Catalysis of a Diels-Alder reaction by amidinium ions

Article abstract of DOI:10.1021/jo991372x

Amidines and guanidines are important functional groups in molecular recognition and host-guest chemistry. Here it is shown that lipophilic amidinium ions catalyze a cycloaddition reaction representing the key step of the Quinkert-Dane estrone synthesis.

Full text of DOI:10.1021/jo991372x

J . Org. Chem. 2000, 65, 1697-1701
1697
Ca ta lysis of a Diels-Ald er Rea ction by Am id in iu m Ion s
Tilmann Schuster, Markus Kurz, and Michael W. Go¨bel*,†
Institut fu¨r Organische Chemie der J . W. Goethe-Universita¨t Frankfurt, Marie-Curie-Strasse 11,
D-60439 Frankfurt am Main, Germany
Received August 30, 1999
Amidines and guanidines are important functional groups in molecular recognition and host-
guest chemistry. Here it is shown that lipophilic amidinium ions catalyze a cycloaddition reaction
representing the key step of the Quinkert-Dane estrone synthesis. Hydrogen-bond-mediated
association with the organic cation leads to an electrophilic activation of the dienophile and to
enhanced rates of the Diels-Alder reaction. The observed effects are similar to those expected
from mild Lewis acids. In competition experiments, amidinium catalysis favors the reaction of the
less electron deficient dienophile.
Sch em e 1
In tr od u ction
Among synthetic receptor molecules based on ami-
dinium and guanidinium ions, good catalysts for phos-
phoryl transfer reactions can be found.1,2 Their mode of
action is to increase substrate electrophilicity by forming
hydrogen bonded ion pairs. It has been shown earlier that
organic cations may also associate with neutral carbonyl
compounds.3 The enhanced guest electrophilicity in those
complexes should accelerate reactions with nucleophiles
and cycloadditions.4 Lewis acids are common catalysts
for Diels-Alder reactions. In lithium perchlorate pro-
moted cycloadditions, again the Lewis acidity of the
cation is exploited.5 In rare cases, even weaker electro-
philes are sufficient for catalysis: Diels-Alder reactions
have been accelerated by complexing dienophiles to
hydrogen-bond-forming receptors.6 Since amidinium ions
are good hydrogen bond donors, they might catalyze
cycloadditions in a similar way. In contrast to lithium
perchlorate, the shape of amidinium salts can be easily
optimized for certain applications, e.g. by introduction
of chirality.7 Here we present two examples of ami-
dinium-catalyzed [4 + 2] cycloadditions. The first reaction
constitutes a simple case in which rac-3 is formed from
maleic anhydride 1 and diene 2 (Scheme 1).8 A more
complex behavior is observed in the cycloaddition of 2
and diketone 5.9 This reaction leads to rac-8, a key
intermediate in the Quinkert-Dane synthesis of estrone
(Scheme 2).10
Resu lts a n d Discu ssion
To favor the association of amidinium groups and
dienophiles, less polar solvents and noncoordinating
counterions are essential. Structure 12 was chosen
instead of simpler amidines to guarantee sufficient
solubility of the salts (Scheme 3). Starting from palmitoyl
amide 10,11 we prepared the amidinium tetrafluoroborate
12a via O-alkylation (11; 86%) and ammonolysis (46%).
Anion exchange then produced the tetraaryl borate salts
12b-d .12 While all three compounds are soluble in CH2-
Cl2, a strong tendency for ion pairing was seen in 12b.
† Phone: +49 69 798 29222. Fax: +49 69 798 29464. E-mail:
M.Goebel@chemie.uni-frankfurt.de.
(1) (a) J ubian, V.; Dixon, R. P.; Hamilton, A. D. J . Am. Chem. Soc.
1992, 114, 1120. (b) J ubian, V.; Veronese, A.; Dixon, R. P.; Hamilton,
A. D. Angew. Chem. 1995, 107, 1343; Angew. Chem., Int. Ed. Engl.
1995, 34, 1237. (c) Smith, J .; Ariga, K.; Anslyn, E. V. J . Am. Chem.
Soc. 1993, 115, 362. (d) Oost, T.; Filippazzi, A.; Kalesse M. Liebigs
Ann. Recl. 1997, 1005. (e) Oost, T.; Kalesse, M. Tetrahedron 1997, 53,
8421.
(2) (a) Muche, M.-S.; Go¨bel, M. W. Angew. Chem. 1996, 108, 2263;
Angew. Chem., Int. Ed. Engl. 1996, 35, 2126. (b) Muche, M.-S.; P.
Kamalaprija, P.; Go¨bel, M. W. Tetrahedron Lett. 1997, 38, 2923.
(3) Bell, D. A.; Anslyn, E. V. J . Org. Chem. 1994, 59, 512.
(4) Raposo, C.; Almaraz, M.; Crego, M.; Mussons, M. L.; Pe´rez, N.;
Caballero, M. C.; Mora´n, J . R. Tetrahedron Lett. 1994, 35, 7065.
(5) (a) Pindur, U.; Lutz, G.; Otto, C. Chem. Rev. 1993, 93, 741. (b)
Flohr, A.; Waldmann, H. J . Prakt. Chem. 1995, 337, 609.
(6) Kelly, T. R.; Meghani, P.; Ekkundi, V. S. Tetrahedron Lett. 1990,
31, 3381.
1
The H NMR signal of the a methylene group (δ ) 2.33
(9) Dane, E.; Schmitt, J .; Rautenstrauch, C. Liebigs Ann. Chem.
1937, 532, 29.
(10) Quinkert, G.; Del Grosso, M.; Do¨ring, A.; Do¨ring, W.; Schenkel,
R. I.; Bauch, M.; Dambacher, G. T.; Bats, J . W.; Zimmermann, G.;
Du¨rner, G. Helv. Chim. Acta 1995, 78, 1345 (and references cited
herein).
(11) Krafft, F.; Stauffer, B. Ber. Dtsch. Chem. Ges. 1882, 15, 1728.
(12) Sodium tetrakis(3,5-bis(trifluoromethyl)phenyl)borate: (a) Iwa-
moto, H.; Sonoda, T.; Kobayashi, H. Tetrahedron Lett. 1983, 24, 4703.
(b) Nishida, H.; Takada, N.; Yoshimura, M.; Sonoda, T.; Kobayashi,
H. Bull. Chem. Soc. J pn. 1984, 57, 2600.
(7) (a) Lehr, S.; Schu¨tz, K.; Bauch, M.; Go¨bel, M. W. Angew. Chem.
1994, 106, 1041; Angew. Chem., Int. Ed. Engl. 1994, 33, 984. (b)
Schuster, T.; Go¨bel, M. W. Synlett 1999, 966.
(8) (a) Dane, E.; Ho¨ss, O.; Bindseil, A. W.; Schmitt, J . Liebigs Ann.
Chem. 1937, 532, 39. (b) Bachmann, W.; Controulis, J . J . Am. Chem.
Soc. 1951, 73, 2636.
10.1021/jo991372x CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/24/2000

1698 J . Org. Chem., Vol. 65, No. 6, 2000
Schuster et al.
Ta ble 1. Diels-Ald er Exp er im en ts w ith Ma leic
An h yd r id e 1
Sch em e 2
catalyst
ratioa
yielda
(no. of equiv)
rac-3:rac-4
acceleration
after 4 h (%)
no catalyst
12d (1)
13 (0.5)
13 (1)
6:1
14:1
11:1
13:1
17:1
1
63b
82
74
82
88
2.5
1.7
2.6
3.9
13 (2)
a
b
The yields and ratios were determined by HPLC. 88% yield
after 24 h.
yields were determined by HPLC using 2-methoxy-6-
methylnaphthalene (14) as internal standard (k2 ) 2.5
× 10-6 mM-1 s-1).16 In 1H NMR titrations with ami-
dinium salt 12d (CDCl3, 7 °C), a low-field shift of the NH
signals was seen upon addition of maleic anhydride.
However, the association must be weak since the shift
remained small and could not be saturated. Electron-poor
anhydrides have been shown previously to possess the
lowest affinities toward organic cations within a series
of carbonyl compounds.3 In consequence, the amidinium
salts 12d and 13 did not induce large rate effects (Table
1). The ratio of the endo and exo isomers rac-3:rac-4 was
shifted in favor of endo.
Sch em e 3
Enolization of diketone 5 would form a cyclopentadi-
enone derivative with antiaromatic properties. As a
result, 5 prefers the keto form, a rare exception within
the class of 1,2-dicarbonyl compounds. In contrast to
maleic anhydride, diketone 5 readily associates with
amidinium salt 12d . Ka could be determined by 1H NMR
titration (37 ( 6 M-1; CDCl3, 7 °C). The stability constant
for the complex 13‚5 was estimated in the same way (70
( 20 M-1; CDCl3, 7 °C). Further evidence for association
came from the mass spectra of 13 when 100 equiv of
diketone 5 was added. In the presence of an equimolar
mixture of maleic anhydride and diketone, only the
complex 13‚5 could be detected. Accordingly, ESI-MS
permits to select good binding partners within a mixture
or small library of weakly coordinating guest molecules.
The Diels-Alder reaction of diene 2 and diketone 5
(Scheme 2) is complicated by the existence of constitu-
tional isomers and by the slow tautomerization to the
final keto enols. In addition, special care is necessary to
avoid artifacts: glass surfaces and silica gel are good
catalysts and influence both rates and product ratios.17
All reactions therefore were run in polyethylene vials
(CH2Cl2, 7 °C). Analytical samples, quenched in aceto-
nitrile and stored in liquid nitrogen, proved to be suf-
ficiently stable. They were analyzed without further
workup by HPLC, again using naphthalene 14 as inter-
nal standard. Pure samples of diene 2, 14, product rac-
8, and isomeric product rac-9 were prepared according
to published procedures and applied to the calibration
of the HPLC system. In the absence of catalysts, the
reaction produced rac-8 and rac-9 (ratio < 0.1:1) slowly
and in low yield (<3%). Under these conditions, most of
the diene 2 decomposed. Good total yields and ratios of
rac-8 and rac-9 around 3:1 were found in all amidinium-
promoted reactions (Table 2). Covalent adducts of amidines
and diketones, which might have formed under basic
conditions, could not be detected. Interestingly, when the
in DMSO-d6) is shifted to 0.95 ppm in CDCl3. We
attribute this shift to a ring current effect from a tightly
bound tetraphenyl borate ion. In compound 12c the effect
was reduced (δ ) 1.43) and completely absent in com-
pound 12d (δ ) 2.33, CDCl3). All kinetical studies,
therefore, used the latter salt.13 The 2-(benzylamino)-
pyridinium salt 13 may be seen as a heterocyclic ana-
logue of an amidinium ion.14 Compared to 12a -d , its
acidity is distinctly increased.
The reaction of maleic anhydride 1 with diene 2 (CH2-
Cl2, 7 °C, 24 h) produced 75% of rac-3 together with 13%
of the exo isomer rac-4 (Scheme 1).15 Reaction rates and
(13) In our initial experiments, isolated yields of rac-8 and rac-9
were determined instead of reaction rates. In accordance with their
tendency of ion pairing, 12b led to the lowest and 12d to the best yields.
The tetrafluoroborate salt 12a was not sufficiently soluble to be tested.
(14) Sprinzak, Y. Organic Syntheses; Wiley: New York, 1963;
Collect. Vol. IV, p 91.
(16) Preparation of compound 14: Biehl, E. R.; Deshmukh, A. R.;
Dutt, M. Synthesis 1993, 885.
(15) Andreev, V. M.; Segal, G. M.; Kucherov, V. F. Bull. Acad. Sci.
USSR Div. Chem. Sci. 1961, 1374.
(17) Veselovsky, V. V.; Gybin, A. S.; Lozanova, A. V.; Moiseenkov,
A. M.; Smit, W. A.; Caple, R. Tetrahedron Lett. 1988, 29, 175.

Catalysis of a Reaction by Amidinium Ions
J . Org. Chem., Vol. 65, No. 6, 2000 1699
Ta ble 2. Diels-Ald er Exp er im en ts w ith Dik eton e 5
ally reduced rates, but the yield and product distribution
remained constant.
catalyst
(no. of equiv)
ratioa
rac-8:rac-9
yielda
after 7 db (%)
In uncatalyzed competition experiments of diene 2 with
both dienophiles 1 and 5, only the cycloadducts of maleic
anhydride were formed (rac-3:rac-4 ) 6:1). Since com-
pound 13 quite selectively accelerates the reaction of
diketone 5, the catalyst might switch the product distri-
bution. As expected, in the presence of 13 the estrone
precursor rac-8 became the main constituent of the
mixture (total yield determined by HPLC: 83%; rac-8:
rac-9:rac-3:rac-4 ) 19:6:12:1).
The results have demonstrated that amidinium ions
can accelerate Diels-Alder reactions with considerable
substrate selectivity. The endo adducts are favored, and
product distributions are shifted in a way typical for
Lewis acid-catalyzed processes. A first report on the
catalysis of Diels-Alder reactions by axially chiral ami-
dinium ions has been published recently.19,20
acceleration
no catalyst
12d (1)
13 (0.25)
13 (0.5)
13 (1)
<0.1:1
2.6:1
3.5:1
3.3:1
3.3:1
1
100
125
300
450
<3
82c
93
∼100
∼100
a
b
The yields and ratios were determined by HPLC. The long
times are due to the slow tautomerization of rac-6 and not to the
Diels-Alder reaction. c The tautomerization was enforced by ad-
dition of H2O in MeCN (7 d, room temperature). Otherwise it was
incomplete even after 14 d.
Exp er im en ta l Section
Methylene chloride was filtered over alumina B, activity I
before use. HPLC column: Merck LiChrospher 100 RP-18 (5
mm), 125 × 4 mm. Mp: hot plate microscope, uncorrected. 1H
NMR spectra were recorded at 200, 250, 270, and 400 MHz;
chemical shifts (δ) are given in ppm, and J in Hz. The diene
2,10 the diketone 5,9 and the reference substances rac-8 and
rac-910 were prepared according to published methods.
(()-3-Meth oxy-16-oxa-14â-gon a-1,3,5(10),9-tetr aen -15,17-
d ion e (r a c-3) a n d (()-3-Met h oxy-16-oxa -8R,14â-gon a -
1,3,5(10),9-tetr a en -15,17-d ion e (r a c-4). These were pre-
pared by a slight modification of Andreev’s method:15 Diene 2
(400 mg, 2.15 mmol) and maleic anhydride 1 (211 mg, 2.15
mmol) were dissolved in 20 mL of CH2Cl2 and stirred overnight
at room temperature. After removal of the solvent in vacuo,
crystallization of the residue from Et2O/n-hexane afforded 410
mg (67%) of rac-3 as colorless needles. Purification of the
mother liquid by semipreparative HPLC (10:2 n-hexane/EtOAc
+ 30% CH2Cl2, MN Nucleosil 50-10, 250 × 10 mm, 10 mL/
min, 254 nm) afforded a further 24 mg (4%; 71% combined
overall yield) of rac-3 and after recrystallization from Et2O/
n-hexane 20 mg (3%) of rac-4 as yellowish needles.
(()-3-Meth oxy-16-oxa-14â-gon a-1,3,5(10),9-tetr aen -15,17-
d ion e (r a c-3): mp 205-206 °C (lit.:8b mp 203-204 °C); 1H
NMR (CDCl3) δ 7.40 (d, 1, J ) 8.9), 6.73 (dd, 1, J ) 8.6, 2.4),
6.64 (d, 1, J ) 2.7), 6.24-6.27 (m, 1), 3.44-3.54 (m, 2), 3.79
(s, 3), 2.91-2.97 (m, 1 H), 2.76-2.81 (m, 1), 2.76-2.81 (m, 1),
2.60-2.73 (m, 2), 2.32-2.39 (m, 1), 2.21-2.31 (m, 1), 2.08-
2.15 (m, 1); IR (KBr) 3080, 2996, 2954, 2940, 2923, 2843, 1841,
1771, 1607, 1494. Anal. Calcd for C17H16O4: C, 71.82; H, 5.67.
Found: C, 71.94; H, 5.80.
F igu r e 1. Diels-Alder reaction of diene 2 with diketone 5:
Typical chromatograms of the reaction mixture before (a) and
after (b) tautomerization of rac-6 and rac-7.
reactions were analyzed after a short time, HPLC sug-
gested high selectivities in favor of rac-9. At the same
time, broad and intense peaks were present which later
disappeared (Figure 1). Simultaneous analysis of reaction
kinetics by HPLC, infrared, and UV-vis spectroscopy
gave good evidence that this effect is caused by the initial
cycloadducts, the electrophilic 1,2-diketones rac-6 and
rac-7. The water content of the HPLC solvent allows
formation of carbonyl hydrates during separation thus
causing peak broadening.18 While rac-7 rapidly tau-
tomerizes producing rac-9, the corresponding reaction of
rac-6 is slow. Reaction rates, therefore, had to be calcu-
lated on the disappearance of diene 2. Good rate en-
hancements were seen in all cases (Table 2). The het-
erocyclic compound 13 turned out to be distinctly more
efficient than palmitamidinium salt 12d . In the presence
of 1 equiv of 13, a 450-fold rate increase was observed.
Substoichiometric amounts of catalyst led to proportion-
(()-3-Meth oxy-16-oxa -8R,14â-gon a -1,3,5(10),9-tetr a en -
15,17-d ion e (r a c-4): mp 189-191 °C (lit.:15 mp 189.5-191 °C);
1H NMR (CDCl3) δ 7.49 (d, 1, J ) 8.7), 6.76 (dd, 1, J ) 8.7,
2.7), 6.66 (d, 1, J ) 2.7), 6.35-6.40 (m, 1), 3.81 (s, 3), 3.24 (td,
1, J ) 9.9, 7.9), 2.84-2.97 (m, 2), 2.73-2.78 (m, 2), 2.46-2.61
(m, 2), 2.27-2.39 (m, 1), 1.60-1.76 (m, 1). Anal. Calcd for
C
17H16O4: C, 71.82; H, 5.67. Found: C, 71.58; H, 5.82.
1-E t h oxy-1-h exa d eca n im in e, Tet r a flu or ob or a t e Sa lt
(11). A suspension of palmitoyl amide 1011 (25 g, 99 mmol) in
300 mL of dry CH2Cl2 was cooled to 0 °C, treated with a
solution of triethyloxonium tetrafluoroborate (26 g, 136 mmol)
in dry CH2Cl2 (50 mL), and stirred overnight at room temper-
ature. After filtration over Celite, the solution was concen-
trated in vacuo to a volume of 100 mL. Addition of 400 mL of
Et2O afforded 31 g (86%) of 11 as colorless crystals: mp 67-
(19) Schuster, T.; Bauch, M.; Du¨rner, G.; Go¨bel, M. W. Org. Lett.
2000, 2, 179.
(20) The observation of significant enantioselectivities gives further
evidence that amidinium catalysis is due to complex formation of cation
and dienophile. Chirality transfer from amidinium ions to cycloadduct
8 cannot be explained by simple acid catalysis.
(18) (a) Buschmann, H. J .; Fu¨ldner, H. H.; Knoche, W. Ber. Bunsen-
Ges. Phys. Chem. 1980, 84, 41. (b) Segretario, J . P.; Sleszynski, N.;
Partch, R. E.; Zuman, P.; Hora´k, V. J . Org. Chem. 1986, 51, 5393.

1700 J . Org. Chem., Vol. 65, No. 6, 2000
Schuster et al.
70 °C; 1H NMR (CDCl3) δ 9.37 (s, 1), 8.85 (s, 1), 4.49 (q, 2, J )
7.0), 2.66 (t, 2, J ) 7.7), 1.64-1.73 (m, 2), 1.51 (t, 3, J ) 7.0),
for each compound were obtained from the UV integrals: cf
) area(14)[compound]/area(compound)[14]: 2, cf ) 0.85 (tR
)
1.19-1.31 (m, 24), 0.88 (t, 3, J ) 6.8). Anal. Calcd for C18H38
-
32.6 min); rac-3, cf ) 0.27 (tR ) 19.1 min); rac-4, cf ) 0.27 (tR
) 21.9 min); rac-8, cf ) 0.22 (tR ) 18.1 min); rac-9, cf ) 0.23
(tR ) 19.4 min); 14, cf ) 1.00 (tR ) 28.4 min). The concentration
of each compound was then calculated: [compound] ) area-
(compound)[14]cf/area(14).
BF4NO: C, 58.23; H, 10.32; N, 3.77. Found: C, 58.01; H, 10.12;
N, 3.90.
Hexa d eca n a m id in e, Tetr a flu or obor a te Sa lt (12a ). 11
(9.46 g, 25 mmol) was dissolved in 100 mL of a saturated
ammonia solution in MeOH and stirred overnight at room
temperature. After filtration over Celite, the solvent was
removed in vacuo. Recrystallization of the residue from 1:2
dioxane/Et2O afforded 3.97 g (46%) of 12a as colorless crys-
tals: mp 108 °C; 1H NMR (DMSO-d6) δ 8.81 (s, 2), 8.31 (s, 2),
2.34 (t, 2, J ) 7.7), 1.59 (m, 2), 1.16-1.24 (m, 24), 0.86 (t, 3, J
) 6.5). Anal. Calcd for C16H35BF4N2: C, 56.15; H, 10.31; N,
8.18. Found: C, 55.92; H, 10.37; N, 8.08.
Hexadecan am idin e, Salt with Tetr aph en ylbor ate (12b).
A solution of 12a (2.60 g, 7.58 mmol) in 60 mL of MeOH was
treated with sodium tetraphenylborate (2.90 g, 8.47 mmol)
dissolved in 60 mL of MeOH. Addition of water precipitated
3.83 g (87%) of 12b as colorless crystals: mp 119 °C; 1H NMR
(DMSO-d6) δ 8.81 (s, 2), 8.36 (s, 2), 7.18 (m, 8), 6.93 (m, 8),
6.78 (m, 4), 2.33 (t, 2, J ) 7.7), 1.55-1.61 (m, 2), 1.20-1.28
(m, 24), 0.85 (t, 3, J ) 6.5); 1H NMR (CDCl3) δ 7.54 (m, 8),
7.03 (yt, 8), 6.83 (t, 4, J ) 7.2), 3.03 (s, 4), 1.11-1.31 (m, 24),
0.95 (t, 2, J ) 7.7), 0.88 (t, 3, J ) 6.8), 0.72 (m, 2). Anal. Calcd
for C40H55BN2: C, 83.60; H, 9.65; N, 4.87. Found: C, 83.36;
H, 9.65; N, 5.15.
Hexa d eca n a m id in e, Sa lt w ith Tetr a k is(4-ch lor op h en -
yl)bor a te (12c). A solution of 12a (0.69 g, 2.02 mmol) and
potassium tetrakis(4-chlorophenyl)borate (1.05 g, 2.12 mmol)
in 30 mL of 1:1 MeOH/H2O was treated with 30 mL of CH2Cl2
and stirred for 30 min. The organic layer was separated, the
aqueous layer was extracted with an additional portion of CH2-
Cl2, and the combined organic layers were dried over MgSO4.
Removal of the solvent in vacuo afforded 1.39 g (97%) of 12c
as a colorless oil: 1H NMR (CDCl3) δ 7.39-7.43 (m, 8), 7.04
(d, 8, J ) 8.2), 3.74 (s, 4), 1.43 (t, 2, J ) 7.4) 1.00-1.31 (m,
26), 0.88 (t, 3, J ) 6.6). Anal. Calcd for C40H51BCl4N2: C, 67.43;
H, 7.22; N, 3.93. Found: C, 67.36; H, 7.17; N, 3.81.
Hexa d eca n a m id in e, Sa lt w ith Tetr a k is(3,5-bis(tr iflu o-
r om eth yl)p h en yl)bor a te (12d ). A suspension of 12a (171
mg, 0.5 mmol) in 5 mL of MeOH was treated with a solution
of NaTFPB‚2H2O12 (461 mg, 0.5 mmol) in 5 mL of MeOH. After
stirring the clear solution for 10 min at room temperature,
the solvent was removed in vacuo and the residue treated with
CH2Cl2. The suspension was washed twice with H2O and the
organic layer dried over Na2SO4. Removal of the solvent in
vacuo afforded 553 mg (99%) of 12d as a colorless oil: 1H NMR
(CDCl3) δ 7.70 (s, 8), 7.55 (s, 4), 7.36 (s, 2), 6.43 (s, 2), 2.33 (t,
2, J ) 8.0), 1.52-1.60 (m, 2), 1.22-1.29 (m, 24), 0.87 (t, 3, J )
6.8). Anal. Calcd for C48H47BF24N2: C, 51.54; H, 4.23; N, 2.50.
Found: C, 51.44; H, 4.52; N, 2.28.
2-(Ben zyla m in o)p yr id in iu m Tetr a k is(3,5-bis(tr iflu o-
r om eth yl)p h en yl)bor a te (13). A solution of 2-(benzylamino)-
pyridine14 (92 mg, 0.5 mmol) in 10 mL of EtOH was treated
with 0.6 mL of 1 N HCl and stirred for 10 min at room
temperature. The solvent was removed in vacuo. The residue
was redissolved in 10 mL of EtOH and a solution of NaTFPB‚
2H2O12 (461 mg, 0.5 mmol) in 5 mL of EtOH was added. After
removal of the solvent in vacuo, the residue was taken up in
CH2Cl2 and washed twice with water. Drying of the organic
layer over MgSO4 and removal of the solvent in vacuo afforded
496 mg (94%) of 13 as a colorless solid: mp 78-84 °C; 1H NMR
(DMSO-d6) δ 13.29 (s, 1), 8.86 (s, 1), 7.93 (d, 1, J ) 6.1), 7.83
(t, 1, J ) 7.3), 7.70 (s, 4), 7.60 (s, 8), 7.29-7.40 (m, 5), 7.04 (d,
1, J ) 8.8), 6.85 (t, J ) 6.6), 4.55 (d, 2, J ) 5.3). Anal. Calcd
for C44H25BF24N2: C, 50.40; H, 2.40; N, 2.67. Found: C, 50.56;
H, 2.70; N, 2.63.
Ca libr a tion of th e HP LC System . A 4 µmol amount of 2,
rac-3, rac-4, rac-8, or rac-9 was dissolved in 10 mL of a 1.5
mM solution of 14.16 These solutions were analyzed by
HPLC: injection volume, 20 µL; solvent, MeCN/H2O; linear
gradient from 30 to 50% MeCN in 20 min, then 5 min of 50%
MeCN, and finally from 50 to 70% MeCN in 15 min; flow rate
1 mL min-1; UV detection at 260 nm. Calibration factors (cf)
Gen er a l P r oced u r e for th e Diels-Ald er Exp er im en ts.
Into a 1.5-mL polyethylene vial (Eppendorf tube) containing
catalysts 12d or 13 (0, 0.25, 0.5, 1, or 2 equiv) were added 250
µL of a 60 mM CH2Cl2 solution of dienophiles 1 or 5 and 250
µL of a 90 mM solution of the diene 2 in CH2Cl2 containing 5
µmol of the standard 14. The final concentrations at t ) 0 were
30 mM of dienophile, 45 mM of diene, 10 mM of 14, and 0,
7.5, 15, 30, or 60 mM of the catalyst. The reaction mixture
was stored at 7-8 °C in a cold room for 14 d. Aliquots (10 µL)
were taken after 1, 5, 15, 30, 60, 120, and 240 min and 1, 2, 4,
7, and 14 d, diluted with 500 µL of MeCN, and stored at -196
°C. The samples were analyzed by HPLC under the conditions
given above. The second-order rate constants k2 were obtained
by fitting the concentrations of diene 2 to the experimental
values of the first 4 h (variables: k2 and kdecomp; d[2] ) -k2-
[2][dienophile]dt - kdecomp[2]2dt). The concentration of 2 de-
creased independently from the Diels-Alder reaction by
decomposition. This minor effect could be empirically described
by the correction term
k
decomp[2]2dt. In the noncatalyzed
reaction of 2 and 5, an upper limit of k2 ) 4 × 10-8 mM-1 s-1
could be estimated. Rate accelerations shown in Table 2 are
based on this number.
Sim u lta n eou s An a lysis of th e Diels-Ald er Rea ction
by HP LC, UV, a n d IR. Into a 3-mL vial containing catalyst
13 (62.91 mg, 60 µmol) were added 1 mL of a 60 mM CH2Cl2
solution of 5 and 1 mL of a 90 mM solution of the diene 2 in
CH2Cl2 containing 20 µmol of the standard 14. The final
concentrations at t ) 0 were 30 mM of dienophile, 45 mM of
diene, 10 mM of 14, and 30 mM of 13. A 500 µL volume was
transferred into a 1 mm UV quartz cell and stored at 7 °C.
The rest of the reaction mixture was also stored at 7 °C and
used for IR and HPLC analysis. IR and UV spectra were
measured without further manipulation. Aliquots (10 µL) for
HPLC analysis were diluted with 500 µL of MeCN and
analyzed under the conditions given above. UV and IR spectra
and HPLC analyses were done after 15, 60, and 240 min and
2, 7, 14, and 21 d. Within 4 h, a broad band around 400 nm
developed in the UV spectra. In the IR spectra, the initial
diketone signals (1720, 1774 cm-1) were replaced by a signal
at 1747 cm-1. At the same time a broad peak appeared in the
HPLC chromatograms. When the concentration of rac-8 in-
creased in the HPLC chromatogram, these signals (UV, IR,
HPLC) bled and completely disappeared after 21 d. The
coincidence of the phenomena in all three analytical methods
gave strong evidence that the same intermediates are respon-
sible for each of the effects: the diketones rac-6 and rac-7.
Com p etition Exp er im en t. Into a 1.5-mL polyethylene vial
containing catalyst 13 (14.15 mg, 13.5 µmol) were added 150
µL of a 90 mM CH2Cl2 solution of 1, 150 µL of a 90 mM CH2-
Cl2 solution of 5, and 150 µL of a 72 mM solution of the diene
2 in CH2Cl2 containing 4.5 µmol of the standard 14. The final
concentrations at t ) 0 were 30 mM of each dienophile, 24
mM of diene, 10 mM of 14, and 30 mM of catalyst 13. The
reaction mixture was stored at 7-8 °C for 4 h. Samples were
taken after 1, 60, 120, 180, and 240 min and analyzed as
described above to determine the yields of rac-3 and rac-4.
After 3 h, diene 2 was no longer detectable. To prevent an
acylation of the keto enols rac-8 and rac-9 by the anhydrides,
an identical sample was quenched with 500 µL of water. After
4 d at room temperature, the tautomerization of rac-6 and
rac-7 was complete and allowed one to determine the yields
of rac-8 and rac-9. Compounds rac-3 and rac-4 were hydrolyzed
by this procedure.
Gen er a l P r oced u r e for th e 1H NMR Titr a tion . Into
eight different NMR tubes containing 200 µL of a 60 mM
solution of catalyst 12d or 13 in CDCl3 were added 0, 5, 10,
20, 40, 80, 160, or 320 µL of a 600 mM solution of dienophile
1 or 5 in CDCl3. The samples were then filled up to 600 µL

Catalysis of a Reaction by Amidinium Ions
J . Org. Chem., Vol. 65, No. 6, 2000 1701
with CDCl3. The final concentration were catalyst 20 mM and
dienophile 0, 5, 10, 20, 40, 80, 160, and 320 mM. The average
of the NH signals (12d ) or the CH2 signals (13) in the proton
spectra (7 °C) were applied for the determination of Ka.
Calculated shifts were adapted to the experimental values in
a nonlinear fitting procedure using Ka as the variable.
Associa tion Stu d ies by ESI-MS. A 1 mM solution of 13
containing 100 equiv of 1 and 100 equiv of 5 was investigated
by ESI-MS (injection temperature: 30 °C): m/z 296.3 (7, 13+‚
5), 295.1 (32, 13+‚5), 273.2 (6), 253.2 (10), 241.2 (10), 221.0
(11, 2 × 5 + H+), 186.2 (15, 13+), 185.1 (100, 13+). No signal
for a complex of 13 and 1 could be observed.
Ack n ow led gm en t. We are grateful to the Fonds der
Chemischen Industrie and to the Schweizerischer Na-
tionalfonds zur Fo¨rderung der wissenschaftlichen For-
schung for financial support. We also wish to express
our gratitude to Mrs. Ilona Priess, University of Frank-
furt, for the ESI mass spectra.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR data for
2, 5, rac-8, and rac-9. This material is available free of charge
via the Internet at http://pubs.acs.org.
J O991372X