Structural and Rate Studies of the 1,2-Additions of Lithium Phenylacetylide to Lithiated Quinazolinones: Influence of Mixed Aggregates on the Reaction Mechanism

Article abstract of DOI:10.1021/ja0305813

The 1,2-addition of lithium phenylacetylide (PhCCLi) to quinazolinones was investigated using a combination of structural and rate studies. ⁶Li, ¹³C, and ¹⁹F NMR spectroscopies show that deprotonation of quinazolinones and

Full text of DOI:10.1021/ja0305813

Published on Web 04/07/2004
Structural and Rate Studies of the 1,2-Additions of Lithium
Phenylacetylide to Lithiated Quinazolinones: Influence of
Mixed Aggregates on the Reaction Mechanism
Timothy F. Briggs,† Mark D. Winemiller,† David B. Collum,*,†
Rodney L. Parsons, Jr.,‡ Akin H. Davulcu,‡ Gregory D. Harris,‡
Joseph M. Fortunak,‡ and Pat N. Confalone‡
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853-1301, and Bristol Myers-Squibb Company,
Process Research and DeVelopment, One Squibb DriVe, New Brunswick, New Jersey 08903
Received October 10, 2003; E-mail: dbc6@cornell.edu
Abstract: The 1,2-addition of lithium phenylacetylide (PhCCLi) to quinazolinones was investigated using
a combination of structural and rate studies. 6Li, 13C, and 19F NMR spectroscopies show that deprotonation
of quinazolinones and phenylacetylene in THF/pentane solutions with lithium hexamethyldisilazide affords
a mixture of lithium quinazolinide/PhCCLi mixed dimer and mixed tetramer along with PhCCLi dimer.
Although the mixed tetramer dominates at high mixed aggregate concentrations and low temperatures
used for the structural studies, the mixed dimer is the dominant form at the low total mixed aggregate
concentrations, high THF concentrations, and ambient temperatures used to investigate the 1,2-addition.
Monitoring the reaction rates using 19F NMR spectroscopy revealed a first-order dependence on mixed
dimer, a zeroth-order dependence on THF, and a half-order dependence on the PhCCLi concentration.
The rate law is consistent with the addition of a disolvated PhCCLi monomer to the mixed dimer. Investigation
of the 1,2-addition of PhCCLi to an O-protected quinazolinone implicates reaction via trisolvated PhCCLi
monomers.
Introduction
Several new classes of potent nonnucleoside reverse tran-
scriptase inhibitors have been developed by DuPont Pharma-
ceuticals and Merck Research Laboratories.1 Efavirenz (1) is
now widely prescribed under the names Sustiva and Stocrin
for the treatment of AIDS and symptomatic HIV-1 infection.2-4
Second-generation drug candidates 2, 3, and 4 possess signifi-
cant activity against wild-type HIV and mutant strains resistant
to currently approved drug regimens,1,5,6 propelling them to
advanced clinical trials.5
by the 1,2-addition illustrated in eq 1. This remarkable reaction
presents a challenging problem in structural and mechanistic
organolithium chemistry. The five lithium saltss6, 7, 8, 9, and
lithium hexamethyldisilazide (LiHMDS)scould aggregate in
countless proportions, connectivities, and stereochemistries that
would likely depend markedly on solvent, temperature, con-
centration, and stoichiometries.7 Previous structural studies have
shown that the underlying aggregation effects are indeed very
complex.8-11
Reverse transcriptase inhibitors 2-4 are prepared on a large
scale, up to 5000 kg, in greater than 98% enantiomeric excess
† Cornell University.
‡ Bristol Myers-Squibb Co.
(1) For a review of the syntheses of nonnucleoside reverse transcriptase
inhibitors, see: Parsons, R. L., Jr. Curr. Opin. Drug DiscoVery DeV. 2000,
3, 783.
(2) Young, S. D.; Britcher, S. F.; Tran, L. O.; Payne, L. S.; Lumma, W. C.;
Lyle, T. A.; Huff, J. R.; Anderson, P. S.; Olsen, D. B.; Carroll, S. S.;
Pettibone, D. J.; O’Brien, J. A.; Ball, R. G.; Balani, S. K.; Lin, J. H.; Chen,
I.-W.; Schleif, W. A.; Sardana, V. V.; Long, W. J.; Byrnes, V. W.; Emini,
E. A. Antimicrob. Agents Chemother. 1995, 39, 2602.
As a key step toward understanding the enantioselective 1,2-
addition, we have investigated the nonenantioselective variant
(3) Pierce, M. E.; Parsons, R. L., Jr.; Radesca, L. A.; Lo, Y. S.; Silverman, S.;
Moore, J. R.; Islam, Q.; Choudhury, A.; Fortunak, J. M.; Nguyen, D.; Luo,
C.; Morgan, S. J.; Davis, W. P.; Confalone, P. N.; Chen, C. Y.; Tillyer, R.
D.; Frey, L.; Tan, L.; Xu, F.; Zhao, D. L.; Thompson, A. S.; Corley, E. G.;
Grabowski, E. J. J.; Reamer, R.; Reider, P. J. J. Org. Chem. 1998, 63,
8536. Thompson, A. S.; Corley, E. G.; Huntington, M. F.; Grabowski, E.
J. J. Tetrahedron Lett. 1995, 36, 8937.
(4) Efavirenz is currently being prepared by an amino alcohol-mediated addition
of alkynyl zinc derivatives. Tan, L.; Chen, C. Y.; Tillyer, R. D.; Grabowski,
E. J. J.; Reider, P. J. Angew. Chem., Int. Ed. 1999, 38, 711. Chen, C. Y.;
Tan, L. Enantiomer 1999, 4, 599.
(5) Tucker, T. J.; Lyle, T. A.; Wiscount, C. M.; Britcher, S. F.; Young, S. D.;
Sanders, W. M.; Lumma, W. C.; Goldman, M. E.; O’Brien, J. A.; Ball, R.
G.; Homnick, C. F.; Schleif, W. A.; Emini, E. A.; Huff, J. R.; Anderson,
P. S. J. Med. Chem. 1994, 37, 2437. Corbett, J. W.; Ko, S. S.; Rodgers, J.
D.; Gearhart, L. A.; Magnus, N. A.; Bacheler, L. T.; Diamond, S.; Jeffrey,
S.; Klabe, R. M.; Cordova, B. C.; Garber, S.; Logue, K.; Trainor, G. L.;
Anderson, P. S.; Erickson-Viitanen, S. K. Antimicrob. Agents Chemother.
1999, 43, 2893.
(6) Kauffman, G. S.; Harris, G. D.; Dorow, R. L.; Stone, B. R. P.; Parsons, R.
L., Jr.; Pesti, J. A.; Magnus, N. A.; Fortunak, J. M.; Confalone, P. N.;
Nugent, W. A. Org. Lett. 2000, 2, 3119.
9
10.1021/ja0305813 CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 5427-5435
5427

A R T I C L E S
Briggs et al.
Table 1. 6Li, 13C, and 19F NMR Spectroscopic Dataa
b
cmpd
δ 6Li (m, JLi-C; m, JLi-N1
)
δ
13C (m, JLi-C
)
δ
19F
8
2.25 (-; d, 2.5 Hz)
1.04 (-; s)
-64.7, -65.2
0.80 (-; s)
c
10
-
-80.6 to -80.3
(br mound)
13 0.59 (d, 8.5 Hz; s)
14 0.48 (t, 8.4 Hz; -)
139.07 (qn, 8.6 Hz) -64.8
139.35 (qn, 8.4 Hz)
-
16 1.75 (d, 9.0 Hz; d, 3.6 Hz) 137.54 (qn, 8.5 Hz) -65.1
0.77 (d, 8.1 Hz; s)
a The JC-Li coupling constants were measured from the 1J(6Li,13C)-
resolved spectrum. The multiplicities are denoted as follows: s ) singlet,
d ) doublet, t ) triplet, qn ) quintet. The 6Li, 13C, and 19F chemical shifts
6
6
are reported relative to 0.3 M LiCl/MeOH at -90 °C (δ Li ) 0.0 ppm),
hexafluorobenzene at -110 °C (δ 19F ) -162.55 ppm), and the CH2O
carbon of THF at -115 °C (δ 13C ) 67.97 ppm), respectively. All J values
are reported in hertz. b The indicated 6Li-15N coupling is only observed
using [1-15N]5. c Complex envelope of peaks between 0 and 2 ppm.
in eq 2. Several simplifications in the protocol have provided
sufficient control over aggregate structures to allow detailed rate
studies. We preface the mechanistic study by foreshadowing
results from the structural studies: Treating THF solutions of
quinazolinone 5 and phenylacetylene with LiHMDS under
judiciously chosen conditions affords mixed dimer 13 and
lithium phenylacetylide (PhCCLi) dimer 14 as the predominant
species. The most central and pressing question is what role
does mixed dimer 13 play in the 1,2-addition? At the outset,
we considered two plausible answers: (1) mixed dimer 13 reacts
via an intramolecular mechanism that exploits the reagent
proximities afforded by mixed aggregation, or (2) mixed dimer
13 is simply a protected form of quinazolinone 5 that undergoes
structures sufficiently to carry out detailed mechanistic studies.
The results from NMR spectroscopic studies are summarized
in Table 1 and are described below. Selected spectra are
illustrated in Figures 1 and 2.
LiHMDS is readily purified through recrystallization,13 serves
as an excellent proton scavenger,9 and does not form mixed
aggregates except under unusual conditions.9,14,15 Indeed, a
mixture of PhCCLi and lithium salt 8 generated in situ with
excess [6Li]LiHMDS13 contains no LiHMDS-derived mixed
aggregates.
PhCCLi proves especially important to the success of the
solution kinetics. Lithium cyclopropylacetylide (6) and other
lithium acetylides typically exist as mixtures of cyclic dimers
and cubic tetramers in THF and THF/hydrocarbon solu-
tions.8,9,16-18 Lithium phenylacetylide in THF/toluene shows a
greater tendency to form dimer 14 than do most lithium
acetylides, but it affords tetramer 15 at low THF concentra-
tions.18 Such a dimer-tetramer mixture would undermine the
rate studies. In contrast, 6Li and 13C NMR spectra recorded on
THF/pentane mixtures of [6Li,13C]PhCCLi reveal exclusively
dimer 14 (Table 1) in solutions containing as little as 20% THF
by volume. The pronounced hydrocarbon effect can be witnessed
by incrementally replacing the toluene in 20% THF/toluene
solutions of PhCCLi with pentane, causing conversion of a 2:1
15:14 mixture to exclusively dimer 14. The origins of this odd
hydrocarbon dependence are being investigated.18 As a practical
matter, THF/pentane mixtures offer important structural control
for the rate studies.
1,2-addition by an external nucleophilic PhCCLi fragment.12
A
comparison with rate studies of the 1,2-addition to O-protected
quinazoline derivative 11 (eq 2) reveals notable similarities and
differences.
Results
Structural Studies. A combination of forethought and
serendipity allowed us to control the organolithium solution
(7) For leading references to structural studies of RLi/R′OLi mixed aggregates,
see: Thomas, R. D.; Huang, H. J. Am. Chem. Soc. 1999, 121, 1, 11239.
Lochmann, L. Eur. J. Inorg. Chem. 2000, 1115. Sorger, K.; Schleyer, P.
v. R.; Fleischer, R.; Stalke, D. J. Am. Chem. Soc. 1996, 118, 6924. Jackman,
L. M.; Rakiewicz, E. F. J. Am. Chem. Soc. 1991, 113, 1202. Kremer, T.;
Harder, S.; Junge, M.; Schleyer, P. v. R. Organometallics 1996, 15, 585.
(8) (a) Thompson, A.; Corley, E. G.; Huntington, M. F.; Grabowski, E. J. J.;
Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1998, 120, 2028. (b) Xu,
F.; Reamer, R. A.; Tillyer, R.; Cummins, J. M.; Grabowski, E. J. J.; Reider,
P. J.; Collum, D. B.; Huffman, J. C. J. Am. Chem. Soc. 2000, 122, 11212.
(9) (a) Parsons, R. L., Jr.; Fortunak, J. M.; Dorow, R. L.; Harris, G. D.;
Kauffman, G. S.; Nugent, W. A.; Winemiller, M. D.; Briggs, T. F.; Xiang,
B.; Collum, D. B. J. Am. Chem. Soc. 2001, 123, 9135. (b) Briggs, T. F.;
Winemiller, M. D.; Xiang, B.; Collum, D. B. J. Org. Chem. 2001, 66,
6291. (c) Sun, X.; Winemiller, M. D.; Xiang, B.; Collum, D. B. J. Am.
Chem. Soc. 2001, 123, 8039.
In situ generation of quinazolinone salt 8 in THF or THF/
pentane affords a complex mixture that is characteristic of
N-lithiated carboxamides in the absence of added lithium
salts.11,19,20 In the presence of excess [6Li,13C]PhCCLi, however,
we observe two species displaying spectral data consistent with
(10) The asymmetric additions are based on work by Jackman and co-workers:
Ye, M.; Logaraj, S.; Jackman, L. M.; Hillegass, K.; Hirsh, K. A.; Bollinger,
A. M.; Grosz, A. L.; Mani, V. Tetrahedron 1994, 50, 6109.
(11) Briggs, T. F.; Winemiller, M. D.; Collum, D. B., unpublished.
9
5428 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

1,2-Addition of PhCCLi to Quinazolinones
A R T I C L E S
Figure 2. J(6Li,13C)-resolved spectrum (A) and 6Li,13C-HMQC spectrum
(B) recorded on 0.133 N [6Li,13C]PhCCLi and 0.067 N [6Li]8 in 7.6 M
THF (pentane cosolvent) at -115 °C. [6Li]LiHMDS is indicated by an
asterisk (*).
multiplets in the J(6Li,13C)-resolved9 and 6Li,13C-heteronuclear
multiple quantum coherence (HMQC)21,22 spectra (Figure 2).23
A
Figure 1. 6Li spectra were recorded at -115 °C. Spectra A and B were
recorded on a 2:1 mixture of [6Li,13C]PhCCLi (0.133 N) and [6Li]8 (0.067
N) in 7.6 M THF/pentane; spectrum C was recorded on a 2:1 mixture of
[6Li]PhCCLi (0.20 N) and [6Li,1-15N]8 (0.10 N) in 9.5 M THF/pentane.
(Residual [6Li]LiHMDS is marked by *.) Spectrum A is 13C broad-band
decoupled.
13C quintet indicates coupling to two symmetry-equivalent
lithium nuclei. The lack of chelation by the quinazoline is
confirmed by the absence of coupling when dimer 13 is
generated from 15N-enriched quinazolinones [1-15N]5 and
[3-15N]5. A single 19F resonance attests to the symmetry of
mixed dimer 13 and provides a means to monitor reaction rates
(vide infra).
mixed dimer 13 and mixed tetramer 16 (Figure 1; Table 1). A
5-fold excess of PhCCLi is required to convert quinazolinone
salt 8 completely to the mixed aggregates.
Mixed tetramer 16 manifests two lithium resonances in a 1:1
ratio (Figure 1A). Both 6Li resonances are coupled to a common
6
Mixed dimer 13 displays a Li doublet due to coupling to a
single 13C nucleus that is readily discerned from the other
(18) Breslin, S. R.; Collum, D. B., unpublished.
(12) Carboxamides are often protected as their N-lithiated forms. For several
representative examples, see: Kelly, T. R.; Lebedev, R. L. J. Org. Chem.
2002, 67, 2197. Basu, A.; Thayumanavan, S. Angew. Chem., Int. Ed. 2002,
41, 716.
(19) Gallagher, D. J.; Du, H.; Long, S. A.; Beak, P. J. Am. Chem. Soc. 1996,
118, 11391. Maetzke, T.; Hidber, C. P.; Seebach, D. J. Am. Chem. Soc.
1990, 112, 8248. Maetzke, T.; Seebach, D. Organometallics 1990, 9, 3032.
Chivers, T.; Downard, A.; Yap, G. P. A. Inorg. Chem. 1998, 37, 5708.
Liddle, S. T.; Clegg, W. Chem. Commun. 2001, 17, 1584. Liddle, S. T.;
Clegg, W. J. Chem. Soc., Dalton Trans. 2002, 21, 3293. Boss, S. R.; Haigh,
R.; Linton, D. J.; Schooler, P.; Shield, G. P.; Wheatley, A. E. H. J. Chem.
Soc., Dalton Trans. 2003, 5, 1001.
(13) Romesberg, F. E.; Bernstein, M. P.; Gilchrist, J. H.; Harrison, A. T.; Fuller,
D. J.; Collum, D. B. J. Am. Chem. Soc. 1993, 115, 3475.
(14) (a) Lucht, B. L.; Collum, D. B. Acc. Chem. Res. 1999, 32, 1035. (b)
Romesberg, F. E.; Collum, D. B. J. Am. Chem. Soc. 1994, 116, 9198.
(15) LiHMDS forms mixed aggregates with (a) hindered lithium enolates in
THF solution: Kim, Y.-J.; Streitwieser, A. Org. Lett. 2002, 4, 573. (b) In
the presence of poorly coordinating solvents: Zhao, P.; Collum, D. B. J.
Am. Chem. Soc. 2003, 125, in press. (c) Ester enolates bearing a â-amino
group.
(20) N-Lithiated carboxamides form mixed aggregates with lithium acetylides
as well as lithium alkoxides (such as 7): Winemiller, M. D.; Collum, D.
B., unpublished.
(21) Xiang, B.; Winemiller, M. D.; Briggs, T. F.; Fuller, D. J.; Collum, D. B.
Magn. Reson. Chem. 2001, 39, 137.
(16) Lithium acetylide solution structural studies: (a) Fraenkel, G.; Stier, M.
Prepr. Pap.-Am. Chem. Soc., DiV. Fuel Chem. 1985, 30, 586. (b) Fraenkel,
G.; Pramanik, P. J. Chem. Soc., Chem. Commun. 1983, 1527. (c) Gareyev,
R.; Streitwieser, A. J. Org. Chem. 1996, 61, 1742. (d) Goralski, P.; Legoff,
D.; Chabanel, M. J. Organomet. Chem. 1993, 456, 1.
(17) (a) Ha¨ssig, v. R.; Seebach, D. HelV. Chim. Acta 1983, 66, 2269. (b) Bauer,
W.; Seebach, D. HelV. Chim. Acta 1984, 67, 1972. (c) Jackman, L. M.;
Scarmoutzos, L. M.; DeBrosse, C. W. J. Am. Chem. Soc. 1987, 109, 5355.
(22) (a) Gu¨nther, H.; Moskau, D.; Bast, P.; Schmalz, D. Angew Chem., Int. Ed.
Engl. 1987, 26, 1212. (b) Braun, S.; Kalinowski, H. O.; Berger, S. 150
and More Basic NMR Experiments, 2nd ed.; Wiley-VCH: New York, 1998;
p 350.
(23) Gu¨nther, H. J. Braz. Chem. 1999, 10, 241. Mons, H.-E.; Gu¨nther, H.;
Maercker, A. Chem. Ber. 1993, 126, 2747. Gu¨nther, H. In AdVanced
Applications of NMR to Organometallic Chemistry; Gielen, M., Willem,
R., Wrackmeyer, B., Eds.; Wiley & Sons: Chichester, 1996; Chapter 9.
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5429

A R T I C L E S
Briggs et al.
Figure 4. 19F NMR spectra recorded at -80 °C of THF/pentane solutions
of 0.04 N PhCCLi and 0.01 N 8 at the THF concentrations indicated.
Figure 3. 19F NMR spectra recorded at -100 °C of 7.6 M THF/pentane
solutions of 0.10 N PhCCLi and 8 (at the normality indicated).
13C nucleus (Figure 1B and 2B). Tetramer 16 prepared from
[1-15N]5 shows a distinct N-Li coupling in one 6Li resonance
(Figure 1C), whereas 16 prepared from [3-15N]5 affords no
coupling. Several isomers of tetramer 16 are consistent with
the spectroscopic data.24 Cubic tetramers, however, are excluded
by the coupling patterns.8,9
Figure 5. Plot of [mixed dimer] versus [mixed oligomer] (measured in
normality of subunit 8) in 8.0 M THF solutions containing 0.10 N PhCCLi
and a range of concentrations of 8 in pentane cosolvent at -100 °C. The
curve depicts an unweighted least-squares fit to [mixed oligomer] ) K′-
[mixed dimer]m (K′ ) (3 ( 2) × 102, m ) 2.0 ( 0.1).
We originally suspected the chelated mixed oligomer to be
mixed dimer 17 rather than mixed tetramer 16. The spectro-
scopically opaque O-Li linkages render 16 and 17 difficult to
distinguish. However, the chelated form is promoted relative
to mixed dimer 13 at elevated total mixed aggregate concentra-
tions (Figure 3) and low THF concentrations (Figure 4),
suggesting the chelated form has a lower solvation number and
higher aggregation state than those of mixed dimer 13.
The relationship of the chelated form to mixed dimer 13 is
described according to eqs 3 and 4, which can be rearranged to
eq 5. Keeping THF at a high fixed concentration, plotting the
relative concentrations of the mixed aggregates (Figure 5)
enables calculation of a relative aggregation number, m ) 2.0
( 0.1, that is consistent with the mixed tetramer assignment.
9
5430 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

1,2-Addition of PhCCLi to Quinazolinones
A R T I C L E S
Figure 6. Plot of [mixed dimer]2/[mixed tetramer] (i.e., [13]2/[16]) versus
[THF] of solutions containing 0.04 N PhCCLi and 0.01 N 8 over a range
of THF concentrations in pentane cosolvent at -80 °C. The curve depicts
an unweighted least-squares fit to [13]2/[16] ) Keq[THF]n (Keq ) (3.0 (
0.5) × 10-5, n ) 2.91 ( 0.07).
In turn, maintaining a fixed concentration of mixed aggregate
and varying the THF concentration (eq 6, Figure 6) affords a
high relative per-lithium solvation number, n ) 2.91 ( 0.07.
Although the measured solvation number seems to implicate
an unsymmetrically solvated tetramer,25 a systemic error causing
a deviation from an integer value of either 2.0 or 4.0 seems
equally plausible.
Keq
(PhCCLi)m(8)m(THF)4m-n
+ nTHF y z
“mixed oligomer”
m(PhCCLi)(8)(THF)4
(3)
“mixed dimer” (13)
Keq ) [mixed dimer]m/{[THF]n[mixed oligomer]} (4)
[mixed oligomer] ) K′[mixed dimer]m
K′ ) 1/Keq[THF]n ≈ constant
(5)
[mixed dimer]m/[mixed oligomer] ) Keq[THF]n (6)
(m ) 2 from Figure 5)
Figure 7. 19F NMR spectra recorded on a neat THF solution of 0.10 N
PhCCLi and 0.005 N 8 at the temperatures indicated.
Although the complexity caused by the pair of mixed
aggregates did not bode well for detailed rate studies, we
observed a somewhat surprising and fortuitous promotion of
dimer 13 relative to tetramer 16 (Figure 7) at elevated temper-
atures. Indeed, monitoring the temperature-dependent 13/16
ratios in six combinations of THF and mixed aggregate
concentrations (Figure 8 is emblematic) affords relatively
invariant (concentration-independent) thermodynamic param-
eters: ∆H° ) 1.8 ( 0.1 kcal/mol and ∆S° ) -11.6 ( 0.9 cal/
(mol‚K).26 Figure 9 shows a plot of the calculated mole fraction
of dimer 13 as a function of THF under conditions that are
representative of those used in the rate studies. Thus, under
(24) Organolithium ladder structures: Gregory, K.; Schleyer, P. v. R.; Snaith,
R. AdV. Inorg. Chem. 1991, 37, 47. Mulvey, R. E. Chem. Soc. ReV. 1991,
20, 167. Beswick, M. A.; Wright, D. S. In ComprehensiVe Organometallic
Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.;
Pergamon: New York, 1995; Vol. 1, Chapter 1. Mulvey, R. E. Chem. Soc.
ReV. 1998, 27, 339 and references cited within.
(25) For examples of lithium aggregates displaying a noninteger per-lithium
solvation number, see: Seebach, D.; Bauer, W.; Hansen, J.; Laube, T.;
Schweizer, W. B.; Dunitz, J. D. J. Chem. Soc., Chem. Commun. 1984,
853. Depue, J. S.; Collum, D. B. J. Am. Chem. Soc. 1988, 110, 5518.
(26) The 19F resonances of 13 and 16 coalesce at -55 °C.
Figure 8. Plot of ln(Keq) versus (1/T) of a solution containing 0.10 N
PhCCLi and 0.005 N 8 in neat THF. The curve depicts an unweighted least-
squares fit to ln(Keq) ) (-∆H/R)(1/T) + ∆S/R (∆H ) 1.89 ( 0.05 kcal/
mol, ∆S ) -11.7 ( 0.2 cal/(mol‚K)).
carefully controlled conditions, dimer 13 is dominant, and
tetramer 16 is largely irrelevant to the rate studies (vide infra).
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5431

A R T I C L E S
Briggs et al.
Figure 10. Plot of the 19F integration versus time for the 1,2-addition of
PhCCLi (0.50 N) to 13 (0.005 M) at 10 °C in a 7.75 M THF/pentane. The
Figure 9. Plot of the mole fraction of 13 (X(13)) versus THF concentration
(M) for a solution of [13] and [16] at 10 °C (total normality ) 0.005 N).
The curve was constructed from the experimentally determined values of
Keq ) 1.20 × 10-4, m ) 2, and n ) 3 (see eqs 3-6) and a complex
expression derived in the Supporting Information.
obsdt
curve depicts an unweighted least-squares fit to integration ) integration0e-k
+ c (kobsd ) (8.19 ( 0.07) × 10-4; integration0 ) 136.1 ( 0.7; c ) 0.0 (
0.2). The inset depicts the ln(integration) versus time.
Unfortunately, the absorbance of 13 is obscured by a product-
derived absorbance at 1505-1545 cm-1 that becomes prominent
at high THF concentrations.
The temperature- and concentration-dependent dimer-tet-
ramer equilibrium requires additional comment. Conventional
wisdom suggests that the more highly solvated lower aggregate
would be favored at low temperatures due to a net enthalpic
stabilization attributed to the stabilizing enthalpy of solvation.27
In contrast, the more highly solvated dimer 13 is enthalpically
and entropically disfavored when compared to tetramer 16. The
dominance of dimer 13 under the conditions used to follow the
reaction rates stems from the substantial deviation from the unit-
molarity standard state due to the high THF concentration and
low mixed aggregate concentration.
The product of the 1,2-addition is depicted generically as
dianion 10; spectroscopic studies suggest that 10 is a mixture
of species as shown by a broad mound of 19F resonances. The
dianion formulation (rather than a monoanion resulting from
protonation by residual hexamethyldisilazane) is supported by
control experiments showing that deprotonations of 2 with 1.0
and 2.0 equiv of LiHMDS afford spectroscopically distinct
product distributions. Most importantly, however, the product
of the reaction is of no consequence under pseudo-first-order
conditions used to study the kinetics.
It is also noteworthy that the formation of 10 in the presence
of excess [6Li]PhCCLi shows no evidence of PhCCLi-containing
mixed aggregates. Thus, the 1,2-addition in eq 2 does not result
in the net consumption of free PhCCLi dimer 14. Consequently,
conversion-dependent changes in mechanism that can dramati-
cally increase the mechanistic complexity under stoichiometric
conditions are unlikely to intervene.
Rate Studies: 1,2-Addition to Mixed Dimer 13. Pseudo-
first-order conditions were established by maintaining low mixed
aggregate concentrations (0.005 N), high PhCCLi concentra-
tions, high THF concentrations (using pentane as the cosolvent),
and LiHMDS in a fixed (0.10 M) excess. Under these condi-
tions, dimer 13 is the dominant mixed aggregate, and tetramer
16 is insignificant (Figure 9). The rates of the 1,2-addition in
eq 2 were initially investigated by monitoring the loss of mixed
dimer 13 (1552 cm-1) using in situ IR spectroscopy.28 (The
IR-derived rate data are archived in the Supporting Information.)
19F NMR spectroscopy proves to be a superior analytical
method for the current case study.29 Monitoring the 19F
resonance of 13 reveals a clean first-order decay (Figure 10),
provided the THF concentrations are elevated (>7.0 M in
pentane). It is normal that a visually clean first-order decay in
conjunction with concentration-independent values of kobsd is
adequate to demonstrate a first-order dependence on the limiting
reagent. However, this complex case demanded that we
determine the order by a best-fit protocol.30 We prefer a method
based on a nonlinear least-squares fit to the function in eq 7.31
The adjustable parameter R corresponds to the reaction order
in 13. The data for the 18 runs used to construct Figure 11
(discussed below) were fit to eq 7 and averaged, affording R )
1.05 ( 0.03.
-(1-R) -1/(R-1)
[13] ) {(R - 1)kobsdt + [13]0
}
(7)
The first-order dependence on the concentration of 13 at high
THF concentration is unequivocal. At lower THF concentrations,
however, tetramer 16 becomes significant (Figure 9), and we
observe considerable deviations from clean first-order behav-
ior.32 Curiously, the deviations are positive (R > 1.0) rather
than negative (R < 1.0). (The order for a mixed-dimer based
reaction would approach R ) 0.5 in the limit of complete
tetramer formation.) It is possible that a new pathway intervenes
at low THF concentration, but we are currently unable to provide
additional insight.
The question of whether 13 reacts directly as a mixed
aggregate or through the intervention of additional PhCCLi is
(28) (a) Sun, X.; Collum, D. B. J. Am. Chem. Soc. 2000, 122, 2452. (b) Rein,
A. J.; Donahue, S. M.; Pavlosky, M. A. Curr. Opin. Drug DiscoVery DeV.
2000, 3, 734.
(29) For a review of structural studies using 19F NMR spectroscopy, see: Gakh,
Y. G.; Gakh, A. A.; Gronenborn, A. M. Magn. Reson. Chem. 2000, 38,
551.
(30) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms; McGraw-
Hill: New York, 1995; p 31.
(31) (a) Equation 7 is a nonlinear variant of the Noyes equation.31b A derivation
is found in the Supporting Information. (b) Casado, J.; Lopez-Quintela,
M. A.; Lorenzo-Barral, F. M. J. Chem. Educ. 1986, 63, 450.
(32) The deviation from a first-order fit appears as a poor fit in a plot of [13]
versus time when fit to the function [13] ) [13]0e-kt. Fitting the data to eq
7 affords reaction orders in 13 as high as 1.4. The deviation is most easily
visualized as an upward curvature in the plot of ln[13] versus time
(Supporting Information).
(27) Deaggregation is normally promoted at low temperatures and is widely
believed to be the result of the dominance of the enthalpy affiliated with
higher solvation numbers. It is unclear why the less highly solvated tetramer
16 (as shown by THF concentration dependencies) is enthalpically preferred
(as shown by the temperature dependencies).
9
5432 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

1,2-Addition of PhCCLi to Quinazolinones
A R T I C L E S
idealized rate law described by eq 8 is consistent with the
mechanism depicted generically in eqs 9 and 10. We do not
intend to suggest the requisite intermediacy of a discrete
disolvated monomer 18 or even a specific sequence of events;
rather, we emphasize the incorporation of the PhCCLi(THF)2
fragment in rate-limiting transition structure 19.
d[13]/dt ) k′[13][PhCCLitot]1/2[THF]0
(8)
(9)
1/2(PhCCLi)2(THF)
(PhCCLi)(THF)
4 h
2
(18)
(14)
(PhCCLi)(8)(THF)4 (PhCCLi)(THF)
+
2 f
Figure 11. Plot of kobsd versus [PhCCLi] in 7.75 M THF/pentane for the
1,2-addition of PhCCLi to mixed dimer 13 (0.005 N of 8) at 10 °C. The
curve depicts an unweighted least-squares fit to kobsd ) a[PhCCLi]b (a )
(1.16 ( 0.01) × 10-3, b ) 0.42 ( 0.01).
(13)
(18)
[(PhCCLi)2(8)(THF)6]q
(19)
(10)
Rate Studies: 1,2-Addition to Quinazoline 11. For com-
parison, we monitored the 1,2-addition to O-protected quinazo-
line 11. The choice of the methoxymethyl protecting group was
based on ease of synthesis and purification.36 We hasten to add
that there is compelling data suggesting the second oxygen
linkage of the ketal can inductively influence reactivities, but
the evidence also reveals that ketals do not function as
polydentate ligands.37,38
The 1,2-additions to 11 were carried out using methods
similar to those described for the additions to 13. Monitoring
the 19F resonance of 11 shows a clean first-order decay
(Supporting Information).39,40 The addition of PhCCLi to 11 is
102-103 times faster than its addition to 13, consistent with
the higher electron density on 13. The rate data for 1,2-addition
to 11 reveal notable similarities and differences with 13. A plot
of kobsd versus [THF] for the addition to 11 displays a first-
order dependence on the THF concentration (Figure 13). A plot
of kobsd versus PhCCLi concentration in 8.6 M THF reveals a
fractional (0.45 ( 0.04) order. The idealized rate law (eq 11)
implicates a mechanism based on trisolvated PhCCLi monomers
(eqs 12 and 13).
Figure 12. Plot of kobsd versus [THF] in pentane cosolvent for the 1,2-
addition of PhCCLi (0.20 N) to 13 (0.005 N of 8) at 10 °C. The curve
depicts an unweighted least-squares fit to kobsd ) k[THF] + k′ (k ) (-2.29
( 0.02) × 10-5, k′ ) (8.0 ( 0.1) × 10-4).
addressed by determining the rate dependencies on the PhCCLi
concentration. A plot of kobsd versus [PhCCLi]33 at 7.75 M THF/
pentane affords an order of 0.42 ( 0.01 (Figure 11). Although
0.42 is lower than expected,34 it is strongly suggestive of the
half-order dependence anticipated for a mechanism requiring
deaggregation of PhCCLi dimer 14.
A plot of kobsd versus THF concentration (Figure 12) reveals
nearly [THF]-independent rates.35 (Recall that the values of kobsd
are only approximate because of deviations from a first-order
decay that are due to the appearance of tetramer 16.) The slight
negative drift does not approach the >4-fold changes in kobsd
expected for first-order or inverse-first-order dependencies on
the THF concentration. Therefore, we conclude that the
predominant pathway displays a zeroth-order THF concentration
dependence.
d[11]/dt ) k′[11][PhCCLi]1/2[THF]1
(11)
(12)
1/2(PhCCLi)2(THF)
(PhCCLi)(THF)
4 THF h
3
(20)
(14)
[(11)(PhCCLi)(THF) ]q
11 + (PhCCLi)(THF)
(20)
3 f
(13)
3
(21)
Discussion
The highly enantioselective addition of lithium acetylides to
quinazolinones illustrated in eq 1 is the key step in the syntheses
Overall, the evidence supports a first-order dependence on
mixed dimer 13, a half-order dependence on PhCCLi dimer 14,
and a zeroth-order dependence on the THF concentration. The
(36) Simple O-alkylations of quinazolinone 5 were problematic. O-Silylation
proceeded without event, but desilylation tended to occur under the
conditions of the 1,2-addition.
(37) (a) Chadwick, S. T.; Rennels, R. A.; Rutherford, J. L.; Collum, D. B. J.
Am. Chem. Soc. 2000, 122, 8640. (b) Lucht, B. L.; Bernstein, M. P.;
Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1996, 118, 10707.
(38) Trauner and co-workers recently reported a crystal structure of an aryllithium
with an ortho methoxymethyl protected phenol in which only the methoxy
group coordinates to lithium: Hughes, C. C.; Scharn, D.; Mulzer, J.;
Trauner, D. Org. Lett. 2002, 4, 4109.
(39) Quenching the reaction with aqueous NH4Cl followed by an extractive
workup affords a crude mixture of two compounds displaying spectroscopic
behavior consistent with tautomeric imino ethers. Alternatively, quenching
the reaction with 2 N HCl affords the dihydroquinazolinone.
(33) The concentration of PhCCLi refers to the concentration of the monomer
unit in units of normality (N).
(34) A low basal reactivity of 13 without intervention of additional lithium
acetylide could explain the reduced fractional order in PhCCLi. Indeed, if
the rate data are fit to kobsd ) a[13]b + c, a small nonzero intercept, c, is
accompanied by a fractional order of b ) 0.54. We hasten to add, however,
that such fits in which fractional orders and nonzero intercepts are adjustable
parameters are inordinately sensitive to the quality of the data. In our
opinion, the closer fit to a one-half order cannot be interpreted beyond
simply a demonstration of the general principle.
(35) The values of kobsd are independent of LiHMDS and free hexamethyldisi-
lazane concentrations (zeroth order).
(40) The data for the 18 runs used to determine the order in PhCCLi were fit to
eq 7 and averaged, affording R ) 1.01 ( 0.09.
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5433

A R T I C L E S
Briggs et al.
experiments show that LiHMDS and its conjugate acid (HMDS)
do not influence the rate. The reaction also appears to be zeroth
order in THF, indicating no additional per-lithium solvation of
13 or 14 occurs en route to the rate-limiting transition structure.
Finally, an approximate half-order dependence on PhCCLi
indicates that PhCCLi dimer 14 dissociates to a disolvated
monomer before the 1,2-addition. The collective concentration
dependencies are consistent with the idealized rate law and
mechanism illustrated in eqs 8-10.
Although, in principle, mixed dimer 13 contains the compo-
nents necessary for a 1,2-addition, the structural and rate data
implicate a transition structure of stoichiometry [(PhCCLi)2-
(8)(THF)6]q (eq 10).41,42 Of course, the rate law does not provide
insight into connectivity or the three-dimensional structure. We
envision [(PhCCLi)2(8)(THF)6]q as a composite of mixed dimer
13 and PhCCLi(THF)2 monomer and provide 22 as a working
model. The most important implication of this model is that
mixed dimer 13 is simply a protected quinazolinone and the
PhCCLi component of mixed dimer 13 is of no mechanistic
consequence.
Figure 13. Plot of kobsd versus [THF] in pentane cosolvent for the 1,2-
addition of PhCCLi (0.05 N) to 11 (0.005 M) at -40 °C. The curve depicts
an unweighted least-squares fit to kobsd ) a[THF]b (a ) (2.1 ( 0.2) ×
10-4, b ) 1.09 ( 0.04).
of nonnucleoside reverse transcriptase inhibitors 2-4.1 Using
unprotected quinazolinones with consequent in situ formation
of lithium salt 8 markedly simplifies the synthesis by avoiding
a formal protection-deprotection sequence. The coexistence of
five lithiated species, however, introduces potentially enormous
structural and mechanistic complexities.7-9
We have taken a step toward understanding the enantiose-
lective addition by investigating the nonenantioselective variant
in eq 2. Omission of the amino alkoxide 7 simplified the
problem. Even so, the structural control necessary to carry out
detailed rate studies required careful planning and an element
of luck. We knew that readily purified LiHMDS13 could be used
in excess as a proton scavenger without the intervention of
LiHMDS mixed aggregates.9,14,15 The control of the lithium
acetylide structure was essential. Although lithium acetylides
in THF and THF/hydrocarbon mixtures normally exist as
mixtures of cyclic dimers and cubic tetramers, we found that
PhCCLi exists as dimer 14 to the exclusion of tetramer 15 in
THF and THF/pentane mixtures. The reluctance to form cubic
tetramers is unique to this particular lithium acetylide/solvent
combination16 for reasons that are unclear.18
Characterizing and controlling the structure of lithium
quinazolinide salt 8 in the presence of PhCCLi was particularly
challenging. At the very low temperatures used to investigate
solution structures, a mixture of excess PhCCLi and 8 in THF/
pentane solutions affords primarily ladder-based mixed tetramer
16 (or an isomeric ladder) along with low concentrations of
mixed dimer 13. However, the conditions that are optimal for
rate studies, low total mixed aggregate concentrations, high THF
concentrations, and elevated temperatures, promote the forma-
tion of mixed dimer 13. Consequently, the rate studies were
carried out with mixed dimer 13 as the dominant mixed
aggregate (g90%; Figure 9), rendering tetramer 16 largely
inconsequential.
We investigated the analogous 1,2-addition to O-protected
quinazoline 11 to understand the nuances of the 1,2-addition to
lithium quinazolinide 8. The methoxymethyl protecting group
was chosen due to its ease of preparation;36 mounting evidence
argues that the often-cited special (polydentate) ligating proper-
ties of acetals are overstated.37,38 Quinazoline 11 is 102-103
times more reactive than mixed dimer 13 toward 1,2-addition,
qualitatively consistent with the relative electron densities on
11 and 13. The first-order dependence on the THF concentration
(Figure 13) and the half-order dependence on the PhCCLi
concentration are consistent with a trisolvated PhCCLi-monomer-
based addition. Conventional thinking might suggest that the
three coordinated THF ligands implicate a transition structure
such as 23 lacking an N-Li interaction (and the affiliated Lewis
acid assistance43) so as not to exceed four-coordinate lithium.
One could further argue that this pathway does not appear in
the 1,2-addition to mixed dimer 13 due to the excessive electron
density that would develop at the transition structure or the steric
demands of the mixed dimer coordination sphere.
Monitoring the rate of the 1,2-addition by in situ IR
spectroscopy and 19F NMR spectroscopy provided self-
consistent results; the high resolution offered by 19F NMR
spectroscopy was particularly convenient.29 Given the complex-
ity of this case study, we took special precautions to show that
the reaction was first order in dimer 13. (A method based on a
nonlinear least-squares best fit to eq 7 is noteworthy.) At low
THF concentrations wherein tetramer 16 becomes a significant
component, deviations from first order are observed. Control
(41) Edwards, J. O.; Greene, E. F.; Ross, J. J. Chem. Educ. 1968, 45, 381.
(42) Rutherford, J. L.; Hoffmann, D.; Collum, D. B. J. Am. Chem. Soc. 2002,
124, 264.
(43) Casado, F.; Pisano, L.; Farriol, M.; Gallardo, I.; Marquet, J.; Melloni, G.
J. Org. Chem. 2000, 65, 322.
9
5434 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

1,2-Addition of PhCCLi to Quinazolinones
A R T I C L E S
Experimental Section
Conversely, we believe that the observed trisolvation does
not necessarily exclude a Li-N interaction. There is evidence
that lithium can exceed four-coordinate.44-48 Three poignant
examples include: (1) octahedral +Li(η2-DME)3, commonly
observed crystallographically;44 (2) a trigonal bipyramidal +Li-
(THF)5 observed crystallographically;45 and (3) (Me3Si)2NLi-
(THF)4 (and the oxetane-solvated analogue) observed spectro-
scopically.46 Therefore, we submit that a contact ion pair such
as 24 is reasonable.47,48 In fact, reflecting back on transition
structure 22, it is not altogether obvious that the six THF ligands
would necessarily be distributed evenly on the three lithiums
as drawn.
Reagents and Solvents. THF and pentane were vacuum-transferred
from degassed blue or purple stills containing sodium benzophenone
ketyl. The pentane still contained 1% tetraglyme to dissolve the ketyl.
Air- and moisture-sensitive materials were manipulated using vacuum
line and syringe techniques. Quinazolinone 5 was prepared using a
literature procedure.1 Quinazolinones [1-15N]5 and [3-15N]5 were
prepared using [15N]chloroaniline (>99%) and [15N]benzyl isocyanate
(99%).1 [6Li]PhCCLi and [6Li,13C]PhCCLi were prepared in situ using
triply recrystallized [6Li]LiHMDS13 and commercially available phe-
nylacetylene and [13C-2]phenylacetylene. The phenylacetylenes were
vacuum-transferred from 4 Å molecular sieves. Lithium salt 8 was
generated in situ from quinazolinone 5 and LiHMDS.
NMR Spectroscopic Analyses. All NMR tubes were prepared using
stock solutions and sealed under partial vacuum. [6Li]LiHMDS was
kept in excess at all times to scavenge the acidic protons of phenyl-
Conclusion
The proximity of an electrophilic imine moiety and potentially
nucleophilic acetylide fragment in mixed dimer 13 is provoca-
tive. It might be tempting to invoke an intraaggregate condensa-
tion. Nonetheless, the rate studies lead us to conclude that mixed
dimer 13 is simply an O-protected quinazolinone that reacts
with an external (additional) disolvated PhCCLi monomer via
transition structure 22. The condensation of PhCCLi is shown
to react analogously through a more highly solvated analogue
suggested to be 23 or 24.
6
acetylene and quinazolinone. Standard Li, 13C, and 19F NMR spectra
were recorded on a 400 MHz spectrometer at 58.84, 100.58, and 376.38
MHz (respectively). The 6Li, 13C, and 19F resonances are referenced to
0.3 M [6Li]LiCl/MeOH at -90 °C (0.0 ppm), the CH2O resonance of
THF at -110 °C (67.35 ppm), and C6F6 (-162.55 ppm set at the
recording temperature), respectively. The spectra were recorded with
a three-channel probe designed to accommodate lithium and carbon
6
1
pulses. Li,13C-HMQC spectroscopy21 and J(6Li,13C)-resolved spec-
troscopy have been reported.9,22-23
Rate Studies by 19F NMR Spectroscopy. All spectroscopic samples
for monitoring the rate of the 1,2-additions were prepared using stock
solutions, sealed under partial vacuum, and stored at -78 °C (for mixed
dimer 13) or at -196 °C (for quinazoline 11) before analysis. Detailed
protocols are outlined in the Supporting Information.
In principle, such studies move us closer to understanding
how the enantioselective addition in eq 1 occurs. In practice,
significant speculation would be premature. We will, however,
ask one rhetorical question: If mixed dimer 13 is a protected
quinazolinone, is it possible that chiral lithium amino alkoxide
7 converts lithium quinazolinide 8 to a mixed aggregate that is
functionally an O-protected form bearing a chiral auxiliary? We
suspect that, despite the complexity of the reaction mixture in
eq 1, the high enantioselectivity likely derives from a single,
compelling control element.
Acknowledgment. The Cornell group thanks the National
Institutes of Health and DuPont Pharmaceuticals for direct
support of this work. T.F.B. thanks the National Institutes of
Health for a Chemistry-Biology Interface Training Grant (T32-
GM08500). The Cornell group also thanks Merck, Pfizer,
Boehringer-Ingelheim, R. W. Johnson, Aventis, and Schering-
Plough for indirect support.
(44) Bock, H.; Beck, R.; Havlas, Z.; Schoedel, H. Inorg. Chem. 1998, 37, 5046.
Bock, H.; Havlas, Z.; Gharagozloo-Hubmann, K.; Sievert, M. Angew.
Chem., Int. Ed. 1999, 38, 2240.
(45) Olmstead, M. M.; Power, P. P.; Sigel, G. Inorg. Chem. 1986, 25, 1027.
(46) Lucht, B. L.; Collum, D. B. J. Am. Chem. Soc. 1995, 117, 9863.
(47) Depue, J. S.; Collum, D. B. J. Am. Chem. Soc. 1988, 110, 5524. Ram´ırez,
A.; Lobkovsky, E.; Collum, D. B., unpublished.
(48) Octahedral sodium cation solvated by six THF ligands has been impli-
cated: Jorgensen, W. L.; Chandrasekhar, J. J. Chem. Phys. 1982, 77, 5080.
Supporting Information Available: NMR spectra, rate data,
and experimental protocols (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
JA0305813
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5435