The "fully Catalytic System" in Mitsunobu Reaction Has Not Been Realized Yet

Article abstract of DOI:10.1021/acs.orglett.6b01894

An investigation of the recently reported "fully catalytic Mitsunobu reaction" using catalytic amounts of a phosphine reagent and an azo reagent has shown that although benzyl 4-nitrobenzoate is formed under the fully catalytic conditions, the same result is obtained if the hydrazine catalyst is omitted, indicating that this is not a Mitsunobu reaction. In addition, when the reaction between (-)-ethyl lactate and 4-nitrobenzoic acid was carried out using the "fully catalytic" method, the corresponding ester was formed but in very low yield and with predominant retention of configuration. Unfortunately, the system catalytic in phosphine reagent is incompatible with that in the azo reagent.

Full text of DOI:10.1021/acs.orglett.6b01894

Letter
pubs.acs.org/OrgLett
The “Fully Catalytic System” in Mitsunobu Reaction Has Not Been
Realized Yet
Daisuke Hirose,† Martin Gazvoda,‡,∥ Janez Kosm
,∥
rlj,* and Tsuyoshi Taniguchi*
,‡
,§
̌
†
Graduate School of Natural Science and Technology, Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan
‡
Faculty of Chemistry and Chemical Technology, University of Ljubljana, Vecn
̌
a pot 113, SI-1000, Ljubljana, Slovenia
§
School of Pharmaceutical Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kakuma-machi,
Kanazawa 920-1192, Japan
S Supporting Information
*
ABSTRACT: An investigation of the recently reported “fully
catalytic Mitsunobu reaction” using catalytic amounts of a
phosphine reagent and an azo reagent has shown that although
benzyl 4-nitrobenzoate is formed under the fully catalytic
conditions, the same result is obtained if the hydrazine catalyst
is omitted, indicating that this is not a Mitsunobu reaction. In
addition, when the reaction between (−)-ethyl lactate and 4-
nitrobenzoic acid was carried out using the “fully catalytic”
method, the corresponding ester was formed but in very low yield and with predominant retention of configuration.
Unfortunately, the system catalytic in phosphine reagent is incompatible with that in the azo reagent.
4
here are myriad methods for the esterification of
using iodobenzene diacetate as a sacrificial oxidizer in 2006.
T
carboxylic acids with alcohols. Most of these do not
affect the stereochemistry potentially present at the chiral
We reported another catalytic Mitsunobu reaction with azo
reagents recyclable by aerobic oxidation in 2013. Ethyl 2-(3,4-
5
secondary alcohol substrate, providing esters with retention of
dichlorophenyl)azocarboxylate has been identified as an
efficient catalyst, enabling the reactions in the presence of a
stoichiometric amount of triphenylphosphine, a catalytic
amount of iron phthalocyanine [Fe(Pc)], and atmospheric
oxygen. Recently, we reported the substantial improvement of
1
configuration. In contrast, an oxidation−reduction condensa-
tion protocol using azo and phosphine reagents, well-known as
the Mitsunobu reaction, is a remarkable exception. The
stereochemistry of a chiral secondary alcohol that is combined
with a carboxylic acid in the Mitsunobu reaction is typically
6
this methodology.
2
inverted in the resulting ester product. The stereochemical
From the standpoint of phosphine reagents, Buonomo and
Aldrich reported the catalytic Mitsunobu reaction with
outcome is rationalized by the Walden inversion via
alkoxyphosphonium intermediates. This is the most important
feature of the Mitsunobu reaction and is utilized as the method
for inversion of stereochemistry of secondary alcohols as well as
stereocontrolled substitution of the hydroxy groups with other
nucleophiles.
Despite overt features that make the Mitsunobu reaction so
attractive, it is associated with serious drawbacks. Typical azo
reagents, such as diethyl azodicarboxylate, are toxic as well as
light and shock sensitive, posing a considerable health and
safety threat. In addition, massive amounts of waste are
produced, including hydrazinedicarboxylates and triphenyl-
phosphine oxide, often making purification of desired products
troublesome.
Creative modifications in reagents and methods have been
realized to circumvent the above-mentioned disadvantages.
Most of them rely on removal of hydrazine and phosphine
oxide byproducts. Ultimately, development of catalytic systems
in the Mitsunobu reaction should be a substantial improve-
ment.
7
recyclable phosphine reagents in 2015. The catalytic concept
is based on in situ reduction of phosphine oxides with
phenylsilane, which has been developed by O’Brien and co-
8
workers in their catalytic Wittig reactions. The authors
succeeded in showing that the catalytic system in phosphines
worked in the Mitsunobu reaction with a stoichiometric
amount of diisopropyl azodicarboxylate (DIAD). Much earlier,
O’Brien and co-workers disclosed the concept of the
Mitsunobu reactions catalytic in phosphines in their patent
8c
report.
In their publication, Buonomo and Aldrich proposed a “fully
catalytic system” of the Mitsunobu reaction combining their
catalytic system in phosphines with ours of the azo reagents
7
recyclable by aerobic oxidation (Scheme 1). If realized, this is
without doubt an outstanding idea, completely resolving the
major drawback of the Mitsunobu protocol, i.e., generation of
the large amounts of problematic wastes, hydrazides, and
phosphine oxides. Under the “fully catalytic system” conditions,
3
From the standpoint of azo reagents, Toy and co-workers
Received: June 29, 2016
reported the first example of the catalytic Mitsunobu reaction
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b01894
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 1. Proposed Concept of the Fully Catalytic
Mitsunobu Reactions
the system catalytic in the phosphine reagent provided the
same product in 68% yield and in a greater than 99% inversion
ratio (Scheme 2, eq 2). We repeated the reaction of Buonomo
7
and Aldrich three times, but our experiments gave ester 1 in
lower yields (38−46%) and moderate inversion (76−87%)
than those described by the authors (Scheme 2, eq 2). When
the reaction was quenched soon after the start (ca. 3 h), no
significant change of the enantiomer ratio (87:13) in ester 1
could be observed (see the Supporting Information (SI)),
ruling out a possibility of gradual racemization of the product
during the course of the reaction. Similarly, we checked the
reaction of chiral 1-phenylethanol (98:2 er) to give the ester
product 2 and again obtained the result that was inferior (51%
yield, 77% inversion) to the reported one (69% yield, 94%
inversion) (Scheme 2, eq 3). It is certain that the “single
catalytic system” of Buonomo and Aldrich provides the
Mitsunobu product, but our results imply that the present
reaction conditions have not been sufficiently optimized as the
Mitsunobu reaction.
however, the authors provided only a few examples with a
limited scope of substrates, demonstrating the condensation of
two benzyl alcohols (4-methoxylbenzyl alcohol and benzyl
alcohol) with 4-nitrobenzoic acid.
It is essential to investigate the stereochemistry in products
of the protocol that is claimed to be the Mitsunobu reaction. In
their “fully catalytic system”, however, Buonomo and Aldrich
have not clarified this point, though the reactions of chiral
secondary alcohols have been properly illustrated in the system
catalytic in the phosphine reagent.
We tested reproducibility of the reaction between benzyl
alcohol and 4-nitrobenzoic acid under the “fully catalytic
7
system” conditions according to the authors’ procedure and
obtained the corresponding ester 3 in 38% average yield
(
Figure 1, entry 1, first and second runs). Although the product
In this letter, we tentatively decided to adapt a model
reaction using (−)-ethyl lactate (>99:1 er) and 4-nitrobenzoic
acid to evaluate the “fully catalytic” Mitsunobu reaction. Our
system catalytic in the azo reagent provided the corresponding
ester 1 in 50% yield and in a 97% inversion ratio (Scheme 2, eq
6
1
). On the other hand, Buonomo and Aldrich reported that
Scheme 2. Results of Catalytic Mitsunobu Reactions in the
Azo Reagent or in the Phosphine Reagent
Figure 1. Results of reactions under the “fully catalytic Mitsunobu”
conditions. Yields are based on isolated products unless otherwise
a
noted. MS 5 Å was activated by heating using heat gun in vacuo.
b
c
MS 5 Å was activated by heating using Bunsen burner in vacuo. The
d
reaction was performed in the absence of the hydrazine catalyst. Yield
was estimated by H NMR analysis of the crude product.
1
yield is somewhat lower than that reported by the authors (68%
yield), we considered that the result was qualitatively
reproduced in the reaction of benzyl alcohol.
Next, we applied the “fully catalytic system” to the model
substrates, (−)-ethyl lactate and 4-nitrobenzoic acid. The
reaction with a chiral alcohol has not been described in the
a
b
Percent yield of isolated products is given. MS 5 Å was activated by
7
c
authors’ publication. Surprisingly, the corresponding ester
heating using heat gun in vacuo. MS 5 Å was activated by heating
using Bunsen burner in vacuo. Three independent reactions were
conducted. (R)-1-Phenylethanol was used instead of (S)-1-phenyl-
ethanol as a starting material in our trial.
d
product was obtained in only 4% isolated yield, and HPLC
analysis showed only a 12% inversion ratio (Figure 1, entry 2,
first run). This result indicates that the Mitsunobu product 1
e
B
DOI: 10.1021/acs.orglett.6b01894
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
was obtained in only 0.5% yield in this system. The same
experiment was repeated twice and almost reproduced the
result (6−7% yield, 9−16% inversion). Incidentally, replacing
Scheme 3. Outline of the Mechanism of the Reaction
between Benzyl Alcohol and 4-Nitrobenzoic Acid under the
“Fully Catalytic” Conditions
1
-phenylphospholane-1-oxide with another cyclic phosphine
oxide (5-phenyldibenzophosphole oxide), described in authors’
7
publication, provided only a trace amount of the product 1
(<5% yield, SI).
Our reaction conditions might partially differ from those of
Buonomo and Aldrich. First, suppliers of the reagents and
solvents used would not be the same in each case. We tested
the reactions using reagents from different suppliers, with no
impact to the outcome (see SI for our case). Second, our
method for the molecular sieves activation differs from that of
the authors’. The authors reported in the Supporting
Information that “molecular sieves were activated by heating
nitrobenzoic acid solely in the presence of molecular sieves, as
well as in the presence of molecular sieves and phenylsilane. No
formation of ester 3 could be observed in any of these cases,
9a
at 200 °C under reduced pressure”, but we have recently
demonstrated that such a method is insufficient for activating
6
molecular sieves in the catalytic Mitsunobu reaction. There-
ruling out S 1 pathways mediated by the benzoic acid
N
fore, we used molecular sieves activated strictly by heating with
a heat gun (ca. 450 °C) or Bunsen burner (>1000 °C) under
reduced pressure (0.1−0.9 mbar), and at least this change
should work advantageously in our experiments. In the present
study, we found out that the impact of molecular sieves was
significantly great in reactions conducted in THF. Activating
molecular sieves by Bunsen burner instead of a heat gun proved
beneficial to our catalytic system in THF (Scheme 2, eq 1).
This is somehow different from our recent observation where
both methods of activation resulted in comparable yields of the
Mitsunobu reaction in toluene and is probably a result of the
hygroscopic nature of THF. On the other hand, the difference
in the molecular sieves activation method affected the
esterification of neither benzyl alcohol (Figure 1, entry 1) nor
derivative or the silane (SI). The reaction between benzyl
alcohol and 4-nitrobenzoic acid in the presence of only Fe(Pc)
and molecular sieves resulted in the formation of benzaldehyde
1
11
(
75% yield by H NMR) and trace amounts of ester 3 (SI).
Similarly, when phenylsilane was added into the reaction
mixture, no ester product was detected (SI).
By replacing benzyl alcohol with 3-phenylpropan-1-ol in the
fully catalytic conditions, the corresponding ester 4 was hardly
produced (<5% yield, Figure 1, entry 3). The above-mentioned
results imply that Fe(Pc)-catalyzed oxidation−reduction
condensation might be ruled out in the formation of ester 3.
A remaining possibility is that an intermediate, generated in situ
from Fe(Pc), the phosphine oxide, and phenylsilane, induces an
SN1 reaction as a Lewis acid. This putative Lewis acid might
slightly include Fischer-type esterification in the reaction of
(−)-ethyl lactate or 3-phenylpropan-1-ol. Incidentally, another
model reaction of benzyl alcohol using Fe(Pc) and
tricyclohexylphosphine provided no ester product, supporting
that phenylsilane might participate in formation of the
intermediate (SI).
(
−)-ethyl lactate (Figure 1, entry 2). Third, the authors used
oxygen generated by sodium hypochlorite (NaClO) and
hydrogen peroxide (H O ). Although we typically employed
2
2
commercially available oxygen filled in a gas cylinder, for
comparison reasons, we repeated the reaction from entry 1 in
Figure 1 by using oxygen generated from NaClO and H O ,
2
2
yet, with no change in the result (SI). Therefore, the above-
mentioned differences in the reaction conditions are not an
essential issue here.
We previously reported that our system catalytic in the azo
reagent did not work well when a trialkylphosphine, such as tri-
n-butylphosphine, was used as a stoichiometric phosphine
6
It could be that the pathway in which ester 3 is produced in
the “fully catalytic system” has been misinterpreted, and it
might come from an inappropriate selection of the model
reagent. Probably, the highly nucleophilic phosphines cause
structural or functional changes in iron phthalocyanine by
strong coordination to the metal center. A green spot having
low polarity was observed on TLC analysis of the mixture of tri-
n-butylphosphine and iron phthalocyanine. It is assumed that
this spot belongs to a complex of iron phthalocyanine with the
phosphine. Under the conditions of the “fully catalytic system”,
a similar spot was observed probably because the cyclic
phosphine that formed in situ also has strong nucleophilicity
(SI). Thus, our system catalytic in the azo reagent might be
incompatible with the system catalytic in the phosphine when
the cyclic phosphine is used.
In conclusion, our preliminary results suggested two
problems for the report by Buonomo and Aldrich. First, the
protocol of the catalytic Mitsnobu reaction in the phosphine
appears insufficiently optimized. Second, the proposed “fully
catalytic system” does not work as the catalytic Mitsunobu
reaction. The former problem might be solved by strict
optimization of the reaction conditions. The latter issue is more
significant because the reaction does not seem to undergo the
Mitsunobu process. Of course, we do not fully rule out the
future feasibility of this ideal concept by further studies. In our
7
alcohol substrates. Namely, these reactions were performed
with benzyl alcohols at elevated temperature (70 °C), which
might induce the formation of a benzyl cation in the presence
of 4-nitrobenzoic acid to provide the ester via the SN1
mechanism (Scheme 3). Indeed, the authors suggested the
possibility of formation of the benzyl cation in the reaction
between 4-trifluoromethylbenzyl alcohol and 4-nitrobenzoic
9b
acid in the presence of phenylsilane. Yet, another pathway is
possible. For instance, typical dehydrated condensation or
1
0
Fe(Pc)-catalyzed oxidation−reduction condensation would
give ester products.
To shed light on this, we conducted control experiments in
the reaction between benzyl alcohol and 4-nitrobenzoic acid. By
performing the reaction in the absence of the hydrazine
catalyst, product 3 was obtained in almost identical yield (36%)
as that of the reaction with the hydrazine catalyst, which
indicates that the hydrazine catalyst does not participate in the
authors’ “fully catalytic system” (Figure 1, entry 1, third run). In
addition, we tested the reaction between benzyl alcohol and 4-
C
DOI: 10.1021/acs.orglett.6b01894
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
opinion, however, it is premature at the moment to support the
claim that “the concept of a fully catalytic Mitsunobu reaction
has finally been realized”.
Kunkel, S. R.; Tellez, J. L.; Doonan, B. J.; Coyle, E. E.; Lavigne, F.;
Kang, L. J.; Przeworski, K. C. Chem. - Eur. J. 2013, 19, 15281−15289.
(c) O’Brien, C. J. PCT Int. Appl. WO2010/118042A2, 2010.
(
S5.
9) See the Supporting Information of ref 7: (a) p S2. (b) pp S3−
ASSOCIATED CONTENT
Supporting Information
■
(10) Taniguchi, T.; Hirose, D.; Ishibashi, H. ACS Catal. 2011, 1,
*
S
1
(
469−1474.
11) Mahyari, M.; Shaabani, A. Appl. Catal., A 2014, 469, 524−531.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b01894.
Full experimental details and analytical data (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*
E-mail: janez.kosmrlj@fkkt.uni-lj.si (J.K.).
E-mail: tsuyoshi@p.kanazawa-u.ac.jp (T.T.).
*
Author Contributions
∥D.H. and M.G. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
D.H. and T.T. are thankful to Profs. Shigeyoshi Kanoh,
Katsuhiro Maeda and Tomoyuki Ikai (Kanazawa University)
for their kind support. This work was supported by MEXT/
JSPS KAKENHI Grant-in-Aid for Scientific Research (C)
(
(
Grant No. 25460011) and Grant-in-Aid for JSPS Fellows
Grant No. 14J02441), and the Ministry of Education, Science
and Sport, Republic of Slovenia, the Slovenian Research Agency
(Grant P1-0230).
REFERENCES
■
(
1) Otera, J.; Nishikido, J. Esterification: Methods, Reactions, and
Applications: Completely Revised and Enlarged, 2nd ed.; Wiley- VCH:
Weinheim, Germany, 2010; Chapter 1, pp 3−157.
2) Representative reviews on Mitsunobu reactions: (a) Mitsunobu,
(
O. Synthesis 1981, 1−28. (b) Hughes, D. L. Org. React. 1992, 42, 335−
56. (c) But, T. Y. S.; Toy, P. H. Chem. - Asian J. 2007, 2, 1340−1355.
d) Swamy, K. C. K.; Kumar, N. N. B.; Balaraman, E.; Kumar, K. V. P.
6
(
P. Chem. Rev. 2009, 109, 2551−2651. (e) Fletcher, S. Org. Chem.
Front. 2015, 2, 739−752.
(
3) For reviews on the modification of Mitsunobu reactions:
(
a) Dembinski, R. Eur. J. Org. Chem. 2004, 2763−2772. (b) Dandapani,
S.; Curran, D. P. Chem. - Eur. J. 2004, 10, 3130−3138. See also refs
2
c−e. For a recent example of modification, see: (c) Figlus, M.;
Wellaway, N.; Cooper, A. W. J.; Sollis, S. L.; Hartley, R. C. ACS Comb.
Sci. 2011, 13, 280−285. (d) Maity, P. K.; Rolfe, A.; Samarakoon, T. B.;
̈
Faisal, S.; Kurtz, R. D.; Long, T. R.; Schatz, A.; Flynn, D. L.; Grass, R.
N.; Stark, W. J.; Reiser, O.; Hanson, P. R. Org. Lett. 2011, 13, 8−10.
e) Vanos, C. M.; Lambert, T. H. Angew. Chem., Int. Ed. 2011, 50,
(
1
2222−12226. (f) Nacsa, E. D.; Lambert, T. H. Org. Lett. 2013, 15,
3
8−41. (g) Tang, X.; Chapman, C.; Whiting, M.; Denton, R. Chem.
Commun. 2014, 50, 7340−7343.
(
4) (a) But, T. Y. B.; Toy, P. H. J. Am. Chem. Soc. 2006, 128, 9636−
637. (b) But, T. Y. S.; Lu, J.; Toy, P. H. Synlett 2010, 1115−1117.
5) Hirose, D.; Taniguchi, T.; Ishibashi, H. Angew. Chem., Int. Ed.
013, 52, 4613−4617.
6) Hirose, D.; Gazvoda, M.; Kosm
016, 7, 5148−5159.
7) Buonomo, J. A.; Aldrich, C. C. Angew. Chem., Int. Ed. 2015, 54,
3041−13044.
8) (a) O’Brien, C. J.; Tellez, J. L.; Nixon, Z. S.; Kang, L. J.; Carter, A.
L.; Kunkel, S. R.; Przeworski, K. C.; Chass, G. A. Angew. Chem., Int. Ed.
009, 48, 6836−6839. (b) O’Brien, C. J.; Nixon, Z. S.; Holohan, A. J.;
9
(
2
(
2
(
1
(
̌
rlj, J.; Taniguchi, T. Chem. Sci.
2
D
DOI: 10.1021/acs.orglett.6b01894
Org. Lett. XXXX, XXX, XXX−XXX