Synthesis of nitric oxide reductase active site models bearing key components at both distal and proximal sites

Article abstract of DOI:10.1021/ol071007p

Porphyrins 1ab and 2ab were successfully synthesized from cis-α₂-bisimidazole-β-imidazole-tail porphyrins and two newly synthesized imidazole pickets containing an aliphatic ester chain following a [2+1] approach. The four compounds possess a d

Full text of DOI:10.1021/ol071007p

ORGANIC
LETTERS
2007
Vol. 9, No. 15
2855-2858
Synthesis of Nitric Oxide Reductase
Active Site Models Bearing Key
Components at Both Distal and
Proximal Sites
James P. Collman,* Ying Yang, and Richard A. Decre´au
Department of Chemistry, Stanford UniVersity, Stanford, California 94305-5080
jpc@stanford.edu
Received April 30, 2007
ABSTRACT
Porphyrins 1ab and 2ab were successfully synthesized from cis-r2-bisimidazole-â-imidazole-tail porphyrins and two newly synthesized imidazole
pickets containing an aliphatic ester chain following a [2 1] approach. The four compounds possess a distal trisimidazole set, a distal
+
carboxylic acid, and a proximal imidazole, which constitute all the key features of the coordination environment of the active site in Bacterial
Nitric Oxide Reductase (NOR) and make them the closest synthetic NOR model ligands to date.
The success in modeling cytochrome c oxidase (CcO)
chemistry1 aroused our interest in biomimetic studies of nitric
oxide reductase (NOR), a distant member of the superfamily
of cytochrome oxidases. NOR, a membrane-bound enzyme,
is involved in bacterial denitrification, the sequential reduc-
tion of nitrate to dinitrogen; it catalyzes the 2e- reduction
of nitric oxide to nitrous oxide with the formation of a N-N
bond.2 Structurally, CcO and NOR have many similar
features, as well as some significant differences. Both active
sites are bimetallic with a monohistidine ligated five-
coordinate heme and a trisimidazole ligated distal metal.
However, FeB occupies the distal position in NOR instead
of CuB in CcO. Also, a tyrosine residue in CcO is missing
in NOR, while a glutamic acid residue is postulated to
coordinate to the FeB center in NOR.
So far, few synthetic models have been developed to study
NOR. A diiron porphyrin lacking the glutamic acid mimic
and the proximal histidine group was able to bind NO to
form dinitrosyl complexes, but did not lead to N2O forma-
tion.3 Recently, a closer model featuring a porphyrin bearing
(1) (a) Collman, J. P.; Herrmann, P. C.; Boitrel, B.; Zhang, X. M.;
Eberspacher, T. A.; Fu, L.; Wang, J. L.; Rousseau, D. L.; Williams, E. R.
J. Am. Chem. Soc. 1994, 116, 9783. (b) Collman, J. P.; Fu, L.; Herrmann,
P. C.; Zhang, X. Science 1997, 275, 949. (c) Collman, J. P.; Rapta, M.;
Broring, M.; Raptova, L.; Schwenninger, R.; Boitrel, B.; Fu, L.; L’Her, M.
J. Am. Chem. Soc. 1999, 121, 1387. (d) Collman, J. P.; Sunderland, C. J.;
Boulatov, R. Inorg. Chem. 2002, 41, 2282. (e) Collman, J. P.; Sunderland,
C. J.; Berg, K. E.; Vance, M. A.; Solomon, E. I. J. Am. Chem. Soc. 2003,
125, 6648. (f) Collman, J. P.; Decre´au, R. A.; Zhang, C. J. Org. Chem.
2004, 69, 3546. (g) Collman, J. P.; Devaraj, N. K.; Decre´au, R. A.; Yang,
Y.; Yan, Y. L.; Ebina, W.; Eberspacher, T. A.; Chidsey, C. E. D. Science
2007, 315, 1565. (h) Decre´au, R. A.; Collman, J. P.; Yang, Y.; Yan, Y.;
Devaraj, N. K. J. Org. Chem. 2007, 72, 2794.
(2) (a) Tavares, P.; Pereira, A. S.; Moura, J. J. G.; Moura, I. J. Inorg.
Biochem. 2006, 100, 2087. (b) Wasser, I. M.; Vries, S. D.; Moenne-Loccoz,
P.; Schro¨der, I.; Karlin, K. D. Chem. ReV. 2002, 102, 1201. (c) Hendriks,
J.; Gohlke, U.; Saraste, M. J. Bioenerg. Biomembr. 1998, 30, 15. (d) Zumft,
W. G. Microbiol. Mol. Biol. ReV. 1997, 61, 533. (e) Averill, B. A. Chem.
ReV. 1996, 96, 2951.
(3) (a) Ju, T. D.; Woods, A. S.; Cotter, R. J.; Moenne-Loccoz, P.; Karlin,
K. D. Inorg. Chim. Acta 2000, 297, 362. (b) Wasser, I. M.; Ju, T. D.; Huang,
H. W.; Moenne-Loccoz, P.; Karlin, K. D. J. Inorg. Biochem. 2001, 86,
483. (c) Wasser, I. M.; Martens, C. F.; Verani, C. N.; Rentschler, E.; Huang,
H. W.; Moenne-Loccoz, P.; Zakharov, L. N.; Rheingold, A. L.; Karlin, K.
D. Inorg. Chem. 2004, 43, 651. (d) Wasser, I. M.; Huang, H. W.; Moenne-
Loccoz, P.; Karlin, K. D. J. Am. Chem. Soc. 2005, 127, 3310.
10.1021/ol071007p CCC: $37.00
© 2007 American Chemical Society
Published on Web 06/20/2007

trisimidazole picket and a carboxylic acid moiety at the distal
site showed that the presence of a carboxylic acid moiety
not only helped stabilize FeB coordination at the distal
binding site, but also mediated the redox properties of both
the heme Fe and FeB centers.4 However, to more faithfully
mimic the active site of NOR, a proximal histidine residue
should be included in the model. Although spectroscopic
studies indicate that the proximal histidine dissociates from
heme iron after its binding of NO,5 histidine binding also
prevents NO association with the metal center on the
proximal side. Also, recoordination of the detached proximal
histidine upon reduction of the diiron center and concomitant
loss of the bridging oxo ligand should stabilize the reduced
heme iron center. Moreover, recent studies on FeII porphyrin
NO complexes suggested that a six-coordinate heme-nitrosyl
intermediate with an imidazole axial ligand is beneficial for
the N-N bond forming step in the proposed mechanism.6
Such a six-coordinate species was detected in the reduction
of NO to N2O by a ba3-type heme-copper oxidase from
resonance Raman spectroscopy.7
In addition, the carboxylic acid moiety in 1ab is on the ortho
position of the phenyl ring that is N-substituted to a distal
imidazole, while in 2ab, the moiety is on the C-2 position
of a distal imidazole. The former may favor an orientation
parallel to the porphyrin plane, while the latter will possibly
point upward, which might provide a different coordination
environment for FeB after metalating these free bases. The
four porphyrins possess all the key features of the coordinat-
ing environment of the active site of the native NORs and
represent the best available free-base NOR ligands.
Given the similarities between the actives sites of CcO
and NOR, modification of our previous CcO models is a
promising strategy to provide functional NOR models. Since
the cis-R2-bisimidazole-â-tail-porphyrins 5ab were previ-
ously synthesized in our laboratory,1f,h and attaching a
phenol-containing imidazole onto them led to free-base CcO
models,1f the key is to synthesize new imidazole pickets
containing a carboxylic acid moiety and to attach them onto
5ab (Scheme 3) to afford free-base NOR models.
The synthesis of new imidazole pickets starts from the
key intermediate 8, which was described previously.1f The
desired alkyl chain carboxylic acid moiety can be introduced
onto either the C-2 position of the imidazole (position 1) or
the ether linkage on the benzene ring (position 2) (Figure
2). Before the imidazole picket is attached onto a porphyrin,
In this report, we disclose the synthesis of NOR ligands
1ab and 2ab which feature a porphyrin with three imidazole
pickets and a carboxylic acid moiety at the distal site and
an imidazole ligand at the proximal position; both the distal
ligand set and the proximal ligand are covalently attached
to the porphyrin (Figure 1). Porphyrins 1-2a have a proximal
Figure 2. Two types of imidazole pickets 9 and 10 containing an
alkyl chain ester and their key precursor 8.
the carboxylic acid needs to be protected as an ester to avoid
its interference in the amide-forming reaction. Since there
are two ester groups (COOEt and the aliphatic chain ester)
in the same imidazole picket, it is necessary to differentiate
them and selectively convert the COOEt to COOH while
keeping the aliphatic chain ester.
We chose to convert the ethyl ester group of 8 to an alde-
hyde, then introduce the aliphatic chain ester and oxidize the
aldehyde to afford the desired ester-containing imidazole acid.9
Therefore, compound 8 was reduced to alcohol 11 by LiAlH4
in THF (Scheme 1). Oxidation of 11 in the presence of excess
MnO2 in CH2Cl2 afforded the aldehyde 12. These two steps
gave very good yields (>90%). Transformation of the phenyl
methyl ether of 12 to phenol was first tried with BBr3. How-
ever, the yield was very low (10-13%) even with over 10
equiv of BBr3 and a prolonged reaction time (room temper-
Figure 1. NOR ligands 1ab and 2ab.
CF3Ph imidazole tail, while 1-2b possess a terminal alkyne
moiety on the tail. The former is to be used as a comparison
to our first-generation NOR model4 for spectroscopic and
NO reactivity studies, whereas the latter is to be covalently
attached onto electrode surfaces by Sharpless Click Chem-
istry8 in order to conduct electrocatalytic studies of 2e-
reduction of NO to N2O under rate-limiting electron flux.
(4) (a) Collman, J. P.; Yan, Y. L.; Lei, J. P.; Dinolfo, P. H. Org. Lett.
2006, 8, 923. (b) Collman, J. P.; Yan, Y. L.; Lei, J. P.; Dinolfo, P. H.
Inorg. Chem. 2006, 45, 7581.
(5) Praneeth, V. K. K.; Na¨ther, C.; Peters, G.; Lehnert, N. Inorg. Chem.
2006, 45, 2795.
(6) Pinakoulaki, E.; Ohta, T.; Soulimane, T.; Kitagawa, T.; Varotsis, C.
J. Am. Chem. Soc. 2005, 127, 15161.
(7) Moeenne-Loccoz, P.; de Vries, S. J. Am. Chem. Soc. 1998, 120, 5147.
(8) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed.
2001, 40, 2004.
(9) Our attempts to selectively hydrolyze an imidazolyl carboxylic acid
ethyl ester in the presence of an aliphatic tert-butyl ester were unsuccessful.
In our control experiments, the tert-butyl ester was always hydrolyzed first.
2856
Org. Lett., Vol. 9, No. 15, 2007

coupling reaction between 15 and an alkylzinc iodide reagent
[IZn(CH2)4COOMe] was tried to produce compound 17.
Although it has been reported that such selective carbofunc-
tionalization of the imidazole moiety at the C-2 atom worked
pretty well and afforded desired products in excellent yields,10
in our hands, the starting material was always recovered
cleanly without any reaction. Finally, a Pd-catalyzed Sono-
gashira coupling reaction was used to react the iodoimidazole
15 with methyl 4-pentynoate yielding 16 in 70% yield.
Subsequent hydrogenation of the alkyne moiety under
atmospheric H2 pressure in the presence of Pd/C led to 17
in 75% yield.11a Oxidation of the aldehyde group of 17 by
NaClO2/NaH2PO4 afforded the imidazole acid 7 in nearly
quantitative yield.
Scheme 1. Synthesis of O-Alkylated Imidazole Acid 6
The imidazole acids 6 and 7 can be easily converted to
their corresponding acyl chlorides 9 and 10 by reaction with
oxalyl chloride or thionyl chloride under anhydrous condi-
tions. While the acyl chloride 9 precipitated during the
reaction and could be separated as a pure solid in high yield,
the acyl chloride 10 remained as an oil. Attempts to
precipitate 10 from various solvent mixtures such as diethyl
ether and acetonitrile failed.
ature, 72 h). A possible reason is that the complex formed
by the coordination of BBr3 with N atoms in compound 12
became nearly insoluble, which may prevent further attack
by BBr3 on the O atom of the ether moiety. The fact that
more than half of the starting material was usually recovered
is consistent with this explanation. The methyl ether was
successfully removed to yield phenol 13 in 62% yield when
refluxing 12 in 48% HBr aqueous solution. Compound 14
was easily obtained by refluxing 13 and methyl 5-iodopen-
tanoate in the presence of excess K2CO3 in dry acetonitrile.
Subsequent oxidation of 14 with NaClO2/NaH2PO4 led
quantitatively to the formation of pure imidazole acid 6.
Synthesis of the second imidazole picket containing an
aliphatic chain ester on the C-2 position of the imidazole
ring is described in Scheme 2. Refluxing 12 with 2 equiv of
Synthesis of R3-trisimidazole-â-tail-porphyrins 3ab and
4ab followed the [2+1] strategy formerly developed in our
group (Scheme 3).1f Freshly prepared acyl chloride 9 reacted
Scheme 3. Synthesis of NOR Models 1ab and 2ab following
a [2+1] Approach
Scheme 2. Synthesis of C-Alkylated Imidazole Acid 7
with acid protected cis-R2-bisimidazole-â-tail-porphyrins 5ab
and led to 3ab without any other byproduct. The [2+1]
products have almost the same polarity as their corresponding
starting materials, which made it impossible to monitor the
progress of the reaction by TLC. Therefore, an excess of 9
(>10 equiv) was added to ensure good conversion of 5ab.
Separation of the reaction mixture that usually contains the
product and the unreacted starting material was achieved by
rotating disk chromatography (chromatotron). Though only
(10) Evans, D. A.; Bach, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1326.
(11) (a) About 10-15% of alcohol was obtained, which was very easily
converted to aldehyde 17 with MnO2 in CH2Cl2. (b) After removing all
excess thionyl chloride, the residue was washed with anhydrous ether (2 ×
1 mL), then left under vacuum overnight.
N-iodosuccinimide in dry THF resulted in the formation of
an iodoimidazole 15 in 90% yield. First, a Pd-catalyzed
Org. Lett., Vol. 9, No. 15, 2007
2857

one band could be observed, the last part of the fraction was
collected separately and it turned out to be the clean [2+1]
product.
synthons.11 We suggest that there are several conformers
present due to the rotation of two functional groups, the alkyl
chain and the methoxyphenyl ring, which is reminiscent of
the phenomenon encountered in our previous CcO model
containing one methoxyphenyl ring that ended up having
several conformers.14 The ratio of the split peaks is always
1:1 for the spectra obtained with different NMR solvents
and on NMR instruments with different magnet field (500
and 600 MHz), suggesting the presence of two sets of
conformers with almost the same free energy. This complex
phenomenon involving several conformers in a racemic
mixture is currently being studied by ROESY and COSY
NMR, Gaussian and Spartan calculations, and also through
the spectroscopic study of simpler models bearing parts of
2 in order to disentangle the contribution of each component
in this process. A thorough spectroscopic analysis of these
complex models (1 and 2) and their metalated versions will
be reported in future work.
The proton signals of both the alkyl ester and alkyl acid
chain in compounds 1-4ab moved upfield (0.5-2.0 ppm)
compared to those in the free imidazole pickets 6 and 7,
implying that they are actually suspended over the porphyrin
plane instead of rotated away from the porphyrin.
Free bases 1ab and 2ab possess all the key groupss
porphyrin, proximal imdazole, three distal imidazoles includ-
ing one imidazole containing an aliphatic carboxylic acids
and represent the closest metal-free models for the natural
NOR active site reported to date. These are to be metalated
with Fe at both porphyrin and the distal site and will be
examined as functional NOR models for biomimetic and
mechanistic studies of reduction of NO to N2O. Spectroscopic
characterization of the diiron porphyrins and investigation
of interaction between diiron porphyrins and O2 and NO are
currently in progress and will be reported later.
The syntheses of 4ab were not as straightforward as that
of 3ab. The reaction of 5a with freshly prepared acyl chloride
10 from oxalyl chloride led to 4a (25%) along with several
polar fractions. Characterization of one of these (compound
2c, structure shown in the Supporting Information) shows
that it bears an oxamic acid picket. This inefficient synthesis
was due to difficulties in working up the acyl chloride
precursor 10, which forms an oil always containing some
oxalyl chloride. When 10 was obtained from the reaction of
7 with thionyl chloride, it could be more easily worked up
than when prepared from oxalyl chloride.11b As a result, the
reaction of a cleaner version of 10 with 5ab led mainly to
the expected product of 4ab along with unreacted starting
material and a very small amount of a polar fraction.
Surprisingly, the [2+1] products 4ab were less polar than
their starting materials 5ab, and it was possible to monitor
the progress of the reaction by TLC and to separate 4ab from
5ab with the chromatotron.
The methyl esters were saponified to afford free bases 1ab
and 2ab in high yields. The inseparable mixture of 3a, 5a
and 3b, 5b can be easily separated after saponification of
3ab to 1ab, since the acid products 1ab are much more polar
than the remaining compounds 5ab. Thus, the unreacted
valuable cis-bisimidazole porphyrin synthon can be cleanly
recovered. Porphyrin compounds 1-4ab were purified and
characterized based on the standards of porphyrinoid and
porphyrin chemistry established previously.12
1
Model 1 is a racemic mixture that has H NMR features
reminiscent of earlier racemic CcO models.1f,h,12b However,
model 2, which is structurally quite similar to 1, is character-
ized by a distinct 1H NMR spectrum with a split in the signals
of N-Me, OMe(ether), OMe (ester), and to a lesser extent
the alkynyl and the NH protons. This was unexpected given
the high degree of purity of our models and precursor
Acknowledgment. This material is based upon work
supported by NIH Grant Nos. GM-69568-01A1. We thank
Dr. Stephen Russell Lynch of the Stanford Univeristy NMR
facilities for NMR experiments and analyses. We also thank
Dr. Allis Chien of the Stanford University Mass Spectrometry
Group for mass spectrometry analysis. R.A.D. is thankful
for a Lavoisier Fellowship.
(12) (a) Collman, J. P.; Broring, M.; Fu, L.; Rapta, M.; Schwenninger,
R.; Straumanis, A. J. Org. Chem. 1998, 63, 8082. (b) Collman, J. P.; Broring,
M.; Fu, L.; Rapta, M.; Schwenninger, R. J. Org. Chem. 1998, 63, 8084. (c)
Collman, J. P.; Brauman, J. I.; Doxsee, K. M.; Halbert, T. R.; Bunnenberg,
E.; Linder, R. E.; Lamar, G. N.; Delgaudio, J.; Lang, G.; Spartalian, K. J.
Am. Chem. Soc. 1980, 102, 4182. (d) Collman, J. P.; Gagne, R. R.; Reed,
C. A.; Halbert, T. R.; Lang, G.; Robinson, W. T. J. Am. Chem. Soc. 1975,
97, 1427. (e) Collman, J. P.; Decre´au, R. A. Org. Lett. 2005, 7, 975.
(13) (a) Careful analyses of the purity of imidazole and porphyrin
Supporting Information Available: Synthetic procedures
and characterization including 1H NMR, 13C NMR, LRMS,
HRMS, and HPLC-MS of compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
synthons in the syntheses of models 2ab were carried out by 1H NMR, 13
C
NMR, and HPLC-MS. Only the C-2 regioisomer was found in the
iodoimidazole precursor 15 and the subsequent synthons 16, 17, and 7.
Substitution at C-4 vs C-2 in imidazoles is known to trigger significant
changes in their 1H and 13C NMR spectra, which is well documented in
the literature.13b Moreover, these analyses did not reveal the presence of
trans-regioisomer contaminant in the cis-bisimidazole porphyrin precursor
5ab. (b) O’Connell, J. F.; Parquette, J.; Yelle, W. E.; Wang, W.; Rapoport,
H. Synthesis 1988, 767.
OL071007P
(14) Collman, J. P.; Decreau, R. A.; Costanzo, S. Org. Lett. 2004, 6,
1033.
2858
Org. Lett., Vol. 9, No. 15, 2007