Base-Mediated Cascade Aldol Addition and Fragmentation Reactions of Dihydroxyfumaric Acid and Aromatic Aldehydes: Controlling Chemodivergence via Choice of Base, Solvent, and Substituents

Article abstract of DOI:10.1021/acs.joc.8b01867

The diester derivative of dihydroxyfumaric acid (DHF) has been used exclusively as an electrophile in organic synthesis. However, the synthetic utility of DHF's nucleophilic reactivity, contained in the ene-diol moiety, has been underexplored. Inspired by recently observed pH-dependent chemodivergent nucleophilic aldol reactions of dihydroxyfumarate (DHF^2-) with glyoxylate and formaldehyde, we report herein the control and synthetic application of base-controlled chemodivergent reactions between dihydroxyfumarate and aromatic and heteroaromatic aldehydes. With hydroxide as the base in a predominantly aqueous medium, aldol addition followed by deoxalation occurs to provide various 3-aryl-2,3-dihydroxypropanoic acids. With triethylamine as the base in THF, 1-aryl-2,3-dihydroxypropanones are the products of the reaction. In order to understand the difference in reactivity between DHF, its dicarboxylate, and its dimethyl ester, we undertook computational and experimental studies that provide a rationale as to why the dihydroxyfumarate (DHF^2-) is a nucleophile while the corresponding diester reacts as an electrophile.

Full text of DOI:10.1021/acs.joc.8b01867

Subscriber access provided by Warwick University Library
Featured Article
Base-Mediated Cascade Aldol Addition and Fragmentation Reactions
of Dihydroxyfumaric Acid and Aromatic Aldehydes: Controlling
Chemodivergence via Choice of Base, Solvent, and Substituents
George Ward, Charles L. Liotta, Ramanarayanan Krishnamurthy, and Stefan France
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.8b01867 • Publication Date (Web): 17 Sep 2018
Downloaded from http://pubs.acs.org on September 18, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Base-Mediated Cascade Aldol Addition and Fragmentation Reactions of Dihydroxyfumaric
Acid and Aromatic Aldehydes: Controlling Chemodivergence via Choice of Base, Solvent,
and Substituents
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
George Warda, Charles L. Liottaa,*, Ramanarayanan Krishnamurthyb,*, and Stefan Francea,*
a
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia,
30332, United States
b Department of Chemistry, The Scripps Research Institute, La Jolla, California 92037, United
States
c
Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, Atlanta,
Georgia 30332, United States
stefan.france@chemistry.gatech.edu
rkrishna@scripps.edu
charles.liotta@carnegie.gatech.edu
i. NaOH, LiOH
THF:H2O (1:>3.5)
23 ˚C, 18 h
OH
OH
CO2H
O
NEt3 (4 equiv)
O
CO2Me
HO2C
Ar
Ar
OH
Ar
H
ii. A-15 resin, 0.5 h
iii. H2SO4, MeOH, 5 h
THF, 65 ˚C, 18 h
OH
OH
OH
13 examples
(up to 64% yield)
24 examples
(up to 72% yield)
deoxalation
pathway
decarboxylation
pathway
Abstract
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 49
1
2
3
4
5
6
7
8
9
The diester derivative of dihydroxyfumaric acid (DHF) has been used exclusively as an
electrophile in organic synthesis. However, the synthetic utility of DHF’s nucleophilic reactivity,
contained in the ene-diol moiety, has been underexplored. Inspired by recently observed pH-
dependent chemodivergent nucleophilic aldol reactions of dihydroxyfumarate (DHF2-) with
glyoxylate and formaldehyde, we report herein the control and synthetic application of base-
controlled chemodivergent reactions between dihydroxyfumarate and aromatic and heteroaromatic
aldehydes. With hydroxide as the base in a predominantly aqueous medium, aldol addition
followed by deoxalation occurs to provide various 3-aryl-2,3-dihydroxypropanoic acids. With
triethylamine as the base in THF, 1-aryl-2,3-dihydroxypropanones are the products of the reaction.
In order to understand the difference in reactivity between DHF, its dicarboxylate and its dimethyl
ester, we undertook computational and experimental studies that provide a rationale as to why the
dihydroxyfumarate (DHF2-) is a nucleophile while the corresponding diester reacts as an
electrophile.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
The reactivity of DHF, a molecule that was reported by Fenton in the 1890s,1,2 has been the subject
of recent studies (Figure 1). While it has been primarily reported as a reducing agent (oxidation to
di-oxosuccinic acid), DHF has also been shown to undergo reduction to tartaric acid.3,4 Due to its
inherent propensity toward oxidation, DHF has been used as an additive in various food and wine
-
preparations as an anti-oxidant.5 DHF also undergoes CO2 group migration (benzilic acid-type
rearrangement) upon oxidation with O2 and hydroxide to form di- and trisubstituted methoxide
anions, which have been used to make ligand complexes and coordination polymers.6 Until
recently, aside from oxidation/reduction of the ene-diol moiety, the only examples reflecting the
2
ACS Paragon Plus Environment

Page 3 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
electrophilic character of DHF involved the conjugate addition of nucleophiles to the a,b-
unsaturated ester portion of DHF.7,8,9 In all cases, the carboxylic acids were derivatized as alkyl
esters.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The physical and chemical properties of dihydroxyfumaric acid vary significantly as one
proceeds from the diacid to the conjugate base of the diacid to the diester. Depending on the
specific form, it can: (a) exhibit nucleophilic or electrophilic character; and/or (b) behave as an
oxidizing or a reducing agent (Figure 1A). For instance, aryl amines provided nitrogen-containing
heterocycles after nucleophilic attack and subsequent ring closure.10 Moreover, protection of the
hydroxy groups has also been used to facilitate nucleophilic (1,4-) attack on the unsaturated ester
portion.7 In contrast, the latent nucleophilic behavior of DHF has been relatively underexplored.
A.
OH
CO2Alk
OH
CO2H
OH
CO2
AlkO2C
HO2C
O2C
OH
OH
OH
DHF dicarboxylate
DHF, 1
DHF diester
OR OH
OR OH
O
O
O
O
OH OR
Nuc
OH OR
E+
electrophilic
behavior
nucleophilic
behavior
OH
B.
OH
CO2H
O
HO2C
[H]
[O]
CO2H
HO2C
HO2C
CO2H
OH
DHF, 1
OH
tartaric acid
O
di-oxosuccinic acid
O2
OH
O
O
+
O2C
O2C
CO2
O2C
CO2
Figure 1. Amphoteric reactivity of dihydroxyfumaric acid (DHF), its diester, and its
dicarboxylate
Coincidentally, the nucleophilic behavior of DHF lies at the heart of its importance in
prebiotic chemistry. Eschenmoser11 originally hypothesized a “glyoxylate scenario,” in which
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 49
1
2
3
4
5
6
7
8
9
DHF (as the formal dimer of glyoxylate) is a central starting material that serves as a source
molecule for the formation of a variety of biogenic molecules (such as sugars, a-amino acids, and
pyrimidines), and various constituents of the citric acid cycle (or their primordial equivalents).
Experimentally, this hypothesis has begun to be explored. It has recently been reported that the
reaction of DHF with glyoxylate (2) in water proceeds by two completely different reaction
pathways depending on the pH of the medium (Figure 2A). The reaction of DHF with glyoxylate
at pH 7-8 produces dihydroxyacetone (4) and pentulosonic acid (5) via the formation of 2,3-
dihydroxy-oxalosuccinate 3 (aldol reaction) followed by decarboxylation.12 Conducting the
reaction at pH 13-14 produces tartrate (6) and oxalate (7) via a hydroxide-mediated deoxalation of
the same intermediate 3.13 The divergent product formation based on the different pH has been
rationalized based on the proclivity of decarboxylation pathway at lower pH verus the dominance
of deoxalation pathway at high pH.14 The reaction of glycolaldehyde and glyceraldehyde as
electrophiles, in place of glyoxylate, with DHF at pH 7-8 was found to produce efficiently the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
corresponding ketoses, tetrulose and pentuloses respectively.12
A. Liotta and Krishnamurthy
O
OH O
pH = 7-8
HO2C
OH OH
OH OH
4
5
Decarboxylation
OH
CO2
CO2
LiOH
H2O
H
CO2
O2C
O2C
CO2
OH
OH
O
OH
O
3
1
2
OH
O
CO2
O
O2C
O
pH = 13-14
OH
Deoxalation
O
6
7
B. Mahrwald (Decarboxylation)
OH
O
OH
O
OH
H
R
i, ii, or iii
HO2C
R
R
R
CO2H
R
dioxane
H2O
O
OH
OH OH
OH
O
OH OH
i. Cs2CO3; ii. brucine, Cs2CO3; iii. LiOAc
Figure 2. Previous examples of nucleophilic DHF chemistry.
4
ACS Paragon Plus Environment

Page 5 of 49
The Journal of Organic Chemistry
1
2
3
4
Mahrwald et al. have used these observations to explore the synthetic utility of the
5
6
decarboxylative aldol reactions of DHF with
a
variety of aromatic aldehydes,
7
8
9
hydroxyacetaldehyde, and several chiral oxygen-containing enolizable aldehydes (Figure 2B).15
Several synthetic protocols were presented, of which in situ generation of the dihydroxyfumarate
salt employing a variety of lithium or cesium bases proved to be the most promising. In the
presence of a slight excess of aldehyde, mono-decarboxylative aldol reactions were observed.
Conversely, in the presence of a large excess of aldehyde, a second addition of aldehyde occurred.
The mechanistic pathways for the mono- and di-decarboxylative aldol processes were discussed.
They also reported that: (1) in many cases mixtures of carbonyl regioisomers were produced; (2)
the yields of the product varied from very good to marginal, and especially for aryl products the
yields were particularly low; and (3) the reactions did not result in clean isolable products for many
of the examples.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While some of the synthetic possibilities of the decarboxylative aldol pathway have been
demonstrated by Mahrwald et al., the synthetic utility and generalization based on the
chemodivergence reported by Sagi et al.12 and Butch et al.13 have not yet been realized. Herein,
we describe our efforts toward controlling chemodivergence in the reactions of aryl and heteroaryl
aldehydes with DHF by the careful choice of base and solvent and the applicability in organic
synthesis (Scheme 1). We also provide rationale for the disparate physical properties and reactivity
of DHF and dihydroxyfumarate (DHF2-) when compared to DHF-diester derivatives, which exhibit
significant deviations ranging from solubility to propensity for self-condensation and aerobic
oxidation.
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 49
1
2
3
4
OH
CO2H
Ar
OH
9
5
6
7
8
deoxalation
pathway
OH
CO2H
OH
O
O
base
HO2C
Ar
CO2
HO CO2
Ar
H
OH
1
8
O
Ar
OH
9
OH
10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
decarboxylation
pathway
Scheme 1. Base-mediated Chemodivergent Aldol Additions of DHF to Aromatic Aldehydes
RESULTS AND DISCUSSION
DHF Nucleophilic Reactivity. From the current literature, there is evidence that deprotonation
of DHF to provide its dicarboxylate (DHF2-) is critical to unveiling its nucleophilicity.
Deprotonation effectively minimizes the orbital overlap between the delocalized carboxylate
anions and the electron-rich ene-diol portion of the molecule. To the best of our knowledge, in
cases where DHF behaves as an electrophile, the carboxylic acids of DHF existed as alkyl esters.6-
10 Whereas when reacting as a nucleophile, the carboxylic acid groups of DHF were always ionized
using basic media.12-15 Even with this clear trend in the literature there has been no focused study
on why seemingly minor changes to the carboxylic acid moieties of DHF have a profound effect
on its reactivity. The relative stabilities of the DHF derivatives, especially toward reaction with
atmospheric oxygen (to form di-oxosuccinic acid) and self-condensation (via the putative keto
form of DHF2- to form sugar acid 5, Scheme 2)16 also seem to be affected by state of ionization of
DHF versus its ester derivatives.
6
ACS Paragon Plus Environment

Page 7 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
OH
CO2
OH
OH
O
OH
CO2H
O2C
O2C
O2C
HO
CO2
CO2
OH
- 3 CO2
HO
O
OH
5
OH
CO2
O2C
keto-enol
tautometization
O
9
Scheme 2. Self-condensation of DHF2- in H2O.16 Reproduced with permission from Sagi, V.
N.; Karri, P.; Hu, F.; Krishnamurthy, R. Diastereoselective Self-Condensation of
Dihydroxyfumaric Acid in Water: Potential Route to Sugars. Angew. Chem. Int. Ed. 2011, 50,
8127-8130. Copyright 2011 John Wiley and Sons.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For example, DiMeDHF (11) is stable under air for months with no degradation or self-
condensation as observed by NMR in the solid state or in solution. It is also surprisingly insoluble
in most common organic solvents giving rise to dilute solutions in dichloromethane or chloroform
at room temperature. Heating at reflux is required to dissolve DiMeDHF in THF or 1,4-dioxane.
Upon cooling, DiMeDHF can be recovered quantitatively. Treating DiMeDHF with an
electrophile, such as benzaldehyde or 4-nitrobenzaldehyde, in chloroform-d at room temperature
resulted in no observed reaction (Figure 3A). Conducting the reaction in THF at reflux or with
added base (i.e., NEt3) also failed to provide any observable reaction by crude NMR. With H2SO4
as the aldehyde activator, only degradation of the starting materials was observed. This suggests
that the ene-diol moiety in ester 11 is unreactive as a nucleophile towards reactive electrophiles
such as aryl aldehydes.
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 49
1
2
3
O
A.
O
H
4
5
6
7
8
9
H
X
8a
8a (X = H)
8l (X = NO2)
No reaction
THF, rt or reflux (~95% aldehyde
OH
O
NEt3
No reaction
(~95% aldehyde
recovery)
MeO
O
OMe
THF, reflux
11
OH
O
recovery)
or CDCl3, 45 ˚C
B.
H
O
H
O
O
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
OH
HO
O
O
H
O
Low conversion
(>95% recovery of
starting materials
OH
H
THF, rt, 24 h
O
OH
DHF H-bonding
model
Figure 3. Probing DiMeDHF (A) vs DHF free acid (B) nucleophilic reactivity
When compared to DiMeDHF, DHF (as the free acid) is oxidized by oxygen but it still can
be stored in a refrigerator without rigorous degassing of the container. DHF is much more soluble
than DiMeDHF in organic solvents which is odd as, in general, esters are more soluble. The
differences in solubility between DiMeDHF and DHF can be rationalized by the fact that in DHF,
the molecule is expected to adopt a very ordered intramolecular hydrogen bond arrangement17
which, ultimately, hides the multiple highly polar functional groups and increases organic
solubility18 (Figure 3B). The reactivity of DHF as the free acid in organic solvents has not been
studied. To probe this, we performed a control reaction in which DHF (free acid) was stirred in
THF and the solution monitored daily by 1H NMR (Figure 3B). After 3 days, only ~10% of DHF
was consumed forming a trace amount of pentulosonic acid via self-condensation. As further proof
of DHF acid's poor nucleophilic behavior, extremely low conversion (<5%) was observed by 1H
NMR when DHF (free acid) was stirred with an electrophile (i.e., benzaldehyde) in THF after 24
h.
Finally, in contrast, the dicarboxylate can be rapidly oxidized by O2 so it needs to be stored
in a freezer and ideally under a nitrogen balloon to retain its purity for any length of time. The
self-condensation of DHF2- in water has been shown by Sagi et. al.16 to proceed cleanly to sugar
8
ACS Paragon Plus Environment

Page 9 of 49
The Journal of Organic Chemistry
1
2
3
4
acid 5 in only a few hours.
5
6
7
8
In order to gain insights into the relative reactivities of dimethyl dihydroxyfumarate
(DiMeDHF), the free acid (DHF), and the dicarboxylate (DHF2-), DFT calculations [B3LYP; 6-
311+G(2df,2p)]19 were employed to determine the physical and electronic structure associated
with each of these species in aqueous media. Figure 4 illustrates the structure of each of these
molecules, the energies of their HOMO and LUMO, and the corresponding molecular orbitals. A
common structural feature is that the three structures are held in a planar conformation by the
presence of strong hydrogen bonds. The di-deprotonation of dihydroxyfumaric acid to form the
corresponding dianion raises the HOMO from -7.53 ev to -5.33 ev suggesting that the
nucleophilicity of the dianion should be enhanced compared to the protonated species. In contrast,
the respective LUMOs for both dihydroxyfumaric acid and the corresponding dimethyl ester are -
2.94 ev and -2.21 ev suggesting that these species are stronger electrophiles than
dihydroxyfumarate (LUMO = -0.76 ev). These relative energies also suggest that the
dicarboxylate (DHF2-) can behave as a reducing agent or nucleophilic agent while the acid and the
diester cannot. These results support the hypothesis that the formation of the dianionic structure
is important in order to facilitate reactions where DHF acts as a nucleophile.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C) DHF2-
A) DiMeDHF
B) DHF free acid
LUMO = -2.21 eV
HOMO = -6.35 eV
LUMO = -2.94 eV
HOMO = -7.53 eV
LUMO = -0.76 eV
HOMO = -5.33 eV
Figure 4. Calculated Frontier Molecular Orbitals of DHF-Diester, DHF Free Acid and DHF
Dicarboxylate
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 49
1
2
3
4
5
6
7
8
9
Optimization the Reaction of DHF with Benzaldehyde and Hydroxide Bases- The
Deoxalation Pathway. With this understanding of the nucleophilic behavior of the dicarboxylate
form of DHF, we sought to expand the scope of the work by Sagi et al.12 and Butch et al.13 by
exploring the reactivity of aryl and heteroaryl aldehydes toward the observed chemodivergence.
Benzaldehyde was selected as the substrate for optimization and various bases and solvent systems
(e.g., organic, aqueous, and mixed) were screened to first determine conditions selective for the
deoxalation pathway. The results are summarized in Table 1. It was previously shown by Sagi et
al.12 and Butch et al.13 that two hydroxide molecules (one as a nucleophile, one as a base) are
necessary to promote the deoxalation pathway (Scheme 3).14
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
OH
O
O2C
O
OH
CO2
R
H
CO2
O
OH
O
O
aldol-type
addition
R
H
OH
H
OH
OH
O
OH
O
deoxalation
CO2
R
R
CO2
HO CO2
OH
CO2
O2C
oxalate, 7
Scheme 3. Deoxalation Pathway and the Importance of Hydroxide
When 4 equivalents (2 equiv to form dicarboxylate and 2 equiv to promote deoxalation) of
either TBAH, NaOH, or LiOH were included in the THF reaction, only Cannizzaro reaction
products20,21 (e.g., benzoic acid, benzyl alcohol, etc.) were observed (Table 1, entries 1-3).
Addition of 1 equivalent of 15-crown-5 (to activate the hydroxide) with NaOH in THF gave the
desired deoxalation product 9a in 56% yield as a diastereomeric mixture of the sodium carboxylate
complexed to the crown ether (entry 4). Reducing the amount of 15-crown-5 to 0.5 equiv resulted
in a severe drop in yield (23%, entry 5). Using of 2 equiv of NaH (to form the DHF dicarboxylate)
10
ACS Paragon Plus Environment

Page 11 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
and 2 equiv of NaOH with 1 equiv of the crown ether improved the yield of the deoxalation product
9a to 75% yield (entry 6). Unfortunately, efforts to remove the 15-crown-5 by acidification or ion
exchange either resulted in degradation or substantial loss of material. Moreover, the presence of
the crown ether made characterization using NMR extremely difficult due to overlapping aliphatic
signals with the product. Given the difficulty in purification of the product and the requirement of
a full equivalent of crown ether, we abandoned this approach and looked for other conditions that
would facilitate the deoxalation reaction without the necessity of crown ether.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Deoxalation Reaction Optimizationa
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 49
1
2
3
4
5
O
OH
O
O
CO2H
OH
base
HO
H
OM
OH
solvent
23 ˚C
OH
OH
CO2H
(3 equiv)
(1 equiv)
6
7
1
8a
9a
10a
entrya
Base (equiv)
Solvent
%Yield (9a/10a)
8
9
0/0c
0/0c
TBAH (4)
NaOH (4)
1
2
THF
THF
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0/0c
56d/0
LiOH (4)
3
4
THF
THF
NaOH:15-c-5 (4:1)
NaOH:15-c-5 (4:0.5)
NaH:NaOH:15-c-5 (2:2:1)
NaOH (4)
5
6
7
THF
THF
H2O
23d/0
75d/0
0/0c
0/0c
NaOH (4)
8
THF:H2O (1:1)
0/0c
0/0c
25e/0
45e/0
72e/0
NaOH:LiOH (4:2)
NaOH:LiOH (4:2)
NaOH:LiOH (4:2)
NaOH:LiOH (4:2)
NaOH:LiOH (4:2)
9
H2O
10
11
12
13
THF:H2O (1:1)
THF:H2O (1:3.5)
THF:H2O (1:30)
THF:H2O (1:47)
a Reactions were performed with 1 equiv. of DHF, 3 equiv. of benzaldehyde, and indicated base in indicated solvent
1
and temperature for 18 h. b Yields determined by quantitative H NMR of the crude reaction mixture with dimethyl
c
d
sulfone as an internal standard. No desired products. Only products from Cannizzaro reaction observed. Isolated
e
1
yield of the 15-crown-5 sodium carboxylate complex. Yield determined by quantitative H NMR with dimethyl
sulfone as an internal standard following acidification with Amberlyst 15 resin and extraction of organic soluble
byproducts.
Given this issue, we went back to the drawing board and decided to directly adapt the
published aqueous method from Sagi et al.12 and Butch et al.13 NaOH (6 equivalents) in H2O or
in a 1:1 THF-H2O mixture only provided Cannizzaro reaction products (entries 7 and 8). Previous
work had indicated that a combination of both LiOH (2 equiv) and NaOH (4 equiv) afforded better
results as the DHF dilithium salt is much more soluble in water than the sodium salt.
Unfortunately, initial combination of these bases either in H2O or in a 1:1 THF-H2O mixture
continued to produce Cannizzaro products (entries 9 and 10). Interestingly, employing a lower
THF to H2O ratio and changing the order of addition was key in the development of a successful
process. Benzaldehyde dissolved in THF (2 mL) was added to a solution of NaOH in H2O (5 mL)
followed by slow addition of DHF and LiOH in H2O (2 mL) to give a 1:3.5 THF-H2O mixture.
Using this sequence, deoxalation product 9a was observed (entry 11). Since the product could not
12
ACS Paragon Plus Environment

Page 13 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
be extracted into the organic phase, the reaction mixture was treated with Amberlyst 15 resin22 to
acidify the solution and remove the sodium and lithium cations and the organic soluble byproducts
were removed by extraction. Upon concentration of the aqueous phase, a 25% yield of deoxalation
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
product 9a was observed by quantitative H NMR along with a collection of other unidentified
peaks. Reducing the amount of THF used to dissolve the benzaldehyde to ~0.23 mL resulted in a
1:30 THF-H2O ratio and gave a 45% 1H NMR yield (entry 12). In the end, the addition of a minimal
amount of THF (0.15 mL) to dissolve benzaldehyde (providing a 1:47 THF-H2O solution) proved
to be the best conditions, with a 72% 1H NMR yield observed for deoxalation the product 9a (entry
13).
Deconvoluting Deoxalation from DHF Self-Condensation Pathways. We hypothesized that the
undesired peaks in the deoxalation reaction arose from DHF self-condensation. To test that
hypothesis, we exposed DHF to the deoxalation reaction conditions in the absence of benzaldehyde
(Figure 5A). Through spiking studies as well as comparison to literature precedent14 we
determined that the signal at 174.3 ppm is glycolic acid (13), the pair of peaks at 173.3 and 172.8
ppm are the two diastereomers (D/L and meso) of tartaric acid (6), and the peak at 170.8 ppm is
tartronic acid (12). 2-D NMR experiments and DEPT further confirm these identifications.
Comparing the control reaction against the same reaction with 3 equiv of benzaldehyde, the signals
at 174.1 and 174.0 ppm represent the diastereomers of the deoxalation product while the same
peaks from DHF self-condensation are also observed (Figure 5B). A signal for oxalic acid can also
be observed at 162.2 ppm.
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 49
1
2
3
4
5
OH
OH
HO2C
CO2H
tartaric acid (6)
(D/L and meso)
6
7
A.
HO2C
CO2H
8
9
CO2H
OH
tartronic acid (12)
OH
glycolic acid (13)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
B.
CO2H
HO2C
OH
CO2H
9a
oxalic acid (7)
Ph
OH
Figure 5. Crude 13C NMR (DMSO-d6) of Reactions of (A) DHF with NaOH:LiOH (4:2) in
THF:H2O (1:3.5) and (B) with 3 equiv of benzaldehyde added after acidification.
We envision a range of different pathways that DHF can undergo to form the observed
byproducts under strong basic conditions (Figure 6). These pathways are distinct from the
previously published self-reactivity of DHF under slightly basic conditions (pH 8). A single DHF
molecule can fragment to form glycolate (13) and glyoxylate (2). Glyoxylate is a highly reactive
species that has already been shown to react with DHF to form tartrate (6) through a deoxalative
mechanism.13 Dimerization of DHF forms the tetra-carboxy intermediate I, which undergoes
decarboxylation to form II. II can also tautomerize to III then proceed through a deoxalation-like
cleavage mechanism to release tartronate (12) and tartrate. Tartronate can also decarboxylate to
produce glycolate. The three observed products are the only non-volatile, stable products from the
self-reactivity of DHF at high pH after acidification. It is important to mention that the observed
peak ratios do not accurately represent how much condensation products are formed. Based on the
proposed mechanism, tartrate should be generated in a greater amount than tartronate. However,
the NMR shows a much larger amount of tartronic acid than tartaric acid. This discrepancy is an
14
ACS Paragon Plus Environment

Page 15 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
artifact of the workup procedure. Some tartrate is extracted into the organic phase and some
undergoes degradation following the acidic workup.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
deoxalation-like
fragmentation
CO2
O
HO
OH
O2C
OH
O2C
H
CO2
DHF, 1
glycolate, 13
glyoxylate, 2
dimerization
O
HO CO2
OH
O
HO CO2
OH
decarboxylation
O2C
O2C
OH CO2
O2C OH
CO2
II
I
deoxalation-like
fragmentation
OH
CO2
HO
HO CO2
HO
O2C
CO2
OH
O2C
CO2
tartrate, 6
OH
tartronate, 12
O
CO2
III
Figure 6. Proposed mechanism for the origin of tartronic acid (12), tartaric acid (6), and glycolic
acid (13)
Examination of Substrate Scope of Deoxalation Reaction. With working conditions for the
deoxalation reaction in place, we next examined the scope and limitations of employing various
aryl and heteroaryl aldehydes (Scheme 4). Instead of relying on quantitative NMR, we decided to
derivatize the products to the corresponding methyl esters for easy isolation via column
chromatography. Heating the crude mixture with catalytic H2SO4 in MeOH readily afforded
esterification products in greater than ~80% yield. For instance, the deoxalation product from
benzaldehyde was converted to its methyl ester in 81% yield. This gave an overall yield of 58%
for the two steps upon isolation. The ester 14a was isolated as a 1:1 mixture of diastereomers.
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 49
1
2
3
4
Scheme 4. Hydroxide-Mediated Cascade Aldol-Deoxalation Reactions
5
6
7
8
i. NaOH, LiOH
THF:H2O
18 h
OH
CO2H
OH
OH
H2SO4 (cat)
MeOH, 5 h
O
HO2C
CO2Me
CO2H
+
R
R
R
H
ii. A-15 resin
0.5 h
OH
OH
OH
1
8
9
14
OH
OH
OH
9
OH
CO2Me
CO2Me
CO2Me
CO2Me
OH
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
OH
OH
Me
MeO
Et
14a, 58%a,b (72%)d
14b, 0%b
14c, 63%a,b
(1:1 dr)
14d, 38%a,b
(1:1 dr)
(1:1 dr)
OH
OH
OH
F
OH
CO2Me
CO2Me
F
CO2Me
CO2Me
OH
OH
OH
OH
i-Pr
F
14e, traceb
14f, 64%a,b
(1:1 dr)
14g, 40%a,b (45%)d
14h, 44%a,b
(2:1 dr)
(2:1 dr)
OH
OH
OH
OH
CO2Me
CO2Me
CO2Me
CO2Me
OH
OH
OH
OH
MeO2C
14k, 25%a,c
Br
O2N
Cl
14i, 61%a,c
(3:2 dr)
14j, 68%a,c
(3:2 dr)
14l, 41%a,c
(1:1 dr)
(2:1 dr)
NH OH
OH
OH
CO2H
N
CO2Me
CO2Me
N
OH
OH
OH
S
14m, 32%a,b (38%)d
14n, 43%a,c
picrasidine Y
9o, — (8%)c,d
(1:1 dr)
(3:2 dr)
a Isolated yield of ester 14 (over two steps)
b THF:H2O (1:47)
c THF:H2O (1.3.5)
d qNMR yield of acid 9
We then probed electronic effects on the reaction by looking at 4-substituted
benzaldehydes. The reaction with 4-anisaldehyde failed to provide any desired deoxalation product,
whereas 63% yield of dihydroxyester 14c was obtained with 4-tolualdehyde (8c). In contrast to
tolualdehyde, the 4-ethylbenzaldehyde 8d gave 38% yield of 14d whereas only trace product was
observed with 4-isopropylbenzaldehyde 8e. 4-Fluorobenzaldehyde (8e) was similarly compatible
with the reaction sequence forming dihydroxy ester 14e in 64% yield. Both 3-fluoro- and 2-
fluorobenzaldehydes (8f and 8g) similarly gave their corresponding methyl esters 14f and 14g in
40% and 44% yield, respectively. 4-Chloro- and 4-bromobenzaldehydes (8i and 8j) are solids and
were not readily solubilized in the amount of THF (0.15 mL) used in the 1:47 THF-H2O mixture.
To ensure solubility, 2 mL of THF was used to dissolve the solids before addition to the aqueous
DHF/base solution (resulting in a 1:3.5 THF-H2O mixture). In this solvent system, the respective
16
ACS Paragon Plus Environment

Page 17 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
chloro and bromo ester products 14i and 14j were obtained in 63% and 68% yield. Stronger
electron-withdrawing groups on the phenyl ring, such as 4-CO2Me and 4-NO2,23 were also
amenable to the reaction, albeit with reduced yields. 4-CO2Me-benzaldehyde (8k) only gave 25%
yield of the ester product 14k while 41% yield of ester 14l was obtained with the 4-NO2-
benzaldehyde (8l). In both cases, increased amounts of Cannizzaro reaction products were
observed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Several heteroaryl aldehydes were also examined under the reaction conditions. 3-
Thiophencarboxaldehyde 8m gave 38% qNMR yield of 9m and a 32% yield of ester 14m over the
two steps. 2-Formylpyridine (8n) performed similarly and gave 43% yield of ester 14n. Finally,
1-formyl b-carboline (8o) was subjected to the reaction conditions to form the natural product,
picrasidine Y24 (9o). 9o was formed in 8% yield by qNMR along with several undesired side
products (including Cannizzaro products). Protection of the carboline would most likely improve
the product yield given the strong basic conditions and the acidic carboline proton. Picrasidine Y
has been previously synthesized in 7 steps24a using tartaric acid. Despite the low yield, this
synthesis thus represents a one-step approach to picrasidine Y which could offer a rapid
opportunity for compound library development by employing substituted 1-formyl-b-carbolines.
Optimization of the Reaction of DHF with Benzaldehyde and Tertiary Amine Bases- The
Decarboxylation Pathway. While searching for a viable protocol for the deoxalation reaction, we
simultaneously probed for conditions to selectively promote the decarboxylation reaction (Table
2). Given that trace decarboxylation products were detected in the reaction of DHF and
benzaldehyde with no added base in THF at room temperature (entry 1), we started by increasing
the reaction temperature. Upon heating the reaction at reflux, 2,3-dihydroxypropanone 10a was
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
obtained in 15% yield as the only decarboxylation product (entry 2). This observation runs counter
to what was expected based on Mahrwald's decarboxylation work.15 In that report for aryl
aldehydes, the isomeric aryl-substituted dihydroxyacetones or diaryl trihydroxybutanones were
predominantly formed. Only three examples of aryl 2,3-dihydroxypropanone formation were
shown. To affect the transformation, brucine (50 mol %) and Cs2CO3 (1.5 equivalents) were
employed, but the yields were low (13-38% yield) in each case. In contrast, our reaction provides
selective mono-aldol additions (despite using excess aldehyde) and formation of the
dihydroxypropanones. It is also important to note that the transformation represents a formal
C(sp2)-H alkylation of the aromatic aldehyde without involving any transition metal-based
chemistry.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Next, bases were screened to facilitate the decarboxylation reaction at 65 ˚C. The use of
hydroxides, alkoxides, or carbonate bases resulted in Cannizzaro reaction (entries 3-5). Next,
tertiary amine bases were added. With Hünig’s base (iPr2NEt, 4 equiv), 10a was obtained in 25%
yield (entry 6). While no product formation was detected with DBU (entry 7), NEt3 afforded 10a
in 68% yield (entry 8). Increasing the amount of NEt3 to 6 equivalents resulted in a 62% yield
(entry 9), whereas reducing the amount of NEt3 to 3 equivalents provided 72% yield of 10a (entry
10). Further efforts to lower the amount of NEt3 failed to improve the yield (entries 11 and 12).
Changing the relative amounts of aldehyde and DHF from 3:1 to either 2:1 or 1:1 resulted in
reduced yields (entries 13 and 14). Other solvents such as dioxane and methylene chloride also
failed to improve the product yield (entries 15 and 16).
18
ACS Paragon Plus Environment

Page 19 of 49
The Journal of Organic Chemistry
1
2
3
4
Table 2. Decarboxylation Optimizationa
5
6
7
8
base
(X equiv)
O
O
OH
OH
CO2H
HO2C
H
solvent, Temp
18 h
OH
OH
10a
8a
1
9
yield (%)b
entry
base (equiv)
solvent (Temp [˚C])
8a:1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
THF (23)
THF (65)
1
2
3:1
3:1
None
None
trace
15
--c
NaOH (4)
THF (65)
THF (65)
3
4
3:1
3:1
--c
--c
25
K2CO3 (4)
NaOEt (4)
THF (65)
THF (65)
5
6
3:1
3:1
i-Pr2NEt (4)
d
DBU (4)
NEt3 (4)
NEt3 (6)
NEt3 (3)
NEt3 (2)
NEt3 (1)
NEt3 (4)
NEt3 (4)
NEt3 (4)
NEt3 (4)
THF (65)
THF (65)
THF (65)
THF (65)
THF (65)
THF (65)
THF (65)
THF (65)
CH2Cl2 (40)
dioxane (100)
7
8
3:1
3:1
3:1
3:1
3:1
3:1
2:1
1:1
3:1
3:1
—
68
62
72
64
55
45
28
23
20
9
10
11
12
13
14
15
16
a Reaction performed with indicated amounts of DHF, benzaldehyde 8a, and base in indicated solvent at the listed
temperature. b Isolated yields after column chromatography. c Formation of benzoic acid via Cannizzaro-type reaction
of 8a. d No desired product. Degradation of DHF observed.
Examination of Substrate Scope of the Decarboxylation Reaction. Having optimized the
conditions for the decarboxylation reaction with benzaldehyde, the scope and limitations of the
transformation were initially examined with the same substituted aryl aldehydes that were
previously utilized in the deoxalation study (Scheme 5). Whereas no deoxalation product was
formed with 4-anisaldehyde (8b), a 10% isolated yield of decarboxylation product 10b was
obtained under the triethylamine conditions. 4-Methyl-, 4-ethyl-, and 4-isopropyl-benzaldehydes
(8c-e) gave dihydroxy-1-phenylpropanones 10c-e in 54%, 41%, and 25% yield, respectively.
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 49
1
2
3
4
Scheme 5. Triethylamine-mediated Decarboxylative Aldol Cascade Reactions of DHF with Aryl
5
6
Aldehydes
7
8
9
NEt3
(4 equiv)
O
O
OH
OH
CO2H
HO2C
H
R
R
THF, 65 ˚C
18 h
OH
OH
8
1
10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
OH
O
OH
O
OH
O
OH
OH
OH
OH
O
OH
O
OH
MeO
OH
Me
Et
10a, 72%a
10b, 10%
10c, 54%
10d, 41%
O
OH
OH
X
O
F
OH
OH
OH
OH
i-Pr
F
10f, 65%
10g, 31%
10e, 25%
10h (X = F), 0%
10p (X = Br), 0%
10q (X = Me), 22%d
O
OH
O
OH
O
OH
O
OH
OH
OH
OH
OH
OH
Cl
Br
MeO2C
O2N
10i, 49%
10j, 45%
10k, 0%c
10l, 0%c
O
O
OH
O
OH
NH
O
OH
N
X
OH
OH
OH
S
N
OH
10m, 23%
10r (X = S), 48%
10s (X = O), 39%
10n, 0%
10o
a Reaction performed with 3 equiv NEt3.
b No Reaction (NR)
c Reduction to benzyl alcohol was observed.
d Obtained as a 2:1 mixture of 10q and its dihydroxyacetone isomer 17q.
e Reaction performed in CH2Cl2 at 40 ˚C.
For the mono-substituted fluorobenzene series, 4-fluoro- and 3-fluoro-benzaldehydes (8f
and 8g) provided decarboxylation products 10f and 10g in 65% and 31% yield, respectively. A
complex mixture of the isomeric diaryl trihydroxybutanones (16h and 17h) was obtained with 2-
fluorobenzaldehyde (8h) (Figure 7). Despite using excess aldehyde, this was the first instance that
double aldol addition products were observed under our conditions. Intrigued by the observation
of the double aldol addition products, we employed 2-bromobenzaldehyde (8p) and 2-
tolualdehyde (8q) and to probe steric and electronic effects. 2-Bromobenzaldehyde gave a similar
complex mixture of double aldol isomers (15p and 16p) as seen with the 2-fluoro derivative. 2-
Tolualdehyde instead produced 22% yield of a 2:1 mixture of carbonyl isomers 10q and 17q.
20
ACS Paragon Plus Environment

Page 21 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
Beyond these complex substrates, 4-chloro- and 4-bromobenzaldehydes (8i-8j) were more
amendable and each gave their respective dihydroxypropanones 10i-10j in 49% and 45% yield.
7
8
9
Me OH OH
X
OH OH
X
O
OH
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
OH
X
OH OH
X
15h (X = F)
15p (X = Br)
16g (X = F)
16p (X = Br)
17q
Figure 7. Dihydroxyacetone and Isomeric Diaryl Trihydroxybutanones Products Obtained from
2-Substituted Benzaldehydes
Interestingly, products 10k and 10l were not observed with 4-CO2Me- and 4-NO2-
benzaldeyde (8k and 8l). Instead, only reduction to form benzyl alcohols 18k and 18l were
observed (Scheme 6). The benzyl alcohols do not arise from Cannizzaro reactions as there are no
hydroxide (or alkoxy) bases present. Therefore, DHF must be playing a direct role in their
formation.
O
O
OH
CO2H
NEt3
OH
HO
H
OH
THF
65 ˚C
X
OH
CO2H
X
X
1
10k (X = CO2Me)
10l (X = NO2)
not observed
18k (X = CO2Me)
18l (X = NO2)
8k (X = CO2Me)
8l (X = NO2)
Scheme 6. Unexpected Aldehyde Reduction
Heteroaryl aldehydes were next explored, but proved troublesome upon initial screening;
substantial degradation of starting materials and/or products was observed and no
dihydroxypropanone products were detected. After some optimization, it was found that products
could be isolated when the reaction was performed in CH2Cl2 at reflux. Under these conditions, 3-
thienyl-, 2-thienyl-, and 2-furyl-2,3-dihydroxypropanones (10m, 10r, and 10s) were obtained in
23%, 48%, and 39% yield, respectively. In contrast, nitrogen-containing heteroaryl rings were
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 49
1
2
3
4
problematic for the method. Reduction to the benzyl alcohols took place with both pyridine-2-
5
6
carbaldehyde (8n) and 1-formyl-b-carboline (8o).
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Application of the Decarboxylation: Synthesis of C-Veratroylglycol. Aryl 2,3-
dihydroxypropanones (10) represent a class of highly oxidized alkyl chains that have garnered the
attention of synthetic chemists for many years.25 They represent a diverse set of compounds that
include both pharmaceutically-relevant lignan and non-lignan natural products and synthetically-
derived compounds.26 Despite the prevalence of (hetero)aryl 2,3-dihydroxylpropanones in nature,
only a small subset has been examined for biological activity due to the low abundance of the
individually isolated compounds.27 For instance, C-veratroylglycol (10t) is a naturally-occurring
2,3-dihydroxy-1-phenylpropanone that has been isolated from a variety of plants, fruits, and tree
saps (Scheme 7).28 It has been shown to demonstrate modest antioxidant,29 antiproliferative
(against human colon cancer cells),30 and COX-2 inhibitory activity31 in various studies. The key
issues in studying the activity of 10t are its low abundance in natural sources which gives rise to
poor isolated yields. For example, <75 mg can be isolated from 1 kg of hazelnut shells after
extensive grinding, milling, sequential extractions, and column fractionations.28 Similarly, while
one study reported that 1.5 mg of 10t was isolated from 20 L of maple syrup, another claimed that
only 0.5 mg was obtained from 1 kg of syrup.29,30,32 C-Veratroylglycol can also be purchased
primarily from contract companies at a cost of >$100/mg.33 Therefore, a cheaper approach toward
C-veratroylglycol would be an enabling endeavor for the scientific community.
O
O
OH
NEt3
MeO
HO
MeO
HO
CO2H
OH
H
HO
THF
65 ˚C
OH
CO2H
14%
Vanillin (8t)
C-Veratroylglycol (10t)
1
22
ACS Paragon Plus Environment

Page 23 of 49
The Journal of Organic Chemistry
1
2
3
4
Scheme 7. One-step Decarboxylative Synthesis of C-Veratroylglycol (10t)
5
6
7
8
9
We were pleased to find that the reaction of vanillin (8t), DHF, and triethylamine (7
equivalents)34 afforded C-veratroylglycol (10t) in 14% yield from the single step reaction (Scheme
7). One true value of this decarboxylation approach lies in the financial savings. Based on raw
material and solvent pricing,35 we can synthesize 25-30 mg of C-veratroylglycol for ~$2.36/mg,
which is ~50-fold cheaper than purchasing it from current commercial sources. To complement
this synthesis, we have begun to assemble a small compound library of various substituted aryl
2,3-dihydroxypropanones by using the decarboxylation method (see Supporting Information for
more details).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Complete Mechanistic Picture of Chemodivergence. The complete mechanistic landscape for
the reactions of DHF with aryl aldehydes is shown in Figure 8. The reaction can take place along
various pathways depending on the choice of base, solvent, and the electron-donating or electron-
withdrawing substituents on the aryl aldehyde component. Upon aldol addition of DHF to the
aldehyde to form intermediate 19, three possible pathways for fragmentation are plausible. In the
first pathway, nucleophilic attack on the carbonyl by hydroxide occurs. Subsequent deprotonation
(mediated by hydroxide) leads to the deoxalative fragmentation pathway to generate 1-aryl 2,3-
dihydroxypropanoates 9 (see Figure 5). In the second pathway, NEt3 mediates the transformations
as a proton transfer agent (NEt3/+NEt3H). Following aldol addition of DHF, aldol addition
intermediate 19 then undergoes decarboxylation to form enediol 21. NEt3-mediated
tautomerization provides 22 followed by another decarboxylation to form enediol intermediate 23,
which can tautomerize to dihydroxyacetone 17. Further isomerizations mediated by NEt3 and its
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 49
1
2
3
4
5
6
7
8
9
conjugate acid provide the observed 2,3-dihydroxypropanone 10. Alternatively, a second aldol-
type reaction can take place between enediol 23 and another aldehyde molecule to provide diaryl
trihydroxybutanones 15 and 16 (after carbonyl isomerization). This second addition arises when
the isomerization from 24 to 10 is slower than tautomerization to enediol 23. In the cases of
electron-poor arenes with NEt3, fragmentation of intermediate 19 can occur such that a resonance-
delocalized benzylic anion is generated along with 2,3-dioxosuccinate (26). Subsequent
protonation yields the alcohols 18. The major impediment toward higher yield is presumably the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
complications that can result from self-condensation of DHF (see Figure 6).
fragmentation
OH
O
OH
NEt3
OH
OH
OH
H-NEt3
Aldehyde
reduction
pathway
CO2
Ar
CO2
Ar
Ar
Ar
H
O
O
O
18
9
OH
O
Deoxalation
pathway
(Figure 3)
CO2
19
O2C
O2C
CO2
26
oxalate, 7
O
O
O
8
aldol-type
addition
Ar
H
H
NEt3
8
Ar
H
NEt3
HO
HO
O
HO
Ar
OH
HO
Ar
OH
decarboxylation
tautomerization
OH
O
O
O
self-condensation
pathway
O
Ar
O2C
CO2
aldol-type
addition
O
H
O
(Figure 5)
O
O
O
O
OH
1
19
22
21
Et3N
H
Decarboxylation
pathway
decarboxylation
HO
Ar
OH
O
OH
OH OH
O
H
tautomerization
tautomerization
tautomerization
O
OH
Ar
Ar
Ar
25
O
OH
10
H
OH
24
23
Double
aldol
O
addition
pathway
Ar
H
H
O
OH
O
OH
OH OH
tautomerization
tautomerization
H
Ar
Ar
Ar
Ar
Ar
27
Ar
OH OH
OH OH
O
OH
17
15
Figure 8. Integrated Mechanistic Proposal for Product Formation
24
ACS Paragon Plus Environment

Page 25 of 49
The Journal of Organic Chemistry
1
2
3
4
CONCLUSIONS
5
6
7
8
In summary, through a combination of control reactions and DFT calculations, we have
demonstrated that DHF is nucleophilic in its dicarboxylate form and electrophilic in both its free
acid form and when derivatized as a diester. The DHF dicarboxylate, under the high pH conditions,
readily undergoes self-condensation to afford tartronic acid, glycolic acid, and tartaric acid. In
contrast, the diester is completely stable and unreactive even at elevated temperatures, while the
free acid undergoes very slow self-condensation. This represents the first efforts to understand and
codify the reactivity of DHF in order to harness its greater synthetic potential.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NaOH/LiOH
OH
(4:2)
CO2H deoxalation
pathway
9
Ar
THF:H2O
(1 : > 3.5)
23 ˚C
OH
CO2H
OH
O
HO2C
Ar
H
O
OH
1
8
NEt3
Ar
10
OH
Ar
OH
THF
65 ˚C
OH
18
decarboxylation aldehyde reduction
pathway
(for electron-rich
aromatics)
pathway
(for electron-poor
aromatics)
Scheme 8. Summary of Observed Chemodivergence
In particular, the reactions of DHF with aryl aldehydes display remarkable base-mediated
chemodivergence (Scheme 8). With excess hydroxide in a THF:H2O (1:47) mixture, DHF reacts
with aryl aldehydes by aldol addition followed by a deoxalative fragmentation to form 3-aryl-2,3-
dihydroxypropanoates. The propanoates were isolated as their methyl esters in up to 64% yield
following esterification of the crude mixture. The nature of the base used to form the dicarboxylate
plays a central role in the nucleophilic reactions of DHF. Hydroxide acts as both a base and a
nucleophile to promote deoxalation. Conversely, with excess NEt3 in THF, the same aldol
intermediate undergoes a series of decarboxylations and carbonyl isomerizations to form 1-aryl-
2,3-dihydroxypropanones in up to 72% yield. NEt3 acts as a proton shuttle to mediate the various
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 49
1
2
3
4
5
6
7
8
9
transformations. Compared to the literature precedent for the DHF reaction with aryl aldehydes
(which predominantly forms substituted dihydroxyacetones or the disubstituted
trihydroxybutanones), selective formation of the 1-aryl-2,3-dihydroxypropanones is achieved. The
overall reaction represents a formal C(sp2)-H alkylation of an aldehyde to form the propanone
products. In both cases, the reaction is generally amenable to 3- and 4-substituted benzaldehydes,
although strongly electron-rich substituents give very low yields or no reaction at all. While
strongly electron-withdrawing substituents are tolerable in the deoxalation reactions with
hydroxide, the desired propanones are not formed with NEt3 and reduction to the corresponding
benzyl alcohols are observed instead. Heteroaryl aldehydes are amenable to the deoxalation
reactions but those containing basic nitrogens give reduction products or other side reactions with
NEt3. Finally, in an application of the decarboxylation method toward target synthesis, C-
veratroylglycol- an expensive, bioactive, and naturally-occurring 2,3-dihydroxy-1-
phenylpropanone- was synthesized cheaply in one step from vanillin with a 14% overall yield.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
General Information. Chromatographic purification was performed as flash chromatography
with Silicycle SiliaFlash P60 silica gel (40-63 µm) or preparative thin-layer chromatography (prep-
TLC) using silica gel F254 (1000 µm) plates and solvents indicated as eluent with 0.1-0.5 bar
pressure. For quantitative flash chromatography, technical grades solvents were utilized.
Analytical thin-layer chromatography (TLC) was performed on Silicycle SiliaPlate TLC silica gel
F254 (250 µm) TLC glass plates. Visualization was accomplished with UV light.
26
ACS Paragon Plus Environment

Page 27 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Infrared Spectra were obtained on a Thermo Nicolet 6700 FTIR and analyzed using
OMNIC software. The IR bands are characterized as broad (br), weak (w), medium (m), and strong
(s). Proton and carbon nuclear magnetic resonance spectra (1H NMR and 13C NMR) were recorded
on a Bruker 500 MHz spectrometer with solvent resonances as the internal standard (1H NMR:
CDCl3 at 7.26 ppm; 13C NMR: CDCl3 at 77.0 ppm). MestraNova was used to analyze NMR
spectra. 1H NMR data are reported as follows: chemical shift (ppm), multiplicity (s = singlet, d =
doublet, dd = doublet of doublets, dt = doublet of triplets, ddd = doublet of doublet of doublets, t
= triplet, m = multiplet, br = broad), coupling constants (Hz), and integration. High resolution mass
spectrometry (HRMS) measurements were obtained either through EI on a Micromass AutoSpec
M [forward geometry (EBE) three sector tandem MS] or through ESI on a Thermo Orbitrap XL.
The accurate mass analyses run in EI mode were at a mass resolution of 10,000 and were calibrated
using PFK (perfluorokerosene) as an internal standard. The accurate mass analyses run in ESI
mode were at a mass resolution of 30,000 using the calibration mixture supplied by Thermo.
Uncorrected melting points were measured with a digital melting point apparatus (DigiMelt MPA
160). DFT calculations performed using B3LYP functional with 6-31G* in the gas phase as the
basis set.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of DiMeDHF (11). In a flame dried flask, DHF hydrate36 (5.0 g, 27.2 mmol) was
dissolved in anhydrous MeOH (35 mL) under an inert atmosphere. The resulting solution was
cooled to 0 ˚C and thionyl chloride (8.8 g, 59.7 mmol, 2.2 equiv) was added. The reaction was
warmed to room temperature and stirred for 2 d while a white precipitate was formed. The
precipitate was recovered by filtration to provide DiMeDHF (11) as a white solid (4.4 g, 74%).
Characterization matches previously published results.37
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 49
1
2
3
4
5
6
7
8
DHF Derivative Self-Reactivity Studies. To determine the relative stability of DHF and DiMeDHF,
150 mg of both were separately dissolved in refluxing THF (7 mL) for 3 d. At which point the
reactions were cooled to room temperature and concentrated to dryness. The crude mixture was
then analyzed by NMR. DiMeDHF was completely recovered while only 10% of DHF had reacted
to give pentulosonic acid and other side products. The di-lithiated DHF stability was confirmed
according to the published self-reactivity of DHF work by Sagi et al.16
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General Procedures for the Synthesis of (Hetero)aryl 2,3-dihydroxypropanoic Methyl Esters
Deoxalation Procedure: A dried round bottom flask is charged with nitrogen and NaOH (130 mg,
3.26 mmol, 4 equiv). To this, deionized water (5 mL) is added and stirred to fully dissolve the
salt. A 16 M solution of aldehyde (2.44 mmol) in distilled THF (0.15 mL) was prepared in a vial.
In a second separate vial, 1 eq of DHF and 2 eq of LiOH•H2O (68 mg, 1.63 mmol) are dissolved
in water to bring the concentration of LiOH to 1M (2 mL). The aldehyde solution is added to the
reaction flask followed by slow addition of DHF/LiOH solution via syringe pump. Volume is
added over 30 min. The reaction is then allowed to stir overnight. After ~18 h, Amberlyst-15
cationic exchange resin is added to the flask. Stirring is continued until the pH of the solution
reaches ~4. Usual equilibration of pH takes approximately 20 min depending on amount of resin
beads added. The mixture was poured over a Buchner funnel to remove the beads and washed
with both water and EtOAc. The biphasic mixture was separated via a separatory funnel and the
aqueous phase was extracted with EtOAc (3x) to remove all undesired Cannizzaro reaction
products. The aqueous phase contains both the dihydroxy acid products as well as DHF
dimerization/fragmentation products (tartrate, glycolate, hydroxy malonate).
28
ACS Paragon Plus Environment

Page 29 of 49
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Esterification Procedure: After concentration of the aqueous phase through rotatory evaporation,
the crude mixture was suspended in MeOH and transferred into another round bottom flask.
Catalytic concentrated sulfuric acid (5 drops) was added and the reaction mixture refluxed for 5 h.
After completion by TLC, the reaction was concentrated onto silica and purified via flash
chromatography (5 % MeOH/DCM).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Methyl 2,3-dihydroxy-3-phenylpropanoate (14a). The general deoxalation and esterification
procedures were followed using DHF dihydrate (150 mg, 0.814 mmol), benzaldehyde 8a (259 mg,
2.44 mmol) in 0.15 mL of THF, NaOH (130 mg, 3.26 mmol) and LiOH•H2O (68 mg, 1.63 mmol)
in deionized water. After purification (5% MeOH/DCM, Rf = 0.2), compound 14a was afforded
as a clear oil (93 mg, 58% yield) as a 1:1 mixture of diastereomers. Characterization matches
previously published results.38
Methyl 2,3-dihydroxy-3-(p-tolyl)propanoate (14c). The general deoxalation and esterification
procedures were followed using DHF dihydrate (150 mg, 0.814 mmol), 4-methyl benzaldehyde 8c
(294 mg, 2.44 mmol) in 0.15 mL of THF, NaOH (130 mg, 3.26 mmol) and LiOH•H2O (68 mg,
1.63 mmol) in deionized water. After purification (5% MeOH/DCM, Rf = 0.25), compound 14c
was afforded as a clear oil (108 mg, 63% yield) in a 1:1 mixture of diastereomers. Characterization
matches previously published results.38
Methyl 3-(4-ethylphenyl)-2,3-dihydroxypropanoate (14d). The general deoxalation and
esterification procedures were followed using DHF dihydrate (150 mg, 0.814 mmol), 4-ethyl
benzaldehyde 8d (328 mg, 2.44 mmol) in 0.15 mL of THF, NaOH (130 mg, 3.26 mmol) and
29
ACS Paragon Plus Environment