Selectivity of N-versus O-alkylation in Mitsunobu reactions with various quinolinols and isoquinolinols

Article abstract of DOI:10.3987/COM-17-13710

Reacting quinolinols and isoquinolinols under Mitsunobu conditions can give rise to N-alkylated products in addition to the normally desired O-alkylated structures. An in-depth study of how the solvent, reagent equivalents, position of the quinoline/isoqu

Full text of DOI:10.3987/COM-17-13710

HETEROCYCLES, Vol. 94, No. 7, 2017
1305
HETEROCYCLES, Vol. 94, No. 7, 2017, pp. 1305 - 1313. © 2017 The Japan Institute of Heterocyclic Chemistry
Received, 5th April, 2017, Accepted, 29th May, 2017, Published online, 14th June, 2017
DOI: 10.3987/COM-17-13710
SELECTIVITY OF N- VERSUS O-ALKYLATION IN MITSUNOBU
REACTIONS WITH VARIOUS QUINOLINOLS AND ISOQUINOLINOLS
Ryan E. Hartung,* Mark C. Wall, Sylvain Lebreton, Martin Smrcina, and
Marcel Patek
Icagen, 2090 E. Innovation Park Drive, Oro Valley, AZ 85755, USA.
rhartung@icagen.com
Abstract – Reacting quinolinols and isoquinolinols under Mitsunobu conditions
can give rise to N-alkylated products in addition to the normally desired
O-alkylated structures. An in-depth study of how the solvent, reagent equivalents,
position of the quinoline/isoquinoline nitrogen and type of reacting aliphatic
alcohol employed affect the ratio of N- versus O-alkylation is described.
The Mitsunobu reaction has been employed within the chemistry community for many years on many
scaffolds and has been the subject of multiple reviews.1,2 One such scaffold of interest is 2-pyridones
(Scheme 1), which are similar to the quinoline and isoquinoline substructure.3 However, an in-depth study
of N- versus O-alkylation of quinolinols and isoquinolinols under Mitsunobu conditions has not been
published. Since the Mitsunobu reaction is primarily utilized for O-alkylation, it would be beneficial to
understand how altering the reaction conditions can affect the N- or O-alkylation outcome.
As with the tautomerism observed in 2-pyridones,4 quinolinones and isoquinolinones (or quinolinols and
isoquinolinols as they will be referred to throughout the paper for clarity) contain the same tautomeric traits
as their simplified precursors as shown in Scheme 1.5 The tautomeric equilibria of 2-hydroxypyridine was
shown to depend on solvent polarity, with a pronounced shift to a greater population of the more polar
lactam in more polar media.6 It is this tautomeric equilibria that allows systems of this nature to react as
ambident nucleophiles through the conjugate base resonance structures, often leading to mixtures of O- and
N-alkylation. However, like with 2-pyridones, we observed that many factors can influence the alkylation
ratio. These include, but are not limited to, solvation of the nucleophile (conjugate base of the heterocycle),
the electrophilicity of the activated alcohol and the structure of the starting quinolinol or isoquinolinol.

1306
HETEROCYCLES, Vol. 94, No. 7, 2017
O
OH
HN
N
2-pyridone
2-hydroxypyridine
(lactam tautomer)
O
OH
N
HN
isoquinolinone
isoquinolinol
(lactam tautomer)
H
N
HO
N
O
quinolinone
quinolinol
(lactam tautomer)
Scheme 1. (Iso)quinolinone tautomerism
In the reaction mixture, assuming tautomerization is allowed, the DEAD-PPh3-adduct will deprotonate the
phenol to the phenoxide and the conjugate base of the phenol will resonate. With oxygen having an
electronegativity of 3.44 and nitrogen 3.04 on the Linus Pauling scale, one would expect the oxygen to bear
more negative charge and the O-alkylation would dominate.
With the above in mind, our initial interest in this area resulted from employing the quinolinol and
isoquinolinol substructures in a medicinal chemistry program related to a rare disease target. Using starting
materials containing these substructures, the Mitsunobu reaction generated mixtures of the desired
O-alkylation products with the unexpected N-akylated adducts. We were thus prompted to explore the
factors that influence the regioselectivity of this process in detail. This methodological query is important
as there are many natural products that contain quinolinone and isoquinolinone motiefs.7 Due to the lack of
experimental data on these relatively simple aromatic systems, we wish to present our observations around
the selectivity of O- versus N-alkylation under Mitsunobu conditions.
The first step in dissecting N- versus O-alkylation was to select a set of heterocycles where tautomerism of
the conjugate base in the Mitsunobu reaction is allowed and not allowed. Based on the nitrogen’s position
within the ring, relative to the hydroxyl group, tautomeric conjugate base forms such as a and b can exist
(Scheme 2). However, as in c, where the nitrogen and hydroxyl groups are not in a 1-2 or 1-4 orientation,
tautomerism is not allowed and only O-alkylation can occur through the phenol’s conjugate base.

HETEROCYCLES, Vol. 94, No. 7, 2017
1307
O
O
N
N
a
c
b
OH
Structure C
tautomerism
is not allowed
N
Scheme 2. Example of allowed and disallowed tautomers
In Table 1, a series of different quinolinols and isoquinolinols were subjected to Mitsunobu conditions (see
Representative Procedure). Although the 1H NMR of the O- and N-alkylated products are quite similar, the
13C NMR spectra in DMSO-d6 show dramatic shifts at the carbon attached to the resultant phenolic ether or
carbonyl, depending on O- or N-alkylation allowing for easy identification of both products.
To begin, entries 3, 5, and 6 are heterocycles that are unable to tautomerize, which leads to only one
conjugate base occurring at the phenol leading to only O-alkylated products. For the rest of the entries the
positional requirements necessary for tautomerism are met and both the N- and O-alkylated products can be
expected. Entries 1 and 2, where the nitrogen is located at ring position 4 and 5 respectively; about a 2:1
ratio of O- versus N-alkylation was observed. However, when the phenolic carbon is also next to the
nitrogen atom, the trend changes (entry 4). 2-pyridones and 2-quinolinols are known to exist primarily in
the imidic tautomeric form and thus the conjugate base of the nitrogen is the dominate nucleophile.8 In this
system, the nitrogen’s lone pair is more available to react and a shift toward 85% N-alkylation is observed.
Table 1. O- versus N-Alkylation ratios with various quinoline and isoquinolinols, and quinazolinol
Entry
1
Structure
%O-Alkylationa
%N-Alkylationa
Yieldc
OH
70
30
91
N
OH
2
3
63
27
0
87
90
N
OH
100
N

1308
HETEROCYCLES, Vol. 94, No. 7, 2017
OH
OH
OH
N
4
5
15
100
100
0
85
0
85
90
N
6
0
93
N
OH
N
7b
8
100
0
100
88
N
OH
100
100
100
81
CF3
N
OH
9
0
96
N
OH
10
11
0
95
Cl
Br
N
OH
19
93
N
c
aRatio was determined by 1H NMR. b100% Alkylation occurred on N-2. Combined isolated yield.
In entry 7, the carbonyl tautomer (lactam) is present almost exclusively due to the electronic effect of the
nitrogen at position 4, rendering both the oxygen and position 4 nitrogen unreactive, giving exclusive
N-alkylation at the 2 position. Entries 8 through 11 show the effect of substituents alpha to the potential
reacting nitrogen or oxygen. In compounds 8-10, regardless of tautomerism being allowed and solvation
effects playing a role, the steric bulk of substituents dictate where alkylation will occur. This is readily
apparent in compound 11, where the R group is moved one position away from the nitrogen and results in a
shift dramatic move toward more N-alkylation. Theoretical orbital energies/electron densities for many of
the unsubstituted scaffolds have been previously calculated and well documented.9 However, when moving
to substituted analogs, such as entries 8-11, calculations at a higher level of theory are required to fully
account for solvent and steric effects. The latter component is largely responsible for driving the O- versus

HETEROCYCLES, Vol. 94, No. 7, 2017
1309
N-alkylation ratio in more decorated scaffolds and thus often makes these more complicated calculations
less predictive of the true experimental outcome.
After the investigation of different spatial arrangements of the quinolinol/isoquinolinol nitrogen in relation
to the phenol, we shifted toward how different reacting alcohols might alter O- versus N-alkylation, while
keeping the remaining Mitsunobu reaction conditions constant. It is known that not just sterics surrounding
the conjugate base can influence the Mitsunobu reaction, but the phosphonium salt intermediates can also
influence the system’s reactivity.
Table 2. O- versus N-Alkylation ratios with various alcohols
R
OH
O
O
R
Mitsunobu
N
N
N
versus
Entry
12
Alcohol
%O-Alkylationa
%N-Alkylationa
Yieldb
96
Si
4
96
HO
HO
13
14
15
16
17
18
33
33
20
---
7
67
67
72
84
80
90
93
96
O
HO
HO
80
HO
>99
93
N
HO
HO
O
O
N
14
86
O

1310
HETEROCYCLES, Vol. 94, No. 7, 2017
F
HO
19
20
21
21
---
15
79
>99
85
89
97
85
F
HO
HO
aRatio was determined by 1H NMR. bCombined isolated yield.
In Table 2, a set of ten different alcohols was employed in the Mitsunobu reaction, using isoquinolone 4 as
our substrate. The alcohols were chosen based on their polarities, steric hindrance to nucleophilic attack
once activated and electronic properties. In all cases, N-alkylation was the prevalent outcome, which came
as no surprise. Alcohols 13, 14, and 15 all led to the reaction producing a larger percentage of O-alkylated
product relative to the others. If one assumes that the bulkiness of the activated alcohol does play a role in
the transition state, then a greater percentage of N-alkylation with the cyclopropyl alcohol 15 can be
expected. While the nitrogen is imbedded in the heterocyclic ring system, the phenol is more accessible.
Thus sterically hindered alcohols such as tert-butyl 13 and phenyl ether 14 should react more at the exposed
phenolic portion of the isoquinolone. However, the more electron dense nitrogen is still the better
nucleophile.
However, this reasoning does not hold true with tri-methylsilyl (TMS)-ethanol 12. The combination of the
elongated ethyl spacer along with the electron rich TMS moiety could however change this reacting
alcohol’s environment enough to warrant almost exclusive N-alkylation. One should note as important as
the nucleophile is (the conjugate base) the nature of the electrophile also becomes important when sterics
are involved. Another possibility is that due to the oxophilic nature of the silicon, a mechanism could be
envisioned where the phenolic oxygen and the silicon from a weak bond interaction giving the nitrogen a
tethered electrophile.
The last interesting observation to note is that the basic (pyridin-3-yl)-methanol 16 gave exclusive
N-alkylation. The basicity and polarity of the pyridine clearly plays a role in the alkylation process, leading
to the observed outcome.
Our last look into the Mitsunobu reaction was with the quinolinol 1 ring system and varying the solvent and
equivalents as shown in Table 3. As previously mentioned, different solvents provide different degrees of
solvation to the conjugate bases leading to changes in the relative nucleophilicity of the nitrogen versus

HETEROCYCLES, Vol. 94, No. 7, 2017
1311
oxygen atoms.10 Benzyl alcohol 21 with THF, which is also described in Representative Procedure, was the
standard set of reaction conditions used in both Tables 1 and 3.
Table 3. O- versus N-Alkylation ratios with different solvents dielectrics and reactant equilibria
R
OH
O
O
Mitsunobu
R = Bn
versus
N
N
N
R
Entry
22
Conditions
%O-Alkylationa
%N-Alkylationa
Yieldd
78
()c
Et2O, 2 eq (PPh3,
DEAD, alcohol)b
85
15
4.3
THF, 2 eq (PPh3,
DEAD, alcohol)b
23
24
25
26
27
70
40
30
60
30
25
---
7.6
37
84
96
95
88
94
DMF, 2 eq (PPh3,
DEAD, alcohol)b
CH2Cl2, 2 eq (PPh3,
DEAD, alcohol)b
70
8.9
7.6
THF, 4 eq (PPh3,
DEAD, alcohol)b
75
THF, 2 eq (PPh3,
DEAD) 8 eq alcohol
100
7.6
c
aRatio was determined by 1H NMR. bDEAD was 40% mixture in toluene ( 2.4). Dielectric constant
() Reflecting the solvent’s ability stabilize charges. dCombined isolated yield.
Using the aforementioned Experimental, we performed the Mitsunobu reaction in solvents of different
polarity, such as Et2O, DMF and CH2Cl2. While CH2Cl2 gave the same ratio of products as THF, the effects
of Et2O favored O-alkylation even more than its precursors. Interestingly, DMF, which is the most polar of
the four solvents (entry 24), reversed the alkylation ratio in favor of N-alkylation. This was expected, as the
higher dielectric solvent will allow the development and stabilization of charges, leading to a
thermodynamically controlled outcome. This table of solvents demonstrated the strong effect a solvent’s
solvation can play on the reacting nucleophile, as more polar solvents tend to lead to a greater ratio of the
N-alkylated product when the electrophile is kept constant.

1312
HETEROCYCLES, Vol. 94, No. 7, 2017
Lastly, while keeping THF as the solvent, we increased the amount of the alcohol, PPh3 and DEAD to four
equivalents instead of two (entry 26). Although not necessarily significant, the amount of O-alkylated
product did increase, leading us to wonder what an even higher equivalent increase in reactants would do.
In the last reaction (entry 27), we increased to eight equivalents of the alcohol, PPh3 and DEAD. In this case,
only the O-alkylated product was observed. By swamping the reaction with available electrophile, the more
electronegative phenolic oxygen wins out. With an overabundance of electrophile, kinetics now determine
the reaction outcome.
In conclusion, we have demonstrated that the location of the ring nitrogen relative to the phenolic oxygen
can have a dramatic effect on O- versus N-alkylation. It tautomerism is not allowed or sterics block the
nitrogen atom, then only O-alkylation is allowed. Also, the solvation of the conjugate base, as well as the
reacting electrophile, can likewise have a large influence on the final ratio of O- versus N-alkylation in the
quinolinol and isoquinolinol ring systems, which until now have been unexplored. By varying the solvent,
reacting heterocycle and electrophilic alcohol, one should be able to shift the reaction toward whichever O-
or N-alkylated product is desired. With the Mitsunobu reaction’s mild conditions and tolerability of a large
variety of functional groups, when compared to other similar reactions such as the Williamson alkylation,12
which have harsher conditions and lower amount of allowable functional groups, the Mitsunobu reaction
continues to be widely used in many modern synthetic schemes due to its ease and versatility.
EXPERIMENTAL
PPh3 (89 mg, 0.34 mmole, 2 eq) and the quinolinol or isoquinolinol (25 mg, 0.17 mmole, 1 eq) were
combined in a glass vial and purged with nitrogen. THF (700 µL) was then added followed by benzyl
alcohol (44 µL, 0.34 mmole, 2 eq). A 40% by weight solution of DEAD in toluene (170 µL, 0.34 mmole, 1
o
eq) was then added dropwise to keep the reaction temperature below 30 C. The reaction mixture was
shaken at room temperature overnight and then purified on a Waters preparative LC/MS system with a
gradient of 0 to 60% MeCN-H2O to give the desired product with yields ranging from 50% to 98%. In
instances where the isomers were not able to be separated, the percentage ratio was determined by 1H NMR.
Purified products were characterized by 1H and 13C NMR.
ACKNOWLEDGMENTS
The authors gratefully acknowledge Sanofi for their support and the resources necessary for the majority of
this work, as well as Dr. Zhongli Gao, Dr. John Hofferberth, Dr. David Hilmey and Dr. Joseph Kim in their
assistance with this manuscript.

HETEROCYCLES, Vol. 94, No. 7, 2017
1313
REFERENCES AND NOTES
1. O. Mitsunobu and M. Yamada, Bull. Chem. Soc. Jpn., 1967, 40, 2380.
2. (a) O. Mitsunobu, Synthesis, 1981, 1, 1; (b) D. L. Hughes, Org. Reactions, 1992, 42, 335; (c) D. L.
Hughes, Org. Prep. Proced. Int., 1996, 28, 12; (d) K. C. Swamy, N. N. Kumar, E. Balaraman, and K.
V. Kumar, Chem. Rev., 2009, 109, 2551; (e) D. P. Christen and J. J. Li, Name Reactions for
Homologations, 2009, 671.
3. (a) M. C. Torhan, N. P. Peet, and J. D. Williams, Tetrahedron Lett., 2013, 54, 3926; (b) D. L. Comins
and G. Jianhua, Tetrahedron Lett., 1994, 35, 2819.
4. (a) P. Yang and W. Duan, Youji Huaxue, 1991, 11, 620; (b) G. Galstyan and E.-W. Knapp, J. Phys.
Chem., 2012, 116, 6885.
5. (a) D. Liu, L. P. Sanow, and C. Zhang, Tetrahedron Lett., 2014, 55, 3090; (b) S. Fletcher, Org. Chem.
Front., 2015, 2, 739; (c) K. C. K. Swamy, N. N. B. Kumar, E. Balaraman, and K. V. P. P. Kumar,
Chem. Rev., 2009, 109, 2551.
6. M. W. Wong, K. B. Wiberg, and M. J. Frisch, J. Am. Chem. Soc., 1992, 114, 1645.
7. (a) J. He, U. Lion, I. Sattler, F. A. Gollmick, S. Grabley, J. Cai, M. Meiners, H. Shunke, K. Schaumann,
U. Dechert, and M. Krohn, J. Nat. Prod., 2005, 68, 1397; (b) M. Arisawa, A. Nishida, and M.
Nakagawa, J. Organomet. Chem., 2006, 691, 5109; (c) T. Nakatsu, T. Johns, I. Kubo, K. Milton, M.
Sakai, K. Chatani, K. Saito, Y. Yamagiwa, and T. Kamikawa, J. Nat. Prod., 1990, 53, 1508.
8. (a) D. Camp, M. von Itzstein, and I. D. Jenkins, Tetrahedron, 2015, 71, 4946.
9. (a) M. W. Wong, K. B. Wiberg, and M. J. Frisch, J. Am. Chem. Soc., 1992, 114, 1645; (b) G. Galstyn
and E. W. Knapp, J. Phys. Chem., 2012, 116, 6885.
10. O. R. Ludek and C. Meier, Synlett, 2006, 324.
11. C. Ahn and P. DeShong, J. Org. Chem., 2002, 67, 1754.
12. (a) A. Williamson, Ann. Chem. Pharm., 1851, 77, 37; (b) A. W. Williamson, J. Chem. Soc., 1852, 4,
229.