Iron(III) corroles and porphyrins as superior catalysts for the reactions of diazoacetates with nitrogen- or sulfur-containing nucleophilic substrates: Synthetic uses and mechanistic insights

Article abstract of DOI:10.1002/chem.200701885

A thorough mechanistic investigation has been performed on the reactions of primary and secondary amines with diazoacetates, which proceed uniquely quickly and efficiently when catalyzed by iron(III) corroles and porphyrins. Two major differences in relation to other metal-based catalysts are that the iron complexes are not poisoned by excess amine and that metal-carbene intermediates are apparently not involved in the reaction pathway. The results instead point towards nitrogen ylide intermediates formed by nucleophilic attack of the amines on diazoacetate-coordinated iron complexes. Nitrogen ylides are also formed when allyl- and propargylsubstituted tertiary amines react with diazoacetates, a scenario that smoothly leads to 2,3-rearrangement reaction products with catalytic amounts of the iron(III) complexes. Similar findings regarding the superiority of the iron-(III) complexes (in terms of catalyst loading, chemical yields, and reaction conditions) were obtained with thiols (S-H insertion) and sulfides (2,3-rear-rangement reactions), which suggest similar mechanisms operate in these cases.

Full text of DOI:10.1002/chem.200701885

FULL PAPER
DOI: 10.1002/chem.200701885
Iron
Diazoacetates with Nitrogen- or Sulfur-Containing Nucleophilic Substrates:
Synthetic Uses and Mechanistic Insights
(III) Corroles and Porphyrins as Superior Catalysts for the Reactions of
AHCTREUNG
Iris Aviv and Zeev Gross*[a]
Abstract: A thorough mechanistic in-
vestigation has been performed on the
reactions of primary and secondary
amines with diazoacetates, which pro-
ceed uniquely quickly and efficiently
pathway. The results instead point to-
wards nitrogen ylide intermediates
formed by nucleophilic attack of the
amines on diazoacetate-coordinated
iron complexes. Nitrogen ylides are
also formed when allyl- and propargyl-
substituted tertiary amines react with
diazoacetates, a scenario that smoothly
leads to 2,3-rearrangement reaction
products with catalytic amounts of the
iron
regarding the superiority of the iron-
(III) complexes (in terms of catalyst
A
when catalyzed by iron
A
ACHTREUNG
and porphyrins. Two major differences
in relation to other metal-based cata-
lysts are that the iron complexes are
not poisoned by excess amine and that
metal–carbene intermediates are appa-
rently not involved in the reaction
loading, chemical yields, and reaction
conditions) were obtained with thiols
ꢀ
(S H insertion) and sulfides (2,3-rear-
Keywords: catalysis
iron porphyrinoids
mechanisms
·
corroles
· reaction
·
rangement reactions), which suggest
similar mechanisms operate in these
cases.
·
Introduction
The catalytic transfer of a carbene moiety from diazo com-
pounds to organic substrates is a very powerful tool in or-
ganic synthesis used for the production of numerous com-
pounds.[1] One example is the use of quite stable a-diazo
esters (ethyl diazoacetate (EDA) is commercially available)
Scheme 1. Insertion of carbene moieties derived from a-diazoacetates
ꢀ
into amine N H bonds.
the 1970s by Paulissen et al.[6] However, both the earlier and
the more recently developed copper-, silver-, gold-, and
ruthenium-based catalysts are usually required in amounts
of 1 to 10 mol% and suffer from one or more of the follow-
ing limitations: gradual addition of the diazo compound to
avoid dimerization byproducts, gradual addition of the
amine to avoid catalyst poisoning, long reaction times, and
low-to-moderate chemical yields.[2,7–9]
ꢀ
for the insertion of CRCO2R’ into X H bonds, in which X=
R3C, R2N, RO, or RS.[2] The N H insertion reaction is par-
ꢀ
ticularly versatile and leads to the synthesis of a-amino
esters, dipeptides, nitrogen-containing heterocycles, and
other protected a-amino acid derivatives that can be used as
precursors to natural products (Scheme 1).[3]
ꢀ
The first reported catalysts for N H insertion reactions
were copper bronze[4] and CuCN.[5] These catalysts were
Decade-long efforts were devoted to the development of
catalysts for enantioselective metal–carbene insertion into
overshadowed by [Rh2(OAc)4], developed as a catalyst in
A
ꢀ
N H bonds. Chiral auxiliaries served quite well for inducing
ꢀ
rhodium-catalyzed asymmetric intramolecular N H inser-
[a] Dr. I. Aviv, Prof. Dr. Z. Gross
Schulich Faculty of Chemistry
Technion - Israel Institute of Technology
Haifa 32,000 (Israel)
tion (optically active proline derivatives were obtained),[10]
but their extension to intermolecular reactions met with
very limited success. Twenty one different chiral RhII cata-
lysts were examined for that purpose in one reaction, but
the highest enantiomeric excess (ee) obtained was 9%.
Even the diastereoselectivity appeared to be low because
the use of chiral amino acid derivatives as substrates result-
ed in diastereomeric excesses (de) of only up to 37%.[11]
Fax : (+972)4829-5703
E-mail: chr10zg@tx.technion.ac.il
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author. It contains physi-
cal data, GC/HPLC retention times, and NMR spectra for all prod-
ucts.
Chem. Eur. J. 2008, 14, 3995 – 4005
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3995

Better results (up to 45% ee)
were reported by Garcia et al.
who developed an asymmetric
ꢀ
intramolecular N H insertion
based on chiral RhII cata-
lysts.[12] In
a
more recent
effort, Jørgensen and co-work-
ers obtained 27% ee with CuI
and 48% ee with an AgI-based
catalyst in the intermolecular
Scheme 2. Corrole- and porphyrin-based catalysts used in the investigations.
ꢀ
N H insertion of diazoacetates
into substituted anilines.[7b] The
only really successful example was reported last year by
Zhou and co-workers: copper complexes of chiral spiro bis-
of triarylcorroles and tetraarylporphyrins (Scheme 2) display
unique features with regards to the insertion of diazoace-
tates into the NH bonds of anilines (Scheme 1). Full and
very fast conversion was obtained by simultaneous addition
of equimolar amounts of the substrates to the catalyst
A
ular insertion of ethyl diazopropionate (EDP) into ani-
lines.[13]
ꢀ
The mechanism of metal–carbene insertion reactions into
(0.1 mol%) and the selectivity towards activation of the N
C H bonds has been extensively explored with rhodium
H bond (vs. C=C, CH, and OH bonds) was absolute.[21] We
ꢀ
acetate as a catalyst and is generally accepted to be a con-
certed process that proceeds through either one of the tran-
sition states proposed by Doyle, Taber, and Nakamura and
their co-workers.[2,14] A concerted mechanism for silver- and
copper-based catalysts postulated from computational
models was also suggested by PØrez and co-workers.[15]
There have also been several mechanistic investigations into
also reported the success of the same catalysts in a much
more challenging reaction, the activation of the N H bond
ꢀ
of ammonia (Scheme 1, R1 =R2 =H, R=H or CH3). In fact,
the iron complexes appeared to be the first catalysts capable
of inducing the formation of nitrogen-free glycine and ala-
nine esters from ammonia and diazo esters.[22] We now
report a full study of these intriguingly facile transforma-
tions that focused on the following aspects: Identification of
the key reaction intermediate involved in the reactions, and
more importantly, the mechanism of its formation; catalysis
in aqueous solutions with the aid of both biological and bio-
mimetic catalysts; extension of the synthetic utility of the re-
action to nonaromatic amines, carbon-substituted diazoace-
tates, and thiols; and the design of very efficient reaction
conditions for obtaining the 2,3-sigmatropic rearrangement
products from diazoacetates and tertiary amines or sulfides.
All of these reactions were found to proceed faster and with
lower catalyst loadings than in previously reported systems.
We also show that, in contrast with the proposals for other
catalysts, the iron-catalyzed reactions apparently do not pro-
ceed through metal–carbene intermediates.
ꢀ
O H insertion reactions in which both concerted and step-
wise pathways were considered.[2,16] A Hammett plot analy-
sis led Wang and co-workers to conclude the reaction pro-
ceeded through a concerted transition state,[17] whereas Hu
and co-workers found very strong evidence in favor of a
stepwise reaction pathway.[18] Aldehydes and imines were
used as electrophilic trapping agents of the proposed reac-
tion intermediate, an oxonium ylide formed by the nucleo-
philic addition of the alcohol to the metal–carbene. This sce-
nario is of synthetic utility for three-component reactions of
aryl diazoacetates, alcohols, aldehydes, and imines.
One shortcoming, which apparently hampers the develop-
ment of catalysts for enantioselective insertion of diazoace-
ꢀ
tates into N H bonds, is that surprisingly little is known
about the mechanistic aspects of this reaction. Stepwise re-
action sequences involving ylides and concerted three- or
four-centered transition states have been proposed in the lit-
erature.[1a,2] Investigations focusing on the stereochemical
Results and Discussion
ꢀ
ꢀ
outcome of the N H insertion process led Davis et al. to
Insertion of the carbene moiety from EDA into amine N H
propose a concerted mechanism,[10] whereas evidence for a
stepwise mechanism involving a nitrogen ylide came from
investigations of three-component experiments with imines
or azodicarboxylates as electrophiles to trap this intermedi-
bonds: The data presented in Table 1 compares the potency
of iron corroles/porphyrins with other metal complexes in
the catalysis of the reaction of aniline with EDA. The iron-
(III) complexes are clearly unmatched by any other catalyst
ate.[1b,19] Importantly, practically all discussions about metal-
in terms of the following parameters: catalyst loading
(0.1 mol%), reaction time (min vs. h), reaction conditions
(aerobic and reagents added as a single portion), and chemi-
cal yields of isolated products.
More detailed information is provided in Table 2, which
summarizes the results obtained for the reactions of many
aniline derivatives with EDA (Scheme 3) utilizing specific
catalyzed reactions of diazoalkanes with compounds that
contain CH, OH, or NH bonds assumed that a metal–car-
bene (M=CRR’, in which M=transition metal, R=H, and
R’=CO2Et in the case of EDA) is formed (usually in the
rate-limiting step) prior to interaction with the substrate.
As part of the large developments in corrole-based appli-
cations,[20] we recently reported that the iron
A
iron(III) corroles and porphyrins as catalysts.
U
3996
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3995 – 4005

Iron(III) Corroles and Porphyrins as Catalysts
A
FULL PAPER
Table 1. Metal-catalyzed reactions of aniline with EDA (at room temperature unless specified differently).
Despite the very simple re-
action conditions (both re-
agents added as a single por-
tion to an open-to-air solution
of catalyst in nonpurified sol-
vent) and the use of only
0.1 mol% of catalyst, N-aryl-
glycine esters 4a–f were isolat-
ed in very high yields in all
cases. The reactions were also
not limited to primary amines.
The N-(4-chlorophenyl)glycine
ester (4b, obtained from 4-
chloroaniline and EDA) was
efficiently converted into bis-
substituted N,N-bis(4-chloro-
phenyl)glycine ester 12b (en-
tries 14 and 15) and N-methyl-
aniline also provided the ex-
pected product 5 (entries 7 and
12). None of the other com-
plexes, which includes the
most commonly used [Rh2-
Metal catalyst
FeIII corrole or
porphyrin
Aniline/EDA/catalyst
1000:1000:1
Reaction conditions
Yield [%]
>92[a]
Time
Ref.
[b]
aerobic, added in one portion
<3 min
[b]
[b]
[b]
A
N
1000:1000:1
1000:1000:1
1000:1000:1
no data
50:50:1
20:20:1
aerobic, added in one portion
aerobic, added in one portion
aerobic, added in one portion
808C
aerobic, syringe pump
anaerobic, ionic liquids, slow
addition of EDA
anaerobic, dropwise addition
50% aniline in CH2Cl2
anaerobic, slow addition of both
reagents
anaerobic, aniline as solvent, slow
addition of EDA
<5[c]
0
0
1 h
24 h
24 h
E
A
(OAc)4]
no data
[6]
[7a]
[23b]
CuI
[Cu
8[a]
8[a]
20 h
20 h
AgI
AuI
10:10:1 (with EDP)
5500:25:1
100:150:1 (with 4-Me-
aniline)
8[a]
>99[d]
64[c]
20 h
24 h
>5.5 h
[7b]
[6b]
[8a]
RuII porphyrin
ReVII (MTO)
270:250:1
89
1 h
[23b]
[a] Isolated yield. [b] This work. [c] Yield determined by in situ NMR spectroscopy. [d] Yield determined by in
situ GC.
Table 2. Metal-catalyzed reactions of ring-substituted anilines with EDA.[a]
Entry
Catalyst
Substrate
Time
Yield [%]
Product
1
2
3
4
5
6
7
8
2a
2a
4-Cl-C6H4NH2
3-CN-C6H4NH2
4-Cl-C6H4NH2
C6H4NH2
4-MeO-C6H4NH2
4-Me-C6H4NH2
N-methylaniline
C6H4NH2
4-Cl-C6H4NH2
4-Cl-C6H4NH2
4-CN-C6H4NH2
N-methylaniline
4-Cl-C6H4NH2
4b
3 min
5 min
3 min
3 min
2 min
2 min
2 min
6 min
3 min
3 min
2 min
2 min
3 min
15 min
3 min
1 h (24 h)
1 h
92
94
92
93
95
90
91
93
90
94
97
96
95
91
92
4b
4 f
4b
4a
4d
4c
5
4a
4b
4b
4e
5
4b
12b
12b
4b
4b
–
2b
2b
2b
2b
2b
2b[b]
2c
3a
3a
3a
A
the reactions to any significant
extent when applied under the
reaction conditions used with
the iron(III) corroles and por-
E
9
phyrins. With one order of
magnitude lower amounts of
aniline and 2 mol% rather
than 0.1 mol% of catalyst, rho-
dium corrole 1a was capable
10
11
12
13
14
15
16
17
18
19
3b
2a[c]
3a[c]
4b
ꢀ
G
(OAc)4]
4-Cl-C6H4NH2
4-Cl-C6H4NH2
4-Cl-C6H4NH2
4-Cl-C6H4NH2
<5 (25)
92
of catalyzing the N H inser-
tion reaction. The expected
product was obtained within
one hour with high selectivity
(only traces of diethyl maleate
and diethyl fumarate were ob-
tained, entry 17). This demon-
strates that the rhodium cor-
role is capable of catalyzing
the reaction, but that only
quite small amounts of (the
not very basic) aniline can be
used; therefore, catalyst 1a is
not suitable for synthetic use.
Iron salen and ruthenium cor-
role did not catalyze the reac-
tion even under these less de-
manding reaction conditions
(entries 18 and 19). The superi-
ority of the iron complexes as
catalysts for NH insertion re-
actions is very different from
the results obtained for reac-
tions between EDA and ole-
G
24 h
24 h
no reaction[e]
no reaction
G
–
[a] Catalyst/EDA/substrate 1:1000:1000 with 0.3–0.75 mm catalyst at room temperature in diethyl ether. EDA
and the substrate were added together in one portion and the reported yields are of isolated products. [b] Cat-
alyst/EDA/substrate 1:1000:500. [c] Catalyst/EDA/substrate 1:100:100. [d] Catalyst/EDA/substrate 1:50:100.
[e] There was also no reaction when this reaction was performed in methanol in the presence of iron powder.
ꢀ
Scheme 3. Catalytic insertion of the carbene moiety derived from EDA into amine N H bonds.
Chem. Eur. J. 2008, 14, 3995 – 4005
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3997

I. Aviv and Z. Gross
Table 3. Metal-catalyzed reactions of nonaromatic amines with EDA.[a]
fins or alcohols.[21,23] In these reactions the iron complexes
are actually much poorer catalysts than [Rh
um corroles, and ruthenium porphyrins.
A
Entry Catalyst Substrate
Time
Yield [%] product
1
2
3
4
5
6
7
3a[b]
3a[b]
3a[b]
2b[d]
3a[d]
3b[d]
3b[d]
piperidine
2-methylpiperidine 5 min
2-methylindoline
morpholine
morpholine
7 min
97
7
8
9
10
10
10
12a
6
The poor performance of [Rh
C
94[c]
93
5 min
4 h
5 min
5 min
15 min 70
30
15 min 95
89[c]
96
95
1a, [Ru(tpfc)(NO)] (tpfc = the 5,10,15-trispentafluorophe-
U
morpholine
propylamine
nylcorrole trianion), and ruthenium porphyrins. Evidence to
prove that coordination of aniline to rhodium (and subse-
quent catalyst poisoning) occurs is given in Figure 1. Addi-
8
9
3a[e]
3a[f]
propylamine
alanine ethyl ester
12a
11
5 min
89
[a] EDA and the substrate were added together in one portion into
0.5 mL of catalyst dissolved in diethyl ether and the reported yields are
of isolated products. [c] In CH2Cl2. [b] Catalyst/EDA/substrate
1:1000:1000. [d] Catalyst/EDA/substrate 1:50:50. [e] Catalyst/EDA/sub-
strate 1:1000:500. [f] Catalyst/EDA/substrate 1:100:100.
and 3b to activate the primary amines alanine ethyl ester
and propylamine was also fruitful and the expected products
were obtained in high overall yields (entries 7–9). The bis-
activated product 12a (a tertiary amine) was the main prod-
uct obtained with propylamine as the substrate even when
the amine/EDA ratio was 1:1. This indicates that, in this
case, monosubstituted product 6 (a secondary amine) under-
goes insertion faster than the original substrate, the primary
propylamine.
Figure 1. Electronic spectrum of [Rh
and in the presence of 4-chloroaniline (gray).
N
G
Use of ethyl diazopropionate and aqueous solutions for pos-
ꢀ
tion of 4-chloroaniline to a solution of the red-colored com-
plex 1a induced a color change to green and the electronic
spectrum changed to one characteristic of bis-amine-coordi-
nated corroles.[24] Similarly, the color of the reaction mixture
immediately changed from brown to green when the sub-
strates were added to the solvated catalyst in reactions in
which the ruthenium corrole was used as the catalyst
sible enantioselective N H activation: One extension of the
above results was the use of EDP instead of EDA, the goal
being to develop routes to nonracemic alanine derivatives
(Scheme 4).[13] The results presented in Table 4 show that
(entry 17). In contrast, the electronic spectra of the iron(III)
T
complexes were not affected by aniline and the same con-
clusion regarding noncoordination of the two was reached
from NMR spectroscopy investigations. The characteristic b-
pyrrole resonance at d=+80 ppm for the five-coordinated
d5 high-spin complex 2b ( 6 mm in CDCl3) was not affected
by a 50 molar excess of aniline, which is in contradiction
with the formation of a complex with a different coordina-
tion environment or oxidation state.[25]
The apparent nonpoisoning of the iron corroles and por-
phyrins by aniline derivatives suggests that much more coor-
dinating amines may also be used in this system (Scheme 3,
Table 3). Piperidine, 2-methylpiperidine, and 2-methylindo-
line were indeed functionalized under the same reaction
conditions as those used for anilines to give the correspond-
ing products in isolated yields of 97, 94, and 93%. The reac-
tions were very fast and were completed within a few mi-
nutes (entries 1–3) with 0.1 mol% of 3a as the catalyst. The
use of 1 mol% of the catalyst with morpholine resulted in
the formation of the desired product and the reaction pro-
ceeded much faster with the porphyrins than with the cor-
role (entries 4–6). The use of porphyrin-based catalysts 3a
Scheme 4. Alanine derivatives from the catalytic insertion of the carbene
moiety derived from EDP into anilines. Mb=myoglobin, SA=serum al-
bumin.
the iron complexes are good catalysts for the transformation
of anilines into expected products 13a–d in high yields. Re-
actions in organic solvents were completed within minutes
with the iron porphyrins and corroles (entries 1–6) com-
pared with hours with the rhodium corroles (entries 7 and
8).
The goal of obtaining enantiomerically enriched products
could have been addressed by either preparing iron com-
3998
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3995 – 4005

Iron(III) Corroles and Porphyrins as Catalysts
A
FULL PAPER
Table 4. Metal-catalyzed reactions of anilines with EDP.[a]
advantage of the noncovalent conjugates formed between
iron corrole 14 (Scheme 4) and serum albumins, a combina-
tion that has been successfully used for chiral induction in
the sulfoxidation of thioanisoles by H2O2.[26] This approach
led to high yields when performed at 258C in aqueous solu-
tions with 0.1 mol% of 14 (Table 4, entries 11–19). However,
despite the milder reaction conditions, the use of five differ-
ent albumins, and the various pH conditions applied, the
alanine derivative was always obtained as a racemic mix-
ture.
To conclude, we found that Mb and iron corrole conjugat-
ed albumins were capable of catalyzing the reactions of
amines and diazo compounds in aqueous solutions. Howev-
er, the reactions did not lead to enantiomerically enriched
products. As previous success in similar approaches to a dif-
ferent reaction indicates that poor chiral induction by the
protein is unlikely to be the reason for the failure in this
case,[26] a mechanistic reason seemed more likely,[1b,19] and
this aspect has been explored.
Entry Catalyst Substrate
Time
T
[8C]
Yield
[%]
Solvent
1[a]
2[a]
3[a]
4[a]
5[a]
6[a]
7[a]
8[a]
9[b]
3a
3a
3a
3b
2b
2b
1a
1b
Mb
4-Cl-C6H4NH2
C6H4NH2
C6H4NH2
C6H4NH2
4-Cl-C6H4NH2
C6H4NH2
4-Cl-C6H4NH2
4-Cl-C6H4NH2
C6H4NH2
3 min 25
4 min 25
2 min 25
3 min 25
30 min 25
30 min 25
91
93
93
89
92
94
86
89
87
Et2O
Et2O
CH2Cl2
CH2Cl2
Et2O
Et2O
10 h
10 h
25
25
CH2Cl2
CH2Cl2
THF/
H2O
THF/
H2O
15 min 65
10[b]
Mb
4-CH3O-
C6H4NH2
15 min 65
85
11[c]
12[c]
13[c]
14[c]
15[c]
16[c]
17[c]
18[f]
19[h]
14/BSA C6H4NH2
14/HSA C6H4NH2
14/RSA C6H4NH2
14/PSA C6H4NH2
14/SSA C6H4NH2
14/BSA C6H4NH2
14/BSA C6H4NH2
14/BSA C6H4NH2
14/BSA C6H4NH2
5 h
5 h
5 h
5 h
5 h
24 h
5 h
5 h
1 h
25
25
25
25
25
4
25
25
25
85
85
H2O[d]
H2O[d]
H2O[d]
H2O[d]
H2O[d]
H2O[d]
H2O[e]
H2O[g]
H2O[i]
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
Mechanistic investigations addressing the formation of an
ylide intermediate on the pathway to the final product: The
most straightforward indication of the formation of ylide in-
termediates is the trapping of their nucleophilic moiety by
electrophiles, of which diethyl azodicarboxylate (DEAD) is
most commonly used (Scheme 5).[1b]
[a] Catalyst/EDP/substrate 1:100:100 with 0.4–0.5 mm catalyst in 2 mL of
solvent at RT. [b] Mb (5 mg), EDP (0.156 mmol), and amine
(0.156 mmol) in solutions of phosphate buffer (pH 7)/10% THF
(1.5 mL). [c] 14 (0.2 mm)/albumin/aniline/EDP 1:1.5:120:120 in phosphate
buffer, n.d.=not precisely determined (>70 yield), BSA, HSA, RSA,
PSA, and SSA=bovine, human, rat, pig, and sheep serum albumin, re-
spectively. [d] pH 7. [e] pH 5.8. [f] Trizma buffer. [g] pH 9. [h] Acetate
buffer. [i] pH 4.
plexes of chiral porphyrins/corroles or by using proteins that
contain natural (heme) or synthetic iron complexes. The
latter approach was adopted despite the fact that it requires
work in aqueous solution, which is challenging for many rea-
sons. The limitation unique to working with diazoacetates is
that they might react with water, especially in the presence
of metal complexes that might catalyze OH insertion reac-
tions. Nevertheless, and in accordance with previous re-
search that showed that the iron-based catalysts display a
very high selectivity towards the activation of NH compared
with OH bonds,[21] aniline did react with EDA at room tem-
perature to form expected product 4a when combined with
0.2 mm of the iron porphyrin containing protein myoglobin
(Mb) in a solution of 10% THF/90% aqueous buffer. The
same reaction, but with EDP instead of EDA (Scheme 4),
worked well when performed at 658C. Alanine-related prod-
ucts 13a and 13d were formed in very high yields and in
very short reaction times (Table 4, entries 9 and 10), but as
racemic mixtures (confirmed by HPLC on a chiral station-
ary phase). One obvious reason for the apparent failure of
the protein with regard to chiral induction could be that it
was denatured under the reactions conditions. This hypothe-
sis was disproved by examination of the CD spectrum of Mb
before and after reaction. As there was no change in the
spectrum, the non-enantioselectivity of the process appa-
rently does not reflect damage to the chirality-inducing host
of the prosthetic iron porphyrin. Another attempt to obtain
enantioselective reactions between aniline and EDP took
Scheme 5. Efficient trapping of ylide intermediate Y by DEAD.
To check for
a possible ylide intermediate (Y in
Scheme 5) in the current case, the reaction between aniline,
EDA, and DEAD was tested with iron catalysts 2b, 3a, and
3b. Equimolar amounts of the three reagents were mixed
and added in a single portion to solutions of the catalyst at
room temperature. The sole product in all cases was 15,
which was fully formed within minutes with iron porphyrins
3a and 3b and within hours with iron corrole 2b. This is
reminiscent of an investigation reported by Hu and co-work-
ers who used rhodium acetate as the catalyst in CH2Cl2 at
reflux under argon with dropwise addition of a dilute EDA
solution over one hour.[1b] To verify the result, the isolated
aniline glycine ester 4a was allowed to interact with DEAD
in the presence of 2b (or 3a). In this case, no reaction oc-
curred even after 18 h. The most important outcome of this
short investigation is that it readily explains the lack of
chiral induction for the reaction of EDP when performed in
the chiral environment provided by proteins (Scheme 4).
Apparently, the lifetime of reaction intermediate Y (R=
CH3 in Scheme 5) is long enough (i.e., internal nitrogen-to-
Chem. Eur. J. 2008, 14, 3995 – 4005
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3999

I. Aviv and Z. Gross
carbon proton transfer is relatively slow) to allow full race-
mization of its negatively charged carbon atom.
the rate-limiting formation of a metal–carbene intermediate
on the pathway to the ylide and the final products. In partic-
ular, it cannot explain why the reactions of amines are so
much faster than those of other substrates. The most ex-
Mechanistic investigations into the events leading to the for-
mation of the ylide intermediate: In spite of many other in-
vestigations (mainly on rhodium-based catalysts), it may not
be too surprising that either free (Y) or metal-bound (D) ni-
trogen ylide intermediates are also formed in the current
system (Scheme 6). Many groups have in fact demonstrated
treme example relates to iron(III) porphyrin 3b in which
A
there was no reaction when 3b and EDA were present in so-
lution for hours, but addition of amine caused the full re-
lease (about 900 catalytic turnovers) of nitrogen gas within
seconds.[22] One could have considered the amines as poten-
tial reducing agents of 3b to the more reactive [FeII
(tpp)]
G
(tpp=tetraphenylporphyrin), but it is known that most
amines used in this work are not capable of doing so (as we
have confirmed for aniline, see above).[29]
To shine some light on the mechanism of N-ylide forma-
tion, the electronic effect on the product-forming step in the
ꢀ
iron corrole/porphyrin catalyzed N H insertion reactions
was deduced from intermolecular competition experiments
between a series of ring-substituted aniline derivatives and
limiting quantities of EDA (5 mol% relative to the amines;
Scheme 7). N-Substituted glycine ethyl esters 4a–k were the
Scheme 6. Plausible steps leading to metal–carbene and nitrogen ylide in-
termediates in the metal-catalyzed reactions of amines with EDA.
ꢀ
Scheme 7. Competitive N H insertion of EDA into phenyl-substituted
anilines.
the synthetic utility of three-component reactions that pro-
ceed via oxygen or nitrogen ylides of late transition metals
formed in situ.[1b,18,19,27] There is also one report of a related
case that involves iron porphyrin 3b as the catalyst in a
three-component reaction between aldehyde, phenyldiazo-
methane, and trimethyl phosphite.[28] In all of these investi-
gations, the precursor of the ylide was proposed to be a
metal–carbene formed by the interaction of the diazoalkane
with the metal (B!C in Scheme 6). In fact, a metal–carbene
complex was also the starting point in research that conclud-
ed that ylides are not formed on the pathway that leads to
the final products.[10]
The fact that a metal–carbene intermediate is unlikely in
the current case is illustrated by Scheme 6, which outlines
the plausible elementary steps that lead from the reactants
to the products. Computational data (mainly for M=Rh)
suggests that the rate-limiting step is the formation of a car-
bene intermediate C from B with an activation energy of
about 17 kcalmolꢀ1 followed by its fast reaction with a sub-
strate (EDA, alcohol, olefin, or amine by routes c–f, respec-
tively, in Scheme 6).[14] The activation energy for forming
the iron–carbene intermediate (M=Fe in Scheme 6) is most
likely to be even larger than that for rhodium because the
sole products, which allowed straightforward determination
of the relative reactivities of the substituted anilines from
the observed product ratios. The corresponding Hammett
plot obtained with 2b as the catalyst correlated much better
with s (r2 =0.9449 for all data and r2 =0.9928 without 4-CN-
aniline) than with s+ (r2 =0.8416 for all data and r2 =0.8879
without 4-CN-aniline).[30] The same trend held for porphyrin
complex 3a, but both 3-CN- and 4-CN-substituted anilines
did not fit well with the correlation. For a fair comparison
between the results obtained for the corrole and the porphy-
rin iron complexes, the data for both 3-CN- and 4-CN-sub-
stituted anilines were ignored, which led to 1=ꢀ1.82 (r2 =
0.9912) for the former and 1=ꢀ1.53 (r2 =0.9741) for the
latter (Figure 2). The smaller value of 1 obtained with 3a
than with 2b is consistent with the experimental observa-
tions that show the porphyrin complex reacts faster than the
corrole. According to the Hammond principle, this should
be reflected in an earlier transition state and a smaller 1
value. The more important information is that the 1 values
of ꢀ1.82 and ꢀ1.53 are much smaller than those obtained
for the protonation or acylation of anilines for which the re-
ported 1 values are ꢀ2.89 and ꢀ3.21, respectively.[31] This
implies that although the nucleophilicity of amines is impor-
tant, it is apparently not the only factor contributing to their
reactivity in the current system.
[Rh
A
much faster than that catalyzed by iron(III) corroles and
N
porphyrins.[21] On the other hand, the synthetic results ob-
tained for the iron-catalyzed reactions between EDA and
amines seem inconsistent with any proposal that involves
4000
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3995 – 4005

Iron(III) Corroles and Porphyrins as Catalysts
A
FULL PAPER
Proposed mechanism for the formation of nitrogen ylides:
Analysis of the data accumulated so far is shown in
Scheme 6 in which all steps except KA and KB may safely be
assumed to be irreversible reactions. The initial step consists
of the coordination of EDA to the metal center to form
either A or B. Literature data may safely be used to suggest
that the equilibrium leading to A is faster than that leading
to B,[33] and that B is the precursor of metal–carbene inter-
mediate C. In most cases, and in practically all mechanistic
discussions in the literature, the formation of C en route to
the final product is taken for granted. The various mechanis-
tic proposals focus on the step that follows step b, that is,
whether the attack on C is concerted (three- or four-cen-
tered, synchronous or not, etc.) or stepwise via ylide inter-
mediates, such as D (and Y), with amines. However, the evi-
dence against the formation of C in the reaction of EDA
with amines is quite strong for the iron-based catalysts em-
ployed in the current investigations. There is no simple way
to explain why the reaction of amines liberates nitrogen
within seconds, whereas in the absence of amines there is
either no reaction (with iron porphyrin catalyst 3b) or a
very slow one with EDA, alcohols, or olefins (routes c–e).
Although one could argue that the metal coordination of
amines possibly accelerates the formation of C, this option
ꢀ
Figure 2. Hammett plots for the N H insertion of EDA into para- and
meta-substituted anilines, catalyzed by the iron corrole 2b (squares) and
the iron porphyrin 3a (triangles).
As an additional probe for evaluating the importance of
nucleophilicity, competition reactions between aniline/mor-
pholine and aniline/propylamine were carried out with 2b
and 3a as the catalysts (Scheme 8). Although the difference
is particularly unlikely for the iron(III) porphyrins (and was
E
ruled out for 3b, see above), which are actually coordina-
tively saturated in intermediate B owing to the chloride
ligand that is trans to the coordinated EDA. The absolute
selectivity for amine-derived products found in the iron-cat-
alyzed competition reactions of amines and olefins with
EDA,[21] a feature not shared by rhodium-based catalysts, is
a further indication that the two systems proceed through
substantially different reaction intermediates.
Scheme 8. Competition between aniline and morpholine (top) and pro-
pylamine (bottom).
Based on all of the above, we propose that intermediate
C is not formed in the reaction with amines, although this
pathway is reasonable for other substrates. We note that
even in reactions with somewhat nucleophilic alcohols
(route d of Scheme 6), the overall reaction was very slow
and route c was highly competitive (30% alcohol-derived
product and 70% from EDA coupling).[21] The amines, how-
ever, apparently adopt a lower-energy pathway that is only
available as a result of their greater nucleophilicity relative
to other substrates. One alternative that is consistent with
the very fast emission of nitrogen is the reaction of amines
with EDA activated by either N- or C-coordination to the
in purely nucleophilic reactions is several orders of magni-
tude in favor of the nonaromatic amines,[32] the results from
this work clearly show that there is no pronounced prefer-
ence for the formation of products resulting from the reac-
tions of morpholine or propylamine. In all cases the aniline-
related products were also formed and sometimes even in
preference. The approximately equal reactivity that aniline
ꢀ
and nonaromatic amines displayed in the N H insertion re-
action adds confidence to the conclusions obtained from the
Hammett plots, that is, that nucleophilicity is not the sole
factor that affects the reactivity of amines towards diazo
compounds under catalysis by iron corrole/porphyrins.
The main clues obtained from the current investigation
iron(III) corrole/porphyrin (Scheme 6; routes a and a’ to in-
A
termediates A and B, respectively). Transition-state struc-
tures proposed for the processes that lead to the ylide inter-
mediates obtained from either A or B are shown in
Scheme 9. They are formed by nucleophilic attack of the
amine on the N2-carrying carbon atom of EDA with a par-
tial hydride-like donation to the iron atom in either a four-
or six-centered fashion. This explains the main findings ob-
tained in the current investigation: the quite non-nucleophil-
ic anilines react very quickly because they are relatively
strong hydride donors, whereas the much more nucleophilic
amines do not react faster than anilines because they are
ꢀ
with regards to the mechanism of the N H insertion reac-
tions are the unlikelihood of a metal–carbene intermediate
and the limited importance of the nucleophilicity of the
amine. In fact it is clear that the results cannot be explained
by assuming that an intermediate is formed in an irreversi-
ble reaction (even if not rate limiting) between the metal
and the diazo compound prior to the interaction with the
amines. Accordingly, we have proposed alternative pathways
for the formation of the ylide intermediates from the iron-
catalyzed reaction of amines with EDA.
Chem. Eur. J. 2008, 14, 3995 – 4005
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4001

I. Aviv and Z. Gross
was mixed with a limited amount of EDA (30 mol% rela-
tive to the amines), only the product derived from N-meth-
ylaniline was obtained. The advantage of this last result is
that it rules out other possibilities regarding the effect of
aniline, such as changing the reactivity of the metal because
of coordination or reduction. All together, the competition
between the two substrates is indicative of the importance
of NH bonds for facilitating the formation of the ylide inter-
mediates.
Scheme 9. Possible transition states for the iron corrole/porphyrin cata-
lyzed formation of ylide Y from the reactions of amines with EDA coor-
dinated to the metal through its carbon (B) or nitrogen (A) atom (R’=
H, R’’=CO2Et, L=corrole3ꢀ or (chloro)porphyrin3ꢀ).
Synthetic use of ylides formed in the reactions of allyl- and
propargyl-substituted tertiary amines with EDA: Both iso-
lated and ylides formed in situ are of great synthetic value
in many cases.[2,35] One use relevant to the current case is
the in situ generation of ammonium ylides by metal-cata-
lyzed reactions of diazo compounds with allylic or propar-
gylic amines followed by 2,3-sigmatropic rearrangement to
weak hydride donors. The relatively low 1 value obtained
for ring-substituted anilines is also consistent with the pro-
posed transition states as the large negative value expected
for nucleophilic attack is reduced by the contribution of a
positive value expected for hydride transfer. In fact, this
proposal might even explain the superiority of porphyrin
ꢀ
produce a new C C bond (Scheme 11). Such reactions with
cobalt-, copper-, rhodium-, and ruthenium-based catalysts
have been previously reported,[2,36] but none with iron com-
plexes. The strong indications for the highly facile formation
of ammonium ylides in the reactions of NH-containing
relative to iron(III) corrole complexes. The reduction poten-
N
tial of the former is much less negative than that of the
latter,[34] and accordingly, partial hydride transfer is more
significant for the former.
amines with EDA under catalysis by iron(III) complexes of
N
porphyrins and corroles suggest that these complexes could
have significant added value when used with tertiary
amines.
The above hypothesis was indeed found to be valid
(Scheme 11 and Table 5). The reaction between 16a and
EDA (1:1) in the presence of 2b (0.5 mol%) proceeded
very well. Full and very fast conversion into the 2,3-sigma-
tropic rearrangement product 17a (entry 1, ꢁ500 catalytic
turnovers, 96% isolated yield) was obtained within one
minute during which time bubbles of nitrogen gas were
A final test for probing the proposed mechanism was per-
formed by comparing the reactivity of a secondary and a ter-
tiary amine with practically identical nucleophilicity
(Scheme 10). If partial hydride transfer was not important,
the rate of formation of the corresponding ylides should be
very similar. Separate experiments carried out with N-meth-
ylaniline and N-allyl-N-methylaniline showed that the reac-
tion of the former was complete within seconds and with ab-
ꢀ
solute selectivity for the N H insertion product, whereas
the reaction of the latter was much slower (about 30 min)
and provided a mixture of ylide-derived and EDA-coupled
products. Moreover, when a 1:1 solution of the two amines
Table 5. Metal-catalyzed reactions of EDA with tertiary amines that are
suitable for the formation of 2,3-sigmatropic rearrangement products.[a]
Entry
Catalyst
Substrate
Time
Yield [%]
Product
1
2
3
4
5
6
3a
3b
2b
16a
16a
16a[b]
16a
16b[b]
18
1 min
24 h
24 h
3 d
30 min
1 min
96
92
91
0
52
96
17a
17a
17a
–
17b
19
A
2(OAc)4]
3a
3a
[a] Catalyst/EDA/substrate 1:500:500 in CH2Cl2 (2 mL) at RT. EDA and
the substrate were added together in one portion. Yields are of isolated
products. [b] 1 mol% of catalyst.
Scheme 10. Reaction products obtained from ylides derived from secon-
dary and tertiary amines.
Scheme 11. Products obtained from ammonium ylides formed in situ.
4002
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3995 – 4005

Iron(III) Corroles and Porphyrins as Catalysts
A
FULL PAPER
Table 6. Metal-catalyzed reactions of EDA/EDP with thiols and sulfides
that are suitable for the formation of 2,3-sigmatropic rearrangement pro-
ducts.[a]
clearly observed. The same reaction with the more electron-
rich porphyrin and corrole iron complexes also led to 17a in
high yield and selectivity, but in a longer time (entries 2 and
3). On the other hand, the reaction completely failed when
Entry
Catalyst
Substrates
Time
Yield [%]
Product
1
2
3
4
5
6
7
8
3a[b]
3b[b]
2a[b]
2b[b]
3a[b]
3a[b]
3a[b]
3a[c]
2b[c]
3a[d]
3a[d]
2b[d]
20a/EDA
20a/EDA
20a/EDA
20a/EDA
21a/EDA
22a/EDA
23a/EDA
20a/EDP
20a/EDP
24a/EDA
26/EDA
<1 min
1 min
2 min
1 min
<1 min
<1 min
<1 min
2 min
2 min
1 min
1 min
20 min
24 h
98
97
95
95
96
92
93
89
90
95
94
93
0
20b
20b
20b
20b
21b
22b
23b
20c
20c
25a
27
[Rh
conditions (entry 4). This finding emphasizes the superiority
of the iron-based catalysts because success with [Rh2(OAc)4]
2(OAc)4] was used under otherwise identical reaction
AHCTREUNG
and other previously reported catalysts required the slow
addition of reagents, higher temperatures, or other limiting
reaction conditions.[36] Even the much less nucleophilic N-
allyl-N-methylaniline (16b) was converted into the expected
product when catalyzed by 3a, although the yield was lower
and EDA coupling products were also obtained in this case
(entry 5). The 3a-catalyzed reaction of EDA with N,N-dime-
thylpropargylamine (18, entry 6) was as efficient in terms of
reaction time as that of allylamine to give allenyl-substituted
a-amino ester 19 in an isolated yield of 96%. A comparison
with literature data for the same reaction with other cata-
lysts revealed that the iron porphyrin is superior in all as-
pects of synthetic efficiency. These results are also fully con-
sistent with the mechanistic aspects discussed earlier with
regards to the very fast formation of ylide intermediates
9
10
11
12
13
14
24a/EDA
24a/EDA
24a/EDP
25a
–
25b
[Fe
A
3a[c]
1 h
91
[a] All reactions were performed in CH2Cl2 (2 mL) at room temperature
and EDA/EDP and the substrate were added together in one portion.
Yields are of isolated products. [b] Catalyst/EDA/substrate 1:1000:1000.
[c] Catalyst/EDP/substrate
1:500:500.
1:100:100.
[d] Catalyst/EDP/substrate
when catalyzed by the iron(III) corroles and porphyrins.
G
ꢀ
S H insertion reactions and 2,3-sigmatropic rearrangements
in ylides formed by the reactions of sulfides with diazoace-
tates: The above results suggested that the same iron-based
catalysts could be highly useful for the formation of sulfur
ylides in situ, which are of highly significant synthetic
value.[37] The iron
(III) corrole and porphyrin complexes
U
were indeed excellent catalysts for the reactions of EDA/
EDP with both thiols and sulfides. The results summarized
in Table 6 (0.1 mol% catalyst) show that both thiophenols
(entries 1–5) and nonaromatic thiols (entries 6 and 7) are
converted very quickly and in very high yields into the ex-
Scheme 12. Reactions of diazoacetates with thiols and sulfides. The pic-
ture (demonstrating the intense nitrogen release) was taken 5 s after the
addition of thiophenol and EDA to a solution of catalyst 3a.
ꢀ
pected S H insertion products, a-thio ethyl ester derivatives
(Scheme 12).[8,38] The reactions were very fast (see picture in
Scheme 12), and even the less reactive EDP reacted quickly
with thiophenol (entries 8 and 9) to provide 20c.
Although a full mechanistic study was not performed for
this reaction, it apparently proceeds in a similar manner to
the reaction of amines. The reactivity of thiols and amines is
also not very different because the reaction of EDA with
equimolar amounts of thiophenol and aniline provided a
mixture that contained products derived from both reagents.
More important, the iron-catalyzed reactions of EDA with
allyl- and propargyl-substituted sulfides were also very
facile, leading to the expected products in very high yields
within very short reaction times (entries 10–12). In a similar
Conclusion
A detailed mechanistic investigation of the outstandingly
high activity of corrole and porphyrin iron(III) complexes in
G
the catalysis of NH-containing amines to amino acid esters
by their reaction with diazoacetates has led to the following
conclusions: Nitrogen ylides are key reaction intermediates,
they are not produced by a reaction pathway that involves
metal–carbenes, and both the nucleophilicity and hydride-
transfer capability of primary and secondary amines are im-
portant aspects involved in the transition states leading to
the ylides. The very fast formation of N-ylides was then ap-
plied to the reactions of tertiary amines that contain either
allyl or propargyl substituents; 2,3-sigmatropic rearrange-
ment products were formed more efficiently when catalyzed
manner to the observations with amines, the iron(III) salen
A
complex was not able to catalyze the reaction (entry 13).
The other end of the reactivity scale is allyl methyl sulfide
(24a; entry 14), which did undergo the corresponding reac-
tion (to give 25b) even with the much less reactive diazo
compound EDP, a reaction that did not succeed with the
analogous amine 16a.
by the iron(III) complexes than with previously reported
catalysts. Similar findings regarding the superiority of the
iron complexes were obtained with thiols and sulfides, which
G
Chem. Eur. J. 2008, 14, 3995 – 4005
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4003

I. Aviv and Z. Gross
Procedure for the quantitative reaction of thiols with EDA (or EDP): A
solution of EDA and the substrate in CH2Cl2 (0.5 mL, 1000-fold excess,
1–1.5 mmol, of each relative to the catalyst) was added in a single portion
to a solution of the catalyst (1 mg, 1–1.5 mmol, in 1.5 mL CH2Cl2). Re-
ported yields are for isolated products; their purities were checked by
GC analysis (>99% for all the products) and identities were confirmed
by NMR spectroscopy analysis and compared with previous reports of
suggests the operation of similar mechanisms in these cases.
Work on using the knowledge acquired in this study to solve
the challenge of turning these reactions into enantioselective
processes is under investigation.
the same compounds. The colorless oil obtained from the reaction of
1
thio
A
2H), 7.26 (m, 3H), 4.15 (q, J=7.2 Hz, 2H), 3.61 (s, 2H), 1.20 ppm (t, J=
7.2 Hz, 3H).
Experimental Section
Chemicals: Standard reagents and solvents (including [Rh2(OAc)4] and
U
the iron porphyrins) were used as received from commercial sources
without any further purification. EDP and N-allyl-N-methylaniline were
synthesized according to published procedures.[39] The iron and rhodium
corroles were prepared as described in previous publications.[40]
Acknowledgement
Quantitative reactions of amines with EDA: A solution of EDA and the
substrate (1–1.5 mmol in 0.5 mL diethyl ether) was added in a single por-
tion to a solution of the catalyst (1 mg, 1–1.5 mmol in 1.5 mL diethyl
ether). The 4-Cl, 4-CN, and 3-CN-substituted aniline-derived products
precipitated from the reaction mixture, whereas the products of the other
amines were obtained by evaporating the solvent and by treatment of the
solid material with cold diethyl ether. Reported yields are for isolated
products (0.2–0.3 g); their purities were checked by GC analysis (>99%
for all the products) and identities were confirmed by NMR spectroscopy
analysis (see the Supporting Information) and comparison with previous
reports of the same compounds.
We thank the German–Israel Cooperation (DIP) for financial support of
this research.
[1] a) A. C. B. Burtoloso, C. R. D. Correia, J. Organomet. Chem. 2005,
690, 5636; b) H. Huang, Y. Wang, Z. Chen, W. Hu, Adv. Synth.
Catal. 2005, 347, 531; c) F. A. Davis, Y. Wu, H. Xu, J. Zhang, Org.
Lett. 2004, 6, 4523; d) M. R. Fructos, T. R. Belderrain, M. C. Nicasio,
S. P. Nolan, H. Kaur, M. M. Diaz-Requejo, P. J. PØrez, J. Am. Chem.
Soc. 2004, 126, 10846; e) G. S. Singh, Curr. Org. Synthesis 2005, 2,
377.
[2] M. P. Doyle, M. A. McKervey, T. Ye, Modern Catalytic Methods for
Organic Synthesis with Diazo Compounds, Wiley, New York, 1998.
[3] a) K. E. Bashford, A. L. Cooper, P. D. Kane, C. J. Moody, S. Muthus-
amy, E. Swann, J. Chem. Soc., Perkin Trans. 1 2002, 1672; b) K. Ya-
mazaki, Y. Kondo, Chem. Commun. 2002, 210; c) E. C. Talor,
H. M. L. Davies, J. Org. Chem. 1984, 49, 113; d) for Merck synthesis
of the antibiotic theinamycin, see: T. M. Salzmann, R. W. Ratcliffe,
B. G. Christensen, F. A. Bouffard, J. Am. Chem. Soc. 1980, 102,
6161; e) A. C. B. Burtoloso, C. R. D. Correia, Tetrahedron Lett. 2004,
45, 3355; f) J. Wang, Y. Hou, P. Wu, J. Chem. Soc., Perkin Trans. 1
1999, 2277; g) J. R. P. Davies, D. Kane, C. J. Moody, Tetrahedron
2004, 60, 3967; h) R. T. Buck, P. A. Clarke, D. M. Coe, M. J. Dry-
sdale, L. Ferris, D. Haigh, C. J. Moody, N. D. Pearson, E. Swann,
Chem. Eur. J. 2000, 6, 2160; i) S. Bertelsen, M. Nielsen, S. Bach-
mann, K. A. Jørgensen, Synthesis 2005, 13, 2234.
Reaction between EDA and aniline, catalyzed by Mb: Mb (5 mg,
0.2 mmol), EDA (16.4 mL, 0.156 mmol), and aniline (14.2 mL, 0.156 mmol)
were stirred in a solution of THF (10%)/aqueous phosphate buffer
(pH 7, 1.5 mL) at room temperature until all of the EDA was consumed
(5 min). Extraction with diethyl ether provided pure 4a as a white solid
(25 mg, 90% isolated yield). 1H NMR (300 MHz, CDCl3): d=7.17 (t,
2H), 6.72 (t, 1H), 6.57 (d, 2H), 4.20 (q, J=7.2 Hz, 2H), 3.86 (s, 2H),
1.26 ppm (t, J=7.2 Hz, 3H).
Reaction between EDP and aniline, catalyzed by Mb: Mb (5 mg,
0.2 mmol), EDP (20 mg, 0.156 mmol), and aniline (14.2 mL, 0.156 mmol)
were stirred in a solution of THF (10%)/aqueous phosphate buffer
(pH 7, 1.5 mL) and placed in a heating bath at 658C until all of the EDP
was consumed (15 min). Extraction with diethyl ether provided pure 13a
as a colorless oil (25.6 mg, 87% isolated yield). 1H NMR (300 MHz,
CDCl3): d=7.16 (t, 2H), 6.73 (t, 1H), 6.59 (d, 2H), 4.17 (q, J=7.2 Hz,
2H), 4.09 (q, J=6.9 Hz, 1H), 1.45 (d, J=6.9 Hz, 3H), 1.23 ppm (t, J=
7.2 Hz, 3H).
[4] P. Yates, J. Am. Chem. Soc. 1952, 74, 5376.
[5] a) T. Saegusa, Y. Ito, S. Kobayashi, K. Hirota, T. Shimizu, Tetrahe-
dron Lett. 1966, 7, 6131; b) J.-F. Nicoud, H. B. Kagan, Tetrahedron
Lett. 1971, 12, 2065.
[6] R. Paulissen, E. Hayez, A. J. Hubert, P. TeyssiØ, Tetrahedron Lett.
1974, 15, 607.
[7] a) M. E. Morilla, M. M. Díaz-Requejo, T. R. Belderrain, M. C. Nica-
sio, S. Trofimenko, P. J. PØrez, Chem. Commun. 2002, 2998; b) S.
Bachmann, D. Fielenbach, K. A. Jørgensen, Org. Biomol. Chem.
2004, 2, 3044.
[8] a) E. Galardon, P. Le Maux, G. Simonneaux, J. Chem. Soc., Perkin
Trans. 1 1997, 2455; b) E. Galardon, P. Le Maux, G. Simonneaux,
Tetrahedron 2000, 56, 615.
[9] a) M. R. Fructos, T. R. Belderrain, P. de FrØmont, N. M. Scott, S. P.
Nolan, M. M. Díaz-Requejo, P. J. PØrez, Angew. Chem. 2005, 117,
5418; Angew. Chem. Int. Ed. 2005, 44, 5284; b) P. de FrØmont, E. D.
Stevens, M. R. Fructos, M. M. Díaz-Requejo, P. J. PØrez, S. P. Nolan,
Chem. Commun. 2006, 2045; c) P. Rodríguez, A. Caballero, M. M.
Díaz-Requejo, M. C. Nicasio, P. J. PØrez, Org. Lett. 2006, 8, 557.
[10] F. A. Davis, B. Yang, J. Deng, J. Org. Chem. 2003, 68, 5147.
[11] R. T. Buck, C. J. Moody, A. G. Pepper, ARKIVOC 2002, 16.
[12] C. F. Garcia, M. A. McKervey, T. Ye, Chem. Commun. 1996, 1465.
[13] B. Liu, S.-F. Zhu, W. Zhang, C. Chen, Q.-L. Zhou, J. Am. Chem.
Soc. 2007, 129, 5834.
Reaction between EDP and aniline, catalyzed by 14/BSA: EDP (3 mg,
23.4 mmol) and aniline (2.1 mL, 23.4 mmol) were added to a small vessel.
A solution of aqueous phosphate buffer (pH 7, 1 mL) containing iron
corrole 14 (0.2 mg) and BSA (20 mg) was then added. The reaction mix-
ture was stirred at room temperature until complete consumption of
EDP was evident by TLC. The solution was extracted with diethyl ether
to provide 13a (3.84 mg, 85% yield). The extracts were concentrated
under a stream of argon and the residue was diluted with HPLC-grade
solvents (hexane/isopropanol 9:1) and analyzed by HPLC for possible en-
antiomeric enrichment. Similar results were obtained when this reaction
was carried out with HSA, PSA, SSA, and RSA as the albumin source
instead of BSA. No enantiomeric excess was observed in any of the reac-
tions.
Procedure for reactions of tertiary amines/sulfides with EDA: A solution
of EDA and the substrate in CH2Cl2 (0.5 mL, 500-fold excess, 0.5–
0.7 mmol, of each relative to the catalyst) was added in a single portion
to a solution of the catalyst (1 mg, 1–1.5 mmol in 1.5 mL CH2Cl2). Report-
ed yields are for isolated products; their purities were checked by GC
analysis (>99% for all the products) and identities were confirmed by
NMR spectroscopy analysis and compared with previous reports of the
same compounds. The colorless oil obtained from the reaction of N,N-di-
methylallylamine with EDA was 17a. 1H NMR (300 MHz, CDCl3): d=
5.72 (m, 1H), 5.04 (m, 2H), 4.14 (q, J=7.2 Hz, 2H), 3.14 (dd, J=8.3,
6.6 Hz, 1H), 2.41 (m, 2H), 2.35 (s, 6H), 1.24 ppm (t, J=7.2 Hz, 3H).
[14] a) M. P. Doyle, L. J. Westrum, W. N. E. Wolthuis, M. M. See, W. P.
Boone, V. Bagheri, M. M. Pearson, J. Am. Chem. Soc. 1993, 115,
4004
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3995 – 4005

Iron(III) Corroles and Porphyrins as Catalysts
A
FULL PAPER
958; b) D. F. Taber, K. K. You, A. L. Rheingold, J. Am. Chem. Soc.
1996, 118, 547; c) E. Nakamura, N. Yoshikai, M. Yamanaka, J. Am.
Chem. Soc. 2002, 124, 7181.
York, 1972; d) For lower 1 values reported for iron porphyrin-cata-
lyzed reactions of anilines with EDA under very high catalytic turn-
over conditions (EDA/anilines=1:1), see L. K. Baumann, H. M.
Mbuvi, G. Du, L. K. Woo, Organometallics 2007, 26, 3995..
[32] a) C. D. Ritchie, M. Sawada, J. Am. Chem. Soc. 1977, 99, 3754;
b) C. D. Ritchie, A. Kawasaki, J. Org. Chem. 1981, 46, 4704.
[33] R. Cohen, B. Rybtchinski, M. Gandelman, H. Rozenberg, J. M. L.
Martin, D. Milstein, J. Am. Chem. Soc. 2003, 125, 6532, and referen-
ces therein.
[34] L. Simkhovich, A. Mahammed, I. Goldberg, Z. Gross, Chem. Eur. J.
2001, 7, 1041.
[35] a) A. Padwa, S. F. Hornbuckle, Chem. Rev. 1991, 91, 263; b) A.-H.
Li, L. X. Dai, V. K. Aggarwal, Chem. Rev. 1997, 97, 2341; c) D. M.
Hodgson, F. Y. T. M. Pierard, P. A. Stupple, Chem. Soc. Rev. 2001,
30, 50.
[36] a) M. P. Doyle, W. H. Tamblyn, V. Bagheri, J. Org. Chem. 1981, 46,
5094; b) F. G. West, K. W. Glaeske, B. N. Naidu, Synthesis 1993, 977;
c) C.-Y. Zhou, W.-Y. Yu, P. Wai Hong Chan, C.-N. Che, J. Org.
Chem. 2004, 69, 7072; d) G. Simonneaux, E. Galardon, C. Paul-
Roth, M. Gulea, S. Masson, J. Organomet. Chem. 2001, 617–618,
360; e) T. Fukudat, R. Irie, T. Katsuki, Tetrahedron 1999, 55, 649;
f) P. Heath, E. Roberts, J. B. Sweeney, H. P. Wessel, J. A. Workman,
J. Org. Chem. 2003, 68, 4083.
[37] a) M. Liao, J. Wang, Green Chem. 2007, 9, 184; b) V. K. Aggarwal,
Asymmetric Synthesis, 2007, 52–56; c) V. K. Aggarwal, J. Richard-
son, Chem. Commun. 2003, 2644; d) H. M. L. Davies in Nitrogen,
Oxygen and Sulfur Ylide Chemistry (Ed.: J. S. Clarke), OUP,
Oxford, 2002, pp. 156–163.
[15] A. A. C. Braga, F. Maseras, J. Urbano, A. Caballero, M. M. Díaz-
Requejo, P. J. PØrez, Organometallics 2006, 25, 5292.
[16] a) H. E. Helson, W. L. Jorgensen, J. Org. Chem. 1994, 59, 3841;
b) D. J. Miller, C. J. Moody, Tetrahedron 1995, 51, 10811; c) M. P.
Doyle, M. Yan, Tetrahedron Lett. 2002, 43, 5929.
[17] Z. Qu, W. Shi, J. Wang, J. Org. Chem. 2004, 69, 217.
[18] a) C.-D. Lu, H. Liu, Z.-Y. Chen, W.-H. Hu, A.-Q. Mi, Org. Lett.
2005, 7, 83; b) H. Huang, X. Guo, W. Hu, Angew. Chem. 2007, 119,
1359; Angew. Chem. Int. Ed. 2007, 46, 1337.
[19] Y. Wang, Y. Zhu, Z. Chen, A. Mi, W. Hu, M. P. Doyle, Org. Lett.
2003, 5, 3923.
[20] a) I. Aviv, Z. Gross, Chem. Commun. 2007, 1987; b) S. Nardis, D.
Monti, R. Paolesse, Mini-Rev. Org. Chem. 2005, 2, 546; c) D. T.
Gryko, J. P. Fox, D. P. Goldberg, J. Porphyrins Phthalocyanines 2004,
8, 1091; d) J.-M. Barbe, G. Canard, S. Brandes, R. Guilard, Chem.
Eur. J. 2007, 13, 2118.
[21] I. Aviv, Z. Gross, Synlett 2006, 6, 951.
[22] I. Aviv, Z. Gross, Chem. Commun. 2006, 4477.
[23] a) Z. Gross, L. Simkhovich, N. Galili, Chem. Commun. 1999, 599;
b) J. R. Wolf, C. G. Hamaker, J.-P. Djukic, T. Kodadek, L. K. Woo, J.
Am. Chem. Soc. 1995, 117, 9194.
[24] a) I. Saltsman, L. Simkhovich, Y. S. Balazs, I. Goldberg, Z. Gross,
Inorg. Chim. Acta 2004, 357, 3038; b) A. Mahammed, I. Giladi, I.
Goldberg, Z. Gross, Chem. Eur. J. 2001, 7, 4259.
[25] G. N. La Mar, F. A. Walker in The Porphyrins, Vol. IV (Ed.: D. Dol-
phin), Academic Press, New York, 1978, Chapter 2.
[26] A. Mahammed, Z. Gross, J. Am. Chem. Soc. 2005, 127, 2883.
[27] M.-Z. Wang, H.-W. Xu, Y. Liu, M.-K. Wong, C.-M. Che, Adv. Synth.
Catal. 2006, 348, 2391.
[28] V. K. Aggarwal, J. R. Fulton, C. G. Sheldon, J. de Vicente, J. Am.
Chem. Soc. 2003, 125, 6034.
[29] C. E. Castro, M. Jamin, W. Yokoyama, R. Wade, J. Am. Chem. Soc.
1986, 108, 4179.
[30] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165.
[31] a) A. I. Biggs, R. A. Robinson, J. Chem. Soc. 1961, 388; b) P. R.
Wells, Chem. Rev. 1963, 63, 171; c) N. B. Chapman, J. Shorter, Ad-
vances in Linear Free Energy Relationships, Plenum Press, New
[38] E. Galardon, S. RouØ, P. Le Maux, G. Simonneaux, Tetrahedron
Lett. 1998, 39, 2333.
[39] L. Benati, D. Nanni, P. Spagnolo, J. Org. Chem. 1999, 64, 5132; M.
Alvaro, C. Aprile, A. Corma, V. Fornes, H. Garcia, E. Peris, Tetrahe-
dron 2004, 60, 8257.
[40] a) L. Simkhovich, I. Goldberg, Z. Gross, Inorg. Chem. 2002, 41,
5433; b) A. Mahammed, I. Goldberg, Z. Gross, Org. Lett. 2001, 3,
3443; c) I. Saltsman, L. Simkhovich, Y. S. Balazs, I. Goldberg, Z.
Gross, Inorg. Chim. Acta 2004, 357, 3038.
Received: November 29, 2007
Published online: March 20, 2008
Chem. Eur. J. 2008, 14, 3995 – 4005
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4005