H-bonding interactions and control of thiolate nucleophilicity and specificity in model complexes of zinc metalloproteins

Article abstract of DOI:10.1021/ic048630f

It is shown in model complexes designed to mimic the binding site of zinc-thiolate proteins that a single hydrogen bond between an amide N-H and a Zn-coordinated thiolate reduces its reactivity toward electrophiles by up to 2 orders of magnitude. In addition, we show that a single N-H...S hydrogen bond is sufficient to achieve near 100% regiospecificity of reaction between a strong, and hence inherently indiscriminate, alkylating agent like trimethyloxonium tetraflouroborate and a single sulfur in a dithiolate construct. The importance of these results in understanding how systems such as the zinc fingers of the GATA family and the E. coli DNA repair protein Ada, which share the same pseudotetrahedral structure and tetracysteinyl ligation around the zinc, can fulfill such widely divergent (structural vs reactive) roles and how specificity of reaction in such multi-thiolate containing systems can be achieved is discussed.

Full text of DOI:10.1021/ic048630f

Inorg. Chem. 2005, 44, 2012−2017
H-Bonding Interactions and Control of Thiolate Nucleophilicity and
Specificity in Model Complexes of Zinc Metalloproteins
Jennifer N. Smith,† Justin T. Hoffman,‡ Zahida Shirin,† and Carl J. Carrano*,‡
Departments of Chemistry and Biochemistry, San Diego State UniVersity,
San Diego, California 92182-1030, and Texas State UniVersity, San Marcos, Texas 78666
Received September 30, 2004
It is shown in model complexes designed to mimic the binding site of zinc-thiolate proteins that a single hydrogen
bond between an amide N H and a Zn-coordinated thiolate reduces its reactivity toward electrophiles by up to 2
orders of magnitude. In addition, we show that a single N ‚‚‚S hydrogen bond is sufficient to achieve near 100%
−
−H
regiospecificity of reaction between a strong, and hence inherently indiscriminate, alkylating agent like trimethyloxonium
tetraflouroborate and a single sulfur in a dithiolate construct. The importance of these results in understanding how
systems such as the zinc fingers of the GATA family and the E. coli DNA repair protein Ada, which share the same
pseudotetrahedral structure and tetracysteinyl ligation around the zinc, can fulfill such widely divergent (structural
vs reactive) roles and how specificity of reaction in such multi-thiolate containing systems can be achieved is
discussed.
Introduction
mental evidence favors a Zn(II)-bound thiolate as the
nucleophile.6-9 Another, as yet, unanswered enigma is how
two systems with essentially identical structure and ligation
(compare for example the zinc fingers of the GATA family
and the DNA repair protein Ada which both share the same
pseudotetrahedral structure and CCCC ligation around the
zinc) can fulfill widely divergent roles. In the latter case,
the zinc thiolate functions as a reactive nucleophile that
removes a methyl group from the phosphotriester backbone
of alkylation damaged DNA while in the former case any
such reactivity would be detrimental to its purely structural
role. A related question requiring an answer is how, in a
thiol rich construct such as that found in the Ada protein, is
one specific zinc thiolate bond made to be reactive (cys 38
in this particular case) despite the similar steric accessibility
of several such bonds.5 Since it is well-known that the zinc-
thiolate centers in the zinc fingers are involved in extensive
N-H‚‚‚S hydrogen bond networks with backbone amides,
it has been proposed that such interactions might be a
controlling factor in regulating reactivity and specificity of
such bonds.10 Although a statistical analysis of many Zn-S
Thiolate coordination is an integral part of the coordination
sphere of many of the hundreds of zinc metalloproteins
known.1 These sulfur rich zinc centers fulfill one of two basic
functional roles, either as structural elements, as for example
in the zinc fingers and some metalloregulatory proteins,2,3
or as a reactive entity for alkyl group transfer as seen in the
methionine synthesizing enzymes MetE and MetH of E. coli
or in the DNA repair protein Ada.4,5 However, a large number
of unanswered questions remain about how the presence of
the zinc modulates the reactivity of a thiolate anion. One
such question is whether the reactive nucleophile in Zn(II)-S
systems is a Zn(II)-bound thiolate or a dissociated thiolate
anion. At issue is the nature of the nucleophile that leads to
the transition state in this overall SN2 reaction. In previous
work, we and others have shown, at least for electroneutral
complexes designed to mimic zinc thiolate protein coordina-
tion in nonpolar solvents, that the majority of the experi-
* To whom correspondence should be addressed. E-mail: carrano@
sciences.sdsu.edu.
† Texas State University.
‡ San Diego State University.
(6) Hammes, B. S.; Warthen, C. R.; Crans, D. C.; Carrano, C. J. J. Biol.
Inorg. Chem. 2000, 6, 82.
(7) Brand, U.; Rombach, M.; Seebacher, J.; Vahrenkamp, H. Inorg. Chem.
2001, 40, 6151-6157.
(8) Chiou, S. J.; Innocent, J.; Riordan, C. G.; Lam, K. C.; Liable-Sands,
L.; Rheingold, A. L. Inorg. Chem. 2000, 39, 4347.
(9) Bridgewater, B. M.; Fillebeen, T.; Friesner, R. A.; Parkin, G. J. Chem.
Soc., Dalton Trans. 2000, 4494.
(1) Lipscomb, W. N.; Straeter, N. Chem. ReV. 1996, 96, 2375.
(2) Berg, J. M.; Shi, Y. Science 1996, 271, 1081.
(3) Outten, C. E.; O’Halloran, T. V. Science 2001, 292, 2488.
(4) (a) Matthews, R. G.; Goulding, C. W. Curr. Opin. Chem. Biol. 1997,
1, 332-339. (b) Hightower, K. E.; Fierke, C. A. Curr. Opin. Chem.
Biol. 1999, 3, 176-81.
(5) Sun, L. J.; Yim, C. K.; Verdine, G. L. Biochemistry 2001, 40, 11596.
2012 Inorganic Chemistry, Vol. 44, No. 6, 2005
10.1021/ic048630f CCC: $30.25
© 2005 American Chemical Society
Published on Web 02/15/2005

Model Complexes of Zn Metalloproteins
Chart 1
Experimental Section
All syntheses were initially carried out under inert atmosphere,
but subsequent workups were conducted in air. The reagents and
solvents were purchased from commercial sources and used as
received unless otherwise noted. EtOH and THF were distilled under
argon over magnesium turnings/I2 and Na/benzophenone, respec-
tively. CDCl3 was purchased from Aldrich and used as received.
Ligands L1-4 and their corresponding methyl and thiophenolate
zinc complexes were synthesized by literature methods.15-17 2,2′-
Dithio-bis(N-phenylacetamide) and 2,2′-dithio-bis(N-phenyl-2,2,2-
trifluoroacetamide) were prepared by reported procedures.18-20
[(L1)Zn(S-o-CH3CONHC6H4)], 1. Solid NaBH4 (0.02 g, 0.53
mmol) was added to a THF-ethanol (9:1, 25 mL) solution of 2,2′-
dithiobis(N-phenylacetamide) (0.17 g, 0.51 mmol). The mixture was
stirred at room temperature for about 2 h to give a clear solution,
to which [(L1)ZnCH3] (0.40 g, 0.89 mmol) was added. The resulting
solution was stirred overnight, solvents were removed, and the solid
was dried under reduced pressure. The resulting solid was redis-
solved in CH2Cl2, filtered through Celite, and crystallized by
layering with hexane. Yield: 0.37 g (69%). Anal. Calcd for
C30H37N5O2S1Zn‚(CH2Cl2)0.12: C, 59.57; H, 6.18; N, 11.53.
Found: C, 59.59; H, 6.27; N, 11.43. FTIR (KBr), cm-1: νNH 3324.
proteins pointed to a strong correlation between electrostatic
screening of the core (predominantly by H-bonding interac-
tions) and reactivity,11 only recently using synthetic analogues
has any experimental confirmation of this notion, or a
determination of its magnitude, been reported.12-14 In the
present report, we study the alkylation of LZnSPh complexes
designed to mimic the tetrahedral binding site found in all
zinc thiolate proteins. In these complexes, the ligand L is
one of a series of N2X tripodal scorpionates designated L1-4
(Chart 1) containing phenolate oxygen, alkoxy oxygen,
carboxylate oxygen, or thiolate sulfur donors, respectively,
in the “X” position.15-17 By substituting o-N-Ac-thiophenol
(known to possess an internal hydrogen bond between the
amide and the thiolate sulfur)18-20 in place of thiophenol itself
as the exogenous fourth ligand, we show that H-bonding can
indeed control reactivity and that even a single such bond is
sufficient to achieve specificity of reaction with a powerful,
and hence inherently indiscriminate, electrophile. These
results have a direct bearing on our understanding of these
important metalloproteins. Portions of this work have already
appeared in abbreviated form.12
1
FTIR (in CH2Cl2): νNH 3327. H NMR (CDCl3): δ 8.85 (s, 1H,
NH), 8.33 (d, 1H, J ) 7.6 Hz, S-ArH), 7.80 (d, 1H, J ) 7.4 Hz,
S-ArH), 7.06 (d, 1H, J ) 2.4 Hz, ArH), 7.02 (t, 1H, J ) 7.8 Hz,
S-ArH), 6.90 (s, 1H, -CH-), 6.78 (t, 1H, J ) 7.8 Hz, S-ArH),
6.68 (d, 1H, J ) 2 Hz, ArH), 5.91 (s, 2H, PzH), 2.46 (s, 6H, Pz-
CH3), 2.22 (s, 3H, S-ArCH3), 2.20 (s, 6H, Pz-CH3), 2.19 (s, 3H,
ArCH3),1.35 (s, 9H, -C(CH3)3). 13C NMR (CDCl3): δ 168.40,
163.02, 150.39, 142.69, 140.68, 138.43, 134.29, 130.34, 129.06,
124.57, 123.28, 120.66, 119.78, 119.19, 106.81, 73.50, 35.40, 29.33,
25.04, 20.40, 12.79, 11.62.
[(L1)Zn(S-o-CF3CONHC6H4)], 2. Solid NaBH4 (0.02 g, 0.53
mmol) was added to a THF-ethanol solution (9:1, 25 mL) of 2,2′-
dithio-bis(N-phenyl-2,2,2-triflouroacetamide) (0.225 g, 0.51 mmol).
The mixture was stirred at room temperature for about 2 h to give
a clear solution, to which [(L1)ZnCH3] (0.40 g, 0.89 mmol) was
added. The resulting solution was stirred overnight, solvents
removed, and the solid was dried under reduced pressure. The
resulting solid was redissolved in CH2Cl2, filtered through Celite,
and crystallized by layering with hexane. Yield 0.29 g (66%). FTIR
1
(KBr), cm-1: νNH 3247. FTIR (in CH2Cl2): νNH 3264. H NMR
(CDCl3): δ 10.00 (s, 1H, NH), 8.34 (d, 1H, J ) 8 Hz, S-ArH),
7.89 (d, 1H, J ) 8 Hz, S-ArH), 7.10 (t, 1H, J ) 7.8 Hz, S-ArH),
7.05 (d, 1H, J ) 2.4 Hz, ArH), 6.94 (t, 1H, J ) 7.6 Hz, S-ArH),
6.90 (s, 1H, -CH-), 6.68 (d, 1H, J ) 2.0 Hz, ArH), 5.93 (s, 2H,
PzH), 2.46 (s, 6H, Pz-CH3), 2.29 (s, 6H, Pz-CH3), 2.18 (s, 3H,
ArCH3) 1.26 (s, 9H, -C(CH3)3). 13C NMR (CDCl3): δ 162.93,
150.37, 142.69, 140.78, 134.98, 130.40, 129.15, 125.28, 125.05,
120.81, 119.75, 119.46, 106.87, 73.53, 35.36, 29.20, 20.43, 12.92,
11.66.
(10) Meyers, L. C.; Verdine, G. L.; Wagner, G. Biochemistry 1993, 32,
14089.
(11) Maynard, A. T.; Covell, D. G. J. Am. Chem. Soc. 2001, 123, 1047.
(12) (a) Smith, J. N.; Shirin, S.; Carrano, C. J. J. Am. Chem. Soc. 2003,
125 (4), 868-869. (b) For a recent summary of synthetic analogues
relevant to the structure and function of zinc metalloproteins in general
see: Parkin, G. Chem. ReV. 2004, 104, 699-767.
(13) Docrat, A.; Morlok, M. M.; Bridgewater, B. M.; Churchill, D. G.;
Parkin, G. Polyhedron 2004, 23, 481-488.
(14) Chiou, S.; Riordan, C. G.; Rheingold, A. L. Proc. Natl. Acad. Sci.
U.S.A. 2003, 100, 3695-3700.
(15) Hammes, B. S.; Carrano, C. J. J. Chem. Soc., Dalton Trans. 2000,
3304.
(16) Hammes, B. S.; Kieber-Emmons, M. T.; Latizia, J. A.; Shirin, Z.;
Carrano, C. J.; Rheingold, A. L. Inorg. Chim. Acta 2003, 346, 227-
238.
(17) Hammes, B. S.; Carrano, C. J. Inorg. Chem. 1999, 38, 4593.
(18) Uyeama, N.; Okamura, T.; Nakamura, A. J. Am. Chem. Soc. 1992,
114, 8129.
(19) Ueyama, N.; Taniuchi, K.; Okamura, T.; Nakamura, A.; Maeda, H.;
Emura, S. Inorg. Chem. 1996, 35, 1945.
(20) Okamura, T.; Takamizawa, S.; Ueyama, N.; Nakamura, A. Inorg.
Chem. 1998, 37, 18-28.
[(L3)Zn(S-o-CH3CONHC6H4)], 3. This complex was synthe-
sized by the same method as described for 1. Colorless crystals
were obtained from CH2Cl2/diisopropyl ether solution in 60% yield.
Anal. Calcd for C26H35N5O3SZn (CH2Cl2)0.1: C, 54.85; H, 6.21;
N, 12.25. Found: C, 54.91; H, 6.16; N, 11.98. FTIR (KBr), cm-1
ν
:
NH 3307. FTIR (in CH2Cl2): νNH 3341. 1H NMR (CDCl3): δ 8.85
(s, 1H, NH), 8.26 (d, 1H, J ) 7.6 Hz, S-ArH), 7.72 (d, 1H, J )
8 Hz, S-ArH), 7.03 (t, 1H, J ) 8 Hz, S-ArH), 6.84 (t, 1H, J )
7.6 Hz, S-ArH), 6.60 (s, 1H, -CH-), 6.05 (s, 2H, PzH), 2.45 (s,
6H, Pz-CH3), 2.26 (s, 3H, S-ArCH3),1.27 (s, 18H, Pz-(CH3)3).
13C NMR (CDCl3): δ 164.72, 141.53, 135.49, 125.85, 123.47,
119.66, 104.84, 67.68, 31.96, 30.23, 25.12, 11.25.
Inorganic Chemistry, Vol. 44, No. 6, 2005 2013

Smith et al.
[(L2)Zn(S-o-CH3CONHC6H4)], 4. The disulfide 2,2′-dithiobis-
(N-phenylacetamide) (107 mg, 0.32 mmol) was reduced with
potassium borohydride (17.4 mg, 0.32 mmol) in 50 mL of a 9:1
mixture of anhydrous THF-ethanol under inert atmosphere. The
solution was allowed to stir overnight when [(L2)ZnCH3] (250 mg,
0.54 mmol) was added to the solution. After stirring for 1 h, the
solution was reduced to dryness and brought into the drybox. The
crude product was dissolved in anhydrous CH2Cl2, and filtered
through a pad of Celite. The resulting clear solution was reduced
to dryness to yield 205 mg (62% yield) of white crystalline product.
X-ray quality crystals of [(L2)Zn(S-o-CH3CONHC6H4)] were
produced by layering a concentrated CH2Cl2 solution of the complex
with hexanes. Anal. Calcd (Found) for [(L2)Zn(S-o-CH3CO-
NHC6H4)], C32H33N5O2SZn‚(CH2Cl2)2‚H2O: C, 50.73 (50.39); H,
diagram for [(L1)Zn(NAcSPh)] is shown in Figure 1, for [(L4)-
Zn(NAcSPh)], it is in Figure 2, for [(L3)Zn(NAcSPh)], it is in
Figure 3, for [(L2)Zn(S-o-CH3CONHC6H4)], it is in Figure 4, and
for [(L4SCH3)Zn(NAcSPh)], it is in Figure 5.
Kinetic Experiments. All experiments were performed under
pseudo-first-order conditions with the alkylating agent in 10-fold
excess as previously described.6 In a typical experiment, 1.8 × 10-5
mol of the metal complex was dissolved in 1 mL of solvent to
give an 18 mM solution. In a few cases, reactions were also run at
5 or 10 mM in metal complex. The solution of the metal complex
was then transferred to a NMR tube, and the alkylating agent added
via a gastight syringe. The NMR tube was then sealed with a
1
septum. The reactions were monitored by H NMR spectroscopy
at 25 °C. The kinetic data were least-squares fit to a single
exponential to determine the pseudo-first-order rate constant. Data
analysis was performed using Sigmaplot 8.0 (Jandel Scientific)
software.
1
4.88 (4.69); N, 8.70 (8.61). H NMR (CDCl3): δ 9.05 (s, 1 H,
-NH-), 8.27 (d, 1 H, SArH), 7.76 (d, 1 H, SArH), 7.54 (dd, 4 H,
ArH), 7.19 (m, 4 H, ArH), 7.15 (m, 2 H, ArH), 7.04 (m, 1 H, SArH),
6.84 (m, 1 H, SArH), 6.60 (s, 1 H, -CH-), 5.67 (s, 2 H, PzH),
2.22 (s, 3 H, Ar-CH3), 2.03 (s, 6 H, Pz-CH3), 2.00 (s, 6 H, Pz-
CH3). 13C NMR (CDCl3): δ 168.67, 149.78, 146.35, 141.27, 139.00,
134.45, 128.04, 126.97, 126.91, 126.85, 124.80, 123.39, 120.13,
105.68, 84.75, 70.96, 25.07, 12.65, 10.84.
Results and Discussion
Synthesis and Structure. We have previously used the
reaction of LZnSPh complexes with CH3I to model the
reaction of Zn-S bonds with an electrophile in solution.6
To mimic the presence of an amide hydrogen to sulfur
H-bond, we have utilized o-N-Ac-thiophenol in place of
thiophenol as the exogenous fourth ligand. o-N-Ac-thio-
phenol and its derivatives are known to possess an internal
hydrogen bond between the amide and the thiolate sulfur,
and extensive studies by Nakamura et al. have demonstrated
the profound effects such bonds can have on redox potentials
and other thermodynamic properties in models for Fe-S and
molybdopterin proteins.18-20 In our case, the desired com-
plexes were readily prepared by borohydride reduction of
the disulfide of the appropriate o-N-acetylated thiophenol
followed by in situ reaction with [LZnCH3] in THF-ethanol.
We have isolated and completely characterized the [LZn(o-
N-Ac-thiophenol)] complexes for the L1-4 heteroscorpi-
onate ligands. For comparative purposes, we have also
isolated and structurally characterized [(L1)Zn(o-N-trifluo-
romethylacetyl-thiophenol)]. As can be seen in the crystal
structures of both the L1 and L4 complexes, an internal
H-bond of the same type and nearly the same geometrical
parameters as those described by Nakamura is present.18-20
These have H‚‚‚S “bond” distances between 2.36 and 2.45
Å and N-H‚‚‚S angles between 113° and 117°. The distance
between the Namide-H and the heteroatom of the heteroscor-
pionate ligand is greater than 5 Å in all these cases as the
amido group is oriented to maximize the hydrogen bonding
interaction with the thiolate sulfur. However, an examination
of the structures of the L2 and L3 complexes reveals a
different story. In the case of the L3 complex, 3, the o-N-
Ac-thiophenol is now oriented in such a way as to maximize
the H-bonding not with the thiolate sulfur but with the
carboxylate oxygen of the heteroscorpionate ligand. Indeed,
the N-H‚‚‚Ocarboxylate distance is very short (1.99 Å), and
the N-H‚‚‚O angle is nearly linear (172°), both of which
are suggestive of a very strong internal hydrogen bond. This
results in a weak interaction with the thiolate sulfur as
reflected in the long N-H‚‚‚S (2.99 Å) “bond” and acute
N-H‚‚‚S angle. The situation with the L2 alkoxy ligand
[(L4)Zn(S-o-CH3CONHC6H4)], 5. The complex was synthe-
sized by the same method as described for 1. Colorless crystals
were obtained in 65% yield. Anal. Calcd for C22H29N5OS2Zn: C,
51.91; H, 5.74; N, 13.76. Found: C, 51.97; H, 5.72; N, 13.68. FTIR
1
(KBr), cm-1: νNH 3324. FTIR (in CH2Cl2): νNH 3326. H NMR
(CDCl3) δ 8.89 (s, 1H, NH), 8.32 (d, 1H, J ) 8.4 Hz, S-ArH),
7.46 (d, 1H, J ) 7.6 Hz, S-ArH), 6.99 (t, 1H, J ) 7.8 Hz, S-ArH),
6.77 (t, 1H, J ) 8 Hz, S-ArH), 5.98 (s, 1H, -CH-), 5.96 (s, 2H,
PzH), 2.39 (s, 6H, Pz-CH3), 2.27 (s, 6H, Pz-CH3), 2.23 (s, 3H,
S-ArCH3), 1.35 (s, 6H, -C(CH3)2). 13C NMR (CDCl3): δ 149.98,
141.03, 133.91, 124.64, 123.19, 119.02, 106.52, 72.95, 48.26, 33.73,
25.16, 12.96, 11.42.
[(L4SCH3)Zn(S-o-CH3CONHC6H4)], 6. A CH3CN solution of
(CH3)3OBF4 (0.08 g, 0.54 mmol) was added dropwise into the
solution of [(L4)Zn(S-o-CH3CONHC6H4)] (0.28 g, 0.55 mmol) in
25 mL of CH3CN. The resulting reaction mixture was stirred for 2
h, filtered, and crystallized by slow diffusion of diethyl ether. Yield
0.27 g (80%). Anal. Calcd for C23H32N5OS2BF4Zn: C, 45.22; H,
5.28; N, 11.46. Found: C, 44.94; H, 5.01; N, 11.62. FTIR (KBr),
cm-1: νNH 3327. 1H NMR (CD3CN): δ 9.18 (s, 1H, NH), 7.67 (d,
1H, J ) 6.8 Hz, S-ArH), 7.52 (d, 1H, J ) 7.6 Hz, S-ArH), 7.14
(t, 1H, J ) 7.6 Hz, S-ArH), 7.07 (t, 1H, J ) 7.2 Hz, S-ArH),
6.21 (s, 2H, PzH), 6.15 (s, 1H, -CH-), 2.45 (s, 6H, Pz-CH3),
2.12 (s, 6H, Pz-CH3), 2.08 (s, 3H, S-ArCH3), 1.67 (s, 3H, SCH3),
1.17 (s, 6H, -C(CH3)2). 13C NMR (CDCl3): δ 152.4, 145.48,
136.66, 125.86, 125.40, 123.42, 107.28, 70.56, 51.42, 25.04, 23.44,
13.11, 12.17, 11.02.
X-ray Crystallography. Crystals of complexes 1-6 were sealed
in thin-walled quartz capillaries or nylon loops (Hampton Research),
and mounted on either a Siemens P4, Nonius Kappa CCD, or Bruker
X8 APEX CCD diffractometer. The structures were solved using
direct methods or via the Patterson function, completed by
subsequent difference Fourier syntheses, and refined by full-matrix
least-squares procedures on F2. All non-hydrogen atoms were
refined with anisotropic displacement coefficients and treated as
idealized contributions using a riding model except where noted.
All software and sources of the scattering factors are contained in
the SHELXTL 6.0 program library (G. Sheldrick, Siemens XRD,
Madison, WI). Crystallographic data collection parameters are given
in Table 1 while selected bond lengths and angles for complexes
1-5 are shown in Table 2 and for 6 in Table 3. The ORTEP
2014 Inorganic Chemistry, Vol. 44, No. 6, 2005

Model Complexes of Zn Metalloproteins
Table 1. Summary of Crystallographic Data and Parameters for [L1ZnS-o-CH3CONHC6H4‚CH2Cl2] (1), [L1ZnS-o-CF3CONHC6H4] (2),
[L3ZnS-o-CH3CONHC6H4] (3), [L2ZnS-o-CH3CONHC6H4] (4), [L4ZnS-o-CH3CONHC6H4] (5), and [L4SCH3Zn S-o-CH3CONHC6H4] (6)
1
2
3
4
5
6
molecular formula
C31H39Cl2-
N5O2SZn
682
293(2)
monoclinic
P21/c
C30H34F3-
N5O2SZn
651.05
293(2)
triclinic
P1h
C26H35N5-
O3SZn
562.95
153
monoclinic
P21/c
C32H33N5-
O2SZn
617.09
273(2)
monoclinic
P21/c
C22H29N5-
OS2Zn
508.99
293(2)
monoclinic
P21/c
C23H32BF4-
N5OS2Zn
610.84
293(2)
monoclinic
P21/n
fw
temp (K)
cryst syst
space group
cell constants
a (Å)
10.915(3)
13.024(5)
23.832(6)
90
102.27
90
13.793(3)
15.131(5)
17.356(9)
90
93.38(3)
114.84(2)
4
3279(2)
0.863
1.319
17.912(2)
9.839(10)
17.074(2)
90
111.6520(10)
90
34.5132(9)
8.4159(2)
20.7128(5)
90
95.0900(10)
90
10.356(2)
13.930(3)
16.680(3)
90
92.29(3)
90
13.780(2)
13.386(2)
15.197(2)
90
91.05
90
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
Z
4
4
6
4
4
V (Å3)
3311(2)
1.002
1.368
2796.87(5)
0.987
1.316
5992.5(3)
1.353
1.436
2404.3(8)
1.219
1.406
2802.8(7)
1.078
1.448
abs coeff, µcalc (mm-1
)
δ
cal (g/cm3)
F(000)
1424
1352
1148
2676
1064
1264
cryst dimens (mm3)
radiation
0.5 × 0.2 × 0.3
Mo KR
(λ ) 0.71073 Å)
0 f 12,
-15 f 0,
-28 f 27b
1.79-25.01
6096
0.8 × 0.6 × 0.2
Mo KR
(λ ) 0.71073 Å)
-16 f 0,
-14 f 16,
-18 f 18
2.06-23.03
9785
0.4 × 0.08 × 0.07
Mo KR
(λ ) 0.71073 Å)
-25 f 25,
-13 f 13,
-22 f 23
2.93-27.50
12135
0.2 × 0.2 × 0.4
Mo KR
(λ ) 0.71073 Å)
-36 f 36,
-8 f 8,
-22 f 22
1.78-22.41
67632
0.29 × 0.13 × 0.12
Mo KR
(λ ) 0.71073 Å)
-13 f 13,
-18 f 18,
-21 f 21
3.08-27.50
10381
0.4 × 0.4 × 0.4
Mo KR
(λ ) 0.71073 Å)
0 f 16,
0 f 15,
-18 f 18
1.98-25.00
5051
h,k,l ranges colled
θ range (deg)
no. reflns colled
no. unique reflns
no. params
5782
379
9051
757
6392
337
7699
739
5465
283
4845
362
data/param ratio
refinement method
R(F)b
15.12
a
0.0563
0.1446
1.031
11.94
a
0.0692
0.1518
18.89
a
0.0476
0.1136
10.42
a
0.0387
0.0982
19.31
a
0.0407
0.0943
13.38
a
0.0748
0.1156
1.031
Rw(F2)c
GOFwd
1.028
1.071
0.995
1
largest diff peak
0.85 and -0.79
0.59 and -0.45
0.63 and -0.63
0.468 and -0.257
0.522 and -0.393
0.41 and -0.36
and hole (e/Å3)
2
a Full-matrix least-squares of F2. b R ) [∑|∆F|/∑|Fo|]. c Rw ) [∑w(∆F)2/∑wFo ]. d Goodness of fit on F2.
Table 2. Selected Bond Lengths (Å) and Angles (deg) for 1-5
Table 3. Selected Bond Lengths (Å) and Angles (deg) for 6
1
2
3
4
5
Zn-N1
Zn-N3
Zn-S1
Zn-S2
Zn-O1
2.047(6)
2.045(7)
2.773(3)
2.254(2)
2.102(6)
Zn-N1
Zn-N3
Zn-X
2.037(3)
2.023(3)
1.887(3)
2.031(7) 2.050(2)
2.014(7) 2.058(2)
1.908(6) 2.002(2)
2.064(3)
2.041(3)
1.910(2)
2.054(2)
2.077(2)
2.257(1)
2.2482(9)
Zn-S1
2.2201(13) 2.241(3) 2.2219(8) 2.223(1)
N1-Zn-N3
N1-Zn-S1
N1-Zn-S2
N1-Zn-O1
N3-Zn-S1
N3-Zn-S2
N3-Zn-O1
S1-Zn-S2
S1-Zn-O1
S2-Zn-O1
91.8(3)
79.9(2)
131.7(2)
98.3(2)
82.6(2)
132.9(2)
96.2(2)
87.91(9)
177.8(2)
94.2(2)
N1-Zn-N3 88.92(13)
90.6(3)
95.9(3)
88.80(10) 89.16(12) 85.81(9)
91.52(9) 93.58(10) 99.00(7)
N1-Zn-X
97.91(14)
N1-Zn-S1 122.92(10) 124.3(2) 124.48(7) 123.15(9) 116.31(7)
N3-Zn-X 101.50(14) 100.8(3) 91.63(9) 91.35(11) 97.04(7)
N3-Zn-S1 117.83(10) 117.1(2) 136.25(7) 131.07(9) 116.73(7)
X-Zn-S1
S1‚‚Hamide
X‚‚‚Hamide
121.27(10) 121.6(2) 112.50(6) 118.59(8) 131.45(3)
2.412(*)
2.450(*) 2.994(*)
1.992(*)
117.0(*) 93.5(*)
172.0(*)
2.865(*)
2.390(*)
95.5(*)
2.358(*)
N5-H‚‚‚S1 117.2(*)
N5-H‚‚‚X
113.0(*)
135.0(*)
complex is between the two extremes in that the o-N-Ac-
thiophenol is oriented in such a way as to allow a bifurcated
H-bonding interaction with both the thiolate sulfur and the
alkoxy oxygen. However, the stronger interaction is clearly
with the heteroscorpionate ligand alkoxy oxygen as indicated
by its shorter (2.39 vs 2.87 Å) and more linear (135° vs 95°)
N-H‚‚‚X interaction.
Effects of Hydrogen Bonding on Kinetics of Thiolate
Alkylation. To determine the effects that this internal H-bond
has on the reactivity of the zinc thiolate, we have compared
the kinetics of reaction between 1 and its non-hydrogen-
bond-containing analogue in CDCl3 with CH3I at 25 °C. The
results (Table 4) show that the presence of a single H-bond
is sufficient to reduce the reactivity of the Zn-S bond toward
the electrophile by approximately 2 orders of magnitude.
Thus, we were initially greatly surprised when repeating this
experiment using the L3 complex 3 and its non-H-bonding
analogue, that the effect was now greatly reduced (reactivity
reduced only by a factor of 7). A solution to this puzzle
became evident however from an examination of the crystal
structure of 3 (vide supra) which clearly revealed that the
H-bond formed in this case was with the carboxylate oxygen
of the scorpionate ligand and not the thiolate sulfur. This
provides a definitive and sensitive test to show that it is the
H-bond itself rather than any peripheral electronic or steric
effect of the N-acetyl group that controls the reactivity of
Inorganic Chemistry, Vol. 44, No. 6, 2005 2015

Smith et al.
Figure 4. ORTEP diagram with 50% thermal ellipsoids of[(L2)Zn-
(NAcSPh)] showing complete atomic labeling.
Figure 1. ORTEP diagram with 40% thermal ellipsoids of [(L1)Zn-
(NAcSPh)] showing complete atomic labeling.
Figure 5. ORTEP diagram with 50% thermal ellipsoids of[(L4SCH3)Zn-
(NAcSPh)] showing complete atomic labeling.
1
Table 4. Pseudo-First-Order Rate Constants (s-1) Determined by H
NMR for Reaction of the Listed Compounds (18 mM) with Methyl
Iodide (180 mM) at 25 °C
Figure 2. ORTEP diagram with 40% thermal ellipsoids of [(L4)Zn-
chloroform
acetonitrile
(NAcSPh)] showing complete atomic labeling.
compd
[(L4)ZnI]
av rate (×10-4
)
av rate (×10-4
)
3.12(8)
3.53(5)
0.017(1)
0.026(2)
0.028(2)
0.004(4)
2.0(1)
27(6)
14.2(3)
0.41(5)
1.3(4)
0.14(4)
0.059(1)
nd
[(L1)Zn(SPh)]
[(L1)Zn(NAcSPh)]
[(L1)Zn(NCOCF3SPh)]
[(L3)Zn(SPh)]
[(L3)Zn(NAcSPh)]
[(L2)Zn(SPh)]
[(L2)Zn(NAcSPh)]
0.23(6)
nd
internal H-bond that is more with the alkoxy oxygen of the
ligand than with the thiolate sulfur. This was confirmed again
by a crystal structure analysis (Figure 4).
If H-bonding can control the reactivity of the Zn-thiolate
with electrophiles, then it can in principle also provide a
means to generate specificity. To test the degree of such
specificity we designed a model system, [(L4)Zn(o-N-Ac-
thiophenol)], 5, which contains two thiolates, one involved
in H-bonding and the other not. Despite the differing nature
of the thiols (i.e., aliphatic vs aromatic), in the absence of
H-bonding effects the thiophenol and heteroscorpionate
thiolate sulfurs react at similar rates (Table 4). To provide a
strict test of the degree of selectivity, we deliberately chose
Figure 3. ORTEP diagram with 40% thermal ellipsoids of[(L3)Zn-
(NAcSPh)] showing complete atomic labeling.
the thiolate. The relative decrease in reactivity between [(L2)-
Zn(SPh)] and [(L2)Zn(S-o-CH3CONHC6H4)] is also only
approximately 1 order of magnitude, consistent with an
2016 Inorganic Chemistry, Vol. 44, No. 6, 2005

Model Complexes of Zn Metalloproteins
the strong and hence inherently indiscriminate alkylating
agent, trimethyloxonium tetrafluoroborate. The result was
in many ways remarkable with only the L4 thiolate sulfur
being alkylated to give the mono-thioether product 6. Even
with a single H-bond, the specificity was near 100% (we
found no trace of the alternative methylated product).
Coordination of the Thioether. Another aspect of certain
enzymatic reactions mimicked by our model system is the
fact that, in many of the former, the thioether produced in
the alkyl transfer remains coordinated to the zinc. Thioether
coordination has been unequivocally demonstrated for the
Ada protein, and spectroscopic evidence consistent with this
has been presented in at least one other case.21 The
coordinated thiolate in a series of zinc complexes of our
tripodal, N2S ligand (L4) has been methylated in solution,
and the coordination properties of the resulting thioether have
been examined.22,23 This system models the reactivity of zinc
containing enzymes involved in alkyl group transfers. While
it is clear that neutral thioethers cannot, in general, compete
with anionic ligands, it is equally clear that a special,
sterically restricted, protein environment is not necessary to
support neutral thioether coordination. Thus, the thioether
in 6 is weakly bound to the Zn as indicated by its position
and its long (2.77 Å) “bond” length. The acetyl-carbonyl
serves to provide the fourth strong ligand donor.
CCCC core of Ada is less screened than average.11 Second,
it suggests that the presence of hydrogen bonds between the
accessible but unreactive cys 69 or 42 and their absence at
cys 38 could be the means by which the well-known
regiospecifity of Ada is maintained. Gratifyingly, the most
recent NMR structure of an active fragment of the Ada
protein, N-Ada10, is reported to contain characteristic
H-bonds to the ligand sulfur atoms for cys-69 and 42 and
but not to the reactive cys-38.24 In this regard, it is notable
that those amide protons on residues near the sulfur atoms
other than cys-38 exchange very slowly as expected for those
engaged in hydrogen bond interactions.25
In conclusion, we have shown that even a single hydrogen
bond to a thiolate sulfur is sufficient to drastically reduce
its nucleophilicity. Thus, nature can use this mechanism to
modulate the reactivity of a zinc-thiolate construct so that it
can be used for either structural or reactive purposes. In
addition, such hydrogen bonding provides another means to
generate the exquisite specificity seen in reactive biomol-
ecules of this type.
Acknowledgment. This work was supported by Grants
CHE-02023535 and CHE-0313865 from the NSF. The NSF-
MRI program Grant CHE-0320848 is acknowledged for
support of the X-ray diffraction facilities at San Diego State
University. The authors thank Dr. Vincent Lynch, University
of Texas at Austin, for data collection on compounds 2 and
3 and Dr. Brian Hammes, St. Joseph’s University, for a
sample of ligand L3.
Biological Implications. What insight into the functioning
of thio-rich zinc cores in metalloproteins can be gleaned from
this work? First, this work supports the notion that zinc-
thiolate cores, which are predicted to be inherently reactive
constructs, can be deactivated by the presence of hydrogen
bonds in complete accord with a recent detailed statistical
analysis.11 Thus, the unreactive CCCC cores found in the
GATA zinc fingers are found to be more electrostatically
screened than average while the correspondingly reactive
Supporting Information Available: X-ray crystallographic files
(in CIF format) for 2 and 4 and ORTEP diagram for 2. This material
is available free of charge via the Internet at http://pubs.acs.org.
X-ray crystallographic files (in CIF format) for 1, 3, 5, and 6 can
be obtained from the Supporting Information available for ref 12.
IC048630F
(21) Huang, C.; Casey, P. J.; Fierke, C. A. J. Biol. Chem. 1997, 272, 20-
(24) (a) Lin, Y.; Doetsch, V.; Winter, T.; Peariso, K.; Myers, L. C.; Penner-
Hahn, J. E.; Verdine, G. L.; Wagner, G. Biochemistry 2001, 40, 4261.
(b) Personal communication.
(25) Habezettl, J.; Myers, C. M.; Yuan, F.; Verdine, G. L.; Wagner, G.
Biochemistry 1996, 35, 9335-9348.
23.
(22) Hammes, B. S.; Carrano, C. J. Inorg. Chem. 2001, 40, 919-927.
(23) Hammes, B. S.; Carrano, C. J. J. Chem. Soc., Chem Commun. 2000,
1635.
Inorganic Chemistry, Vol. 44, No. 6, 2005 2017