Cyclometalated gold(III) iminophosphoranes which incorporate carbohydrate groups

Article abstract of DOI:10.1016/j.jorganchem.2011.11.009

Iminophosphoranes with organic groups derived from d-glucose, d-galactose and l-arabinose have been used to prepare gold(III) dichloride complexes via mercurated intermediates, since direct cyclometallation was unsuccessful. Structures and full spectroscopic data are reported. Replacement of one or more of the chloride ligands by PPh₃, or by thiosalicylate gave new derivatives. Biological screening showed no enhanced activity relative to other alkyl or aryl analogues.

Full text of DOI:10.1016/j.jorganchem.2011.11.009

Journal of Organometallic Chemistry 702 (2012) 1e9
Contents lists available at SciVerse ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Cyclometalated gold(III) iminophosphoranes which incorporate carbohydrate
groups
*
Bevan P. Jarman, Brian K. Nicholson
Department of Chemistry, University of Waikato, Private Bag 3105, Hamilton, New Zealand
a r t i c l e i n f o
a b s t r a c t
Article history:
Iminophosphoranes with organic groups derived from D-glucose, D-galactose and L-arabinose have been
Received 12 September 2011
Received in revised form
31 October 2011
used to prepare gold(III) dichloride complexes via mercurated intermediates, since direct cyclo-
metallation was unsuccessful. Structures and full spectroscopic data are reported. Replacement of one or
more of the chloride ligands by PPh3, or by thiosalicylate gave new derivatives. Biological screening
showed no enhanced activity relative to other alkyl or aryl analogues.
Accepted 3 November 2011
Ó 2011 Elsevier B.V. All rights reserved.
Keywords:
Gold complexes
Cyclometalated ligands
Iminophosphoranes
Sugar ligands
OAc
1. Introduction
1a
Iminophosphoranes, R3P]NeR0, are readily available with
a wide variety of R and R0 groups [1,2]. Alper first showed that it was
possible to produce cyclopalladated species by reacting Ph3P]
NC6H4CH3 with Na2[PdCl4] [3]. Direct cyclometallation has also
been carried out with Pd(OAc)2 or Hg(OAc)2 [4] and with
PhCH2M(CO)5 (M ¼ Mn or Re) [5]. An indirect route where the
iminophosphorane is initially lithiated and then transmetalated has
been used to prepare gold [6,7], rhodium, iridium [8], aluminium
and gallium [9] cyclometalated iminophosphoranes. We have been
particularly interested in cycloaurated iminophosphoranes, since
the ligand is effective at stabilising the gold(III) oxidation state.
Sugar iminophosphoranes have long been known. Acetylated
glycosyl azides react with PPh3 in Et2O, to give species such as 1a
[10]. These sugar iminophosphoranes are useful in organic
synthesis; for example they can be used to reduce azides, to
produce amides or to synthesis imines via the AzaeWittig reaction
[11]. However their use as ligands has not been investigated. Since
other gold(III) iminophosphoranes have been shown to be cata-
lytically and biologically active [12], the introduction of a carbohy-
drate moiety is interesting, since carbohydrates not only introduce
chirality but can enhance biological activity as well. This paper
reports the synthesis and characterisation of some novel cycloau-
rated iminophosphoranes.
O
AcO
AcO
Ph
N
P
OAc
Ph
Ph
2. Results and discussion
2.1. Complexes of glycosyl iminophosphoranes
The Staudinger reaction [13] between an azide and a phosphine
seemed the best route to iminophosphoranes so sugar azides were
needed. Starting from
D-glucose, D-galactose and L-arabinose,
literature methods [14e16] readily produced acetylated glycosyl
azides 2aec, by acetylation of the sugars with acetic anhydride and
catalytic amounts of iodine, bromination with HBr in acetic acid,
and nucleophilic substitution with NaN3, Scheme 1.
The reaction can be extended to many other sugars such as
disaccharides but is only applicable to 1,2-trans sugars, so mannose
with the O2 oxygen in the axial position would not produce the
azide product. Azides are notorious for their explosive nature, but
as the percentage of nitrogen in the azide decreases they become
less explosive. The Smith rule [11] is an indication of whether an
* Corresponding author. Fax: þ64 7 838 4219.
E-mail address: b.nicholson@waikato.ac.nz (B.K. Nicholson).
0022-328X/$ e see front matter Ó 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2011.11.009

2
B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
O
O
(i)
AcO
HO
OAc
OH
OAc
OH
(ii)
O
O
(iii)
AcO
OAc
AcO
N3
OAc
OAc
Br
OAc
OAc
OAc
O
O
O
AcO
AcO
N3
N3
N3
AcO
AcO
OAc
OAc
OAc
2b
2c
2a
Scheme 1. General reaction scheme for the production of glycosyl azides. (i) Acetic anhydride, cat. I2; (ii) HBr, acetic acid; (iii) NaN3, acetone, H2O.
organic azide will be explosive. Since the nitrogen content of all the
sugar azides was low no special precautions were taken with the
manipulation of them.
[30]. The phosphine was reacted with an azide to produce an
intermediate mercurated iminophosphorane, which could be
transmetalated with Au(III). The overall reaction is shown in
Scheme 3.
Acetyl glucosyl azide was reacted with PPh3 with a slight
modification of the previous method [10] to produce 1a, which
was used without purification. The iminophosphorane Ph3P]
NePh, 3, has been directly cyclometalated [5e7,17], so 1a was
tested to see if direct metallation was possible. Pd(OAc)2, Hg(OAc)2
and PhCH2Mn(CO)5 were all tried but no cyclometalated product
could be isolated or detected by high resolution electrospray ion-
isation mass-spectrometry (ESI-MS) of an aliquot of the reaction
mixture. To understand why, a structure determination of 1a was
carried out, shown in Fig. 1.
The azide sugar 2a was reacted with 4. Monitoring of the reac-
tion by 31P NMR and ESI-MS showed that the intermediate
mercurated iminophosphorane, 5a, had fully formed after 1.5 h.
Removal of the solvent gave 5a which was used directly, after
characterisation by 31P NMR and ESI-MS. The ion [AuCl4]ꢁ has been
used to produce gold(III) iminophosphoranes [6,17,30]. To 5a was
added [NMe4][AuCl4] and dry acetonitrile and the mixture stirred
for several days in the dark. Monitoring by ESI-MS showed the slow
disappearance of the mercurated sugar iminophosphorane and the
appearance of the cyclometalated gold(III) compound. After sepa-
ration of HgCl2 by filtration, the solvent was removed under
vacuum and recrystallisation of the residue gave air- and moisture-
stable yellow crystals of 6a.
Compared with other iminophosphoranes with alkyl groups on
the nitrogen the CeN (1.409(1) Å) and N]P (1.5744(10) Å) bond
lengths are similar as is the CeN]P bond angle of 121.16(7)ꢀ
[18,19]. The CeN]P bond angle for the aryl-substituted 3 is
however larger at 130.4(3)ꢀ [20]. While the phosphorus is
approximately tetrahedral there is a slight distortion in the N]PeC
angles, with the carbon atoms cis to the nitrogen substituent having
slightly larger angles. This is also seen in the structure of 3 [20] and
for some other alkyl iminophosphoranes [18,19]. The torsion angle
for C2eC1eN1eP1 is 176.8(1)ꢀ. The sugar portion of the molecule is
as expected. A space filling model indicated that the nitrogen was
sterically crowded in the solid state. Scheme 2 shows the proposed
mechanism for the direct metallation by Pd(OAc)2 [21e23] which
starts with the coordination to the metal by the nitrogen.
Presumably for 1a the lack of availability of the nitrogen hinders the
direct metallation. In support of this when 1a was stirred with
PdCl2 in acetonitrile, under conditions where Ph3P]NePh forms
the N-donor complex, Pd[N(Ph)]PPh3]2Cl2 [24], no product could
be isolated or detected by ESI-MS. While the steric explanation for
the lack of reactivity seems plausible, it should perhaps be noted
that the solution NMR spectrum of 1a does not show inequivalence
of the Ph groups which would arise from hindered rotation about
the CeN bond on the NMR time scale, so in solution conformations
with the N exposed could be generated by twisting (c.f. Fig. 6 dis-
cussed below).
OAc
Ph
Ph
O
P
AcO
AcO
N
OAc
Au
Cl
Cl
6a
At room temperature after one month some decomposition was
seen, with the development of the purple colour from colloidal
elemental gold. However at ꢁ15 ꢀC, crystals were stable for over
a year. The other glycosyl azides 2b and 2c reacted similarly to give
the analogues 6b and 6c.
2.1.1. ESI-MS
ESI-MS of the compounds 6aec showed several characteristic
ions. The main ion was from the loss of a chloride to give [M ꢁ Cl]þ
with a less intense ion from [M þ Na]þ. This is a different mode of
ionisation from that seen for the mercurated intermediate 5, which
forms the ion [M þ H]þ. This difference is because of the availability
of the nitrogen to be protonated on 5, since mercury only weakly
interacts with the nitrogen while gold forms a strong bond.
Since another route was needed, pre-metallation of the phos-
phine was investigated. Transmetallation from mercury to gold(I)
and gold(III) has been used in a number of cases where direct
auration is not feasible [25e29], and the mercurated diphosphine,
4, was used by Kilpin et al. to produce a gold(III) iminophosphorane

B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
3
Fig. 1. The geometry of 1a showing the atom labelling scheme and thermal ellipsoids at the 50% probability level. Phenyl hydrogen atoms have been omitted for clarity. Selected
bond lengths (Å): C1eN1 1.4093(12), N1]P1 1.5744(9); bond angles (ꢀ): C1eN1]P1 121.16(7), N1]P1eC11 108.98(5), N1]P1eC21 113.33(5), N1]P1eC31 115.26(5); torsion angle
(ꢀ): C2eC1eN1eP1 176.8(1).
2.1.2. 31P, 1H and 13C NMR
This smaller angle leads to small coupling constants for H3, H4
and H5. This is also seen in the 1He1H COSY with no cross peak
seen between H4 and H5. The spectrum for 6c was harder to
interpret since H1 and H2 were very broad signals. 6c also has an
equatorial proton at C4 and small coupling constants were also
seen.
Fig. 2 shows the clear changes in the 31P chemical shift of the
series of iminophosphoranes using the sugar azide 2a. With the
free phosphine 4 the shift is 0.38 ppm. Upon reaction with the azide
to form 5a there is a clear shift downfield to 21.2 ppm. Once the
compound has been cyclometalated and the phosphorus incorpo-
rated into a cyclic system, 6a, there is another large downfield shift
to 60.3 ppm, a diagnostic indication that the phosphorus is in a 5-
membered cyclic system [31,32]. For 4 and 5a 199Hg satellite peaks
could be seen from 3JHg,P coupling.
In the 13C NMR spectrum the sugar signals were as expected; the
2
anomeric carbon (C1) showed splitting from JC,P coupling of
w5 Hz. The aromatic region was complex with many slightly
inequivalent carbon atoms, some with coupling to the phosphorus
atom, and this portion was not assigned.
1H NMR also has some diagnostic signals from the anomeric
proton, H1. Fig. 3 shows a series of 1H NMR spectra showing the
change in the anomeric proton signal for galactose (b). In the
gold(III) chloride iminophosphorane the normal doublet signal
becomes a doublet of doublets at 6.21 ppm. This is due to the three
2.1.3. Crystal structures of the gold(III) complexes 6a and 6c
Recrystallisation of 6a and 6c by slow diffusion of hexane into
solutions in dichloromethane or chloroform, respectively, produced
X-ray quality crystals. Monoclinic crystals of 6c had two indepen-
dent molecules in the asymmetric unit, together with two chloro-
form molecules (one disordered) and a hexane molecule. Figs. 4 and
5 show the geometry of 6a and 6c respectively. The structure
determination for 6c was less precise than usual because of crystal
quality/lattice solvent loss so the following discussion is based on
the determination for 6a, but is equally applicable to the close
analogue 6c which has the same overall conformation.
3
bond coupling to the phosphorus atom and gives a JH1,P coupling
constant of 18.1 Hz. The
b anomer has been retained.
For the rest of the 1H NMR spectrum the typical sugar signals
are seen and could be readily assigned. The difference in the
position of the hydroxyl group at C4 for glucose (a) and galactose
(b), was obvious in the 1H NMR spectrum. H4 in glucose is in
the axial position and has a vicinal angle of 180ꢀ with H3 and
H5, while for galactose it is equatorial and the vicinal angle is 60ꢀ.
Ph
N
Ph
N
Ph
Ph
Ph
Ph
P
P
P
-HOAc
N
MeC6H4
MeC6H4
MeC6H4
Pd
Pd(OAc)2
Pd
H
X
O
AcO
X = OAc
O
2
Scheme 2. Proposed mechanism for cyclometallation of MeC6H4N]PPh3 by palladium acetate [21e23].

4
B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
Ph
Ph
Ph
Ph
P
P
Ph
N
Ph
R
P
2R N3
2N2
+ 2[AuCl ]-
4
Hg
Hg
2
N
R
R
N
Au
Cl
Cl
P
P
Ph
Ph
Ph
Ph
6
R = Acetylated sugar
4
5
Scheme 3. General reaction scheme for the cycloaurated sugar iminophosphoranes from 4 (R ¼ Sugar).
In both compounds the gold(III) is square planar. If a plane is
The strength of the nitrogen bond to the gold can be seen by the
lengthening of the N1eP1 and C1eN1 bond as electron density is
removed from the nitrogenephosphorus bond by the gold centre.
For example if 1a and 6a are compared N1eP1 has increased to
1.612(6) Å from 1.5744(10) Å, while C1eN1 has increased to
1.447(9) Å from 1.4093(12) Å. The CeN]P bond angle for 6a is
similar to 1a showing little change upon cyclometallation.
The CeAueN bite angle is 85.6(3)ꢀ for 6a. This is comparable
to known cyclometalated iminophosphoranes such as [2-
(CO)4MnC6H4]Ph2P]NePh (83.5(2)ꢀ [5]) and [2-Cl2AuC6H4]Ph2P]
NePh (84.9(2)ꢀ [6]).
calculated through Au1, Cl1, Cl2, N1 and C20, for 6a the largest
deviation is 0.065(3) Å for C20. The trans influence can be seen in
the lengthening of the AueCl bond trans to the carbon compared
with that opposite the N (2.348(2) Å compared with 2.272(2) Å) for
6a. The bonding parameters involving Au are similar to those re-
ported for (2-Cl2AuC6H4)Ph2P]NePh, as are the CeN and N]P
bond lengths [6].
Both metallocyclic rings show a puckered envelope conforma-
tion, with the nitrogen atom out of the plane. The plane defined by
AueCeCeP is essentially planar with no atom deviating by more
than 0.008(4) Å. The nitrogen is displaced from this plane by
0.411(7) Å. While there are examples of iminophosphoranes having
a near-planar metallocyclic ring, such as [2-(CO)4MnC6H4]Ph2P]
NePh [5], all reported Au(III) compounds give envelope confor-
mations with the nitrogen atom out of plane [6,17].
Each of the sugar portions of the molecules show the 4C1
conformation. In comparison to the
D-glucose structure the L-
arabinose compound lacks the hydroxymethyl group on C5 and O4
is axial instead of equatorial (As a note, even though 6c has the
same configuration as the glucose version, it is assigned the
The torsion angle of C2eC1eN1eP1 is 93.1(6)ꢀ for 6a. For 1a it is
176.8(1)ꢀ showing that the triphenylphosphine group has been
twisted substantially which points the lone pair of electrons on the
nitrogen away from the acetate group on C2 of the sugar. It is
unclear why this rotation could not happen in 1a to allow direct
metallation to occur if steric hindrance of the nitrogen prevents it
from happening. Fig. 6 shows this twisting with a comparison
between 1a and 6a. This puts one of the phenyl rings above the
carbohydrate ring. Despite having an axial acetate group at C4, 6c
has the phenyl ring in a similar position.
a
configuration since it is an
D
L-sugar rather the b configuration for
the -sugar of 6a).
2.2. Reactions at the Au(III) centre
For C,N-cyclometalated gold(III) dichlorides,
different ligands can displace the chlorides on the gold(III) atom.
Two examples were examined in the present study.
a
variety of
Fig. 2. 31P NMR spectra comparing the changes of chemical shift for; A) mercurated
diphosphine 4; B) mercurated iminophosphorane intermediate 5a; C) C,N-cycloau-
rated iminophosphorane 6a.
Fig. 3. 1H NMR spectra showing the change in the anomeric proton signal by
comparing; A) ReOAc; B) ReBr; C) 2b; D) C,N-cycloaurated iminophosphorane 6b.
R ¼ Per-O-acetyl galacosyl. The anomer proton is marked with an asterisk.

B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
5
Fig. 4. The geometry of 6a showing the atom labelling scheme and thermal ellipsoids at the 50% probability level. Selected bond lengths (Å): Au1eCl1 2.348(2), Au1eCl2 2.272(2),
AueC 2.073(7), AueN1 2.040(5), C1eN1 1.447(8), N1eP1 1.612(6); bond angles (ꢀ): CeAueN 85.6(3), CeN]P 122.2(4); torsion angles (ꢀ): P1]N1eC1eC2 93.1(6).
2.2.1. Reaction with PPh3
There are large number of C,N-cyclometalated gold(III)
cyclometalated iminophosphoranes is strong and resists displace-
ment by PPh3 [17].
a
dichloride compounds [12]. The strength of the goldenitrogen
bond can be gauged by reaction with various reagents to see if it is
a goldechloride or the goldenitrogen bond that is broken; PPh3 has
been used to test this [12]. Scheme 4 shows the two possible
pathways; either one of the chlorides is displaced to form a cation
or the nitrogen is displaced to form a neutral compound. Previous
work has shown that the goldenitrogen bond in other
Following the reported procedure [17], the cycloaurated imi-
nophosphorane 6a, PPh3 and NH4PF6 were stirred in CH2Cl2 to
produce 7a$PF6. ESI-MS of the product showed a very intense ion
corresponding to the cation. The 31P NMR spectrum of 7a
confirmed the assignment. The cation showed two signals, one at
63.5 ppm from the iminophosphorane and one at 37.5 ppm from
PPh3. The high downfield position of the iminophosphorane
Fig. 5. The geometry of one of the independent molecules in the asymmetric unit of 6c showing the atom labelling scheme and thermal ellipsoids at the 50% probability level.

6
B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
Fig. 6. Comparison of the torsion angle C2eC1eN1eP1, for C,N-cycloaurated iminophosphorane 6a (A) and iminophosphorane 1a (B), showing the twisting of the C1eN1 bond.
Acetate groups have been omitted for clarity.
phosphorus indicated that it is still in a five-membered metal-
locyclic ring [31,32].
a process precluded in the present study because of potential
cleavage of the acetate groups under these conditions. Therefore an
alternative method using Ag2O in CH2Cl2 [34] was chosen for the
present study, and gave the products in high yields, Scheme 5.
These new compounds were more stable than the dichlorides, with
the thiosalicylate derivatives indefinitely stable at room tempera-
ture. ESI-MS of the compounds showed only the ions [M þ H]þ,
[M þ Na]þ and [M þ K]þ. This compares with the parent dichlorides
which only showed a weak [M þ Na]þ peak; the thiosalicylate
ligand provides additional ionisation sites, and discourages alter-
native ionisation via anion dissociation.
The NMR spectra of the thiosalicylate complexes were similar
to those of the starting dichlorides. The main changes were from
the phosphorus atom signal and the anomeric proton. The
anomeric proton has shifted upfield as the new ligand changes the
electron density on the gold. The same effect is seen in the 31P
NMR, as the phosphorus signal for 8b is 51.7 ppm compared to
60.3 ppm for 6b. Similar shifts were seen for the other
compounds. The anomeric carbon still showed coupling to phos-
phorus (6.1 Hz for 8b).
The 1H NMR spectrum showed large changes. The carbohydrate
signals were very broad so the multiplicity was difficult to deter-
mine, and the anomeric proton had been shifted upfield 1.80 ppm.
In the 13C NMR spectrum the anomeric carbon barely shifted, only
moving 0.2 ppm downfield, it also no longer had any coupling to
phosphorus. Interestingly C2 now showed 3J coupling to the
phosphorus (15.8 Hz for 7b). The rest of the spectrum was similar to
that of the dichloride.
Although crystals of X-ray quality could not be grown, it was
assumed that the PPh3 was trans to the nitrogen as in previous
examples [17]. This is caused by antisymbiosis which is the pref-
erence for a hard donor ligand trans to a soft ligand on a soft metal
[33]. Since ligands trans to each other compete for the same
s and p
metal orbitals, having the phosphine and carbon trans to each other
would destabilise the complex.
2.2.2. Reaction with thiosalicylic acid
There is interest in replacing the chlorides on C,N-cycloaurated
compounds with the dianionic thiosalicylate ligand to form
a second six-membered chelate ring [34,35]. This is because such
derivatives show improved anti-tumour effects [35] compared to
the parent dichloride [17]. Previous reactions with other complexes
were carried out in methanol with triethylamine as base [35],
Although crystals of X-ray quality could not be grown, it was
assumed that the sulphur atom of the thiosalicylate was trans to the
nitrogen because a previous structure determination of a thio-
salicylate complex of a gold(III) iminophosphorane showed that
configuration [17], consistent with antisymbiosis predictions.
OAc
Ph
Ph
O
P
OAc
AcO
AcO
Ph
N
Ph
O
Au
OAc
P
AcO
AcO
Cl
PPh3
N
7a
Au
OAc
Cl
Cl
OAc
Ph
PPh3
Ph
O
P
AcO
AcO
N
Ph3P
OAc
Au
Cl
Cl
Scheme 4. Two possible products of reacting 6a with PPh3.

B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
7
Ph
SH
Ph
O
P
COOH
Ph
AcO
N
O
Ph
P
O
Au
OAc
S
AcO
N
Au
OAc
Ag2O
O
Cl
Cl
6
8
Scheme 5. Reaction of gold(III) dichloride iminophosphoranes with thiosalicylic acid.
2.3. Biological activity
4.2. Preparation of ReN]PPh3 (R ¼ per-O-acetyl glucosyl) (1a)
Gold(III) compounds have shown biological activity [12]. To see
if the carbohydrate had any effect 6a and its thiosalicylate deriva-
tive 8a were tested for anti-tumour activity against a P388 Murine
Leukaemia cell line. 6a was essentially inactive with an IC50 value of
>62,500 ng mLꢁ1, which can be compared with (2-Cl2AuC6H4)
Ph2P]NePh which had an activity of 7546 ng mLꢁ1 [17]. 8a had
much better activity with a value of 1665 ng mLꢁ1, although this is
less than for the thiosalicylate derivative of (2-Cl2AuC6H4)Ph2P]
NePh at <487 ng mLꢁ1 [17]. Hence replacing the aryl substituent
on the nitrogen atom with a carbohydrate moiety does not enhance
the activity to the iminophosphorane, though again the thio-
salicylate does boost the activity significantly, compared with the
dichloride.
The sugar azide 2a (200 mg, 0.536 mmol) and PPh3 (140.5 mg,
0.536 mmol) were dissolved in dry CH2Cl2 (15 mL) under nitrogen
and stirred for 1.5 h. Removal of the solvent gave 1a (322.4 mg, 99%)
which was used for reactions without further purification. Crystals
suitable for X-ray diffraction were grown by cooling a CH2Cl2/
hexane solution of 1a.
1H NMR:
d 7.67e7.10 (m, 15H, arom.), 5.93 (dd, 1H, JH1,H2 8.4 Hz,
JH1,P 19.6 Hz, H1) 5.08 (t, 1H, H3), 4.25 (t, 1H, H2), 4.16 (t, 1H, H4),
3.99 (dd, 1H, JH5,H6b 6.3 Hz JH6a,H6b 12.2 Hz, H6b), 3.80 (m, br, 1H,
H5), 3.70 (m, br, 1H, H6a), 2.19, 2.17, 1.93, 1.86 (s, 12H, 4ꢂ OAc) ppm.
13C NMR:
d 170.5, 170.2, 169.6, 169.5 (C]O), 88.5 (d, JC1,P 5.8 Hz, C1),
74.7 (C5), 72.1 (C2), 72.6 (C3), 68.4 (C4), 62.2 (C6), 21.4, 20.9,
20.5 ꢂ 2 (4ꢂ CH3) ppm. 31P NMR:
d 21.1 ppm.
4.3. Preparation of (2-Cl2AuC6H4)Ph2P]NeR (R ¼ per-O-acetyl
glucosyl) (6a)
3. Conclusions
The reaction of acetylated sugar azides proceeded smoothly
with PPh3 to produce iminophosphoranes. All attempts at direct
metallation of the iminophosphoranes failed, presumably because
of steric crowding of the nitrogen atom. The pre-metallation of
PPh3 with mercury gave access to C,N-cycloaurated iminophos-
phoranes via transmetallation with [Me4N][AuCl4]. The reaction
was successful with different acetylated glycosyl sugars and so it
should be possible to extend this reaction to the wide range of
acetylated glycosyl sugars available.
The sugar azide 2a (200 mg, 0.536 mmol) and the diphosphine 4
(193.8 mg, 0.268 mmol) were dissolved in dry CH2Cl2 (15 mL) under
N2 and stirred for 2 h, solvent was removed under vacuum and
[NMe4][AuCl4] (221 mg, 0.536 mmol) was added and the solids
dissolved in dry acetonitrile (15 mL) and left to stir under N2 and
protected from light for 3 days. Solvent was removed under vacuum
and the residue dissolved in CH2Cl2 and filtered through Celite.
Addition of petroleum spirits and cooling to ꢁ15 ꢀC gave yellow
crystals of 6a (397 mg, 88%). Crystals suitable for X-ray diffraction
were grown by cooling a CH2Cl2/hexane solution of 6a.
1H NMR:
d 8.16e7.10 (m, 14H, arom.), 6.21 (dd, 1H, JH1,H2 8.7 Hz,
4. Experimental
JH1,P 20.4 Hz, H1) 5.09 (t, 1H, H3), 4.27 (t, 1H, H2), 4.16 (t, 1H, H4),
3.98 (dd, 1H, JH5,H6b 6.3 Hz JH6a,H6b 12.2 Hz, H6b), 3.80 (bm, 1H, H5),
3.70 (bm, 1H, H6a), 2.17, 1.93 ꢂ 2, 1.86 (s, 12H, 4ꢂ OAc) ppm. 13C
4.1. General procedures
NMR:
d
170.4, 170.2, 169.6 ꢂ 2 (C]O), 86.9 (d, JC1,P 5.6 Hz, C1), 74.8
Acetylated glycosyl azides were prepared from their parent
sugars [14e16]. 2-LiC6H4PPh2 [36] and Hg(2-C6H4PPh2)2 [37] were
prepared via literature methods. Dry CH2Cl2, THF, ether and hexane
were produced from a Pure Solv solvent purification system [38]. 1H
and 13C NMR spectra were recorded on a Bruker Avance III 400 FT
NMR spectrometer at 400.13 and 100.61 MHz respectively at 300 K,
while 31P{H} were from a Bruker DRX 300 FT NMR spectrometer at
121.5 MHz. Samples were run in CDCl3; 1H and 13C NMR spectra
were referenced to the solvent line at 7.26 and 77.16 ppm respec-
tively, while 31P{H} NMR spectra were referenced using an external
standard of H3PO4. High resolution ESI-MS was carried out on
a Bruker Daltonics MicrOTOF instrument. Samples were dissolved
(C5), 72.9 (C2), 72.6 (C3), 68.4 (C4), 62.3 (C6), 21.3, 20.9, 20.6 ꢂ 2
(4ꢂ CH3) ppm. 31P NMR:
d 61.6 ppm. Anal. Calcd for C32H33NO9P-
Cl2Au (874.452): C, 43.95; H, 3.80; N, 1.60. Found: C, 44.07; H, 3.80;
N, 1.70. ESI-MS: m/z: 838.122 (strong, [M ꢁ Cl]þ, calc. 838.124),
896.086 (weak, [M
þ
Na]þ, calc. 896.083), 947.193 (weak,
[M þ NMe4]þ, calc. 947.191).
4.4. Preparation of (2-Cl2AuC6H4)Ph2P]NeR (R ¼ per-O-acetyl
galacosyl) (6b)
C,N-Cycloaurated 6b was made using the same method as 6a,
with the sugar azide 2b (200 mg, 0.536 mmol), diphosphine 4
(193.8 mg, 0.268 mmol) and [NMe4][AuCl4] (221 mg, 0.536 mmol).
Work up gave 6b (378 mg, 81%).
in methanol and were infused at 180 m
L minꢁ1. The instrument was
calibrated over the appropriate mass range with sodium formate,
and all spectra were acquired in positive ion mode. Microelemental
analysis was carried out at the Campbell Microanalytical Laboratory
at the University of Otago. Biological testing details have been
described earlier [17].
1H NMR:
d 8.12 (m, 1H, arom.), 8.00e7.21 (m, 12H, arom.), 7.05
(m, 1H, arom.), 6.21 (dd, 1H, JH1,H2 8.8 Hz, JH1,P 18.2 Hz, H1) 5.22 (dd,
1H, JH4,H5 1.4 Hz, JH3,H4 3.4 Hz, H4), 4.98 (dd, 1H, JH3,H4 3.4 Hz, JH2,H3

8
B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
10.5 Hz, H3), 4.69 (dd, 1H, JH1,H2 8.8 Hz, JH2,H3 10.5 Hz, H2), 4.06 (m,
1H, H5), 3.81 (dd, 1H, JH5,H6b 7.5 Hz JH6a,H6b 11.4 Hz, H6b), 3.65 (dd,
1H, JH5,H6a 5.2 Hz, JH6a,H6b 11.4 Hz, H6a), 2.24, 1.90, 1.87, 1.69 (s, 12H,
d 170.8, 170.4, 169.9, 169.7 (C]O), 86.9 (d,
JC1,P 4.9 Hz, C1), 74.2 (C5), 70.7 (C3), 70.5 (C2), 67.3 (C4), 61.4 (C6),
(4ꢂ CH3) ppm. 31P NMR:
d 52.8 ppm. Anal. Calcd for
C39H37NO11PSAu (955.72): C, 49.01; H, 3.90; N, 1.47. Found: C,
49.04; H, 4.24; N, 1.53. ESI-MS: m/z: 978.141 (strong, [M þ Na]þ,
calc. 978.138), 956.159 (weak, [M þ H]þ, calc. 956.156), 1933.289
(weak, [2M þ Na]þ, calc. 1933.287).
4ꢂ OAc) ppm. 13C NMR:
21.6, 20.8, 20.7, 20.6 (4ꢂ CH3) ppm. 31P NMR:
d 60.3 ppm. Anal.
Calcd for C32H33NO9PCl2Au (874.452): C, 43.95; H, 3.80; N, 1.60.
Found: C, 44.10; H, 3.85; N, 1.84. ESI-MS: m/z: 838.123 (strong,
[M ꢁ Cl]þ, calc. 838.124), 896.083 (weak, [M þ Na]þ, calc. 896.083),
947.187 (weak, [M þ NMe4]þ, calc. 947.191).
4.8. Preparation of {2-[C6H4(S)CO2]AuC6H4}Ph2P]NeR (R ¼ per-
O-acetyl galacosyl) (8b)
Similarly C,N-cycloaurated 6b (50 mg, 0.057 mmol), thiosalicylic
acid (8.8 mg, 0.057 mmol) and Ag2O (67.2 mg, 0.29 mmol) in CH2Cl2
(10 mL) gave 8b (50.4 mg, 93%).
4.5. Preparation of (2-Cl2AuC6H4)Ph2P]NeR (R ¼ per-O-acetyl
arabinosyl) (6c)
1H NMR:
d 8.16 (m, 1H, arom.), 8.06e7.05 (m, 17H, arom.), 5.82
(dd, 1H, JH1,H2 8.6 Hz, J1,P 19.1 Hz, H1), 5.25 (dd, 1H, JH4,H5 1.4 Hz,
JH3,H4 3.4 Hz, H4), 5.06 (dd, 1H, JH3,H4 3.4 Hz, JH2,H3 10.4 Hz, H3), 4.61
(dd, 1H, JH1,H2 8.6 Hz, JH2,H3 10.4 Hz, H2), 4.23 (ddd, 1H, JH4,H5 1.4 Hz,
JH5,H6b 4.9 Hz, JH5,H6a 7.6 Hz, H5), 3.81 (dd, 1H, JH5,H6a 7.6 Hz, JH6a,H6b
11.4 Hz, H6a), 3.69 (dd, 1H, JH5,H6b 4.9 Hz, JH6a,H6b 11.4 Hz, H6b), 1.89,
Similarly 6c was made using the same method as 6a, with the
sugar azide 2c (200 mg, 0.664 mmol), diphosphine 4 (240 mg,
0.332 mmol) and [NMe4][AuCl4] (274 mg, 0.664 mmol). Work up
gave 6c (456 mg, 86%). Crystals suitable for X-ray diffraction were
grown by cooling a CH2Cl2/hexane solution of 6c.
1.82, 1.69, 1.67 (4s, 12H, 4ꢂ OAc) ppm. 13C NMR:
d 170.5, 170.4,
1H NMR:
d
8.15 (m, 1H, arom.), 7.99e7.54 (m, 9H, arom.),
169.4 ꢂ 2 (C]O), 169.4 (C]O, thiosalicylate), 84.0 (d, JC1,P 6.0 Hz,
7.42e7.25 (m, 3H, arom.), 7.04 (m, 1H, arom.), 6.06 (b, 1H, H1) 5.13
(dd, 1H, JH4,H5 1.7 Hz, JH3,H4 3.5 Hz, H4), 4.96 (dd, 1H, JH3,H4 3.5 Hz,
JH2,H3 10.3 Hz, H3), 4.73 (b, 1H, H2), 3.76 (mm, 2H, H5a, H5b), 2.20,
C1), 74.1 (C5), 71.0 (C2), 70.9 (C3), 67.0 (C4), 61.8 (C6), 20.8, 20.7,
20.6, 20.5 (4ꢂ CH3) ppm. 31P NMR:
d 51.7 ppm. Anal. Calcd for
C39H37NO11PSAu (955.72): C, 49.01; H, 3.90; N,1.47. Found: C, 49.14;
H, 3.98; N, 1.56. ESI-MS: m/z: 978.140 (strong, [M þ Na]þ, calc.
978.138), 956.160 (weak, [M þ H]þ, calc. 956.156).
1.89, 1.72 (s, 9H, 3ꢂ OAc) ppm. 13C NMR:
d 170.3, 170.1, 169.5 (C]O),
88.7 (d, JC1,P 5.2 Hz, C1), 70.4 (C3), 68.7 (C2), 67.7 (C4), 65.9 (C5),
21.0, 20.8, 20.7 (3ꢂ CH3) ppm. 31P NMR:
d 60.6 ppm. Anal. Calcd for
C29H29NO7PCl2Au (802.389): C, 43.41; H, 3.64; N, 1.75. Found: C,
43.63; H, 3.91; N, 1.83. ESI-MS: m/z: 766.106 (strong, [M ꢁ Cl]þ, calc.
766.103), 1569.178 (weak, [2M ꢁ Cl]þ, calc. 1569.174).
4.9. X-ray crystallography
X-ray crystallography data is summarised in Table 1. Intensity
data and unit cell dimensions were collected on a Bruker APEX II
CCD diffractometer. Structures were solved by the direct methods
option of SHELXS97 [39]. All non-hydrogen atoms were either
initially located or found in subsequent difference maps. Full-
matrix least-squares refinement (SHELXL97 [40]) was based on
4.6. Preparation of [(2-Cl(PPh3)AuC6H4)Ph2P]NeR]$PF6 (R ¼ per-
O-acetyl glucosyl) (7a$PF6)
The C,N-cycloaurated compound 6a (50 mg, 0.0572 mmol), PPh3
(15.0 mg, 0.0572 mmol) and [NH4][PF6] (9.3 mg, 0.0572 mmol)
were suspended in dry CH2Cl2 (10 mL) under N2 and stirred for 4 h.
The mixture was filtered through Celite and addition of petroleum
spirits to the filtrate gave 7a$PF6 (47.4 mg, 67%).
2
Fo with all non-hydrogen atoms anisotropic, unless otherwise
stated. Hydrogen atoms were refined using a riding model. All
calculations were carried out by the SHELXL97 suite of programs
[40] and were run under WinGX [41]. Crystal structure graphics
1H NMR:
d 8.28 (m 1H, arom.), 7.95e7.21 (m, 26H, arom.), 6.82
(m 1H, arom.), 6.66 (m 1H, arom.), 6.12 (t, 1H, H2), 5.13 (t, 1H, H4),
4.99 (t, 1H, H3), 4.61 (bd, 1H, H6b), 4.12 (m, 1H, H1), 3.85 (bd, 1H,
H6a), 3.3.36 (bm, 1H, H5), 1.97, 1.94, 1.89, 1.77 (s, 12H, 4ꢂ OAc) ppm.
Table 1
Crystal data and refinement details.
13C NMR:
d
170.6 ꢂ 2, 170.3, 169.3 (C]O), 87.1 (C1), 73.2 (C5), 72.9
(1a)
(6a)
(6c)
Formula
C32H34NO9P
C32H33NO9PCl2Au
2CH2Cl2
1044.28
89(2)
Monoclinic
P21
9.1905(2)
12.6953(3)
17.4863(4)
99.486(1)
2012.34(8)
2
C29H29NO7PCl2Au
CHCl3$1/2C6H14
1929.64
89(2)
Monoclinic
P21
13.4315(5)
13.4103(4)
20.9413(7)
90.785(2)
3771.6(2)
4
(C3), 72.4 (C2), 68.5 (C4), 61.2 (C6), 21.0, 20.8, 20.7, 20.6 (4ꢂ
CH3) ppm. 31P NMR:
Calcd for C50H48NO9F6P3ClAu (1246.249): C, 48.19; H, 3.88; N, 1.12.
Found: C, 48.29; H, 3.91; N, 1.26. ESI-MS: m/z: 1101.288 (strong,
[7a]þ, calc. 1101.284).
d
63.5, 37.5 (PPh3), ꢁ140.4 (PFꢁ6 ) ppm. Anal.
Mr
T(K)
Crystal system
Space group
a (Å)
607.57
83(2)
Monoclinic
P21
8.6161(8)
9.9944 (10)
18.1458(18)
90.705(1)
1562.5(3)
2
b (Å)
c (Å)
4.7. Preparation of {2-[C6H4(S)CO2]AuC6H4}Ph2P]NeR (R ¼ per-
O-acetyl glucosyl) (8a)
b
(deg)
V (Å3)
Z
r
The C,N-cycloaurated compound 6a (50 mg, 0.057 mmol) and
thiosalicylic acid (8.8 mg, 0.057 mmol) were dissolved in CH2Cl2
(10 mL). Ag2O (67.2 mg, 0.29 mmol) was added and the suspension
stirred for 1 h, protected from light. The suspension was filtered
through Celite and the solvent removed from the filtrate under
vacuum. The residue was recrystallised from CHCl3 and petroleum
spirits to give 8a (48.0 mg, 88%).
(g cmꢁ3
)
)
1.291
0.142
1.723
4.145
1.699
4.343
m
(mmꢁ1
Size (mm3)
F (000)
0.66 ꢂ 0.24 ꢂ 0.13 0.49 ꢂ 0.29 ꢂ 0.23 0.70 ꢂ 0.41 ꢂ 0.10
640
34.8
32710
0.982, 0.912
12240(0.031)
1032
28.0
32579
0.450, 0.296
9497(0.029)
1908
30.7
101907
0.433, 0.193
23064(0.058)
qmax (deg)
Refln collected
Tmax, min
Unique reflns
(Rint
1H NMR:
d 8.16e7.20 (m, 18H, arom.), 5.78 (dd, 1H, J1,2 8.7 Hz, J1,P
)
21.2 Hz, H1), 5.13 (t, 1H, H3), 4.25 (m, 1H, H4), 4.17 (m, 1H, H2), 3.98
(m, 1H, H6b), 3.92 (m, 1H, H5), 3.68 (dd, 1H, H6a), 1.95, 1.93, 1.81,
d 170.3, 170.2, 169.8, 169.7
(C]O), 169.5 (C]O, thiosalicylate), 83.8 (d, JC1,P 6.5 Hz, C1), 74.6
R1 [I > 2
s
(I)]
0.0357
0.0906
1.028
0.0431
0.1122
1.134
0.053(7)
þ2.35/ꢁ2.94
0.0594
0.1675
1.035
wR2 (all data)
GOF on F2
1.57 (4s, 12H, 4ꢂ OAc) ppm. 13C NMR:
Flack parameter 0.02(4)
ꢁ0.011(7)
Final
D
e (e Åꢁ3
)
þ0.44/ꢁ0.23
þ8.94/ꢁ2.74
(C5), 73.1 (C2), 72.8 (C3), 68.4 (C4), 62.4 (C6), 20.9, 20.7, 20.7, 20.3

B.P. Jarman, B.K. Nicholson / Journal of Organometallic Chemistry 702 (2012) 1e9
9
[11] S. Bräse, C. Gil, K. Knepper, V. Zimmermann, Angew. Chem. Int. Ed. 44 (2005)
5188.
were generated using ORTEP-3 [42]. The chiral compounds crys-
tallised in non-centrosymmetric space groups and the correct
absolute structures were determined using the Flack-x parameter
[43,44] and were also confirmed from the known stereochemistry
of the sugar moiety. For the refinement of 6a the N and C atoms
bonded to the Au atom were refined isotropically as their thermal
ellipsoids were prolate, presumably from interference from the
adjacent gold atom. The final difference map for 6c showed one
very large peak (8.9 e Åꢁ3) adjacent to one of the gold atoms, too
close to be anything other than an artefact, perhaps associated with
difficulties with the large absorption correction to the data. Some
other large residual peaks (2e4 e Åꢁ3) were associated with the
disordered CHCl3 molecule.
[12] W. Henderson, Adv. Organomet. Chem. 54 (2006) 207.
[13] H. Staudinger, J. Meyer, Helv. Chim. Acta 2 (1919) 635.
[14] K.P.R. Kartha, R.A. Field, Tetrahedron 53 (1997) 11753.
[15] B. Mukhopadhyay, K.P.R. Kartha, D.A. Russell, R.A. Field, J. Org. Chem. 69
(2004) 7758.
[16] F.M. Ibatullin, K.A. Shabalin, Synth. Commun. 30 (2000) 2819.
[17] K.J. Kilpin, W. Henderson, B.K. Nicholson, Inorg. Chim. Acta 362 (2009)
3669.
[18] J. Szmuszkovicz, M.P. Kane, L.G. Laurian, C.G. Chidester, T.A. Scahill, J. Org.
Chem. 46 (1981) 3562.
[19] M.T. Reetz, E. Bohres, Chem. Commun. (1998) 935.
[20] E. Boehm, K. Dehnicke, J. Beck, W. Hiller, J. Straehle, A. Maurer, D.Z. Fenske,
Naturforsch., B: Chem. Sci. 43 (1988) 138.
[21] D.L. Davies, S.M.A. Donald, S.A. Macgregor, J. Am. Chem. Soc. 127 (2005)
13754.
[22] D. Aguilar, R. Bielsa, M. Contel, A. Lledós, R. Navarro, T. Soler, E.P. Urriolabeitia,
Organometallics 27 (2008) 2929.
[23] R. Bielsa, R. Navarro, E.P. Urriolabeitia, A. Lledós, Inorg. Chem. 46 (2007)
10133.
Acknowledgements
[24] D. Aguilar, F. Aznárez, R. Bielsa, L.R. Falvello, R. Navarro, E.P. Urriolabeitia,
Organometallics 26 (2007) 6397.
We thank Dr. Tania Groutso, University of Auckland, and Dr. Jan
Wikaira, University of Canterbury for collection of X-ray intensity
data as well as Dr Kelly Kilpin for helpful discussions. Wendy
Jackson and Pat Gread provided technical assistance.
[25] J. Vicente, A. Arcas, M.T. Chicote, J. Organometal. Chem. 252 (1983) 257.
[26] J. Vicente, M.-T. Chicote, M.D. Bermudez, M.J. Sanchez-Santano, P.G. Jones,
C. Fittschen, G.M. Sheldrick, J. Organometal. Chem. 310 (1986) 401.
[27] J. Vicente, M.-D. Bermudez, F.-J. Carrion, P.G. Jones, Chem. Ber. 129 (1996)
1395.
[28] R.V. Parish, J.P. Wright, R.G. Pritchard, J. Organometal. Chem. 596 (2000) 165.
[29] P.A. Bonnardei, R.V. Parish, R.G. Pritchard, J. Chem. Soc., Dalton Trans. (1996)
3185.
[30] K.J. Kilpin, R.A. Linklater, W. Henderson, B.K. Nicholson, Inorg. Chim. Acta 363
(2010) 1021.
[31] P.E. Garrou, Chem. Rev. 81 (1981) 229.
[32] P. Ona-Burgos, I. Fernández, L. Roces, L.T. Fernández, S. García-Granda,
F.L. Ortiz, Organometallics 28 (2009) 1739.
Appendix A. Supplementary material
CCDC 843659e843661 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
[33] R.G. Pearson, Inorg. Chem. 12 (1973) 712.
References
[34] M.B. Dinger, W. Henderson, J. Organomet. Chem. 560 (1998) 233.
[35] W. Henderson, B.K. Nicholson, S.J. Faville, D. Fan, J.D. Ranford, J. Organomet.
Chem. 631 (2001) 41.
[36] S. Harder, L. Brandsma, J.A. Kanters, A. Duisenberg, J.H. van Lenthe,
J. Organomet. Chem. 420 (1991) 143.
[37] M.A. Bennett, M. Contel, D.C.R. Hockless, L.L. Welling, A.C. Willis, Inorg. Chem.
41 (2002) 844.
[38] A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, F.J. Timmers, Organo-
metallics 15 (1996) 1518.
[39] G.M. Sheldrick, SHELXS-97: A Program for Automatic Solution of Crystal
Structures, University of Göttingen, Germany, 1997.
[40] G.M. Sheldrick, SHELXL-97: A Program for the Refinement of Crystal Struc-
tures, University of Göttingen, Germany, 1997.
[1] A.W. Johnson, Ylides and Imines of Phosphorus, John Wiley and Sons, New
York, 1993.
[2] A. Steiner, S. Zacchini, P.I. Richards, Coord. Chem. Rev. 227 (2002) 193.
[3] H. Alper, J. Organomet. Chem. 127 (1977) 385.
[4] J. Vicente, J.-A. Abad, R. Clemente, J. López-Serrano, M.C. Ramírez de Arellano,
P.G. Jones, D. Bautista, Organometallics 22 (2003) 4248.
[5] M.A. Leeson, B.K. Nicholson, M.R. Olsen, J. Organomet. Chem. 579 (1999) 243.
[6] S.D.J. Brown, W. Henderson, K.J. Kilpin, B.K. Nicholson, Inorg. Chim. Acta 360
(2007) 1310.
[7] D. Aguilar, M. Contel, R. Navarro, E.P. Urriolabeitia, Organometallics 26 (2007)
4604.
[41] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837.
[42] L.J. Farrugia, J. Appl. Crystallogr. 30 (1997) 565.
[43] H.D. Flack, Acta Crystallogr., Sect. A 39 (1983) 876.
[44] G. Bernardinelli, H.D. Flack, Acta Crystallogr., Sect. A 41 (1985) 500.
[8] K.T.K. Chan, L.P. Spencer, J.D. Masuda, J.S.J. McCahill, P. Wei, D.W. Stephan,
Organometallics 23 (2004) 381.
[9] P. Wei, K.T.K. Chan, D.W. Stephan, Dalton Trans. (2003) 3804.
[10] A. Messmer, I. Pintér, F. Szegö, Angew. Chem. Int. Ed. 3 (1964) 228.