Reactions of pyrazolylborate-zinc-hydroxide complexes related to β-lactamase activity

Article abstract of DOI:10.1021/ic050220j

Simple β-lactams and their hydrolysis products, the β-amino acids, react with Tp*Zn-OH under deprotonation. The latter become semibidentate carboxylate ligands with a NH·O hydrogen bond, and the former become N-bound β-lactamide ligands. Likewise the antibiotic derivatives 6-aminopenicillanic acid and 7-aminocephalosporanic acid are incorporated as carboxylate ligands. β-Lactams bearing nitrophenyl or acyl substituents at the nitrogen atoms are opened hydrolytically by Tp*Zn-OH, and the resulting N-substituted β-amino acids are attached to zinc by their carboxylate functions. Only with trifluoroacetyl as the N-substituent does the hydrolytic cleavage occur at the external amide bond, yielding the free β-lactam and Tp*Zn-trifluoroacetate. The kinetic investigation of the opening reactions has shown them to be of second order like all other Tp*Zn-OH-induced hydrolytic cleavages, thereby supporting the four-center mechanism for the monozinc β-lactamases.

Full text of DOI:10.1021/ic050220j

Inorg. Chem. 2005, 44, 4433−4440
Reactions of Pyrazolylborate
-Lactamase Activity
−Zinc−Hydroxide Complexes Related to
â
Florian Gross and Heinrich Vahrenkamp*
Institut fu¨r Anorganische und Analytische Chemie der UniVersita¨t Freiburg, Albertstrasse 21,
D-79104 Freiburg, Germany
Received February 10, 2005
Simple
The latter become semibidentate carboxylate ligands with a NH‚‚‚O hydrogen bond, and the former become N-bound
-lactamide ligands. Likewise the antibiotic derivatives 6-aminopenicillanic acid and 7-aminocephalosporanic acid
are incorporated as carboxylate ligands. -Lactams bearing nitrophenyl or acyl substituents at the nitrogen atoms
are opened hydrolytically by Tp*Zn OH, and the resulting N-substituted -amino acids are attached to zinc by
their carboxylate functions. Only with trifluoroacetyl as the N-substituent does the hydrolytic cleavage occur at the
external amide bond, yielding the free -lactam and Tp*Zn trifluoroacetate. The kinetic investigation of the opening
reactions has shown them to be of second order like all other Tp*Zn OH-induced hydrolytic cleavages, thereby
supporting the four-center mechanism for the monozinc -lactamases.
â-lactams and their hydrolysis products, the â-amino acids, react with Tp*Zn−OH under deprotonation.
â
â
−
â
â
−
−
â
Introduction
catalytic activity while the other can provide additional sub-
strate activation. If one breaks down this complex biochem-
ical scenario to the processes actually occurring at the metal
center3,5 this means that we are again faced with the hydro-
lytic function of (N3)Zn-OH2 or (N3)Zn-OH complexes.
We, together with G. Parkin, have provided the first mono-
nuclear example of such a (N3)Zn-OH complex in the form
of TptBu,MeZn-OH6 (TptBu,Me is hydrotris(3-tert-butyl-5-meth-
ylpyrazolyl)borate) and subsequently applied Tp*Zn-OH
complexes (Tp* denotes 3- and 5-substituted pyrazolylbor-
ates) for various hydrolytic reactions.7 During these studies
it was found that the Tp*Zn-OH nucleophiles are not strong
enough to cleave amide or peptide bonds. The amide bonds
in â-lactams, however, are more reactive due to the ring
tension, and they can be activated further by placing
appropriate substituents on the â-lactams. We were therefore
optimistic that with the right choice of reagents â-lactam
hydrolysis by Tp*Zn-OH complexes would be achieved.
A few bioinorganic chemistry papers dealing with this
subject have already appeared. Beyond the basic observation
that divalent and trivalent metal ions promote the opening
of the lactam ring, they involve the use of mono- and dizinc
complexes of polydentate ligands in the hydrolysis of
penicillin-derived â-lactams.8-11 Kinetic and mechanistic data
The â-lactamases belong to a new class of enzymes
resulting from the evolutionary process that is making
bacteria resistant against penicillin and cephalosporin deriva-
tives.1 They act by opening the â-lactam rings of these
antibiotics. Many of them are metalloenzymes containing
one or two zinc ions in their active centers. In the simplest
case one zinc ion is fixed to the protein by three histidine
residues and bears an aqua ligand which is deprotonated at
physiological pH,2 and this binding motif is also present in
all dizinc â-lactamases.3 Thus the â-lactamases represent the
classical (N3)Zn-OH nucleophile which is the hydrolytically
active ingredient in so many zinc enzymes including carbonic
anhydrase and the large group of the matrix metallopro-
teases.3
The high medicinal relevance has stimulated many studies
on the structure and activity of the metallo-â-lactamases as
well as on the function of the zinc ions and their amino acid
environment.1,3,4 The essence of these studies in terms of
the metal content is that one zinc ion is necessary for the
* Author to whom correspondence shoud be addressed. E-mail:
vahrenka@uni-freiburg.de.
(1) Page, M. I.; Laws, A. P. J. Chem. Soc., Chem. Commun. 1998, 1609.
(2) Bounaga, S.; Laws, A. P.; Galleni, M.; Page, M. I. Biochem. J. 1998,
331, 703.
(5) Kaim, W.; Schwederski, B. Bioinorganic Chemistry, 2nd ed.; Teub-
ner: Stuttgart, Germany, 1995.
(6) Alsfasser, R.; Trofimenko, S.; Looney, A.; Parkin, G.; Vahrenkamp,
H. Inorg. Chem. 1991, 30, 4098.
(3) Parkin, G. Chem. ReV. 2004, 104, 699.
(4) De Seny, D.; Prosperi-Meys, C.; Bebrone, C.; Rossolini, G. M.; Page,
M. I.; Noel, P.; Frere, M.; Galleni, M. Biochem. J. 2002, 363, 687
and references therein.
(7) Vahrenkamp, H. Acc. Chem. Res. 1999, 32, 589.
10.1021/ic050220j CCC: $30.25
Published on Web 05/19/2005
© 2005 American Chemical Society
Inorganic Chemistry, Vol. 44, No. 12, 2005 4433

Gross and Vahrenkamp
Table 1. Bond Lengths (Å) in â-lactams and Amides
C-N
1.32
were obtained for one mononuclear and one dinuclear model
complex,8,10 and theoretical calculations were performed for
a mononuclear situation modeled after the enzyme.11 Con-
sidering this early stage of development of the field, it
seemed worthwhile to us to make some contributions by
Tp*Zn-OH chemistry. For this purpose we used complexes
A and B. This paper describes our results.
CdO
1.24
1.226(1)
1.208(2)
1.200(2)
amides (av)13
â-propiolactam14
1.333(2)
1.376(2)
1.385(2)
4a
4b
at the nitrogen atom the amide C-N bond gets longer and
the amide C-O bond gets shorter. The resulting weakening
of the C-N bond reflects the diminished amide resonance.
Comparably weak C-N bonds are observed in â-lactams
with N-bromophenyl substituents.12 In line with this the Cd
O stretching frequency rises from amides (1600-1680
cm-1)13 to â-propiolactam (1727 cm-1) and further to 4a
(1755 cm-1) and 4b (1785 cm-1).
Two natural derivatives of penicillin and cephalosporin
were included in this investigation, the deacylated forms
6-aminopenicillanic acid (6) and 7-aminocephalosporanic
acid (7) which both are amino acids.
Results and Discussion
â-Amino Acids and â-Lactams. â-Amino acids are the
hydrolysis products of the â-lactams, and as such, they should
be bound to zinc as carboxylates after Tp*Zn-OH-induced
â-lactam cleavage. We therefore used three of them, â-ala-
nine, 3-methyl-â-alanine, and 3-phenyl-â-alanine for refer-
ence purposes.
Among the lactams, we started with the basic species
â-propiolactam (1a) and 4-phenyl-â-propiolactam (1b).
Blocking of their NH function was done by methyl groups
in N-methyl-â-propiolactam (2a) and N-methyl-4-phenyl-â-
propiolactam (2b). A slight activation of the lactams’ peptide
bonds was expected in N-phenyl-â-propiolactam (3a) and
N-phenyl-4-phenyl-â-propiolactam (3b). Nitrophenyl sub-
stituents were used for a stronger activation in N-p-
nitrophenyl-4-phenyl-â-propiolactam (4a) and N-2,4-dini-
trophenyl-4-phenyl-â-propiolactam (4b). Finally, activating
acyl groups were applied in N-acetyl- (5a), N-tert-butyryl-
(5b), N-benzoyl- (5c), and N-(trifluoroacetyl)-4-phenyl-â-
propiolactam (5d).
â-Amino Acid Complexes. These complexes were pre-
pared for reference purposes and for the purpose of finding
out whether â-amino acids would be analogous to R-amino
acids in acting as O,N-chelators toward zinc in the Tp*Zn
environment.15 Both Tp*Zn-OH complexes reacted smoothly
with the three amino acids producing complexes 8a-c and
9a-c in good yields.
1
The IR and H NMR data of complexes 8 and 9 clearly
distinguish them from the corresponding R-amino acid
complexes.15 The carboxylate CO stretching bands in the
1600-1620 cm-1 range are in the typical region for
complexes of simple carboxylates16,17 but 40-50 cm-1 higher
1
than those of the R-amino acid complexes.15 The H NMR
spectra show well-resolved sets of doublets and triplets for
the protons of the CHR-CH2 units indicating that the amino
acid ligands are fixed rather rigidly in similar positions.
(8) Koike, T.; Takamura, M.; Kimura, E. J. Am. Chem. Soc. 1994, 116,
8443.
(9) Ogawa, K.; Nakata, K.; Ichikawa, K. Chem. Lett. 1998, 797.
(10) Kaminskaia, N. V.; Spingler, B.; Lippard, S. J. J. Am. Chem. Soc.
2000, 122, 6411.
(11) Diaz, N.; Suarez, D.; Merz, K. M. J. Am. Chem. Soc. 2001, 123, 9867.
Diaz, N.; Scordo, T. L.; Suarez, D.; Mendez, R.; Martin-Villacorta, J.
New J. Chem. 2004, 15.
(12) Fujiwara, H.; Varley, R. L.; van der Veen, J. M. J. Chem. Soc., Perkin
Trans. 2 1977, 547.
(13) Barton, D., Ollis, W. D., Eds. ComprehensiVe Organic Chemistry;
Pergamon Press: Oxford, U.K., 1979; Vol. 2, p 990.
(14) Yang, Q. C.; Seiler, P.; Dunitz, J. D. Acta Crystallogr., Sect. C 1987,
43, 565.
(15) Rombach, M.; Gelinsky, M.; Vahrenkamp, H. Inorg. Chim. Acta 2002,
334, 25.
(16) Darensbourg, D. J.; Holtcamp, M. W.; Khandelwal, R.; Klausmeyer,
K. K.; Reibenspies, J. H. Inorg. Chem. 1995, 34, 2389.
(17) Ruf, M.; Vahrenkamp, H. Chem. Ber. 1996, 129, 1025.
To obtain a measure of their C-N activation, the structures
of 4a,b were determined. Details are given in the Supporting
Information. Table 1 lists the relevant bond lengths in
comparison to those of reference compounds. It is evident
that upon attachment of electron-withdrawing substituents
4434 Inorganic Chemistry, Vol. 44, No. 12, 2005

Reactions of Pyrazolylborate-Zn-Hydroxide Complexes
Figure 2. Molecular structure of 10a.
Figure 1. Molecular structure of 9b. Relevant bond lengths (Å): Zn-N1
2.024(5), Zn-N2 2.026(5), Zn-N3 2.093(5), Zn-O1 1.923(4), Zn‚‚‚O2
2.610(5), N7‚‚‚O2 2.895(5).
Table 2. Relevant Bond Distances (Å) and Angles (deg) in 10a,b
10a
10b
While this did not rule out chelation by Zn-N coordination,
it could be interpreted more easily by a NH‚‚‚O hydrogen
bond as the origin of rigidity. This was confirmed by the
structure determination of 9b; see Figure 1.
Zn-N
1.903(2)
3.310(3)
1.351(3)
1.480(3)
1.219(3)
1.907(3)
3.429(4)
1.351(4)
1.503(4)
1.225(4)
Zn‚‚‚O
N-C(CO)
N-C(CHR)
C-O
To a first approximation 9b contains a monodentate
carboxylate ligand. Yet the elongation of the Zn-N3 bond
and the Zn‚‚‚O2 distance of 2.61 Å indicate a weak Zn-O2
interaction which corresponds with the fact that N3 and O2
lie on the axial positions of a presumed trigonal bipyramid
around zinc. Thus, the carboxylate ligand can be termed
semibidentate, just like the ones in TpPhZn-benzylate,16 TpPh-
Zn-2-aminobenzoate,18 TpPh,MeZn-GlyPheBoc,15 and
TpCum,MeZn-2-hydroxypropionate.17 The rigidity of the amino
acid ligand is further established by the expected N-H‚‚‚O
hydrogen bond, which due to the N7‚‚‚O2 distance of 2.90
Å is of moderate strength. The amino acid residue fits well
into the pocket provided by the three cumenyl substituents,
thereby confirming that lactam hydrolysis leading to such a
zinc-â-amino acid complex should owe part of its driving
force to product stabilization.
Complexes of Intact â-Lactams. All nonactivated â-lac-
tams 1-3 were treated with the hydroxide complexes A and
B. The NH-containing lactams 1a,b did react with A. But
the products, which according to previous observations17
were expected to be â-amino acid complexes, were identified
by structure determinations as the lactamide complexes
10a,b. The related lactamide complex 11a, derived from B
and 1a, had already been obtained by us from TpCum,MeZn-
OMe and 1a.19 We had previously assigned the product of
the reaction between B and 1a as the â-amino acid complex
9a,17 which would be in accord with the spectra and
analytical data. However, the present findings have rendered
this assignment incorrect.
Zn-N-C(CO)
125.1(2)
141.1(2)
130.0(2)
136.6(2)
Zn-N-C(CHR)
relevant bonding details are given in Table 2. The â-lacta-
mides act as monodentate ligands, the Zn‚‚‚O separations
being beyond the bonding range. The Zn-N bonds are rather
short; cf. the average Zn-N(pyrazolylborate) bond lengths
of 2.05 Å. With about 1.90 Å they are at the lower end of
the range of bonds between zinc and anionic N-heterocy-
cles,19 and they can be compared with the Zn-O bond
lengths in Tp*Zn-hydroxides and -alkoxides.
The â-lactam rings are planar within less than 0.01 Å,
and the planes contain the zinc and oxygen atoms. Compared
to free â-propiolactam (see Table 1) their shape is almost
unchanged, the C-N bond being slightly longer and the
C-O bond being barely shorter. This means that the negative
charge of the lactam and its binding to zinc do not influence
the amount of single bond and double bond character in the
NCO array which represents the amide resonance in â-pro-
piolactam. In contrast to this, however, the CO vibration
bands, which for the free lactams are in the 1750 cm-1
range,21 are lowered to 1693-1699 cm-1 in complexes 10
and 11. This feature is also the major spectroscopic property
distinguishing the â-lactamide complexes from the â-amino
acid complexes (10a vs 8a and 10b vs 8c), which exhibit
their carboxylate bands in the 1610-1620 cm-1 range.
To avoid deprotonation and to direct the Tp*Zn-OH re-
actions to lactam ring opening, the N-methylated derivatives
of 1a,b, namely 2a,b were applied. But even under forcing
reaction conditions and in polar solvents they did not undergo
reactions with A. The same was the case with the N-
(18) Kremer-Aach, A.; Kla¨ui, W.; Bell, R.; Strerath, A.; Wunderlich, H.;
Mootz, D. Inorg. Chem. 1997, 36, 1552.
(19) Brombacher, H.; Vahrenkamp, H. Inorg. Chem. 2004, 43, 6054.
(20) Brombacher, H.; Vahrenkamp, H. Inorg. Chem. 2004, 43, 6042.
(21) Birkhofer, L.; Schramm, J. Liebigs Ann. Chem. 1975, 2195.
The structure of 10a is depicted in Figure 2 (for the similar
structure of 10b, see Supporting Information), and the
Inorganic Chemistry, Vol. 44, No. 12, 2005 4435

Gross and Vahrenkamp
phenylated lactams 3a,b. This means that the mild activation
of the amide bonds due to the N-phenyl resonance is not
sufficient.
In the antibiotic derivatives 6 and 7 the electron-withdraw-
ing carboxylic acid functions could be envisaged to act as
activators for amide bond cleavage in their â-lactam rings.
But at the same time they are good ligating units for the
Tp*Zn moieties. Reactions of TpPh,MeZn-OH (A) with 6 and
7 proved the preference for zinc-carboxylate formation.
Complexes 12 and 13 resulted in good yields under very
mild conditions.
Figure 3. Molecular structure of 14a. Relevant bond lengths (Å): Zn-
N1 2.082(2), Zn-N2 2.036(2), Zn-N3 2.024(2), Zn-O1 1.935(2), Zn‚‚‚
O2 2.584(2), N8‚‚‚O2′ 2.929(3).
The structural assignment of 12 and 13 rests in their
spectra. The 1H NMR spectra show the presence of the intact
units 6 and 7. The IR spectra show the amide bands around
1770 cm-1, the carboxylate bands around 1630 cm-1, and
in addition the ester band of 13 at 1740 cm-1. The presence
of the uncharged â-lactam ring in 12 and 13 induced us to
try to open it by reaction with another 1 equiv of A. Yet
there was no reaction at room temperature, and upon boiling
of the sample in toluene, unidentified decomposition products
resulted.
in a quasi-parallel fashion across a center of symmetry and
their ends are linked by N-H‚‚‚O bonds, generating a 12-
membered (OCCCNH)2 ring.
The reactivity of the N-acylated â-lactams 5a-c toward
TpPh,MeZn-OH (A) was intermediate between that of 4a and
4b. Within a few hours in boiling toluene they were
converted to the amino acid complexes 16a-c.
Lactam Ring Opening. The activation by electron-
withdrawing N-substituents in the â-lactams 4 and 5 was
strong enough to make them susceptible to nucleophilic
attack by Tp*Zn-OH and subsequent ring opening. Reac-
tions between the nitrophenyl-substituted lactams 4a,b and
the zinc complexes A and B proceeded in boiling toluene,
and the â-amino acid complexes 14a,b and 15b were isolated
in good yields. Lactam 4b reacted about 50 times faster than
lactam 4a.
The spectroscopic data showed that complexes 16 are
relatives of 14 and 15. The amide IR band of the free lactams
5 near 1790 cm-1 has disappeared and is replaced by the
carboxylate band between 1600 and 1640 cm-1, while the
amide band of the exocyclic acyl group near 1690 cm-1 is
shifted by only ca. 25 cm-1 to lower wavenumbers in the
1
complexes. Characteristic shifts of the H NMR signals
correspond to those in complexes 14 and 15. Hence, there
should be no doubt about the identity of 16a-c. Relatively
weak and low-wavenumber IR bands for the NH function
indicate that there are also N-H‚‚‚O hydrogen bonds, which
however cannot be located in the absence of a structure
determination.
The presence of two N-acyl functions in the lactams 5
implies the possibility of hydrolytic cleavage of the external
N-COR bonds. It turned out that this takes place for the
most strongly electron-withdrawing COR unit, the trifluo-
roacetyl function in 5d. The reaction between A and 5d at
room temperature took only a few minutes to produce the
free lactam 1b and the known complex TpPh,MeZn-
OCOCF3.17 The easiest way to follow this reaction was by
19F NMR, the spectra showing only the signals for 5d at
75.8 ppm and for the trifluoroacetate complex at 75.3 ppm.
Kinetic Investigation. The course of reaction of both A
and B with the â-lactam 4b was followed by 1H NMR. The
Tp*Zn-OH complexes were treated under pseudo-first-order
conditions with a 5- to 9-fold excess of 4b at 40 °C in CDCl3.
Complexes 14 and 15 were identified by their spectra. The
IR spectra show the NH band near 3300 cm-1 and the
carboxylate band near 1600 cm-1, and they lack the lactam
band near 1750 cm-1. The 1H NMR spectra display the NH
resonance, and all other resonances are shifted considerably
with respect to their position for the free lactams.
Complex 14a was subjected to a structure determination;
see Figure 3. The attachment of the amino acid to zinc
conforms with that in complex 9b; i.e., the carboxylate unit
is semi-bidentate. Bond lengths and angles in the TpZn
moiety as well as in the amino acid are normal. The major
difference between 9b and 14a consists of the hydrogen-
bonding pattern. While it is intramolecular in 9b, it is
intermolecular in 14a in such a way that the amino acid
chains of two neighboring complex molecules are aligned
4436 Inorganic Chemistry, Vol. 44, No. 12, 2005

Reactions of Pyrazolylborate-Zn-Hydroxide Complexes
cleavages of tris(p-nitrophenyl) phosphate by A and B are
characterized by k′′ values of 1.55 and 0.45 M-1 s-1,
respectively, and the cleavage of p-nitrophenyl acetate by B
has yielded a k′′ of 0.35 × 10-2 M-1 s-1.22 In all cases the
strong activation by nitrophenyl substituents is necessary to
enable hydrolysis, in accord with the hidden nature of the
Zn-OH nucleophile in the Tp* ligand pocket and the
completely nonpolar reaction environment. When one wants
to compare phosphate, ester, and lactam hydrolysis by
Tp*Zn-OH, one should use a k′′ for the mononitrophenyl-
substituted 4a which could not be determined here but which
can be estimated (see above) to be close to 2 orders of
magnitude smaller than that for 4b. If taken as such, it is in
the same range as that for the cleavage of p-nitrophenyl
actate. Then the sequence of reactivities (phosphate . ester
≈ â-lactam) corresponds to that which has classically been
established for group transfers from these species.13
Figure 4. Logarithmic plot of the intensity of the benzylic proton 1H NMR
signal in 14b for the cleavage of 4b by A.
Conclusions
As observed before it was found in this work too that the
available zinc-hydroxide complexes are not potent enough
as nucleophiles to effect ring opening of simple, nonactivated
â-lactams. When the lactams contain a NH function, though,
they are deprotonated by Tp*Zn-OH and attached to zinc
as the novel â-lactamide ligands. By using the hydrolysis
products of the â-lactams, the â-amino acids, as such, it could
be shown that they make stable Tp*Zn complexes in which
the shape of the carboxylate-bound amino acid ligands is
fixed by NH‚‚‚CO hydrogen bonds. Likewise the â-lactam
antibiotic derivatives 6-aminopenicillanic acid and 7-ami-
nocephalosporanic acid are attached to the Tp*Zn moieties
as carboxylates.
To have the lactams cleaved by Tp*Zn-OH they must
be activated by N-bound nitrophenyl or acyl substituents.
The products of ring opening end up as zinc-bound â-amino
acid ligands, which can easily be identified by their
characteristic spectral features. As a rule the N-acyl activated
â-lactams undergo hydrolytic cleavage only at the lactam
acyl function. Only in case of N-trifluoroacetyl activation is
the exocyclic acetyl function cleaved yielding the free lactam
and a zinc-trifluoroacetate complex.
The kinetic investigation of the Tp*Zn-OH-induced ring
opening reaction has shown it to be a clean second-order
process whose rate depends noticeably on the accessibility
of the Zn-OH function inside the Tp* ligand pocket. This
is in agreement with the generally accepted four-center
mechanism for Zn-OH-induced hydrolytic reactions and our
specific formulation of it for Tp*Zn systems.7,22 For the
specific case of â-lactams the essential intermediate should
look like the one depicted below. The relevant conclusion
from this is that â-lactam hydrolysis, like all other zinc
enzyme catalyzed hydrolytic processes, needs only one zinc
ion. This finding complements similar observations from
biochemical1,3 as well as from model complex studies.8,10
The fact that several such enzymes, specifically the metallo-
Figure 5. Plot of the kobs values for the cleavage of 4b by B against five
concentrations of the lactam.
The intensities of the signals for the benzylic proton in the
products 14b and 15b were recorded as a function of time.
Pseudo-first-order rate constants were obtained according to
ln(1 - It/I∞) ) -kobst. The resulting kobs values, plotted
against the lactam concentration, define a regression line,
the slope of which is the second-order rate constant. The
quality of the measurements, expressed by correlation
coefficients around 0.995, can be judged from Figures 4 and
5. It should be noted that the plot of Figure 5 has an intercept
at 0.5 × 10-5 s-1, indicating that there is some independent
lactam hydrolysis which is zero order in lactam concentra-
tion. This, however, does not affect the calculation of k′′.
The two second-order rate constants k′′ obtained are 0.57
M-1 s-1 for A and 0.13 M-1 s-1 for B. The fact that the
phenyl-substituted Tp*Zn-OH complex A reacts about four
times faster than the cumenyl-sunstituted B is in agreement
with the more hindered access to the Zn-OH center in the
latter. The scarcity of investigations on zinc complex induced
â-lactam hydrolyses makes it difficult to compare our kinetic
data with reported ones. The only study reporting a second-
order rate constant for a related reaction system is that by
Kimura using a zinc-cyclen complex.8 The reported k′′ value
for the hydrolysis of benzylpenicillin in water (4 × 10-2
M-1 s-1) is about 1 order of magnitude smaller than the k′′
values observed here, which (taking into account the strongly
accelerating influence of the water solvent) reflects the
significantly reduced reactivity of benzylpenicillin in com-
parison to 4b.
On the other side, the second-order rate constants observed
here can be compared with those for other hydrolytic
reactions induced by Tp*Zn-OH complexes. Thus, the
(22) Rombach, M.; Maurer, C.; Weis, K.; Keller, E.; Vahrenkamp, H.
Chem.sEur. J. 1999, 5, 1013.
Inorganic Chemistry, Vol. 44, No. 12, 2005 4437

Gross and Vahrenkamp
Complex 8a. A solution of A (200 mg, 0.35 mmol) in
dichloromethane (20 mL) was added to a suspension of â-alanine
(50 mg, 0.56 mmol) in methanol (20 mL). After the mixture was
stirred for 15 h and filtered, the solvent was removed in vacuo.
The raw product was purified three times by dissolving in
dichloromethane and precipitating with n-heptane. A 142 mg (62%)
amount of 8a remained as a colorless powder, mp 124 °C. IR (KBr;
cm-1): 2545 w (BH), 1611 s (CO). 1H NMR (CDCl3): δ 1.53 (br,
4H, H2O + NH2), 1.82 (t, J ) 5.9 Hz, 2H, CH2), 2.41 (t, J ) 5.9
Hz, 2H, CH2). Anal. Calcd for C33H34BN7O2Zn‚H2O (Mr ) 636.88
+ 18.02): C, 60.52; H, 5.54; N, 14.97. Found: C, 60.45; H, 5.71;
N, 14.72.
Complex 8b. This was synthesized like 8a from A (300 mg,
0.53 mmol) and D,L-3-methyl-â-alanine (66 mg, 0.64 mmol).
Purification by dissolution in dichloromethane and slow addition
of n-heptane yielded 254 mg (78%) of 8b as a colorless powder,
mp 169 °C. IR (KBr; cm-1): 2553 w (BH), 1598 s (CO). 1H NMR
(CDCl3): δ 0.81 (d, J ) 6.5 Hz, 3H, CH3), 1.52 (br, 2H, NH2),
1.58 (dd, J ) 16.7 and 9.4 Hz, 1H, CH), 1.81 (dd, J ) 16.7 and
3.1 Hz, 1H, CH), 2.62 (m, 1H, CH). Anal. Calcd for C34H36BN7O2-
Zn (Mr ) 650.91): C, 62.74; H, 5.57; N, 15.06. Found: C, 62.64;
H, 5.62; N, 14.93.
â-lactamases, contain two zinc ions in the active center is
not contradictory to this, as various supporting functions for
the second zinc ion can be envisaged.
Experimental Section
General Data. All experimental techniques and the standard IR
and NMR equipment were as described previously.23 The Tp*Zn-
OH complexes were prepared as described.24,25 The lactams 1b,26
2a,27 2b,27 3a,28 3b,29 4a,28 5a,29 5b,29 and 5c29 were prepared by
following the literature procedures. Lactams 4b and 5d had not
been described yet. All other organic reagents were purchased from
Merck and Aldrich. The IR, 1H NMR, and 13C NMR spectral data
for the TpPh,Me and TpCum,Me ligands in the new complexes vary
only negligibly between themselves and the reference com-
pounds.15,17,19,20,24,25 Therefore, only the data for the coligands X
in the Tp*Zn-X complexes are reported here.
Lactam 4b. A solution of 1b (1.00 g, 6.79 mmol), 2,4-
dinitroiodobenzene (2.00 g, 6.79 mmol), and K2CO3 (939 mg, 6.79
mmol) in anhydrous dioxane (120 mL) was treated with 20 mg
(0.11 mmol) of CuI and heated to 80 °C for 3d. After being cooled
to room temperature, the mixture was filtered and the filtrate
evaporated to dryness. The residue was dissolved in 30 mL of ethyl
acetate and the product precipitated by slow addition of 60 mL of
cyclohexane. The precipitate was dissolved in 10 mL of dichlo-
romethane and filtered through a 1 cm layer of silica gel, and 30
mL of a 2:1 mixture of cyclohexane and ethyl acetate was added.
Slow evaporation in an open flask yielded 750 mg (35%) of 4b as
yellow crystals, mp 143 °C. IR (KBr; cm-1): 1785 vs (CO), 1388
Complex 8c. This was synthesized like 8a from A (200 mg,
0.35 mmol) and D,L-3-phenyl-â-alanine (66 mg, 0.40 mmol).
Purification by slow evaporation of a acetonitrile/dichloromethane
solution yielded 204 mg (82%) of 8c as a colorless powder, mp
1
168 °C. IR (KBr; cm-1): 2543 m (BH), 1616 s (CO). H NMR
(CDCl3): δ 1.50 (br, 2H, NH2), 2.03 (m, 2H, CH2), 3.64 (dd, J )
9.7 and 3.4 Hz, 1H, CH), 7.21 (m, 5H, Ph). Anal. Calcd for C39H38-
BN7O2Zn (Mr ) 712.98): C, 65.70; H, 5.37; N, 13.75. Found: C,
65.55; H, 5.67; N, 14.58.
Complex 9a. This was synthesized like 8a from B (200 mg,
0.29 mmol) and â-alanine (28.5 mg, 0.32 mmol). Purification by
dissolution in dichloromethane and slow addition of n-heptane
yielded 150 mg (67%) of 9a as a colorless powder, mp 198 °C. IR
1
vs (NO). H NMR (CDCl3): δ 3.21 (dd, J ) 15.9 and 3.3 Hz,
1H), 3.70 (dd, J ) 15.9 and 6.0 Hz, 1H), 5.34 (dd, J ) 3.3 and 6.0
Hz, 1H), 7.35 (s, 5H, Ph), 7.96 (d, J ) 9.0 Hz, 1H, C6H3), 8.35
(dd, J ) 9.0 and 2.5 Hz, 1H, C6H3), 8.60 (d, J ) 2.5 Hz, 1H,
C6H3). Anal. Calcd for C15H11N3O5 (Mr ) 313.27): C, 57.51; H,
3.54; N, 13.41. Found: C, 57.37; H, 3.45; N, 13.33.
1
(KBr; cm-1): 2533 w (BH), 1612 m (CO). H NMR (CDCl3): δ
1.59 (br, 2H, NH2), 1.82 (t, J ) 5.8 Hz, 2H, CH2), 2.44 (t, J ) 5.8
Hz, 2H, CH2). Anal. Calcd for C42H52BN7O2Zn (Mr ) 763.12):
C, 66.11; H, 6.87; N, 12.85. Found: C, 65.47; H, 6.99; N, 12.52.
Complex 9b. This was synthesized like 8a from B (200 mg,
0.29 mmol) and D,L-3-methyl-â-alanine (33 mg, 0.32 mmol).
Purification by slow evaporation from a solution in dichloromethane
and n-heptane yielded 160 mg (71%) of 9b as a colorless crystals,
mp 197 °C. IR (KBr; cm-1): 2550 m (BH), 1618 s (CO). 1H NMR
(CDCl3): δ 0.83 (d, J ) 6.5 Hz, 3H, CH3), 1.58 (dd, J ) 16.1 and
9.3 Hz, 1H, CH), 1.82 (dd, J ) 16.1 and 3.4 Hz, 1H, CH), 2.68
(m, 1H, CH). Anal. Calcd for C43H54BN7O2Zn (Mr ) 777.15): C,
66.46; H, 7.00; N, 12.62. Found: C, 66.19; H, 7.11; N, 12.45.
Complex 9c. This was synthesized like 8a from B (200 mg,
0.29 mmol) and D,L-3-phenyl-â-alanine (53 mg, 0.32 mmol).
Purification by slow evaporation from a solution in dichloromethane
and n-heptane yielded 182 mg (75%) of 9c as a colorless crystals,
mp 207 °C. IR (KBr; cm1): 2542 m (BH), 1612 s (CO). 1H NMR
(CDCl3): δ 1.51 (br, 2H, NH2), 2.01 (m, 2H, CH2), 3.82 (dd, J )
8.8 and 4.5 Hz, 1H, CH), 7.23 (m, 11H, therein C6H5).
Lactam 5d. A solution of 1b (300 mg, 2.04 mmol) in anhydrous
dichloromethane was treated dropwise with stirring first with 453
mg (630 µL, 4.48 mmol) of triethylamine and then upon cooling
with ice with a solution of 471 mg (312 µL, 2.24 mmol) of
trifluoroacetic anhydride in 5 mL of dichloromethane. After the
solution was stirred for 15 h without renewing the ice, a solution
of 2 g of NH4Cl in 15 mL of water was added, and the mixture
was shaken and extracted with three times 20 mL of dichlo-
romethane. The combined extracts were dried over Na2SO4. After
filtration the filtrate was evaporated to dryness. The remaining
colorless oil was chromatographed with dichloromethane over a 5
× 20 cm silica gel column (Rf ) 0.62). The resulting colorless oil
had a purity of 95% according to 1H NMR. The yield was 267 mg
1
(54%). H NMR (CDCl3): δ 3.17 (dd, J ) 17.0 and 3.9 Hz, 1H),
3.68 (dd, J ) 17.0 and 7.0 Hz, 1H), 5.16 (dd, J ) 7.0 and 3.9 Hz,
1H), 7.38 (m, 5H, Ph). 19F NMR (CDCl3): δ -75.84.
Complex 10a. A solution of 1a (50 mg, 0.71 mmol) in
dichloromethane (10 mL) was added dropwise with stirring to a
solution of A (400 mg, 0.71 mmol) in dichloromethane (20 mL).
After the solution was stirred for 2 h, the solvent was removed in
vacuo. Crystallization from acetonitrile yielded 337 mg (73%) of
10a as a colorless crystals, mp 212 °C. IR (KBr; cm-1): 2538 m
(23) Fo¨rster, M.; Burth, R.; Powell, A. K.; Eiche, T.; Vahrenkamp, H. Chem.
Ber. 1993, 126, 2643.
(24) Ruf, M.; Burth, R.; Weis, K.; Vahrenkamp, H. Chem. Ber. 1996, 129,
1251.
(25) Ruf, M.; Vahrenkamp, H. Inorg. Chem. 1996, 35, 6571.
(26) Graf, R. Liebigs Ann. Chem. 1963, 661, 111.
(27) Yamawaki, J.; Ando, T.; Hanafusa, T. Chem. Lett. 1981, 1143.
(28) Takahata, H.; Ohnishi, Y.; Jamazaki, T. Heterocycles 1980, 14, 467.
(29) Bergmann, H. J.; Otto, H. H. Arch. Pharm. (Weinheim, Ger.) 1986,
319, 635.
1
(BH), 1693 vs (CO). H NMR (CDCl3): δ 1.86 (t, J ) 4.1 Hz,
5H, CH2), 2.00 (s, 3H, CH3CN), 2.32 (t, J ) 4.1 Hz, 2H, CH2).
4438 Inorganic Chemistry, Vol. 44, No. 12, 2005

Reactions of Pyrazolylborate-Zn-Hydroxide Complexes
Table 3. Crystallographic Data
4a
4b
9b
10a
10b
14a
formula
MW
space group
Z
C15H12N2O3
268.27
P21/n
C15H11N3O5
313.27
P21/c
C43H54BN7O2Zn
777.15
P21/c
C33H32BN7OZn‚CH3CN
659.91
P1h
2
C39H36BN7OZn‚CH3CN
C39H34BN8O4Zn
757.95
P1h
2
736.01
P21/n
4
4
4
4
a (Å)
b (Å)
c (Å)
6.092(1)
19.191(3)
11.121(2)
90
95.130(3)
90
1294.9(4)
1.38
0.10
10.364(2)
10.198(2)
13.386(2)
90
92.835(4)
90
1413.0(4)
1.47
0.11
8.946(5)
22.526(12)
21.761(11)
90
110.30(2)
90
4113(4)
1.26
0.64
11.759(1)
11.963(1)
13.543(1)
71.861(2)
75.386(2)
67.943(2)
1658.0(2)
1.32
11.818(1)
16.134(1)
19.442(1)
90
93.135(1)
90
3701.5(4)
1.32
0.71
8.401(2)
12.294(3)
18.272(4)
102.644(4)
93.796(4)
98.556(4)
1811.3(7)
1.39
R (deg)
â (deg)
γ (deg)
V (Å3)
d(calcd) (g cm-3
)
µ(Mo KR) (mm1)
R1 (obsd reflcns)
wR2 (all reflcns)
0.78
0.041
0.109
0.73
0.037
0.124
0.042
0.100
0.041
0.107
0.075
0.284
0.054
0.171
Anal. Calcd for C33H32BN7OZn‚CH3CN (Mr ) 618.86 + 41.05):
C, 63.70; H, 5.35; N, 16.98. Found: C, 63.60; H, 5.37; N, 16.64.
Complex 10b. This was synthesized like 10a from A (141 mg,
0.25 mmol) and 1b (37 mg, 0.25 mmol). Yield: 120 mg (65%) of
10b as a colorless crystals, mp 238 °C. IR (KBr; cm-1): 2561 w
4.56 (m, 1H, CH), 6.51 (d, J ) 9.7 Hz, 1H, C6H3), 7.95 (dd, J )
9.7 and 2.8 Hz, 1H, C6H3), 9.04 (d, J ) 2.8 Hz, 1H, C6H3), 9.13
(d, br., J ) 6.5 Hz, 1H, NH). Anal. Calcd for C45H40BN9O6Zn (Mr
) 879.07): C, 61.48; H, 4.59; N, 14.34. Found: C, 60.88; H, 4.94;
N, 14.53.
Complex 15b. This was synthesized like 14a from B (150 mg,
0.22 mmol) and 4b (68 mg, 0.22 mmol) in 20 mL of toluene for
10 h. Yield: 141 mg (64%) of 15b as a yellow powder, mp 221
°C. IR (KBr; cm-1): 3312 w (NH), 2548 w (BH), 1617 vs (CO),
1335 vs (NO). 1H NMR (CDCl3): δ 2.26 (d, J ) 5.3 Hz, 2H, CH2),
4.68 (m, 1H, CH), 6.62 (d, J ) 9.5 Hz, 1H, C6H3), 7.95 (dd, J )
9.5 and 2.8 Hz, 1H, C6H3), 9.13 (d, J ) 2.8 Hz, 1H, C6H3), 9.13
(d, J ) 2.8 Hz, 1H, C6H3), 9.75 (d, br, J ) 6.8 Hz, 1H, NH). Anal.
Calcd for C54H58BN9O6Zn (Mr ) 1005.31): C, 64.52; H, 5.81; N,
12.54. Found: C, 64.38; H, 6.23; N, 12.41.
Complex 16a. A solution of A (100 mg, 0.18 mmol) and 5a
(34 mg, 0.18 mmol) in toluene (25 mL) was refluxed for 24 h. The
solvent was removed in vacuo and the residue dissolved in 5 mL
of dichloromethane. n-Heptane was added in such an amount as to
barely avoid a precipitation. Then the mixture was left to evaporate,
yielding a precipitate of 77 mg (56%) of 16a, mp 202 °C. IR (KBr;
cm-1): 2544 m (BH), 1674 vs (CO), 1605 s (CO). 1H NMR
(CDCl3): δ 1.58 (s, br, 2H, H2O), 1.64 (s, 3H, CH3), 2.20 (dd, J )
17.1 and 5.5 Hz, 1H, CH), 2.35 (dd, J ) 17.1 and 4.9 Hz, 1H,
CH), 4.93 (m, 1H, CH), 6.97 (d, br, J ) 7.5 Hz, 1H, NH). Anal.
Calcd for C41H40BN7O3Zn‚H2O (Mr ) 755.01 + 18.02): C, 63.70;
H, 5.48; N, 12.68. Found: C, 64.26; H, 5.38; N, 12.84.
Complex 16b. This was synthesized like 16a from A (200 mg,
0.35 mmol) and 5b (82 mg, 0.35 mmol). Yield: 177 mg (62%) of
16b as a colorless powder, mp 212 °C. IR (KBr; cm-1): 2557 m
(BH), 1664 s (CO), 1603 m (CO). 1H NMR (CDCl3; cm-1): 0.85
(s, 9H, t-Bu), 1.54 (s, 2H, H2O), 2.13 (dd, J ) 16.8 and 5.5 Hz,
1H, CH), 2.40 (dd, J ) 16.8 and 4.5 Hz, 1H, CH), 5.02 (m, 1H,
CH). Anal. Calcd for C44H46BN7O3Zn‚H2O (Mr ) 797.09 +
18.02): C, 64.84; H, 5.94; N, 12.03. Found: C, 65.23; H, 5.83; N,
11.85.
1
(BH), 1699 vs (CO). H NMR (CDCl3): δ 1.96 (s, 3H, CH3CN),
2.08 (m, 1H, CH), 2.86 (m, 1H, CH), 3.79 (br., 1H, CH), 6.10 (m,
2H, C6H5), 6.73 (m, 2H, C6H5), 6.84 (m, 1H, C6H5). Anal. Calcd
for C39H36BN7OZn‚CH3CN (Mr ) 694.96 + 41.05): C, 66.91; H,
5.34; N, 15.22. Found: C, 66.46; H, 5.44; N, 14.78.
Complex 12. 6-Aminopenicillanic acid (97 mg, 0.45 mmol) was
added to a solution of A (196 mg, 0.35 mmol) in dichloromethane
(40 mL). After the solution was stirred for 5 h, the solvent was
removed in vacuo and the residue dissolved in 3 mL of dichlo-
romethane. Slow addition of 25 mL of a n-heptane/diethyl ether
(4:1) mixture precipitated 226 mg (85%) of 12 as a colorless
powder, mp 170 °C. IR (KBr; cm-1): 2547 m (BH), 1768 vs (CO),
1638 s (CO). 1H NMR (CDCl3): δ 0.84 (s, 3H, CH3), 1.30 (s, 3H,
CH3), 3.77 (s, 1H, CH), 4.34 (d, J ) 4.1 Hz, 1H, CH), 5.12 (d, J
) 4.1 Hz, 1H, CH). Anal. Calcd for C38H39BN8O3SZn (Mr )
764.04): C, 59.74; H, 5.14; N, 14.67; S, 4.20. Found: C, 59.47;
H, 5.60; N, 14.79; S, 3.86.
Complex 13. This was synthesized like 12 from A (200 mg,
0.35 mmol) and 7-aminocephalosporanic acid (106 mg, 0.39 mmol).
Yield: 197 mg (68%) of 13 as a light brown powder, mp 130 °C.
IR (KBr; cm-1): 2552 w (BH), 1779 s (CO), 1740 s (CO), 1616
m (CO). 1H NMR (CDCl3): δ 1.93 (s, 3H, CH3), 3.12 (d, J ) 17.7
Hz, 1H, CH), 3.30 (d, J ) 17.7 Hz, 1H, CH), 4.11 (d, J ) 13.9
Hz, 1H, CH), 4.27 (d, J ) 13.9 Hz, 1H, CH), 4.50-4.90 (m, 2H,
CHCH). Anal. Calcd for C40H39BN8O5SZn (Mr ) 820.07): C,
58.59; H, 4.79; N, 13.66; S, 3.91. Found: C, 58.39; H, 5.22; N,
13.30; S, 4.20.
Complex 14a. A solution of A (200 mg, 0.35 mmol) and 4a
(67 mg, 0.35 mmol) in toluene (40 mL) was refluxed for 6d. Then
the solvent was removed in vacuo. Crystallization from acetonitrile
yielded 178 mg (67%) of 14a as yellow crystals, mp 218 °C. IR
(KBr; cm-1): 3319 m (NH), 2541 m (BH), 1593 s (CO), 1303 vs
Complex 16c. This was synthesized like 16a from A (200 mg,
0.35 mmol) and 5c (89 mg, 0.35 mmol). Yield: 230 mg (78%) of
16c as a colorless powder, mp 204 °C. IR (KBr; cm-1): 2548 m
(BH), 1659 s (CO), 1639 m (CO). 1H NMR (CDCl3): δ 2.27 (dd,
J ) 16.9 and 5.8 Hz, 1H, CH), 2.45 (dd, J ) 16.9 and 3.9 Hz, 1H,
CH), 5.25 (m, 1H, CH), 8.15 (d, br., J ) 8.8 Hz, 1H, NH). Anal.
Calcd for C46H42BN7O3Zn (Mr ) 817.08): C, 67.62; H, 5.18; N,
12.00. Found: C, 67.28; H, 5.58; N, 11.56.
1
(NO). H NMR (CDCl3): δ 2.02 (t, J ) 5.9 Hz, 2H, CH2), 2.88
(m, 2H, CH2), 5.03 (m, 1H, NH), 6.17 (d, J ) 9.3 Hz, 2H, C6H4),
8.01 (d, J ) 9.3 Hz, 2H, C6H4). Anal. Calcd for C39H37BN8O4Zn
(Mr ) 757.97): C, 61.80; H, 4.92; N, 14.78. Found: C, 61.46; H,
5.14; N, 14.87.
Complex 14b. This was synthesized like 14a from A (200 mg,
0.35 mmol) and 4b (111 mg, 0.35 mmol) in 20 mL of toluene for
3 h. Yield: 209 mg (68%) of 14b as a yellow powder, mp 120 °C.
IR (KBr; cm-1): 3294 w (NH), 2549 m (BH), 1618 vs (CO), 1333
Reaction of A with 5d. In an NMR tube 5.4 mg (0.02 mmol)
of 5d and 8.8 mg (0.015 mmol) of A were dissolved in 600 µL of
CDCl3. After 15 min the 1H NMR signals of the starting materials
1
vs (NO). H NMR (CDCl3): δ 2.32 (d, J ) 6.0 Hz, 2H, CH2),
Inorganic Chemistry, Vol. 44, No. 12, 2005 4439

Gross and Vahrenkamp
17
had been replaced by those of TpPh,MeZn-OCOCF3 and 1b, and
diffractometer with Mo KR radiation and subjected to empirical
absorption corrections (SADABS). The structures were solved with
direct methods and refined anisotropically using the SHELX
program suite.30 Hydrogen atoms were included with fixed distances
and isotropic temperature factors 1.2 times those of their attached
atoms. Parameters were refined against F2. Drawings were produced
with SCHAKAL.31 Table 3 lists the crystallographic details.
in the 19F NMR spectrum the resonance for 5d at 75.78 ppm had
been replaced by that for TpPh,MeZn-OCOCF3 at 75.29 ppm.
Kinetic Measurements. Stock solutions were prepared of A,
B, and 4b in CDCl3 (99.8%). All reagents and the cavity of the
NMR spectrometer were thermostated to 314 K before the measure-
ments. The reagents were combined immediately prior to the
measurements. The concentrations of A or B were adjusted to 0.030
M for all measurements and to 0.15, 0.18, 0.21, 0.24, and 0.27 M
Acknowledgment. This work was supported by the Fonds
der Chemischen Industrie. We are indebted to Drs. B.
Kersting, H. Brombacher, and C. Perez Olmo for help with
the structure determinations.
1
for 4b, respectively. The intensity of the H NMR resonance for
the benzylic proton of 14b (during the reaction with A) and the
intensity of the 1H NMR resonance for the pyrazolyl-CH proton of
15b (during the reaction with B) were recorded automatically every
30 s and stored for digital data processing. Each kinetic run was
repeated once to ensure reproducibility. The averaged data were
used for the calculations. The resulting kobs values for the given
Supporting Information Available: Fully labeled ORTEP plots
and X-ray crystallographic files in CIF format for the six structure
determinations. This material is available free of charge via the
Internet at http://pubs.acs.org.
concentrations of 4b were 0.95 × 10-4, 1.08 × 10-4, 1.29 × 10-4
,
1.45 × 10-4, and 1.62 × 10-4 s-1 for the reaction with A and 2.51
× 10-5, 2.84 × 10-5, 3.29 × 10-5, 3.54 × 10-5, and 4.07 × 10-5
s-1 for the reaction with B.
IC050220J
(30) Sheldrick, G. M. SHELX, program package of the Bruker AXS Smart
CCD diffractometer, version 5.1; Bruker: Madison, WI, 2002.
(31) Keller, E. SCHAKAL for Windows; Universita¨t Freiburg: Freiburg,
Germany, 2001.
Structure Determinations. Crystals resulting from the workup
procedures (see above) were used for the structure determinations.
Data sets were obtained at 240 K with a Bruker AXS Smart CCD
4440 Inorganic Chemistry, Vol. 44, No. 12, 2005