In all the above Zn2þ enzymes, the Zn2þ-bound H2O is activated by the Zn2þion with various degrees of assistance from a proximate general base such as Glu, for deprotonation or polarization fornucleophilic attack at the electrophilic substrates. With a pKa value of 6.9, the Zn2þ-bound H2O in CA is deprotonated to attack at substrate CO2 to produce the hydrated HCO3
in the simplest electrophilic reaction. Apparently, the Zn2þcomplex of the three neutral imidazole ligands is sufficiently Lewis acidic to activate the coordinated H2O. Targeting CA is probably the easiest entry to the model study of Zn2þenzymes.
A numberof simple zinc(II) complexes have been designed in orderto gain basic knowledge of Zn2þ ion chemistry in the zinc(II) hydrolases.202–213 Most of them feature tridentate or tetra- dentate ligands with vacant sites. In 1975, Wooley reported the first well-defined zinc(II) complex (33) with a macrocyclic tetramine as a model for CA (Scheme 27).214,215It was demonstrated that the Zn2þ-bound HO species is a reactant that attacks CO2to afford HCO3
, with a second- order rate constant for hydration, khcat, of 2102M1 s1 at 25C. On the basis of the pH-dependent hydration rates, a pKavalue of 8.1 was assigned to the Zn2þ-bound H2O, which was lowerthan the values of 9–10 forH2O bound to free Zn2þ.
pKa=9~10
pKa=8.1 (32a)
–H+
(32b) Scheme 27
In 1990, the zinc(II) complex of 1,5,9-triazacyclododecane (Zn2þ–[12]aneN3) (33) was found to be more structurally and functionally relevant to the active center of CA (Scheme 28).216–223 Its basic form (33b) was isolated as a trimeric complex ((33b)3HClO4(ClO4)3) from pH 8 aqueous solution.216 The X-ray structure of this complex, which contained three (33b) molecules, disclosed a tetrahedral coordination environment about the zinc center, with a Zn2þ–OH length of 1.94 A˚ at the fourth coordination site, shorter than the Zn2þN bonds (average 2.02 A˚). It dissociates into monomeric (33b) in aqueous solution.216Potentiometric pH titration measurements were used to show that the Zn2þ-bound H2O in (33a) is deprotonated to give (33b) with a pKa value of 7.3 at 25C and Iẳ0.1 (NaClO4). The data are interpreted to indicate that (33) is a strong anion binder; i.e., hydroxide anion binds with a logKs value of 6.4 (atIẳ0.2), whereKsis the affinity constant. The Zn2ỵ–OHspecies (33b) turned out to be a good nucleophile and showed CA-like catalytic activity in the hydration of acetaldehyde (at 0C), as well as in the hydrolysis of an unactivated ester, methyl acetate (at 25C). The second-order rate (first order in [(33a)] and [acetaldehyde orCH3CO2Me]) constant-pH plots were sigmoidal, yielding the kinetic pKa values of 7.3 at 25C, which are identical to the thermodynamic pKa
values of 7.3.
pKa=7.3
pKa=7.8 (33a)
(34a)
–H+
–H+
(33b)
(34b) in aqueous solution
in aqueous solution
Zn2+–[12]aneN3 Zn2+–[12]aneN3
Zn2+–cyclen Zn2+–cyclen
(HO–form)
(HO–form) (H2O form)
(H2O form)
Scheme 28
The catalytic hydration of CO2 also occurred with (33) and followed pH-dependent, second- order kinetics. Enzymatic CO2hydration and the reverse (HCO3
dehydration) were mimicked by (33).217 The reaction was followed by measuring the evolution of OH forthe reaction (HCO3!CO2þOH). The CO2hydration rates (the second-order constant is khcat) decreased and HCO3
dehydration rates (the second-order constant is kdcat) increased with lowering pH (Figure 1). The kinetic data agreed with the reaction shown inEquation (1), being first order in [(33b)] and [HCO3
]. The second-order constantkdcatforHCO3
dehydration was found to be 5 M1s1and the kinetically obtained pKavalue for(33a) – (33b)þHþwas 7.3 at 25C. Conclu- sions forthe HCO3
dehydration with the catalyst (33) were: (i) the reactive species is the zinc(II)–OH2form, and (ii) substitution of the zinc(II)-bound H2O by HCO3is rate determining.
6
4
2
0 600
400
200
0
6.0 7.0 8.0 9.0 10.0
pH
(kh cat)obs[M–1 s–1 ] for (a) (kd cat)obs[M–1 s–1 ] for (b)(b)
(a)
Figure 1 The rate–pH profile for (a) a CO2 hydration; (b) HCO3
dehydration, catalyzed by Zn2þ–[12]aneN3(33).
Although the magnitudes of khcat and kdcat differed significantly, the two pH-dependent rate curves (Figures 1aand 1b) are symmetrical (with scales adjusted) with an inflection point at pH ca. 7.4, a value close to the pKavalue of 7.3. Thus, the macrocyclic complex (33) was the first to mimic the pH-dependent behavior of reversible CO2hydration catalyzed by CA. This fact implies that in CA, too, the CO2hydration/HCO3 dehydration should be essentially controlled by the zinc(II)–OH/zinc(II)–OH2equilibrium at the active center.
A zinc(II) complex of the 12-membered macrocyclic tetraamine, cyclen (1,4,7,10-tetraazacyclo- dodecane) (34), was latershown to be a useful Zn2þ-enzyme model, although the pKavalue (7.8) of the Zn2þ-bound water(for(34a)!(34b)þHþ) was a little higherthan that (7.3) forthe triamine complex (33) (Scheme 28).218These findings demonstrated that simple zinc(II) complexes can be used to activate H2O at physiological pH forcatalytic reactions, in analogy to the monometallic zinc(II) enzymes.
A series of tris(pyrazolyl)hydroborate zinc(II)–OHcomplexes (such as (35)) was synthesized to model the catalytic centerof CA and otherZn2þ enzymes.224–238 Indeed, the zinc(II)–OH in (35b) was highly nucleophilic to CO2, carboxylesters, activated amides (-lactam orCF3CONH2), and phosphonates. Those reactions, were run in organic solvents and, therefore, were stoichio- metric. The high nucleophilicity of (35b) was attributed to monomeric character of the zinc(II)- bound OH due to steric hindrance and the hydrophobic environment. The pKavalue for(for (35a)!(35b)þHþ) was estimated to be 6.50.4.209The zinc(II) complex (35b) reacted with CO2
to yield a bridging carbonato complex (37), possibly via formation of (36) (Scheme 29).228,234 Thermodynamic and kinetic parameters for CO2 hydration and CO3dehydration reactions in water-containing solvents were reported.
pKa=6.5
(35a)
–H+
(35b) (36)
(37) CO2
CO2
No reaction
Scheme 29
A similarly water-insoluble Zn2þcomplex of a tris(3,5-diisopropyl-1-pyrazolyl)hydroborate (38) reacted with CO2 instantaneously to give the (-carbonato)-dizinc(II) complex (39), where the binding mode of the bicarbonate bridge was different from that of (37) (Scheme 30).239,240CO2
hydration was also reported for the water-soluble Zn2þcomplexes such as (40).241
1) –H+ 2) CO2
(39)
(40) (38)
Scheme 30
8.23.5.2 Peptidase Models
Amide bonds are extremely stable and their half-life for hydrolysis in neutral aqueous solution is estimated to be seven years.242 Zinc(II)–OH species (derived from mononuclear Zn2þ com- plexes) alone are incapable of hydrolyzing amide bonds. Rather, Zn2þ–OHspecies may often act as a base to deprotonate amides.243
There are only a few dinuclear Zn2þcomplexes known to exhibit peptidase activity. A dimeric Zn2þ complex (41) was reported to hydrolyze L-Leu-4-nitropyhenylanilide (Scheme 31).244The reaction was first order with respect to the substrate and the catalyst (41). The otherdimeric Zn2þ complex (42) cleaves a peptide bond of Gly–Gly.245It was proposed that the amide carbonyl of Gly–Gly coordinates to one of the Zn2þions in (42) to be activated forhydrolysis by the other Zn2þ–OH. Forothermetals, several complexes of Cu2þ, Ni2þ, Pd2þ, and Co3þwere shown to cleave amide bonds.210, 246–251
It is of interest to note that while carboxamides are substrates for CPD, they are inhibitors for CA.252A recent model study has addressed this issue.243
Scheme 31
8.23.5.3 Phosphatase Models
As described in Section 8.23.3.5, many phosphatases are dinuclear metalloenzymes with Zn2þ, Fe3þ, Mn2þ, Mg2þ, etc. Typically, alkaline phosphatases (AP) are dimetallic Zn2þ hydrolases.
As shown in Scheme 10, Ser102 of AP is apparently activated by Zn2þto attack at phosphate substrates polarized by another Zn2þion.
The interaction and hydrolysis of phosphate esters were mimicked by the CA models, (33)253and (34),254and an alcohol-pendant Zn2þcomplex (45)255(Schemes 32and33). It was found that (33) and (34) hydrolyzed bis(4-nitrophenyl) phosphate (BNPP) to give stable complexes (43) and (44) of the product, mono(4-nitrophenyl) phosphate (4-NPP). The second-order rate constants are 8.5105M1s1and 2.1105M1s1, respectively, at 25C (Scheme 32). An alcohol-pendant Zn2þcomplex (45) provided a model for the Zn2þ-activated serine of AP (see Scheme 10).255An alkoxide group in (45b) attacked BNPP to give a phosphorylserine intermediate (46), which was susceptible to further hydrolysis by the intramolecular Zn2þ-bound hydroxide in (46b) to give (47) (Scheme 33). For the first step, the alcohol group is deprotonated by the proximate Zn2þ (pKaẳ7.5) to an alkoxide complex (45b), which was 125 times more effective as nucleophile to the phosphate substrate than was the Zn2þ-activated water of the reference compound (34). Forthe subsequent step, the nucleophilic Zn2þ species (46c) was generated with a pKa value of 9. This intramolecular hydrolysis is 45,000 times faster than the intermolecular hydrolysis of ethyl 4-nitro- phenyl phosphate (ENP) with (34b). These results imply an advantage of the intramolecular arrange- ment of two Zn2þions in AP (as shown inScheme 10) overmononuclearZn2þhydrolases. Toward any mononuclearZn2þphosphatase model, phosphomonoesters were not substrates, but instead were inhibitors, as shown by isolation of the stable complexes (43), (44), and (47). The tris(pyrazolyl)- hydroborate complexes such as (35) and (38) hydrolyze phosphodiesters to phosphomonoesters, which similarly bind to Zn2þto become inert to further hydrolysis.233,239
in aqueous solution
in aqueous solution k2=2.1x10–5M–1s–1
k2=8.5x10–5M–1s–1
(34b)
(34)–4-nitrophenyl phosphate complex (33b)
+4-NP
+4-NP
(44)
(33)–4-nitrophenyl phosphate complex
(43)
BNPP (BNPP) bis(4-nitrophenyl)-
phosphate 2
Scheme 32
The requirement of the dimetallic complexes to hydrolyze phosphomonoesters was demon- strated by a dizinc(II) cryptate (48) with an alkoxide-bridge (Scheme 34).256Complex (48) reacted with phosphomonoesters such as mono(4-nitrophenyl) phosphate (4-NPP) at pH 5–7 in aqueous solution. It should be noted that (48) reacts with 4-NPP exclusively, not with diester, BNPP, nor triester, tris(4-nitrophenyl) phosphate (TNPP). A second-order dependence of the rate constant (first order with respect to both [(48)] and [(4-NPP)] was determined. The rate–pH profile exhibited a bell-shaped relationship with the maximum rate at pH 5.9. On the basis of X-ray crystal-structure analysis of (48) and (50), a mechanism was proposed (Scheme 34). Initially, two Odonors of 4-NPP interact with the two Zn2þions, which are suitably separated (3.42 A˚) (49).
Concomitantly, two of the axial NH groups of (49) begin to dissociate and attack the phosphorus atom of the activated 4-NPP, to displace 4-nitrophenolate and yield stable (50). These reactions have analogies to the alkaline phosphatase reactions shown inScheme 10, where the Zn2þ-bound Ser102 reacts with the incoming phosphomonoester.
in aqueous solution k2=2.1x10–5M–1s–1 (45b)
(46b) (46c)
(45a)
(46a)
(47)
+ BNPP
4-NP
4-NP
pKa=9 pKa=7.5
Scheme 33
k2=1.5x10–3M–1s–1 at pH 5.9
(48) (49) (50)
4-NP 4-NPP
pH 4.9–9.5
Phosphoryl transfer
Scheme 34
Many metal complexes have been designed and synthesized as catalysts fortransesterification of phosphate diesters to model mechanisms of RNases. Typical examples are monomeric Zn2þ complexes (53)257 and (54)258 and dimetallic Zn2þ complexes such as (55)259 (Scheme 35). For (53), the pseudo-first-order rate constant,k, for hydrolysis of BPP and for transesterification of
2-hydroxypropyl 4-nitrophenyl phosphate (HPNP) (51) yielding (52) was 1.5102 h1in 10%
DMSO solution at pH 7.0 and 37C. Among dinuclearZn2þmacrocyclic complexes tested, the dimeric Zn2þcomplexes (55), in which two Zn2þ ions are well separated, were found to be the best catalysts for the hydrolysis of BPP and HPNP. A trinuclear Zn2þcomplex (56) was reported to promote transesterification of adenyl(30-50)adenosine (ApA).260,261 Formodels utilizing other metal complexes, see references 262–269.
(53)–(56)
(53) (55)
(54) (56)
2-hydroxypropyl 4-nitrophenyl phosphate (52) (HPNP)(51)
4-NP
R=H,
Scheme 35
8.23.5.4 b-Lactamase Models
The coordinating ligands to Zn2þin -lactamases (Scheme 9) are essentially similar to those of CA (Scheme 11). In addition to acting as a CA model, Zn2þ–cyclen (34) was found to catalyze the hydrolysis ofN-benzylpenicillin (Scheme 36).270Interestingly, (34) also catalyzed epimerization of the hydrolyzed product, (5R)-benzylpenicilloate, to (5S)-benzylpenicilloate, unlike the enzymatic reaction.
With relevance to dimetallic zinc(II) -lactamases, a dinuclearzinc(II) complex (57) was designed and synthesized (Scheme 37).271,272 It has been found that (57) catalyzes a hydrolysis of nitrocefin, a substrate for the dimetallic zinc(II)-lactamase fromB. fragilis, to give a complex (58) of the hydrolyzed product and (57).
8.23.5.5 Type-II Aldolase Models
Zn2þcomplexes273,274have been prepared that model type-II aldolases, which catalyze reversible aldol addition reactions. Enolizations are essential for the carbon–carbon bond formation in an aldol reaction, and Zn2þ may play a role in such processes. For example, activation of the carbonyl of DHAP in the active site of type-II aldolases was performed by a Zn2þcomplex of 1-(4-bromophenacyl)- cyclen (59) (Scheme 38).273The pKavalue for the acidic protons of the carbonyl-position was 8.4.
The deprotonation yielded an enolate complex (60a) and a Zn2þ–OHspecies (60b) in a 23:77 ratio in aqueous solution. The results imply that the Zn2þ-coordination of the carbonyl group of DHAP certainly facilitates enolization by lowering the pKaof ca. 19 to a physically feasible pH.
8.23.5.6More Recent Developments
8.23.5.6.1 CA-mimetic zinc(II) fluorophores
The concentration of free Zn2þ within cells varies from about 1 nM in the cytoplasm of many cells, to about 1 mM in some vesicles. As the significance of Zn2þphysiology and homeostasis has
hydrolysis
benzylpenicillin
5R
(5S)-benzylpenicilloate (5R)-benzylpenicilloate
eqimerization
5S N-
Scheme 36
in DMSO
(58) nitrocefin
(57)
Scheme 37
increased, the need for useful zinc fluorophores to quantify trace Zn2þflux in biological dynamics has become more urgent.
In a development since the first discovery of sulfonamides as inhibitors of CA,183 Chen and Kernohan studied the interaction of bovine erythrocyte CA with dansylamide (61) and saw a highly fluorescent complex (62) with a dissociation constant Kd of 2.5107M at pH 7.4 (Scheme 39).184
carbonic anhydrase (CA) pH 7.4 dansylamide
(61)
(62) EM: 580nm
EM: 320nm
EM: 468nm EM: 326nm
Hydrophobic environment
Scheme 39
By combining established knowledge of CA-sulfonamide complexes185–188,275,276
and CA-model Zn2þcomplexes,216–222,252–256,277a new type of selective and efficient zinc fluorophore, a dansylami- doethylcyclen (63), was designed (Scheme 40).278–282 The 1:1 complexation constant for(64) (log Kẳlog ([Zn2ỵ–H1L]/[L2Hỵ][Zn2ỵ])) was as large as 20.8 M1at 25C withIẳ0.10 (NaNO3) as determined from potentiometric pH titrations.278The response of (63) (5mM) to Zn2þ(0.1–5mM) (at 528 nm) was linearuntil it reached a 1:1 [Zn2þ]/[63] ratio, and then became a plateau. Alternatively, Cu2ỵ linearly diminished the fluorescence emission until complete quenching at [Cu2ỵ]/[63]ẳ1, although Cu2ỵ forms the most stable five-coordinate complex Cu2ỵ–H1L (log Kẳlog ([Cu2ỵ– H1L]/[Cu2þ][H1L])>30 at 25C). New macrocyclic Zn2þfluorophores have been found, including anthrylmethyl-cyclen (65),283–286Zinquin (66),287–289ACF-1 and -2 (67aand67b),290orZinpyr-1 (68)291,292 (Scheme 41).281
8.23.5.6.2 Artificial receptors by Zn2+enzyme models
As substrates and inhibitors, phosphate anions bind as reversibly monodentate ligands to zinc(II) ions in CPA,293 alkaline phosphatase,139 phospholipase C,73 oralkaline phosphatase A.222 The
(59) (60a) (60b)
pKa=8.4 –H3O+
in H2O
23% 77%
–H2O
Scheme 38
zinc(II)-enzyme-model complexes, (33) and (34), likewise strongly bind to dianionic phosphates such as HPO42, phenyl phosphate (PP), and 4-nitrophenyl phosphate (4-NPP) to yield 1:1 complexes (43) and (44), while metal-free cyclen (similarly dicationic with two protons (L2Hþ)
at neutral pH) ineracted only weakly with these phosphate dianions. The 1:1 complexation constant logKsfor(44) was 3.1 in aqueous solution withIẳ0.2 (NaClO4) at 25C (Scheme 42).218 A bis(Zn2þ–cyclen) complex linked with a m-xylene spacer(69)294 and tris(Zn2þ–cyclen) (70)295 are much better phosphate receptors, with logKsof 4.0 and 5.8 (with 4-NPP).296,297
Much higher emission
λex=320nm λem=528nm Zn-H–1L(64) in H2O
at neutral pH dansylamidoethylcyclen
L.2H+(63) λem=555 nm
+Zn2+
–3H+
Scheme 40
(65)
Em: 416nm (66)
(68)
(67a) (67b)
Zinquin
Em:490nm Em:370nm
Em: 525nm
Em: 515 nm Em: ca. 525nm(with Zn2+)
Ex: 515nm (without Zn2+) Ex: 570nm (with Zn2+) Zinpyr-1
R=CI: ACF-2 R=H: ACF-1
Scheme 41
8.23.5.6.3 Guanidine coordination to Zn2+-enzyme-model complexin aqueous solution at neutral pH
Contrary to a prevalent perception that an arguinine residue always acts as a protonated guanidinium cation in an enzyme, a neutral form of guanidine was discovered to coordinate to Zn2þ in a cyclen functionalized with a guanidinylethyl pendant (71) (pKa>12).298 The guanidi- nium was deprotonated in the vicinity of Zn2þwith pKaof 5.9 to yield ZnL (72b). Moreover, the Zn2þ-bound guanidine in (72b) was displaced by a phosphate monoesterin aqueous solution at pH5.5 to form (73) via Zn2þ–OH2species (Zn(LHþ)H2O) (72a) (Scheme 43). These findings with the Zn2þ-enzyme model suggest transient coordination of arguinine residues in the transition state of the Zn2þ-promoted phosphatase reaction.