New H-bond accepting tris(pyrazolyl)borates: Stabilization of metal aquo species as models for the vicinal oxygen chelate enzyme superfamily

Article abstract of DOI:10.1039/b009621k

The hydrogen-bond accepting ligand K[Tp^CO2Et,Me] was prepared and its complexes with divalent Ni, Co, Mn and Cu reveal a stabilization of penta- or hexa-coordinate metal aquo complexes of the general form [Tp ^CO2Et,MeM(H₂O)_x](x = 2 or 3) which are unprecedented for the alkyl substituted Tp analogs. These complexes make good structural models for many of the enzymes of the vicinal oxygen chelate superfamily.

Full text of DOI:10.1039/b009621k

View Article Online / Journal Homepage / Table of Contents for this issue
New H-bond accepting tris(pyrazolyl)borates: stabilization of metal
aquo species as models for the vicinal oxygen chelate enzyme
superfamily
Brian S. Hammes, Mary W. Carrano and Carl J. Carrano*
Department of Chemistry and Biochemistry, Southwest Texas State University, San Marcos,
TX 78666 USA. E-mail: cc05@swt.edu
Received 30th November 2000, Accepted 6th February 2001
First published as an Advance Article on the web 10th April 2001
The hydrogen-bond accepting ligand K[TpCO Et,Me] was prepared and its complexes with divalent Ni, Co, Mn and Cu
2
reveal a stabilization of penta- or hexa-coordinate metal aquo complexes of the general form [TpCO Et,MeM(H2O)x]
2
(x = 2 or 3) which are unprecedented for the alkyl substituted Tp analogs. These complexes make good structural
models for many of the enzymes of the vicinal oxygen chelate superfamily.
tural and/or reactivity models for metalloenzymes, surprisingly
few well characterized systems of the kinetically labile divalent
Introduction
The resting state of the catalytically active metal center in
first row transition metals containing multiple adjacent water
many metalloenzymes contains one or more water molecules in
molecules are known.
addition to the ligands supplied by the protein backbone. These
Based on the enormous success of the tris(pyrazolyl)borate,
ubiquitous metal bound water molecules generally provide one
TpR, platform6,7 as a ligand for metalloenzyme modeling
of two major functions: (1) a ligand that is labile, and thus
easily replaceable for inner sphere coordination at the metal by
studies, we describe here a first generation of a simple H-bond
accepting analog which simultaneously provides both steric
incoming substrates; (2) as a nucleophile in the actual catalytic
bulk and hydrogen bonding groups which can be directed
reaction. The number of metalloenzymes containing such sites
towards the center of a metal binding cavity. The presence of
are extremely numerous and include enzymes involved in DNA
strong intramolecular H-bonding between the ligand and water
(phosphodiester), amide, or ester hydrolysis, oxygen activation,
molecules does indeed lead to the stabilization of discrete, well
isomerizations etc. This diversity is well illustrated by the so
characterized, divalent metal aquo complexes which are both,
unprecedented in Tp chemistry, and good structural models for
many of the enzymes of the VOC superfamily.
called vicinal oxygen chelate (VOC) superfamily.1 The enzymes
of this family use a variety of different divalent metal ions
in catalysis (Co, Ni, Zn, Mn and Fe) but from an inorganic
perspective, the common thread in this structurally related
group is that they all provide a metal coordination environment
Results and discussion
with two or three “open” (i.e. water containing) coordination
sites, which promote direct electrophilic participation by the
metal ion in catalysis. While mechanisms of action for many of
these enzymes are readily proposed, little hard data concerning
the various steps centered at the metal ion are available.
Based on ease of synthesis and the desire to stabilize the
H-bond donating ligand, water, we chose ester substituted
tris(pyrazolyl)borates as our target. Thus the carboxyethyl
substituted, three-fold symmetric, TpCO Et,Me ligand 1, has been
2
prepared from the corresponding pyrazole and potassium
borohydride in one step using standard melt conditions.6
The ligand has been completely characterized by standard
methodology and its structure verified in the metal complexes
reported below.
One of the well established methods to study the reactivity
inherent in a particular structure and/or ligand environment of
complex molecules such as metalloenzymes is the synthesis,
structural characterization, and reactivity studies of small
synthetic analogs to these proteins i.e. the so called “model
compound” approach. This approach has been highly suc-
cessful in generating spectroscopic, structural and reactivity
models for many metalloproteins.2,3 At the heart of the model
compound approach is the design and synthesis of creative
ligands to mimic metal ion binding sites. Good models for
enzymes of the VOC superfamily could be based on a facially
coordinating tripodal chelate which leaves two or three sites
open at the metal center where water could be bound. However
such water molecules would normally be extremely labile,
making isolation and characterization of such species problem-
atic. Nature herself often stabilizes such reactive or labile
structures through various secondary interactions, the most
important of which is hydrogen bonding. While the use of
sterically restricted ligands to control the size and shape of
metal binding cavities is well established, it has proven more
difficult, despite its importance, to incorporate hydrogen
bonding into synthetic small molecule binding site analogs.4,5
Thus despite the importance of metal aquo complexes, as struc-
We were extremely gratified to find that simple cationic, metal
aquo-complexes of the general type [TpCO Et,MeM(H2O)x]ϩ (M =
2
Ni() 2, Co() 3, Mn() 4 or Cu() 5) where x = 3 for octa-
hedral favoring Ni(), Co() or Mn() and x = 2 for square
pyramidal Cu(), were in fact formed when first row transition
metal salts of poorly coordinating counter ions were reacted
with the ligand in MeOH or THF. The perchlorate salts of
complexes 2–5 have all been crystallographically characterized
and the water molecules in these complexes are strongly
hydrogen bonded to the ester carbonyl moieties, although the
exact H-bonding pattern is different for each. The structure of
the representative octahedral Ni() complex 2 is shown in Fig. 1
where all the water hydrogens were found on the difference map.
Important bond lengths and angles in the coordination sphere
of the Ni are shown in the figure caption. One hydrogen of each
water molecule is hydrogen bonded to an ester carbonyl oxygen
while the other is hydrogen bonded to an oxygen of the
perchlorate group (not shown). The resulting rather tight ion
pair accounts for the high solubility of this complex in polar
1448
J. Chem. Soc., Dalton Trans., 2001, 1448–1451
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b009621k

View Article Online
Fig. 1 Thermal ellipsoid diagram of the cationic portion of 2,
[(TpCO Et,Me)Ni(H2O)3]ϩ. The ellipsoids are drawn at the 30% prob-
2
ability level and hydrogens, except those involved in H-bonding are
removed for clarity. Selected bond distances (Å) and angles (Њ): Ni(1)–
N(1) 2.083(4), Ni(1)–N(3) 2.085(4), Ni(1)–N(5) 2.111(4), Ni(1)–O(7)
2.082(4), Ni(1)–O(8) 2.091(4), Ni(1)–O(9) 2.064(4); N(1)–Ni(1)–N(3)
90.4(2), N(1)–Ni(1)–N(5) 88.0(2), N(1)–Ni(1)–O(7) 177.8(2), N(1)–
Ni(1)–O(8) 88.4(2), N(1)–Ni(1)–O(9) 92.2(2), N(3)–Ni(1)–N(5) 87.3(2),
N(3)–Ni(1)–O(7) 89.0(2), N(3)–Ni(1)–O(8) 91.3(2), N(3)–Ni(1)–O(9)
176.9(2), N(5)–Ni(1)–O(7) 93.6(2), N(5)–Ni(1)–O(8) 176.5(2), N(5)–
Ni(1)–O(8) 91.0(2), O(7)–Ni(1)–O(8) 89.6(2), O(7)–Ni(1)–O(9) 88.5(2),
O(8)–Ni(1)–O(9) 90.6(2).
Fig. 3 Thermal ellipsoid diagram of [(TpCO Et,Me)Cu(H2O)2]ClO4 5.
The ellipsoids are drawn at the 30% probability level and hydrogens,
except those involved in H-bonding are removed for clarity. Selected
bond distances (Å) and angles (Њ): Cu(1)–N(1) 1.981(5), Cu(1)–N(3)
1.991(4), Cu(1)–N(5) 2.322(4), Cu(1)–O(7) 1.986(5), Cu(1)–O(8)
1.981(5); N(1)–Cu(1)–N(3) 85.1(2); N(1)–Cu(1)–N(5) 94.7(2), N(1)–
Cu(1)–O(7) 93.0(2), N(3)–Cu(1)–N(5) 86.1(2), O(8)–Cu(1)–N(1)
175.2(2), O(8)–Cu(1)–N(5) 88.7(2), O(7)–Cu(1)–N(5) 96.8(2), O(8)–
Cu(1)–O(7) 89.9(2), O(8)–Cu(1)–N(3) 91.7(2), O(7)–Cu(1)–N(3)
176.7(2).
2
lengths. In 4 one water molecule (O8) is H-bonded to two ester
carbonyl oxygens (O3 and O1), another water (O7) is H-bonded
to ester carbonyl O5 and a solvent water (O15) while the final
water (O9) interacts with ester carbonyl O1 and perchlorate
oxygen O13. The average Ocarbonyl–Owater distance is 2.71 Å and
Ocarbonyl–H–Owater angle is 153.1Њ. The structure of the square
pyramidal Cu complex 5 is shown in Fig. 3. In this case only
one of the water molecules (O8) is H-bonded to an ester
carbonyl (O5) with an Ocarbonyl–Owater distance of 2.69 Å and
a Ocarbonyl–Hwater–Owater angle of 162.4Њ. The other water is
H-bonded only to the perchlorate.
These complexes are good structural models for some of
the members of the VOC enzyme family. The Ni complex is a
model for Escherichia coli glyoxalase I, the Mn complex a model
for the fosfomycin resistance protein and a Mn extradiol
dioxygenase and the Co complex for methylmalonyl-CoA
epimerase.1 In these enzymes the metal is typically bound to
two histidine nitrogens, one or two aspartate carboxyl oxygens
and two or three mutually cis waters in an overall octahedral
geometry. The Cu complex is an interesting structural model
for the binding site in copper amine oxidases, whose square
pyramidal Cu() ion is surrounded by three histidine nitrogens
and two water molecules.8
To be other than structural models for enzymes it is import-
ant that the water molecules be easily displaced by “substrate”
ligands. In solution there is evidence that the water molecules
on these metal complexes are in fact labile. Thus in pure frozen
CH2Cl2 the Mn complex 4 is not very soluble but shows a six
line pattern indicative of a single Mn environment. Addition of
acetonitrile to increase the solubility gives the frozen solution
EPR spectrum shown in Fig. 4 where both a major and a minor
species can be observed. Other preliminary results indicate
that the waters are easily displaced by bidentate ligands such as
catechol, amino acids etc. to give stable “enzyme-substrate”
analogs as well. Details of these solution based studies will be
reported in full later.
Fig. 2 Thermal ellipsoid diagram of [(TpCO Et,Me)Mn(H2O)3]ClO4ؒ
2
H2O 4. The ellipsoids are drawn at the 30% probability level and
hydrogens, except those involved in H-bonding are removed for clarity.
Selected bond distances (Å) and angles (Њ): Mn(1)–N(1), 2.254(4),
Mn(1)–N(3) 2.273(4), Mn(1)–N(5) 2.307(4), Mn(1)–O(8) 2.165(4),
Mn(1)–O(9) 2.151(4), Mn(1)–O(7) 2.156(4); N(1)–Mn(1)–N(3) 83.6(1),
N(1)–Mn(1)–N(5) 83.36(12), N(1)–Mn(1)–O(8) 93.5(2), N(1)–Mn(1)–
O(9) 92.8(2), N(3)–Mn(1)–N(5) 85.90(13), N(3)–Mn(1)–O(8) 92.3(2),
N(3)–Mn(1)–O(9) 175.5(2), N(5)–Mn(1)–O(8) 176.6(2), N(5)–Mn(1)–
O(9) 91.0(2), O(8)–Mn(1)–O(9) 90.5(2), O(9)–Mn(1)–O(7) 89.6(2),
O(7)–Mn(1)–O(8) 93.2(2).
organic solvents and its low solubility in water. The average
Ocarbonyl–Owater distance of 2.69 Å and the average Ocarbonyl
–
Hwater–Owater angle of 165.4Њ are commensurate with strong
hydrogen bonding between the water ligands and the Tp ester
carbonyls. The fact that all of the ester carbonyls are pointed
into the metal binding cavity gives the tripodal ligand in this
complex a rather open or “splayed out” conformation as com-
pared to that seen in structures lacking the H-bonding inter-
action where the carbonyls are all pointed out. Despite this, no
evidence for the bis-ligand, “sandwich” complex is observed.
The perchlorate salt of the Co complex 3 is isomorphous and
isostructural with the Ni analog.
In conclusion, a series of divalent metal aquo complexes have
been prepared that are stabilized by internal H-bonding with a
tris(pyrazolyl)borate based tripodal ligand. These complexes are
unprecedented for simple Tp ligands which almost invariably
yield tetrahedral species, TpMX, if the ligand is highly sterically
Mn() also forms an octahedral complex with three facially
coordinated water molecules (Fig. 2). Again all the relevant
hydrogens were located on the difference map and these were
refined independently but constrained to have similar bond
J. Chem. Soc., Dalton Trans., 2001, 1448–1451
1449

View Article Online
(CH3CN): 386 (10), 612 (10). Magnetism (solid state, room
temperature): µeff = 2.9 µB.
[(TpCO Et,Me)Co(H2O)3]NO3 (3)
2
A slurry of potassium tris(3-methyl-5-carboxyethyl)pyrazolyl-
borate 1 (1 g, 2.1 mmol) and Co(NO3)2ؒ6H2O (0.6 g, 2.1 mmol)
in 100 mL of THF was stirred for 6 h and dried under reduced
pressure. The purple-pink solid was taken up into CH2Cl2 and
extracted with water. The organic layer was dried and the
solvent removed by rotary evaporation. Crystals for an X-ray
diffraction study were grown by the slow evaporation of
an acetone solution (pink) containing the Co() complex.
Anal. Calc. (found) for [(TpCO Et,Me)Co(H2O)3]NO3, C21H34N7-
2
Fig. 4 Frozen solution X-band (9.6435 GHz) EPR spectrum of 4 in
mixed CH2Cl2–CH3CN solvent, 10 K, 20 µW/40 dB, 2 G mod. amp.
O12BCo: C, 39.02 (39.30); H, 5.31 (5.21); N, 15.17 (15.23%).
FTIR (KBr, cmϪ1): ν = 2563 (B–H), 1708 (C᎐O). λmax/nm
᎐
(ε/MϪ1 cmϪ1): 504 (20). Magnetism (solid state, room temper-
ature): µeff = 4.8 µB. The perchlorate salt can be prepared using
cobalt() perchlorate as the metal source.
encumbered or simple (Tp)2M “sandwich” complexes if not.6
They represent good structural models for enzymes of the VOC
superfamily.
[(TpCO Et,Me)Mn(H2O)3]ClO4 (4)
2
Experimental
A slurry of potassium tris(3-methyl-5-carboxyethyl)pyrazolyl-
borate 1 (0.31 g, 0.61 mmol) in 20 mL of CH3OH was treated
dropwise with a solution of Mn(ClO4)2ؒ6H2O (0.22 g, 0.61
mmol) in CH3OH. The resulting solution was stirred for 6 h and
dried under reduced pressure. The off-white solid was taken up
into CH3CN and filtered to remove a small amount of white
solid. Crystals for an X-ray diffraction study were grown by the
slow diffusion of diethyl ether into a CH3CN solution. FTIR
All syntheses were carried out in air and the reagents and
solvents purchased from commercial sources and used as
received unless otherwise noted. Methanol was distilled under
argon over CaH2. Ethyl 2,4-diketopentanoate was prepared
using previously reported procedures. CAUTION! The per-
chlorate salts used in this study are potentially explosive and
should be handled with due caution.
(KBr, cmϪ1): ν = 2575 (B–H), 1715 (C᎐O). Magnetism (solid
᎐
3-Methyl-5-carboxyethylpyrazole
state, room temperature): µeff = 6.27 µB.
Ethyl 2,4-diketopentanoate (85.5 g, 0.54 mol) was dissolved in
400 mL of MeOH and treated with an excess (48 g, 0.70 mol) of
aqueous hydrazine monohydrochloride. The reaction mixture
was stirred at room temperature for 15 min and extracted
with CH2Cl2, dried over MgSO4 and volatiles removed under
reduced pressure. The crude product was recrystallized from
boiling hexane to yield 45 g, 54% of 3-methyl-5-carboxy-
ethylpyrazole. 1H NMR (CDCl3) δ 6.60 (s, 1 H, PzH), 4.35 (q, 2
H, J = 7 Hz, –OCH2–), 2.40 (s, 3 H, Pz-CH3), 1.34 (t, 3 H, J = 7
Hz, –CH3). 13C NMR (CDCl3) δ 161.16, 143.19, 141.04, 107.41,
61.10, 14.13, 11.32.
[(TpCO Et,Me)Cu(H2O)2]ClO4 (5)
2
A slurry of potassium tris(3-methyl-5-carboxyethyl)pyrazolyl-
borate 1 (0.31 g, 0.61 mmol) in 20 mL of CH3OH was treated
dropwise with a solution of Cu(ClO4)2ؒ6H2O (0.23 g, 0.61
mmol) in CH3OH. The resulting solution was stirred for 6 h
and dried under reduced pressure. The blue-green solid was
taken up into CH3CN and filtered to remove a small amount of
white solid. Crystals for an X-ray diffraction study were
grown by the slow diffusion of diethyl ether into a CH3CN
solution of [(TpCO Et,Me)Cu(H2O)2]ClO4. Anal. Calc. (found) for
2
[(TpCO Et,Me)Cu(H2O)2]ClO4ؒ0.25CH3CN,
C21.5H35.5N6.25O12-
2
Potassium tris(3-methyl-5-carboxyethyl)pyrazolylborate (1)
BClCu: C, 37.94 (38.12); H, 4.86 (4.81); N, 12.86 (12.65%).
A total of 8 g (0.052 mol) of 3-methyl-5-carboxyethylpyrazole
was mixed with 0.93 g (0.0173 mol) of potassium borohydride
and heated slowly to 150 ЊC until hydrogen generation ceased.
The melt temperature was then raised to 170 ЊC and the tem-
perature maintained until hydrogen evolution again ceased and
the initial fluid reaction mixture had solidified. The mixture
was cooled briefly and quenched with toluene. Filtration of the
white solid, brief washing with toluene, followed by hexane,
FTIR (KBr, cmϪ1): ν = 2531 (B–H), 1735 (C᎐O), 1701 (C᎐O).
᎐
᎐
λmax/nm (ε/MϪ1 cmϪ1) (CH3CN): 736 (34). Magnetism (solid
state, room temperature): µeff = 1.54 µB.
Physical methods
Elemental analyses were performed on all compounds by
Quantitative Technologies, Inc., Whitehouse, NJ. All samples
were dried in vacuo prior to analysis. The presence of solvates
was corroborated by FTIR, 1H NMR or X-ray crystallography.
1H and 13C NMR spectra were collected on a Varian UNITY
INOVA 400 MHz NMR spectrometer. Chemical shifts are
reported in ppm relative to an internal standard of TMS. The
13C quaternary carbon peaks that are not observed are a result
of either poor solubility and/or overlapping signals. IR spectra
were recorded from KBr disks on a Perkin-Elmer 1600 Series
FTIR spectrometer and are reported in wavenumbers. Elec-
tronic spectra were recorded using a Hewlett-Packard 8452A
diode array spectrophotometer. Room-temperature magnetic
susceptibility measurements of the metal complexes were
determined using a magnetic susceptibility balance MSB-1
manufactured by Johnson Matthey and calibrated with mer-
cury() tetrathiocyanatocobaltate() (χg = 16.44(8) × 10Ϫ6 cm3
1
and drying gave 6.2 g (70%) of the desired product. H NMR
(CDCl3) δ 6.30 (s, 1 H, PzH), 4.26 (q, 2 H, J = 7 Hz, –OCH2–),
2.42 (s, 3 H, Pz-CH3), 1.32 (t, 3 H, J = 7 Hz, –CH3). 13C NMR
(CDCl3) δ 164.78, 144.22, 142.66, 106.13, 60.18, 14.36, 12.66.
FTIR (KBr, cmϪ1): ν = 2523 (B–H), 1715 (C᎐O). MS (ESI,
᎐
methanol) m/z 471 (MϪ).
[(TpCO Et,Me)Ni(H2O)3]ClO4 (2)
2
A slurry of potassium tris(3-methyl-5-carboxyethyl)pyrazolyl-
borate 1 (0.32 g, 0.63 mmol) in 20 mL of CH3OH was treated
dropwise with a solution of Ni(ClO4)2ؒ6H2O (0.23 g, 0.63
mmol) in CH3OH. The resulting solution was stirred for 6 h and
dried under reduced pressure. The light blue solid was taken up
into CH3CN and filtered to remove a small amount of white
solid. Crystals for an X-ray diffraction study were grown by the
slow diffusion of diethyl ether into a CH3CN solution. FTIR
g
Ϫ1).3 Diamagnetic corrections were taken from those reported
by O’Connor.9 Low temperature EPR spectra were obtained on
(KBr, cmϪ1): ν = 2559 (B–H), 1708 (C᎐O). λmax/nm (ε/MϪ1 cmϪ1
)
a Bruker X-band spectrometer with Oxford helium cryostat
᎐
1450
J. Chem. Soc., Dalton Trans., 2001, 1448–1451

View Article Online
in the Department of Biophysics at the Medical University
of Lübeck, Germany. We thank Dr Volker Schünemann for
assistance in acquiring these spectra.
Acknowledgements
This work was supported in part by Grants AI-1157 from the
Robert A. Welch Foundation and CHE-9726488 from the NSF.
The NSF-ILI program grant USE-9151286 is acknowledged for
partial support of the X-ray diffraction facilities at Southwest
Texas State University.
Crystallography
Crystal data for 2. C21H34BClN6O13Ni, M = 683.51, a =
9.419(2), b = 18.220(2), c = 18.166(2) Å, β = 93.378(12)Њ,
V = 3112.1(7) Å3, monoclinic, space group P21/c, Z = 4,
T = 293(2) K, final R1 = 0.0525, wR2 = 0.1391, GOF (on
F2) = 0.976.
References
1 N. Armstrong, Biochemistry, 2000, 39, 13625.
2 N. Kitajima, K. Fujisawa, C. Fujimoto, Y. Moro-oka, S. Hashimoto,
T. Kitagawa, K. K. Toriumi, M. Tatsumi and A. Nakamura, J. Am.
Chem. Soc., 1992, 114, 1277.
3 N. Kitajima, K. Fujisawa, M. Tanaka and Y. Moro-oka, J. Am.
Chem. Soc., 1992, 114, 9232.
Crystal data for 4ؒCH2Cl2ؒH2O. C22H34BCl3N6O14Mn,
M = 778.65, a = 11.8559(12), b = 12.9935(13), c = 12.9957(13)
Å, α = 72.828(9), β = 67.473(10), γ = 187.706(9)Њ, V = 1760.7(3)
3
¯
Å , triclinic, space group P1, Z = 2, T = 293(2) K, final
4 B. S. Hammes, V. G. Young and A. S. Borovik, Angew. Chem., Int.
Ed., 1999, 38, 666.
R1 = 0.0579, wR2 = 0.1656, GOF (on F2) = 1.043.
5 L. M. Berreau, R. A. Allred, M. M. Makowska-Grzyska and A. M.
Arif, Chem. Commun., 2000, 1423.
6 S. Trofimenko, Scorpionates: The Coordination Chemistry of
Polypyrazolylborate Ligands, Imperial College Press, London, UK,
1999.
Crystal data for 5ؒ2H2O. C21H30BClN6O15Cu, M = 716.31,
a = 12.106(3), b = 12.6519(11), c = 13.012(2) Å, α = 106.982(11),
β = 108.75(3), γ = 103.97(2)Њ, V = 1676.7(5) Å3, triclinic, space
¯
group P1, Z = 2, T = 293(2) K, final R1 = 0.0602, wR2 = 0.1766,
7 G. Parkin, Chem. Commun., 2000, 1971.
GOF (on F2) = 1.131.
8 M. C. Wilce, D. M. Dooley, H. C. Freeman, J. M. Guss,
H. Matsunami, W. S. McIntire, C. E. Ruggiero, K. Tanizawa and
H. Yamaguchi, Biochemistry, 1997, 36, 16116.
CCDC reference numbers 159090, 161749 and 161750.
See http://www.rsc.org/suppdata/dt/b0/b009621k/ for crystal-
lographic data in CIF or other electronic format.
9 C. J. O’Connor, Prog. Inorg. Chem., 1982, 29, 203.
J. Chem. Soc., Dalton Trans., 2001, 1448–1451
1451