic acid (1 mL) was electron 6-311++G(3df,2pd) basis set and the functional
stirred at room temperature for 2 hours. Then methanol M062X, plus the ZPE, thermal, and entropic energies (T =
(3.6 mL) was added, and the reaction mixture was kept under 310.15 K, P = 1 bar), calculated with the B3LYP functional but
magnetic stirring for three days. The solvents were removed in with the 6-31G(d) basis set and the dispersion corrections with
the rotary evaporator, leaving a brownish oil consisting of a vir- the DFT-D3 methodology.34,35 This value was also added to the
tually pure mixture of compounds 13a and 8a, in a 1 : 1 ratio contribution of a conductor-like polarizable continuum model
and quantitative yield.
using the integral equation formalism variant (IEF-PCM),36–38
as implemented in Gaussian 09, to simulate the different sol-
vents employed in the experimental work. To this end,
different dielectric constants were applied: water, 78.355;
acetic acid, 6.253 and 1,4-dioxane, 2.210.
6. Cleavage of the benzyl group in
compound 8a to give 8e
All the activation and reaction energies provided in the text
To the benzylated L-homoproline (8a) (20 mg, 0.075 mmol) in and figures refer to free energy differences calculated at the
MeOH (4 mL) in the presence of a catalyst (Pd/C 10 mol%) was M06DX/6-311++G(3df,2pd) level of theory.
added 1,1,2-trichloroethane (8 µL, 1.1 equiv.). The mixture was
stirred for 2 days under a hydrogen atmosphere (1 atm). The
mixture was filtered through a glass-fiber filter. To the filtrate
was added 0.01 M aq. solution of KOH until pH = 12. The
8. Conclusion
solvent was removed in the rotary evaporator to give an yellow The study describes the development of a very simple, high
21
oil, compound 8e (12 mg, 0.061 mmol); η = 81%; [α]D
=
;
yielding strategy for the synthesis of (3S,4R)-dihydroxy
L-homoprolines from D-erythrose. Trivial reagents, cascade
−32.3 (c 0.01, MeOH); IR (nujol) νmax 3328, 3257, 1722 cm−1
1H NMR (400 MHz, D2O) δ 2.41 (dd, J = 7.6 Hz, 15.2 Hz, 1H, reactions, no purification needs, and completely selective reac-
CH2CO2H)(a), 2.58 (dd, J = 5.2 Hz, 14.8 Hz, 1H, CH2CO2H), 3.23 tions make this process highly appealing for the synthesis of
(q, J = 5.2 Hz, 1H, H-2), 3.63–3.68 (m, 1H, H-4), 3.23 (t, J = 4.8 (3S,4R)-dihydroxy
homoprolines
of
L-stereochemistry.
Hz, 1H, H-3), 3.76–3.84 (m, 2H, H-5) ppm; 13C NMR (100 MHz, Considering the biological importance of the final products,
D2O) δ 32.2(CH2CO2H), 61.8 (C-2), 62.4 (C-5), 71.5 (C-3), 72.6 as new molecular scaffolds of an important type, conjugated
(C-4), 179.2 (CvO) ppm;* HRMS (ESI), calcd for C6H12NO4: with the simplicity of the chemical process, an insight on the
162.0761 (M + H+); found: 162.0758.
(a) Contaminated with DMSO.
energetics of reactions involved became a major demand. As
the cyclization mechanism of 6-carbon atom D-erythrosyl
derivatives to a related type of 1,4-lactone intermediate had
been previously studied, the theoretical work was now focused
on the transformation of these intermediates into (3S,4R)-dihy-
droxy L-homoprolines.
7. Computational calculations
(methodology)
The computational results reveal that the mechanism of the
All geometry optimizations were performed with Gaussian formation of the N-benzyl-(3S,4R)-dihydroxy L-homoprolines 8a
09 22 by applying density functional theory23 with the B3LYP from 1,4-lactones 12a diverges under acidic or basic conditions,
functional24–28 together with the 6-31G(d) basis set,29–32 with a although they share some similarities. In both, acid and basis
fine integration grid.
cases, the catalysts favor the proton shuttle that is required for
In all geometry optimizations, we first searched for the tran- the cleavage of the lactones. The main difference in the mecha-
sition state starting from a structure like the reactant model. nism is that with KOH a zwitterionic reaction intermediate in
This was generally obtained with unidimensional scans along which K+ takes part is involved. This effect favors the intra-
the particular reaction coordinate in which we were interested. molecular hydrogen bond network between the compound and
Once a putative transition structure was located, it was fully the two water molecules required for the next step. This effect is
optimized and characterized. Subsequently, the reactants and not observed in acid catalysis, which leads to a significant
products associated with the calculated transition state struc- increase in the calculated activation free energies.
10062 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
14 I. D. Kuntz, J. Am. Chem. Soc., 1972, 94, 4009–4012.
Conflicts of interest
15 F. Sussman, V. M. Sánchez-Pedregal, J. C. Estévez, R. Balo,
J. Jiménez-Barbero, A. Ardá, A. Gimeno, M. Royo,
M. C. Villaverde and R. J. Estévez, Chem. – Eur. J., 2018, 24,
1–6.
There are no conflcts of interest to declare.
16 Y. J. Chung, L. A. Christianson, H. E. Stanger, D. R. Powell
and S. H. J. Gellman, Am. Chem. Soc., 1998, 120, 10555–
10556.
Acknowledgements
We thank the NMR Portuguese network (PTNMR, Bruker
Avance III 400-Univ. Minho), and FCT and FEDER (European 17 F. Schumann, A. Müller, M. Koksch, G. Müller and
Fund for Regional Development)-COMPETE-QREN-EU for N. Sewald, J. Am. Chem. Soc., 2000, 122, 12009–12010.
financial support to CQ/UM. We also thank Search-ON2 for the 18 S. R. Baker, D. W. Clissold and A. McKillop, Tetrahedron
revitalization of the HPC infrastructure of UMinho,
(NORTE-07-0162-FEDER-000086), co-funded by the North
Lett., 1988, 29, 991–994; P. Zimmermann and
R. R. Schmidt, Liebigs Ann. Chem., 1988, 663–667.
Portugal Regional Operational Programme (ON.2-O Novo 19 M. Fengler-Veith, O. Schwardt, U. Kautz, B. Krämer and
Norte), under the National Strategic Reference Framework V. Jäger, Org. Synth., 2002, 78, 123.
(NSRF), through the European Regional Development Fund 20 J. Ferreira, V. C. M. Duarte, J. Noro, A. Gil Fortes and
(ERDF) and the doctoral scholarship SFRH/BD/87292/2012.
M. J. Alves, Org. Biomol. Chem., 2016, 14, 2930–2937.
21 J. F. Rocha, D. S. Freitas, J. Noro, C. S. S Teixeira,
C. E. A. Sousa, M. J. Alves and N. M. F. S. A. Cerqueira,
J. Org. Chem., 2018, 83, 8011–8019.
References
22 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. J. Bearpark, J. Heyd, E. N. Brothers,
K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand,
K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene,
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
Gaussian 09, Gaussian, Inc, Wallingford, CT, USA, 2009.
1 S. Hannessian, Total Synthesis of Natural Products: The
Chiron Approach, Pergamon, Oxford, 1983.
2 M. Miyashita, N. Chida and A. Yoshikoshi, J. Chem. Soc.,
Chem. Commun., 1984, 195.
3 R. C. Bernotas and B. Ganem, Tetrahedron Lett., 1985, 26,
4981.
4 (a) F. Zanardi, L. Battistini, M. Nespi, G. Rassu, P. Spanu,
M. Cornia and G. Casiraghi, Tetrahedron: Asymmetry, 1996,
7(4), 1167–1180; (b) M. Fujii, T. Miura, T. Kajimoto and
Y. Ida, Synlett, 2000, 1046; C. M. Taylor, C. E. Jones and
K. Bopp, Tetrahedron, 2005, 61, 9611–9617.
5 (a) T. Nakajima and B. E. Volcani, Science, 1969, 164, 1400–
1401; (b) I. L. Karle, J. W. Daly and B. Witkop, Science, 1969,
164, 1401–1402; (c) I. L. Karle, Acta Crystallogr., Sect. B:
Struct. Crystallogr. Cryst. Chem., 1970, 26, 765–770.
6 A. Buku, H. Faulstich, T. Wieland and J. Dabrowski, Proc.
Natl. Acad. Sci. U. S. A., 1980, 77, 2370–2371.
7 H. Faulstich, A. Buku, H. Bodenmueller and T. Wieland, 23 P. Hohenberg and W. Kohn, Inhomogeneous Electron Gas,
Biochemistry, 1980, 19, 3334–3343. Phys. Rev., 1964, 136(3b), B864.
8 K. Csatayov, S. G. Davies, A. L. A. Figuccia, A. M. Fletcher, 24 A. D. Becke, A new mixing of Hartree–Fock and local
J. G. Ford, J. A. Lee, P. M. Roberts, B. G. Saward, H. Song
and J. E. Thomson, Tetrahedron, 2015, 71, 9131–9142;
density-functional theories, J. Chem. Phys., 1993, 98(2),
1372–1377.
S. G. Davies, E. M. Foster, J. A. Lee, P. M. Roberts and 25 A. D. Becke, Density-functional thermochemistry. III. The
J. E. Thomson, Tetrahedron, 2013, 69, 8680.
role of exact exchange, J. Chem. Phys., 1993, 98(7), 5648–
9 D. S. Freitas, J. Noro, A. Drogalin, M. C. S. Fernandes,
5652.
A. V. M. Baptista, J. F. C. Parente, J. E. Rodríguez-Borges, 26 C. T. Lee, W. T. Yang and R. G. Parr, Development of the
A. Gil Fortes and M. J. Alves, Synthesis, 2019, 2720–2272.
10 E. Vass, M. Hollósi, F. Besson and R. Buchet, Chem. Rev.,
2003, 103, 1917–1954.
Colle-Salvetti correlation-energy formula into a functional
of the electron density, Phys. Rev. B: Condens. Matter Mater.
Phys., 1988, 37(2), 785–789.
11 J. A. Smith and L. G. Pease, Crit. Rev. Biochem., 1980, 8, 27 S. H. Vosko, L. Wilk and M. Nusair, Accurate spin-depen-
315–399.
dent electron liquid correlation energies for local spin
density calculations: a critical analysis, Can. J. Phys., 1980,
58(8), 1200–1211.
12 L. M. Gierasch, C. M. Deber, V. Madison, C. H. Niu and
E. R. Blout, Biochemistry, 1981, 20, 4730–4738.
13 G. D. Rose, L. M. Gierasch and J. A. Smith, Adv. Protein 28 P. J. Stephens, F. J. Devlin, C. F. Chabalowski and
Chem., 1985, 37, 1–109.
M. J. Frisch, Ab Initio Calculation of Vibrational
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem., 2019, 17, 10052–10064 | 10063

View Article Online
Paper
Organic & Biomolecular Chemistry
Absorption and Circular Dichroism Spectra Using Density 34 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, A consistent
Functional Force Fields, J. Phys. Chem., 1994, 98(45),
11623–11627.
29 V. A. Rassolov, M. A. Ratner, J. A. Pople, P. C. Redfern and
and accurate ab initio parametrization of density func-
tional dispersion correction (DFT-D) for the 94 elements
H-Pu, J. Chem. Phys., 2010, 132(15), 154104.
L. Curtiss, 6-31G* basis set for third-row atoms, J. Comput. 35 S. Grimme, S. Ehrlich and L. Goerigk, Effect of the
Chem., 2001, 22(9), 976–984.
damping function in dispersion corrected density func-
30 R. Ditchfield, W. J. Hehre and J. A. Pople, Self-Consistent
tional theory, J. Comput. Chem., 2011, 32(7), 1456–1465.
Molecular-Orbital Methods. IX. An Extended Gaussian- 36 G. Scalmani and M. J. Frisch, Continuous surface charge
Type Basis for Molecular-Orbital Studies of Organic
polarizable continuum models of solvation. I. General
Molecules, J. Chem. Phys., 1971, 54(2), 724.
formalism, J. Chem. Phys., 2010, 132(11), 114110.
31 P. C. Hariharan and J. A. Pople, Accuracy of AH n equili- 37 E. Cances, B. Mennucci and J. Tomasi, A new integral
brium geometries by single determinant molecular orbital
theory, Mol. Phys., 1974, 27(1), 209–214.
32 M. S. Gordon, The isomers of silacyclopropane, Chem.
Phys. Lett., 1980, 76(1), 163–168.
equation formalism for the polarizable continuum model:
Theoretical background and applications to isotropic and
anisotropic dielectrics, J. Chem. Phys., 1997, 107(8), 3032–
3041.
33 N. Cerqueira, P. A. Fernandes and M. J. Ramos, Protocol for 38 J. Tomasi, B. Mennucci and R. Cammi, Quantum mechani-
Computational Enzymatic Reactivity Based on Geometry
Optimisation, ChemPhysChem, 2018, 19(6), 669–689.
cal continuum solvation models, Chem. Rev., 2005, 105(8),
2999–3093.
10064 | Org. Biomol. Chem., 2019, 17, 10052–10064
This journal is © The Royal Society of Chemistry 2019