A contribution to the rational design of Ru(CO)3Cl2L complexes for in vivo delivery of CO

Dalton

Transactions

View Article Online

View Journal

PAPER

A contribution to the rational design of

Ru(CO) Cl L complexes for in vivo delivery of CO†

3

2

Cite this: DOI: 10.1039/c4dt02966f

a,b

c

a

João D. Seixas, Marino F. A. Santos, Abhik Mukhopadhyay, Ana C. Coelho,

a

d

b

Patrícia M. Reis, Luís F. Veiros, Ana R. Marques, Nuno Penacho,

b

c

b

Ana M. L. Gonçalves, Maria J. Romão, Gonçalo J. L. Bernardes,‡

Teresa Santos-Silva* and Carlos C. Romão*

c

a,b

A few ruthenium based metal carbonyl complexes, e.g. CORM-2 and CORM-3, have therapeutic activity

attributed to their ability to deliver CO to biological targets. In this work, a series of related complexes

with the formula [Ru(CO)

3

Cl

(CH

H (11); CNCMe

2

L] (L = DMSO (3), L-H

3

CSO(CH

Na (7); 4-NC (CH

Me (12); CNCMeEtCO

2

)

2

CH(NH

SO Na (8); PTA (9); DAPTA (10); H

Me (13); CN(c-C )CO Et) (14)) were

2

)CO

2

H) (6a); D,L-H

3

CSO(CH

2

)

2

CH-

2

(NH )CO

2

H (6b); 3-NC

5

H

4

2

)

2

SO

3

5

H

4

2

)

2

3

CS-

(CH

2

)

2

CH(OH)CO

2

CO

2

H

3 4

2

designed, synthesized and studied. The eﬀects of L on their stability, CO release proﬁle, cytotoxicity and

anti-inﬂammatory properties are described. The stability in aqueous solution depends on the nature of L

as shown using HPLC and LC-MS studies. The isocyanide derivatives are the least stable complexes, and

the S-bound methionine oxide derivative is the more stable one. The complexes do not release CO gas to

the headspace, but release CO

2

instead. X-ray diﬀraction of crystals of the model protein Hen Egg White

II

Lysozyme soaked with 6b (4UWN) and 8 (4UWV) shows the addition of Ru (CO)(H

2

O)

4

at the His15

+

binding site. Soakings with 7 (4UWU) produced the metallacarboxylate [Ru(COOH)(CO)(H

2

O)

3

]

bound to

the His15 site. The aqueous chemistry of these complexes is governed by the water–gas shift reaction

initiated with the nucleophilic attack of HO⁻on coordinated CO. DFT calculations show this addition to

be essentially barrierless. The complexes have low cytotoxicity and low hemolytic indices. Following i.v.

administration of CORM-3, the in vivo bio-distribution of CO diﬀers from that obtained with CO inhalation

or with heme oxygenase stimulation. A mechanism for CO transport and delivery from these complexes

is proposed.

Received 26th September 2014,

Accepted 12th November 2014

DOI: 10.1039/c4dt02966f

www.rsc.org/dalton

Introduction

1

The biological role of CO as a signaling molecule rapidly

opened the way to the discovery of its biological applications.

The cytoprotective, anti-inflammatory, anti-proliferative and

a

Instituto de Tecnologia Química e Biológica-António Xavier da Universidade Nova

de Lisboa, Av. da República, EAN, 2780-157 Oeiras, Portugal.

E-mail: ccr@itqb.unl.pt; Tel: +351 214 469 751

anti-apoptotic properties of CO suggested its therapeutic use.

Soon after the seminal report on the successful use of CO gas

2

in a rodent model of LPS induced sepsis, the concept of pro-

b

Alfama Ltd., Instituto de Biologia Experimental e Tecnológica, IBET, Av. da

drugs capable of delivering CO, circumventing the short-

comings of CO gas inhalation, appeared in the patent

República, EAN, 2780-157 Oeiras, Portugal

c

UCIBIO@REQUIMTE, Departamento de Química, Faculdade de Ciências e

3

–5

Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal.

E-mail: tsss@fct.unl.pt; Tel: +351 212 948 300 (10940)

literature.

Among these, the derivatives of the fragment

] introduced by Motterlini, Mann and co-workers

first illustrated the use of metal-based CO-releasing molecules

CORMs) and became the golden standard in this novel thera-

II

[

Ru (CO)

3

d

Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa,

Av. Rovisco Pais, 1049 Lisboa, Portugal

(

†

Electronic supplementary information (ESI) available: FTIR spectra and HPLC

13

6,7

traces for metal complexes. C NMR spectra of CORM-2 in DMSO solution. peutic area. In fact, over 200 papers have been published in

ESI-MS spectra for selected LC-MS fractions. Atomic coordinates for all DFT opti-

the last 10 years on the biological and therapeutic use of the

mized species. See DOI: 10.1039/c4dt02966f

two cornerstones of this family: [Ru(CO)

3 2 2

Cl ] (CORM-2) and

‡

Present address: Department of Chemistry, University of Cambridge, Lensfield

2

its water soluble glycinate derivative fac-[Ru(CO)

3

Cl(κ -

H NCH CO )] (CORM-3). In spite of a recent contradictory

Road, Cambridge CB2 1EW, UK and Instituto de Medicina Molecular, Faculdade

de Medicina da Universidade de Lisboa, Av. Prof. Egas Moniz, 1649-028 Lisboa,

Portugal.

8

2

9

2

report, these CORMs seem to have a low toxicity and produce

This journal is © The Royal Society of Chemistry 2014

Dalton Trans.

View Article Online

Paper

Dalton Transactions

a variety of beneficial therapeutic eﬀects while keeping the retained in the main organs. The authors also confirm the

values of carboxyhemoglobin (COHb) in circulation close to reaction of these CORMs with plasma proteins and that “every

2

8

baseline values (≤5% COHb). Some remarkable examples of CORM has several forms in blood due to hydrolysis”.

this therapeutic eﬃcacy are given by CORM-3 in the protection

From the existing data it is undeniable that: (i) CORM-2,

1

0,11

against myocardial infarct and heart failure

servation of tissues for transplantation and by CORM-2 in in vivo; (ii) this activity is compatible with that expected for CO

the protection of allogeneic aortic transplants in mice, but and often validated with independent CO gas treatments;

many other results can be found in several reviews.

and the con- CORM-3 and ALF492 are biologically active both in vitro and

1

2

1

3

14

8

,14,15

(iii) solutions of CORM-3 release CO , not CO, and become

2

7

,26

This capacity to act in all these diﬀerent organs and tissues biologically inactive upon aging;

(iv) CO is detected in cells

at similar concentrations is somewhat surprising, and several treated with CORM-3 by the CO specific, fluorescent organo-

2

9

questions regarding its mode of action are still unanswered. In metallic probe, COP-1 (CO probe 1); and (v) the metal

1

6,17

fact, recent studies indicate that the mode of action of scaﬀold may also play a physiologically significant role.

CORM-2 goes beyond its CO releasing activity, due to non- These apparently contradictory findings show that our under-

1

6

II

specific hydrophobic interactions or other biologically sig- standing of the chemistry of [Ru (CO) ] derived complexes

3

1

7

nificant side-eﬀects, e.g. ROS production.

Until 2014, the derivatives of the [Ru (CO)

under biological conditions warrants a further study in order

] fragment to enable the design of analogues equipped with drug-like pro-

II

3

studied under the perspective of their biological activity were perties. In this manuscript we present the synthesis of a series

limited to the methyl β-D-thiogalactoside (Gal-S-Me), the of new [Ru(CO) Cl L] complexes where the variation of the

(Gal-S-Me)] (ALF492), the thiazole nature of L is intended to provide information on the influ-

3

2

1

8

derivative fac-[Ru(CO)

3 2

Cl

1

9–21

derivative [Ru(CO) Cl (thiazole)]

and some analogues of ence of the inner coordination sphere on the solubility, stabi-

3

2

4

CORM-3 bearing other amino acid ligands. ALF492 is much lity, reactivity, CO release profile, cytotoxicity, anti-

more eﬃcient than CORM-3 in rescuing mice in a model of inflammatory activity and other pharmacologically relevant

cerebral malaria without decreasing parasitemia, and this properties of the complexes. We screen the interaction of some

enhanced activity has been linked to selective accumulation in of these CORMs with proteins and examine important

the liver. Indeed, it is reasonable to expect that the biological mechanistic features of their CO release profiles using protein

activity of the [Ru(CO) Cl L] complexes can be modulated by X-ray crystallography and LC-MS spectrometry. Structural and

3

2

the nature of the ancillary ligand L. Understanding the nature reactivity results are also rationalized using DFT (density func-

of modulation for a given metal–carbonyl core is very impor- tional theory) calculations. Finally we present a detailed study

tant for the design of drug-like CORMs, as we have discussed of the bio-distribution of CO delivered by CORM-3 in mice.

2

2,23

elsewhere.

ADMET (Absorption, Distribution, Metabolism,

Excretion and Toxicity) and CO-release properties of a given

CORM are both directly influenced by their inner and outer

coordination spheres. The CO delivery mechanism of CORM-2

Results and discussion

Synthesis and characterization of Ru(CO) Cl L complexes

3 2

2

4

and CORM-3, once labeled “fast CO releasers”, remains

elusive, now that it has been proven that they are unable to The synthesis of new complexes of the formula [Ru(CO)

Cl

transfer CO to deoxyMb as previously accepted. The last involves the cleavage of the chloride bridges of the commer-

observation is in agreement with the absence of CO in the cially available dimer [Ru(CO) Cl (CORM-2) as depicted in

L]

3

2

5

]

2 2

3

0

headspace of solutions of CORM-3 or [Ru(CO) Cl (thiazole)], eqn (1).

3

2

1,26

as ascertained using highly sensitive GC methods.

Besides

½

RuðCOÞ Cl2ꢀ þ 2L ! 2½RuðCOÞ Cl2Lꢀ

ð1Þ

3

2

3

the biodistribution study of ALF492, the only other ADME

II

related observations made in the [Ru (CO) ] based CORM lit-

This cleavage is usually performed in solvents (Solv) such

3

erature were the rapid reaction of CORM-3 and [Ru- as DMSO, THF, acetone or MeOH, where solvated species of

CO) Cl (thiazole)] with proteins (e.g., lysozyme, serum the formula [Ru(CO) Cl (Solv)] react with diﬀerent ligands (L)

albumin and transferrin) without CO release and the to generate the target complexes. Since the aim of this study is

complex pH dependent speciation of these same complexes in the modulation of the physical, chemical and biological pro-

(

3

2

3

2

1,26,27

II

aqueous biologically compatible media. However, during the perties of the derivatives of the [Ru (CO)

3

] fragment, we

selected a collection of ligands with C, N, O, P or S donor

based CORMs with diﬀerent ligands such as amino acid atoms covering a broad range of coordinative properties. Iso-

esters, amino-acidates, acetylacetonate, and pyridine derived cyanide ligands of the type CNCR COOR′ (compounds 12, 13

and 14 in this study) are isoelectronic with CO and have been

The influence of the ancillary ligands on some of the pro- successfully used in other organometallic drugs, namely in the

II

writing of the present article a collection of fifteen Ru (CO)

3

2

8

ligands had their ADMET properties extensively examined.

9

9m

+ − 31

perties of the complexes, including cell and in vivo toxicity, cationic Cardiolite ([ Tc (CNCH

2 2 6

CMe OMe) ] A ) and the

3

2

tissue absorption (Ru contents) and Ru biodistribution profile, liver active CORM [Mo(CO) (CNCMe CO H) ] (ALF794).

3

2

3

is confirmed in the results reported. Moreover, this first study These are weaker π-acids and stronger σ-donors than CO. The

of the metabolism of this type of CORMs shows that Ru–CO water soluble phosphines PTA (1,3,5-triaza-7-phosphaadaman-

complexes are absent in the urine and are generally poorly tane) and DAPTA (diacetylPTA; 3,7-diacetyl-1,3,7-triaza-5-phos-

Dalton Trans.

This journal is © The Royal Society of Chemistry 2014

View Article Online

Dalton Transactions

Paper

phabicyclo[3.3.1]nonane) continue this trend as they are still corresponding ligands in acetone. The isocyanide complexes

weaker π-acids than isocyanides (compounds 9 and 10 in this were prepared in CHCl since the ligands are strong nucleo-

3

study). The pyridine ligands can be considered to represent philes and readily cleave the Ru–Cl–Ru bridges at room temp-

biologically meaningful N-heterocyclic ligands, with modest erature. The pale yellow or white solid products were isolated

π-acidity. Their substitution patterns were chosen to impart in moderate (53–63%) to good (73–85%) yields and were fully

solubility through the sulfonate functionality (compounds 7 characterized using NMR, FTIR and elemental analyses. It is

and 8 in this study). The methionine oxide (compounds 6a or however important to note that the tricarbonyl complexes are

6

b in this study) was selected to improve the biocompatibility not amenable to purification by recrystallization due to their

and aqueous solubility relative to the DMSO complexes that lability in solution. In fact, they were all characterized after

result from dissolving CORM-2 in DMSO for biological appli- precipitation from their mother liquors by filtration followed

cations (see below). Finally, the thioether (compound 11 in by washing and drying. In the case of sulphonate complexes 7

this study) represents a type of biologically relevant ligand and 8 we were unable to obtain a correct analysis due to the

which is regarded as electronically flexible but is usually rather presence of residual amounts of water or the solvent of crystal-

3

4

labile in classical organometallic substitution reactions.

A

lization, which could not be entirely removed under vacuum

1

8

sugar based thioether was successfully used in ALF492. The (see the Experimental section). Some properties of the com-

only O ligand used is DMSO in compound 3 and in some of plexes are given in Table 1.

the isomers of 4 and 5 in the series of known DMSO deriva-

3

5

tives [RuCl

2

(CO)

x

(DMSO)4−x] (x = 1, 2, 3). The structures of

Cl L] (1–14) are depicted in and CO₂released by CORM-2, CORM-3 and the other complexes

Table 1 CO vibrations in FTIR, water solubility and equivalents of CO

the new complexes of type [Ru(CO)

3

2

Fig. 1, together with CORM-2 and CORM-3. The details of described in this work (Fig. 1)

syntheses are given in the Experimental section.

Reaction times and solvents used depended on the solubi-

lity of the selected ligand. Whenever solubility permitted, Compound cm

ν

CO

Water

solubility

(mg mL

a

(

KBr;

Equivalents of

CO released

Equivalents of

−

1

−1

d

)

CO released

2

acetone was the solvent of choice. In the case of complexes 8

and 9 MeOH was used to circumvent the low solubility of the

c

CORM-2

CORM-3

3

2144(s)

5

Not detected

1.80

2

090(s)

069(vs)

g

e

2139(s)

>20

Not detected

0.68

2

1

057(s)

981(w)

2134(s)

068(s)

2077(s)

020(s)

2

Not detected

0.71

2

b

4

<3

Not detected

2

b

5

2001(vs)

2131(s)

>5

2

Not detected

1.20

6

7

8

9

a

2

055(s)

2131(s)

055(s)

2137(s)

053(s)

2137(s)

053(s)

2134(w)

b

2

>5

Not detected

—

2

0.96

1.00

0.28

2

>5

2

Ins

2

1

060(s)

994(s)

10

11

12

13

14

2135(w)

2

2141(s)

2

2145(s)

2

2149(vs) Ins (dec)

2

>5

Not detected

0.19

0.40

067(s)

001(s)

c

>5

077(s)

063(s)

f,g

f

Ins (dec)

Not detected

0.98

093(s)

058(s)

f,g

f

0.85

090(s)

061(vs)

f

2144(s)

Ins (dec)

0.21

2

093(s)

072(s)

a

b

c

In H

2

O, rt, except where noted. A mixture of isomers is used. 10%

d

DMSO–H O. Headspace of a H O solution; GC-TCD; dark; rt; N

2

e f

atmosphere; 24 h. PBS 7.4. 25 mM in acetone–PBS 7.4 after 5 h.

g

Fig. 1 Structures of CORM-2, CORM-3 and the new complexes

Trace amounts of CO were detected after 4 h using GC-RCP. Ins =

described in this work.

insoluble; dec = decomposition.

This journal is © The Royal Society of Chemistry 2014

Dalton Trans.

View Article Online

Paper

Dalton Transactions

1

The H NMR spectra are very simple and show that the more stable all cis-4b. The pure complex 3, prepared indepen-

3

5

6

ligand protons shifted downfield due to the electron withdraw- dently, when dissolved in d -DMSO gives exactly the same

ing character of the Ru(CO) Cl fragment. In the case of com- dicarbonyl products with the liberation of CO gas (see Fig. S2

plexes 6a, b, 7 and 8 the signals of the two CH groups of the in ESI†). However, this CO loss does not progress beyond the

linear chain appear superimposed, aﬀording one singlet in the dicarbonyl stage and after four days both isomers of 4 are still

3

2

1

31

1

H NMR spectrum of each complex. The P NMR spectra of present in the H NMR spectrum without formation of the

the phosphine compounds 9 and 10 present chemical shifts in monocarbonyl complex 5. In fact, the dicarbonyl 4 and mono-

3

3,36,37

the range reported for other Ru complexes.

the deshielding of the P resonance of the DAPTA ligand, and no substitution of the carbonyl groups is observed at

caused by coordination, is remarkable since it moves from room temperature up to four days (NMR evidence not shown).

In particular, carbonyl 5 complexes are highly stable in the DMSO solution

3

1

3

1

δ (ppm) −78.5 in the free ligand to δ (ppm) 5.87 in 10. The

C

The other peaks visible in the H NMR spectrum correspond

NMR spectrum of 12 shows at low fields two CO resonances to the exchange of DMSO (or chloride) with residual H

2

O

6

1

(

δ 183 and 182 ppm) and one CN resonance (δ 169 ppm), in present in d -DMSO. A shift in the water resonance in the H

agreement with a fac-[Ru(CO) Cl (CNCMe CO Me)] configur- NMR spectrum can be detected over time (see Fig. S3 in ESI†).

3

2

1

3

ation. The analogous complexes 13 and 14 have similar

C

2

This easy exchange between DMSO and H O is similar to that

described for RuCl (DMSO) .

2 4

3

9,40

resonance values.

The FTIR spectra of most compounds present the usual ν_C_O

stretching band pattern corresponding to the fac-M(CO) frag-

ment: a sharp, strong vibration at ca. 2135 cm and a very

3

−

1

Solubility and stability in aqueous media

−

1

strong, broader band at ca. 2055 cm . The latter can be split The solubility of the CORMs described above, in water or

as in compounds 9 and 10, which show a weaker band at ca. aqueous media compatible with their biological applications,

135 cm⁻and a splitting of the lower wavenumber vibration was determined by visual inspection of solutions prepared

1

2

−

1

−1

in two strong bands at ca. 2060 cm and 2000 cm . This is with weighed amounts of the compounds in determined

probably due to the lowering of the local symmetry caused by amounts of the solvent, at room temperature. A compound

−

1

the very bulky ligands. It is interesting to note the diﬀerent was considered soluble at a given concentration (mg mL ) if

coordination modes of the sulfoxide ligands in complexes 3 the solution obtained was visually clear and transparent. The

−

1

and 6a (or 6b) (see ESI S2 and S3† for details). While 3 shows a tests were done at 1 mg mL intervals. Since values above

band at 903 cm⁻which is assigned to the νSO vibration of an 5 mg mL are usually appropriate to provide biologically and

1

−1

oxygen bound Me SO ligand, both 6a and 6b have a band at therapeutically useful working concentrations no higher solu-

2

1

017 cm⁻corresponding to a νSO vibration of a sulphur bility limits were determined except for CORM-3 as seen in

1

bound Me

sulfoxide coordination chemistry.

borated by DFT calculations (see DFT calculations) on complex ligands coordinated to the metal. In spite of being used to

. The isomer with dimethylsulfoxide coordinated by the O- improve the water solubility of metal complexes in biological

2

SO ligand in agreement with the versatility of the Table 1. As expected, the majority of CORMs prepared are also

3

5,38

Those results are corro- quite soluble in water due to the nature of the ancillary

3

−

1

41

atom is 5 kcal mol more stable than the isomer with S-co- applications, the phosphine PTA is not able to guarantee the

ordinated sulfoxide, and the latter has a νSO vibration water solubility of 9, which is completely insoluble. In con-

30 cm⁻higher. The nature of the donor atoms (C, N, O, S or trast, the diacetylated version of PTA, DAPTA, imparts a satis-

1

33

2

P) of the diﬀerent ligands used does not cause major diﬀer- factory solubility to [Ru(CO) Cl (DAPTA)] (10). The isocyanide-

3

2

ences in the νCuO stretching vibrations of the corresponding derived complexes 12, 13 and 14 are only slightly soluble in

complexes. Interestingly, the highest νCuO wavenumbers are water where they readily decompose, darkening to an orange-

−

1

found for the isocyanide complexes (e.g. 13 ν_C_u_Ocm : 2149, brown color and releasing gas bubbles.

−

1

2

090, 2061) and the thioether complex in 11 (νCuO cm : 2141,

In contrast, the aqueous solutions of all other complexes

2

077, 2063). The νCuN vibration in the isocyanide complexes remain clear and almost colorless, a fact that tells nothing

−

1

appears at ca. 2250 cm

.

about their stability. For instance, solutions of 7 and 8 in dis-

tilled water have a pH between 2.5 and 2.9, indicating that

there is a reaction that increases [H ]. This eﬀect is similar to

Speciation of CORM-2 in DMSO

+

4

2

In all the cases reported so far, the addition of a ligand L to that discussed at length for fac-[Ru(OCOCF ) (CO) (H O)],

3

2

3

2

4

3

21

the Ru(CO)

3 2

Cl core resulted in the octahedral complexes CORM-3, and [Ru(CO) Cl (thiazole)] and is the result of the

Ru(CO) Cl L]. However, some deviations from this pattern nucleophilic addition of HO to one of the CO ligands in the

−

[

3

2

II

have been observed. In fact, the reaction of [Ru(CO) Cl ]

[Ru (CO) ] fragment (see eqn (2)).

3

2 2

3

(

CORM-2) with DMSO leads to complex 3 and the dicarbonyl

ꢁ

HO

1

3

complex cis,trans,cis-RuCl

NMR of a sample of CORM-2 measured over 29 min after dis-

2

(DMSO)

2

(CO)

2

(4a) as shown by

C

½RuðCOÞ Cl Lꢀ ꢁ)*ꢁ ½RuðCOÞ ðCOOHÞCl Lꢀ

ð2Þ

3

2

þ

H

6

solution in d -DMSO. This spectral evolution is depicted in

the ESI, Fig. S1.† Over a period of 18 h the tricarbonyl complex in aqueous media, HPLC, LC-MS and H NMR analyses were

is fully converted into the dicarbonyl isomers 4a and the carried out for the following complexes: 3, 4, 6a or 6b, 10. For

In order to further investigate the stability of the complexes

1

3

Dalton Trans.

This journal is © The Royal Society of Chemistry 2014

View Article Online

Dalton Transactions

Paper

this purposes complexes 6a and 6b were considered equivalent in ESI†). The m/z = 392.5 peak can be assigned to a species with

+

species since 6b is a mixture of 6a and its optical isomer.

the formulation [Ru(CO) (CO )(DMSO) Cl] (calcd m/z = 392.88).

2 2 2

Aqueous solutions of the CORMs were prepared and ali- The more abundant peak at m/z = 314.9 corresponds to

+

quots eluted through a reverse phase HPLC column using the [Ru(CO)

2 2

(CO )(DMSO)Cl] (calcd m/z = 314.87). These data are

H O–MeOH gradient as described in the Experimental section in agreement with the favorable attack of water on one of the

2

and as shown in the ESI Fig. S4–S9.† A run was done every CO ligands to generate a CO ligand. The lability of the complex

1

5 minutes starting immediately after the dissolution of the explains the easy exchange of chloride and DMSO ligands. This

1

compound (t = 0) and ending 60 min or 75 min later. None of is also in agreement with the similar exchange observed by H

the compounds tested can be considered stable in this time NMR with CORM-2 in DMSO solution (Fig. S3 in ESI†). No

period since several peaks are present in the HPLC traces useful MS was obtained in negative ion mode.

immediately after the dissolution. Some of these traces are

shown in the ESI (Fig. S4–S9†).

The complex 6a (or 6b) possesses a methionine oxide

ligand and is therefore very similar to 3 except for the fact that

The three DMSO derivatives 3, 4 and 5 show a peak in the it is coordinated through the S atom and not through the sulf-

initial chromatogram (t = 0) with a retention time (RT) of ca. oxide oxygen. Interestingly, HPLC reveals that 6b is signifi-

7

–9 min, which gradually disappears along time and which cantly more stable than 3 in solution, as can be seen in Fig. 2

was assigned to the parent complex. Of the three, the more (more details in ESI†). In fact, the peak at RT = 3.0 min is

stable complex is the monocarbonyl 5, which has a half-life of already present in the chromatogram of 6b taken at t = 0 but it

ca. 50 min and a 65% reduction 75 min after dissolution only grows a little until 60 min. Accordingly, the peak assigned

(Fig. S6 in ESI†). The dicarbonyl complex 4 decomposes rather to the parent 6b complex (RT = 10.2 min) only decreases

quickly since the parent peak area (RT = 8.9 min) decreases 18- slightly during 60 min. An LC-MS analysis was carried out for

fold within the first 15 min post dissolution (see Fig. S5 in 6a in order to characterize the products in solution, both in

ESI†). These decay processes are certainly related to aquation positive and negative ion modes.

reactions and do not produce any release of CO or CO₂as

described in the next section and Table 1. In the case of the tri- (2.0–2.4 min) in positive ion mode is dominated by the peak

carbonyl complex [Ru(CO) Cl (DMSO)] (3) the peak assigned to corresponding to the free methionine oxide (Met-O) (m/z =

The MS of the peak with the lower retention time

3

2

the parent complex (RT = 8.8 min) is already the third largest 166.0) and a stronger signal at m/z = 487.0 (see the MS spec-

in terms of the Area Under the Curve measured in the initial trum in ESI, Fig. S11†). This can be assigned to the cation

+

(

t = 0) chromatogram (Fig. 2 and Fig. S7 in ESI†). A peak with [Ru(CO)

2

(Met-O)

2

]

(calcd m/z = 486.98). These are clearly

RT ≈ 3.0 min increases along time for the three complexes. decomposition products or even eventual impurities of the

The HPLC chromatogram of CORM-2, dissolved in acetone, initial sample. The immediate next elution fraction (RT =

and eluted with the same MeOH–H

2

O gradient actually con- 2.6–3.2 min) shows three main signals at m/z = 240.88, 281.65,

sists essentially of a peak at RT = 3.0 min (Fig. S4 in ESI†). It is 298.67 (in decreasing order of intensity). These masses are

+

the most important one in all chromatograms of 3 but smaller than those of [Ru(CO)

increases very little in the case of 4 and only appears at ca. must correspond to smaller decomposition fragments. The

0 min post dissolution in the case of 5. The remarkable bio- peak at m/z = 349.87 found in the fraction with RT =

2

(Met-O)] (m/z = 321.93) and

3

+

logical activity of CORM-2 and its close relationship to complex 6.1–6.5 min can be assigned to [Ru(CO)

3 2

(Met-O)Cl ] -2Cl (m/z

3

prompted our attempted identification of the dominant = 349.93). Another strong peak at m/z = 730.72 seems to corre-

species in aqueous solutions of 3. This was carried out using spond to polynuclear decomposition products that aggregate

LC-MS analysis of a solution of 3 in 10% (v/v) methanol–water. more than one Ru ion. The positive ion mode ESI-MS spec-

The MS of the species observed at RT = 2.6 min has two large trum of the fraction that carries 6a (RT = 10.1–10.7 min) does

peaks at m/z = 314.9 and 392.5 which dominate over a large not contain the parent ion peak. The LC-MS analysis carried

number of small signals spread up to m/z = 1000 (see Fig. S10 out in negative ion mode presents one signal at m/z = 384.8 in

the fraction with RT = 2.7–3.1 min (the fraction that increases

along time), which can be tentatively formulated as [Ru-

−

(

COOH)(CO)

2

{SO(CH

3

)CH

2

CH

2

CH(NH

2

)CO

2

}(OH

2

)] (calc. m/z

=

384.94) (see Fig. S12 in ESI†). Importantly, a peak assignable

to the parent anion [RuCl (CO) (SO(CH )CH CH CH(NH )-

2

3

2

−

2

CO )] (calc. m/z = 421.86) is found at m/z = 421.7 in the frac-

tion with RT = 9.9–10.9 min, the fraction that slowly decreases

along time in HPLC (see Fig. S13 in ESI†). The higher stability

of 6a, relative to 3, is therefore corroborated by both the slow

decay of its HPLC peak and the detection of its parent anion

using ESI-MS. Finally, a peak at m/z = 549.7 found in the same

elution fraction (RT = 9.9–10.9 min) can be assigned to the ion

Fig. 2 HPLC traces of 3 (left) and 6b (right) along time. Initial concen-

−

−1

[RuCl(CO)

{SO(CH

)CH

CH

2

CH(NH

)CO

}

]

(calcd m/z

=

tration: 1 mg mL in 10% DMSO–H

2

O. Column: Nucleosil C18 (5 μm)

3

2

1

50 mm × 4.6 mm. Elution gradient: 10% MeOH–H

2

O; ﬂow: 1 mL min⁻

.

549.94), a plausible decomposition product of 6a. Somewhat

This journal is © The Royal Society of Chemistry 2014

Dalton Trans.

View Article Online

Paper

Dalton Transactions

unexpectedly, the phosphine derivative 10 proved relatively Table 2 for refinement statistics). As expected the obtained

unstable and its HPLC trace shows many signals some of models are highly similar to each other and to the native

which increase in the region of RT = 2.5–3.0 min while other protein.

peaks decrease steadily (see ESI, Fig. S9†).

Common to the three structures is the presence of Ru

A brief survey of the stability of several of these complexes species bound to His15 in an octahedral geometry (Fig. 3), as

1

by H NMR revealed the formation of species containing Ru–H observed for the previously published HEWL CORM-3 complex

·

2

1,26

bonds, which were detected after 24 hours in D

2

O for 3 models.

More Ru ions are also found at the surface of the

(

δ −13.97 ppm), 6a (δ −13.96 ppm), 7 (δ −13.94 ppm), and 8 protein bound to aspartate residues, which in some cases

δ −13.94 ppm). These hydrides are similar to the one already bridge two metal ions (see Table 3). In the obtained structures,

identified before in aqueous solutions of CORM-3 not all the Ru atoms and the ligand atoms are fully occupied

2

6

(

δ −14.5 ppm). The formation of these hydride complexes is due to disorder at the ruthenium binding sites. Nevertheless,

accompanied by the formation of CO

next two sections.

2

as will be shown in the it is clear from the electron density that in the three CORM

·

protein complexes obtained, one or two of the initial CO

ligands as well as the ancillary ligands are absent and have

been replaced by water molecules.

CO and CO release to the headspace in aqueous media

2

The rate of CO release from the compounds to the headspace

In the crystal structures of the 7 HEWL complex, a less

·

of their solutions was evaluated in PBS7.4 or H O, at room common electron density is observed in the coordination

2

temperature, in the dark, using gas chromatography (GC) with

a TCD detector (see the Experimental section for details).

Under these conditions none of the tricarbonyl complexes

released CO to the headspace of their solutions (ca. 10 mg of

Table 2 Data collection and reﬁnement statistics for the organometal-

lic·protein complexes formed between HEWL and 6b, 7 and 8 (values in

the compound in 2 mL of medium, see Table 1). Only the use parentheses correspond to the highest resolution shell)

of the very sensitive GC-RCP chromatography enabled the

6

b HEWL

7 HEWL

8 HEWL

·

detection of traces of CO after 4 h in solution. Instead, all com-

plexes with [Ru(CO) ] fragments produced CO . After 24 h the

3

2

X-ray source

ID14-1

(ESRF)

ID14-2

(ESRF)

ID14-2

(ESRF)

amount of CO detected was found to vary with the nature of

2

the ancillary ligands. It was smaller for the phosphine deriva-

tives 9, 10 (≈0.2 equiv.) and increased through the thioether

Crystal data

Space group

P4

a = b = 79.93, a = b = 78.66, a = b = 80.22,

c = 36.89 c = 37.00 c = 37.16

α = β = γ = 90 α = β = γ = 90 α = β = γ = 90

3

2

1

2

P4

3

2

1

2

3 1

P4 2 2

1

1 (≈0.4 equiv.) and CORM-2 (≈0.7 equiv.) to reach values of 1

equiv. and higher for the pyridine 7, 8 and methionine oxide

a, 6b derivatives. Two of the isocyanide complexes 12 and 13

released close to 1 equiv. CO₂after 5 h in a solution of

acetone–PBS (Table 1). The di-carbonyl complex RuCl

Unit cell parameters

(

Å, °)

Molecules per ASU

6

1

0.25

2.03

1

0.35

1.97

1

0.31

2.06

Mosaicity (°)

Matthews coeﬃcient

2

-

3

−1

(

Å Da

)

(CO)

2 2 2 3

(DMSO) (4) and the mono-carbonyl RuCl (CO)(DMSO)

Solvent content (%)

39.50

37.71

40.37

(

5) do not release CO or CO to the headspace of their solu-

2

Data collection

Wavelength (Å)

Resolution range (Å)

2

tions in PBS or H O at room temperature, in the dark, accord-

ing to GC-TCD detection methods.

0.933

30.89–1.67

0.933

27.81–1.78

(1.87–1.78)

0.933

28.36–1.77

(1.86–1.77)

(

1.76–1.67)

Interactions with proteins

<I/σI>

Multiplicity

Number of observed

19.3 (5.1)

7.0 (6.8)

100 389

(14 122)

14 435

(2064)

2.2 (14.8)

99.9 (100)

The interaction between CORM-3 and a number of proteins

has been previously described. These studies include Human reflections

Number of unique

reflections

Serum Albumin (HSA), human Transferrin (h-Tf), hemoglobin

(

Hb), myoglobin (Mb) and Hen Egg White Lysozyme

Rpim (%)

2

6,27

HEWL).

Using single crystal X-ray diﬀraction, the 3D

Completeness (%)

crystal structures of the model protein HEWL bound to

CORM-3 and [Ru(CO) Cl (thiazole)] have been determined.

3 2

Refinement statistics

Resolution range (Å)

2

1,26

30.89–1.67

16.57

21.15

27.81–1.78

16.13

19.52

0.010

1.29

28.36–1.77

18.30

22.20

0.019

1.81

In the present study we applied the same methodology to com-

plexes 6b, 7 and 8 in order to understand the mode of inter-

action of these newly synthesized CORMs with HEWL. Other

complexes did not produce useful X-ray diﬀraction data.

Crystals of HEWL previously soaked with complexes 6b, 7

and 8 were measured and complete data sets were collected at

high resolution (see Table 2).

R

work (%)

free (%)

RMSD bond length (Å) 0.013

RMSD bond angle (°)

Ramachandran plot (%)

Residues in favored

1.53

96.06

3.94

0

95.28

4.72

0

regions

Residues in additionally 3.94

allowed regions

Residues in disallowed

regions

0

The structures were solved by MR using the native protein

4

as the reference model (PDB code 193L ) and refined with all

residues in the allowed regions of the Ramachandran plot (see PDB code

4UWN

4UWU

4UWV

Dalton Trans.

This journal is © The Royal Society of Chemistry 2014

View Article Online

Dalton Transactions

Paper

Fig. 3 Structural representation of the CORM HEWL complexes at the site with the highest Ru occupation, obtained by soaking HEWL crystals with

·

6

Å

b, 7 and 8. Representation of the experimentally observed electronic density for each protein complex at the His15 site [map contoured at 0.3927 e

−3

(rmsd: 1), 0.4562 e Å (rmsd: 1) and 0.4003 e Å (rmsd: 1) for complexes 6b, 7 and 8, respectively]. The three CORM HEWL complexes present

an octahedral geometry at the Ru atom: in addition to the histidine residue (His15), the metal coordination sphere is completed with CO, COOH and

O molecules depending on the compound the HEWL crystals were treated with.

·

H

2

sphere of Ru at the His15 binding site. This electron density This equilibrium can only be pushed to the left at very low pH

4

3

has been interpreted as a metallacarboxylate with a covalent values and temperatures ≈0 °C as described for CORM-3 and

2

+ 42,45

Ru–C bond. Thus, the interaction of the tricarbonyl complex 7 for the extremely acidic fac-[Ru(CO)

3

(H

2

O)

3

] .

When the

with HEWL crystals led to the formation of a [Ru(CO)(COOH)- pH rises above 3–4, HO⁻addition pushes the equilibrium to

+

(H

2

O)

3

]

species, which is coordinated to the protein by the the right and the metallacarboxylate species starts to appear.

imidazole group of His15. Such a species is formed through Inspection of the FTIR spectrum in the CO stretching region

−

the addition of HO to the dicationic fragment [Ru- and the sudden pH variation registered upon dissolution in

+

water indicate that the addition of HO⁻to the Ru(CO)

(CO)

2

(H

2

O)

3

] .

3

frag-

ment of these compounds is highly favorable and fast. Accord-

ingly, DFT modeling of the reaction of fac-[RuCl (CO) (MeIm)]

DFT studies on structure and reactivity

2

3

(

MeIm = methylimidazole) with HO⁻in the presence of H

2

O

The results described in the preceding sections, taken together

with other studies previously reported,

suggest that the chemistry of the [Ru(CO)

aqueous solution was initiated by the first step of the water–

gas-shift reaction sequence, depicted in eqn (2). The initial tri-

one molecule explicitly considered) revealed an extremely low

2

1,26,42,43,45

strongly

‡

−1

kinetic barrier (ΔG = 0.4 kcal mol ) and a rather large nega-

2

+

3

]

complexes in

−1

tive free energy balance for the reaction (−30.8 kcal mol ) as

depicted in Fig. 4. The mechanism occurs in a single step,

_w_i_t_h_H_O− attacking the C-atom of one CO ligand and yielding

−

carbonyl complexes add HO in a pH dependent equilibrium.

_a_C_O_O_H− ligand bonded to the metal in the final product.

Other computational studies of this reaction for several metal

Table 3 Metal binding sites found in the crystal structures of HEWL carbonyl complexes also confirm this addition to be essentially

4

6–52

soaked with 6b, 7 and 8, describing the ligands at each Ru binding site. barrierless.

Values in parentheses correspond to the occupancy of the ruthenium

_T_h_i_s_k_i_n_e_t_i_c_p_r_o_f_i_l_e_h_i_n_t_s_t_h_a_t_H_O− addition must be much

faster than halide displacement. This is confirmed in Fig. 5

where the putative substitution of chloride by hydroxide in the

CORM·HEWL

adduct ligands

6b·HEWL

7·HEWL

8·HEWL

2 3

complex [RuCl (CO) (MeIm)] is modeled using DFT calcu-

His15 site

Ru (1)

CO

Ru (0.8)

CO

Ru (0.7)

CO

4 H O

2

lations. The reaction occurs in two consecutive steps. First,

−

4

2

H O

COOH⁻

there is release of one Cl ligand, from C to D, indicating a dis-

−

3

2

H O

sociative mechanism. In the second step, from D to E, OH

Asp18 site

Asp52 site

Asp101 site

Asp119 site

Ru (0.65)

CO

Ru–Ru (0.5)

8 H

binds the metal atom, occupying the coordination position

created by the loss of the chloride ligand. The overall barrier

2

O

3

2

H O

‡

−1

Ru (0.5)

Ru (0.4)

CO

obtained (ΔG = 19.3 kcal mol , with respect to A) indicates

3

2

H O

that this substitution reaction is considerably less favorable

3

2

H O

−

than HO attack on coordinated CO.

Ru–Ru (0.7)

Asp 101

Ru–Ru (0.65)

5 H

2

O

The results above strongly suggest that aqueous chemistry

6

2

H O

of Ru(CO) Cl L complexes eﬀectively takes place from [Ru-

3

2

Ru (0.5)

Ru (0.7)

CO

−

(

2 2

COOH)Cl (CO) L] and species of this kind are the ones that

3

2

H O

4

2

H O

actually interact with the proteins and other biological entities

This journal is © The Royal Society of Chemistry 2014

Dalton Trans.

View Article Online

Paper

Dalton Transactions

Fig. 4 DFT calculated free energy proﬁle (kcal mol⁻) for the water

1

−

assisted attack of HO on the model complex fac-[Ru(CO)

3

Cl

2

(MeIm)].

Bond lengths in Å.

−1

Fig. 6 DFT calculated relative energy diﬀerences (kcal mol ) between

2+

isomers of [Ru(CO)(H

2 4 2 2 3

O) (MeIm)] (I, J) and [Ru(CO) (H O) (MeIm)]

(

F, G, J) where MeIm models the His15 ligand. The more stable isomers F

and I have the same stereochemistry as that observed in similar CORM·-

HEWL complexes in Fig. 3.

MeIm at adjacent positions (cis), exactly as found in the cases

of the organometallic·protein complexes obtained in the incu-

Fig. 5 DFT calculated free energy proﬁle (kcal mol⁻) for the water

1

assisted substitution of Cl⁻by HO⁻on the model complex fac-[Ru-

bation of HEWL with [Ru(CO)

3 2

Cl (thiazole)], 6b and 8.

(

3 2

CO) Cl (MeIm)]. Bond lengths in Å.

2+

In contrast to what happens with [Ru(CO) (H O) ] , litera-

3

2

3

ture reports on the preparation and water exchange reactions

2

+ 45

+ 45,53

soon after incubation starts. Nevertheless, from the data gath- of [Ru(CO)

2

(H

2

O)

4

]

and [Ru(CO)(H

2

O)

5

]

do not

ered in this work, it seems that halides and other ligands L are describe nucleophilic additions of water to CO in aqueous

+

easily displaced because they are absent in many of the peaks media. However, the fac-[His-Ru(COOH)(CO)(H

O) ] observed

3

2

studied by LC-MS and are absent in all the organometallic· in the crystal structure of the HEWL crystals incubated with 7

−

protein complexes identified using crystallography.

corresponds to the structure expected for the addition of HO

2

+

In the presence of proteins, the donor residues may substi- to the [His15-Ru(CO)

(H

O)

3

]

complex. Loss of CO

from this

] , the more often

2

−

2+

tute ligands from [Ru(COOH)Cl (CO) L] or may simply bind adduct leads to [His15-Ru(CO)(H

2

O)

4

2

Ru fragments that are generated during its decay in solution observed mono-carbonyl motif at His15 and other binding

without the participation of the protein. So far, and regarding sites of the CORM HEWL complexes (see Table 3). It is impor-

only CO containing species, the data gathered from the inter- tant to take into account that the pH at which the soaking

·

2

6

21

3 2

action of HEWL with CORM-3, [Ru(CO) Cl

(thiazole)] and experiments took place (ca. 4.5) is much higher than the pH at

2

+

the above complexes 6b, 7 and 8 led to the identification of the which [Ru(CO)

(H

2

O)

]

and [Ru(CO)(H O) ] were prepared

2 5

2

4

following [His15-Ru(CO) (H O) L] coordination motifs: [His15- and studied.

x

2

y

Ru(CO)

CO)(H

modeled with the methylimidazole ligand (MeIm) instead of The cytotoxicity of the CORMs was measured using the con-

the histidine yielded the results depicted graphically in Fig. 6.

centration that causes an inhibition of 50% in the survival of

According to these calculations, it is clear that the isomers cells upon incubation (IC ). The results reported in Table 4

2

(H

2

O)

3 2 4

], [His15-Ru(CO)(H O) ] and [His15-Ru(COOH)-

Cytotoxicity studies

(

2

O)

3

]. DFT calculations on similar Ru(II) complexes

5

0

identified in the protein crystal structures correspond to the show that all the compounds investigated in this study were

thermodynamically most stable ones. The fac-[Ru(CO) (MeIm)- not toxic to RAW264.7 cells up to 100 µM. The isocyanide com-

isomer (F) is considerably more stable than the other plexes were not tested due to their rapid decomposition in the

cis-Ru(CO) (G) and trans-Ru(CO) (H) alternative isomers and presence of water.

2

+

2 3

(H O) ]

2

corresponds to the isomer of the adduct [His15-Ru-

CO) (H O) ] present in the structure obtained from incu- in this study were able to decrease the NO production by LPS-

Anti-inflammatory evaluation in vitro. All the CORMs tested

(

2

3

2

6

bation of HEWL and CORM-3. Also, the more stable isomer stimulated RAW264.7 cells in a dose dependent manner

of the complex [Ru(CO)(MeIm)(H

2

+

2

O)

4

]

(I) places CO and (Table 4; Fig. S14 in ESI†). The most active compound in redu-

Dalton Trans.

This journal is © The Royal Society of Chemistry 2014

View Article Online

Dalton Transactions

Paper

Table 4 Cytotoxicity of CORMs in RAW264.7 (MTT assay; 24 h incu-

bation; IC50) and their eﬀect at 100 µM on the inhibition of NO pro-

duction (% control) in LPS-induced RAW264.7 cells. Hemolytic index (%)

−1

at [CORM] = 1 mg mL

CORM

IC50 (µM)

Nitrite (% control)

Hemolysis index (%)

CORM-3

CORM-2

3

>100

58

83

42

52

50

40

11

20

75

<10

70 (0.5 mg mL⁻

1

)

70 (0.5 mg mL⁻

10

50

<10

1

Fig. 7 CO levels in blood and tissues of CD-1 female mice treated with

CORM-3, 50 mg kg⁻, iv (black bars), measured using GC-RCP. Control

values for untreated, healthy CD-1 female mice are shown (grey bars).

The GC-RCP results are expressed in pmol of CO per fresh weighed

tissue (FWT) as means for two independent measurements. Animals

were sacriﬁced 10 minutes after injection.

1

6

7

8

9

1

a

0

1

of the treated animals is only slightly more than that registered

cing NO production was the thioether derivative 11, which was in the control animals. This is in agreement with the repeated

able to reduce nitrite levels in culture by 75% relative to observation that COHb levels in systemic circulation are barely

1

0,55

control cells. This value is slightly lower than that found for aﬀected by administration of CORM-3.

CORM-2. However, it is important to note that CORM-2 has data show that there is no apparent organ specificity in the

two times as many [Ru(CO)

] fragments as all other shown bio-distribution of CORM-3. This is not unexpected given its

Altogether, these

3

CORMs at the same molar concentration. In accord, the NO simple and quite labile chemical structure. Of course, the

inhibitory activity of 3 (42%), which is expected to be the modification of ancillary ligands can change this bio-distri-

major species in a fresh solution of CORM-2 in DMSO, has bution pattern giving it a higher bias towards specific target

half the inhibitory activity registered for CORM-2 in these tests organs, as demonstrated in the case of the analogue

(

83%). The other sulfoxide complex, 6a, as well as the pyridine [Ru(CO) Cl (methylthiogalactoside)] (ALF492) which shows a

3 2

derivative 7 and CORM-3, are all stronger inhibitors of NO higher liver specificity albeit measured by the concentration of

18

than 3. The phosphine derivatives 9 and 10 were the least Ru instead of that of total CO. The bio-distribution in Fig. 7

active compounds in this group.

diﬀers considerably from that obtained with CO inhalation

5

4

Blood hemolysis. Before entering any in vivo studies, CORM and reported by Vreman and coworkers. Taking as 100% the

candidates must be tested with regard to their hemolytic blood concentration of CO generated by inhalation, Vreman

activity. The results of this test for some of the CORMs prepared and coworkers found that the CO concentrations in the other

are also given in Table 4. These show that the majority of the organs were ca. 1 or 2 orders of magnitude below the blood

CORMs under study are much less hemolytic than CORM-2 or concentration (BC): lung (9.4% of BC), spleen (8.6% of BC),

−

1

-

3 which induce hemagglutination at 1 mg mL and are hemo- kidney (4.5% of BC), liver (4.3 of BC), heart (3.8% of BC) and

−1

lytic at 0.25 mg mL . In fact, only 7 and 8 showed a hemolytic brain (0.7% of BC). In the case of CO delivered by CORM-3

index of 50% while the remaining tested CORMs present 10% (Fig. 7), the accumulation in the other organs relative to blood

−1

or less hemolytic activity at 1 mg mL concentration.

is higher: liver (15.6% of BC), kidney (12.8% of BC), spleen

CO bio-distribution after administration of CORM-3 (10.3% of BC), heart (8.3% of BC), brain (5.0% of BC) and lung

in vivo. A successful CORM should be able to specifically (4.7% of BC). The stimulation of endogenous CO generation

deliver CO to tissues or organs under pathophysiological con- via heme oxygenase induction also gives the blood as the main

ditions. CO bio-distribution studies in mice after inhalation of CO loaded organ, followed by the heart and the spleen in an

5

4

air with 500 ppm CO for 30 min showed an increase in CO unspecific fashion.

concentration in the kidney, heart, spleen, liver, lung and

These results indicate that CO delivery in vivo, via CORMs,

brain. Given the wide use and promising results of CORM-3 originates a bio-distribution of CO which is clearly diﬀerent

5

4

1

4

in a variety of pre-clinical studies, we decided to evaluate the from those obtained by endogenous CO production or CO

ability of CORM-3 to deliver CO specifically to organs and inhalation. Considering that CORMs can be equipped with

tissues and compare it with CO inhalation. The results are organ targeting features, it is obvious that they have a clear

graphically expressed in Fig. 7 where the control is the concen- advantage over other methods of triggering CO protection.

tration of endogenous CO present in untreated, healthy mice. This has been argued before, but we believe that it is shown

Although our measurements do not distinguish between free, here quantitatively for the first time.

Hb-bound or Ru-bound CO, we can say that CO accumulates

mainly in the liver, kidney and spleen. In fact, the accumu-

lation in these organs increased by a factor of ca. 3, while that

in the heart increases by a factor of 2 and the lungs and brain

Conclusions

remain essentially at the base values. The accumulation in The undisputed activity of several Ru(CO) complexes as pro-

3

blood is not very important since the value of CO in the blood drugs for the delivery of CO to biological systems in vitro and

This journal is © The Royal Society of Chemistry 2014

Dalton Trans.

View Article Online

Paper

Dalton Transactions

in vivo contrasts with their lack of acceptable pharmacological barrier, in agreement with the fact that aqueous solutions of

II

z±

properties and the limited understanding of their chemistry [Ru (CO) L ] compounds acquire a pH ≤ 3 depending on the

3

and mode of action. The present work represents an attempt ancillary ligands. In other words, as pointed out by Mann and

4

3

−

to overcome this situation and contribute to the design of coworkers, these complexes are able to react with HO at

concentrations ≤10⁻

11

M. Therefore, once dissolved in

useful drug-like CORMs.

Acceptable solubility and stability in aqueous or biological aqueous medium, the chemistry of the compounds [Ru -

media are forefront conditions to be met by a drug and (CO) Cl L] is governed by the reactivity of the metallacarboxy-

CORMs are no exceptions. Such stability will enable the study late species [Ru (CO) (COOH)Cl L] . Under biological con-

II

3

2

II

−

2

of the pharmacokinetic behavior of the compound and ulti- ditions, with pH ≫ 3, the equilibrium of eqn (2) is strongly

mately its biological mode of action. However, as a prodrug, a driven to the right and the Ru(CO) fragment is completely

CORM must also inscribe the chemistry that triggers its absent in solution as shown by FTIR in the cases of CORM-3

3

2

1

3 2

decomposition at the target organ, tissue, cell or biomolecule. and Ru(CO) Cl (thiazole) solutions. It is under these con-

This apparent contradiction can be solved by means of an ditions that it becomes important to ascertain the stability of

appropriate combination of the inner and the outer coordi- the metallacarboxylate intermediate. If it is stable for a time

2

nation spheres of the organometallic CORM candidate. For a span compatible with the time scale of the biological exper-

given organometallic core, the design pathway starts by estab- iment, it is possible that it behaves as a stable prodrug with a

lishing the inner coordination sphere that provides appropri- recognizable and quantifiable pharmacokinetic profile.

ate stability and reactivity control, as exemplified in the According to the HPLC tests performed with some selected

development of the liver rescuing [Mo(CO)

3

(CNCMe

2

COOH)

3

]

2

complexes in MeOH–H O the stability of the initially formed

3

2

−

(ALF794). The ADME and targeting profiling is carried out at [Ru(CO) (COOH)Cl L] complexes is very limited for most

2

a later stage through manipulation of the outer coordination complexes studied, as exemplified in Fig. 2 and in the ESI,

sphere.

Fig. S4–S9.† Complex 3 or its HO⁻adduct are tentatively

With these provisos in mind a series of complexes of the assigned to a peak showing itself in the HPLC trace obtained

II

3 2 2

general formula fac-[Ru (CO) Cl L] was prepared with a broad immediately after dissolution in MeOH–H O (RT = 8.8 min),

range of donor functions L, namely O-sulfoxides, S-sulfoxides, but we could not obtain definitive proof of that using LC-MS

pyridines, phosphines, thioethers and isocyanides (Fig. 1 and with ESI detection. The closest peak identified that can be

Table 1). The complexes are readily synthesized and character- related to the metallacarboxylate derivative of

ized by standard procedures of organometallic chemistry. In [Ru(CO) (CO

2 2

3

is

)(DMSO)Cl] which appears in the highly water-

spite of the diﬀerent electronic abilities of the ligands used, soluble, first-eluting fraction of the chromatogram.

the values of the νCtO vibrations of the complexes (Table 1) do The case of CORM-2 is particularly interesting and worth

+

not vary widely except for the isocyanide (CNR) congeners, discussing due to its extensive use in the literature. Our results

which are clearly more elevated. As the electronic properties of show that according to the method of preparation of CORM-2

the Ru(CO) fragment remain essentially constant for most solutions for biological administration, 3 is the most abundant

3

ligands the rest of the coordination sphere must accommodate species in the DMSO stock solution of the dimer. Depending

the diﬀerent electronic constraints imposed by the L ligands. on the age of this stock solution, variable amounts of 4 are

The CNR ligands are strong enough σ donors and π acceptors also present. When the solution is diluted in aqueous buﬀer at

to change this situation and reduce back-donation to the CO pH 7.4 for biological administration, both these species start

ligands.

immediate decomposition, making it impossible to assign the

Introducing suﬃcient water solubility by means of polar nature of the active biological species, the PK of the drug and

functionalization of the ligands was successful as can be seen the overall mechanism of action. The lipophilicity of both 3

in Table 1. The PTA complex 9 is a surprising exception. This and 4 may account for non-specific hydrophobic interactions

3

7,41

ligand is considered a water-solubilizing phosphine

but in observed with CORM-2, which operate beyond its CO delivery

1

6

this regard it is much less eﬃcient than its DAPTA derivative activity. Interestingly, the methionine oxide derivative 6a or

in complex 10. The isocyanoacetate ester ligands were not 6b showed a much higher stability and the parent molecule

expected to improve solubility per se, and indeed failed to do and its HO⁻adduct were found to be stable for 1 h along

2

so. This lack of solubility was expected to be corrected at later several consecutive HPLC runs in the same MeOH–H O gradi-

+

stages by the use of their anionic forms CNCR COOM (M = ent as that used for 3. The parent ion [RuCl (CO) (SO(CH )

Li , Na ).

2

3

+

+• 56,57

−

2 2 2 2

However, the instability of the CNR complexes CH CH CH(NH )CO )] was identified using LC-MS in nega-

in aqueous media discouraged such experiments.

The stability of the complexes in aqueous solution is rather possible to generate an inner coordination sphere that stabil-

more diﬃcult to control than solubility. In fact, the extremely izes the Ru(CO)

fragment or its HO⁻adduct for a period of

tive ion mode. This important result demonstrates that it is

3

−

facile addition of HO to one of the CO ligands of the time compatible with biological testing. However, this is not

Ru (CO) fragment is certainly the main cause of instability yet easily predictable. The proverbial stability imparted by

II

3

and an intrinsic property of this scaﬀold (eqn (2)). As pointed phosphine ligands to metal carbonyl complexes might have

out by DFT calculations made on the model complex [Ru- suggested improved properties for the PTA and DAPTA deriva-

(

CO) Cl (MeIm)], this addition has virtually no activation tives, which was not the case. Isocyanide derivatives became

3 2

Dalton Trans.

This journal is © The Royal Society of Chemistry 2014

View Article Online

Dalton Transactions

Paper

When this complicated chemistry takes place under biologi-

cal conditions, two situations may happen: (i) it takes place

independently of the biological species present, which may

then react with the reaction/decomposition products; (ii) the

presence of the biological molecules modulates the decompo-

sition process.

Since this is crucial information for the study of the CO

delivery mechanism of potential CORMs of the type

II

2+

Scheme 1 Reactivity of [Ru (CO)

3

L

3

]

CORMs in aqueous, aerobic

Ru(CO)

pounds with the model protein Hen Egg White Lysozyme

HEWL).

The results obtained show that Ru(CO)

3 2

Cl L we studied the interaction of several of these com-

solutions. All types of species presented have been identiﬁed for at least

one such CORM.

(

x

fragments bind

intolerant to water due to excessive activation of CO. Surpris- several residues of the HEWL protein (Table 3) but the adduct

ingly, sulfoxide and thioether ligands, which are considered with His15 residue is the one that systematically presents the

very labile in classical organometallic chemistry, provided the highest occupancy. As depicted in Fig. 3 both complexes 6b

II

best examples, eventually as

functionalization.

a

result of their extra and 8 generated the fragment Ru (CO)(H O) bound to His15

2

4

2

1

as found in similar experiments with Ru(CO) Cl (thiazole).

3

2

Of course, when the metallacarboxylate species ages in The chloride ligands are absent in the final adduct, attesting

solution it decomposes irreversibly, generating CO , metal their lability. The ultimate fate of the initial Ru(CO) Cl L

2

3

2

−

hydrides, hydroxyl radicals and other species that have been complex is the loss of two CO ligands, the two Cl ligands and

identified in this and previous studies. They can all be accom- the ancillary ligand L: pyridine (8), methionine oxide (6b) or

modated in the chemistry of Scheme 1.

thiazole.

We have previously identified a cis-Ru(CO) (H O) fragment

In a typical water–gas-shift reaction (WGSR) pathway, the

2

3

metallacarboxylate is initially formed, forming a cis-Ru(CO)

2

bound to the His15 residue in similar experiments with

moiety which is identified using FTIR spectroscopy and lower- CORM-3. In that case one CO and all other ligands were lost

ing the pH of the solution, as observed experimentally. This following the first addition of HO⁻to the Ru(CO) fragment.

3

2 2

evolves to produce CO and Ru–H complexes. Both CO and Quite interestingly, the soaking of HEWL with 7 led to the

Ru–hydrides were detected chromatographically and spectro- identification of the metallacarboxylate intermediate [His15-

II

−

scopically, respectively, confirming this reactivity pattern. Ru (COOH)(CO)(H O) ] derived from the addition of HO to

2

3

2

+

According to the example reported by Fachinetti and co- the dicarbonyl dication cis-[Ru(CO) (H O) ] . Crystallographic

workers, the complex [RuH(CO) (H O) ] , which can be taken resolution does not allow distinguishing the metallacarboxy-

as a model for the Ru–hydrides derived from the complexes late from the corresponding Ru(η -CO ) species, but at pH 4.5

Cl L], can react with bases to lose H and lead to the we favor the presence of the Ru–COOH fragment. Indeed, the

3 2

formal reduction of Ru(II) → Ru(0). Such very electron rich stability and fate of this metallacarboxylate is pH dependent:

Ru (CO) L ] complexes will react with O to reform Ru(II) higher pH will drive it to CO and other Ru species, and lower

2

3

+

2

3

1

2

+

[Ru(CO)

4

2

0

[

2 3 2

2

•

58,59

species and produce (HO ) radicals,

which have also been pH will revert it to the parent dicarbonyl complex. It is now

observed experimentally using ESR spectroscopy in CORM-2 obvious that the use of HEWL has allowed us to capture a

1

7,60

and CORM-3 aqueous aerobic solutions.

can also expect that such hydrides [RuH(CO)

with protons to produce H . This was indeed detected in the very clear. The loss of the last CO has not been well documen-

Of course, one series of snapshots along the pathway of the decomposition of

+

2

2 3

3 2

(H O) ] react [Ru(CO) Cl L] complexes in aqueous medium, which is now

2

6

case of CORM-3 aqueous solutions but not quantified. More- ted here beyond some Ru ions bound to surface aspartates,

over, many Ru(CO) complexes, including CORM-2, are WGSR but we believe that it can be easily modeled from the analogue

3

6

1

2+

catalysts. The vibrational signature of the cis-Ru(CO) species [Ru(CO)(H O) ] . The wavenumber of the CO vibration in this

2

5

that are formed by both modes of decomposition of the Ru–H complex (1971 cm⁻) suggests that it has a much lower electro-

1

2

+

intermediates is present in the FTIR spectrum of iCORM-3, the philicity at its Ru bound C atom than the [Ru(CO) (H O) ]

2

4

biologically inactive residue left after ageing CORM-3 solutions (2089, 2023 cm⁻) and [Ru(CO) (H O) ] (2170, 2084 cm⁻¹)

1

2+

3

2

3

in aqueous buﬀer at pH 7.4 for 24 h in air. The fact that such analogues, and the complex most likely decomposes by CO dis-

iCORM solutions do not release CO may be due to either their sociation, which is likely to happen through the oxidation of

strong stability or to the involvement of another addition of Ru(II) to Ru(III).

−

HO to those Ru(CO) species. Such addition was not reported

The structures of all these Ru(CO) ·HEWL complexes at the

2

+

x

2

for [Ru(CO)

2

(H

2

O)

4

]

since it was prepared and characterized His15 binding site (Fig. 3) are those thermodynamically pre-

in acidic medium, but has been well documented in this work dicted by DFT calculations. It is not possible to ascertain the

2

+

(

see Fig. 2) and with the dication cis-[Ru(bpy) (CO) ] . This role (if any) of the protein in the decomposition of the

2 2

−

+

reacts with HO to form first [Ru(bpy)

2

3 2 x

(CO)(COOH)] and then Ru(CO) Cl L complexes and formation of the Ru(CO) ·HEWL

[

Ru(bpy) (CO)(CO )] at higher pH, and behaves as a WGSR species. However, the reaction of CORM-3 with HEWL moni-

2

6

2–64

catalyst producing CO and H from CO and H O.

tored using ESI-MS suggests that the protein actually acceler-

2

This journal is © The Royal Society of Chemistry 2014

Dalton Trans.

View Article Online

Paper

Dalton Transactions

erly targeted. Tissue targeting was not addressed in the

II

present study but its eﬃcacy has been demonstrated for [Ru -

1

8

(

CO)

CNCMe

3

Cl

2

(methylthiogalactoside)] (ALF492)

and [Mo(CO)

3

-

3

2

COOH) ] (ALF794). Of course, as pointed out by one

3

of the referees, the very rich and complex chemistry of these

II

Ru (CO)

3

based CORMs, summarized in Scheme 1, may also

II

Scheme 2 The proposed mechanism for the interaction of [Ru - contribute to some of their biological eﬀects. The formation of

(

3 2

CO) Cl L] with cells or biomolecules.

ROS species downstream of the water–gas shift reaction may

actually be relevant in some cases, since these species are

strongly involved in signaling processes, including those invol-

ates the first loss of CO and other ligands leading to the exclu- ving CO. Some of these possibilities have been raised and dis-

1

6,17,60

sive formation of a Ru(CO) complex, Ru(CO) ·HEWL, within a cussed elsewhere,

few minutes after incubation.

but the present data do not provide

2

6

any evidence favoring or disfavoring them.

A relevant and often asked question still remains to be

The complexes described in this study present an accepta-

addressed: if these molecules are so labile and do not release bly low cytotoxicity profile and the majority is not hemolytic.

CO gas to their biological environment, how do they deliver The classical test for anti-inflammatory activity, the NO inhi-

CO to cells and elicit CO dependent biological responses?

We believe that the chemistry described in Schemes 1 and 2 ordinary activity, but reveals some dependence on the nature

provides an easy general answer to this question.

of the ligand L. The thioether derivative 11 markedly inhibits

If the metallacarboxylate lasts long enough in the biological the

NO production of the LPS stimulated

bition of LPS-induced macrophages, does not reveal any extra-

medium, it gets to cells where it can exchange its added HO⁻RAW264.7 macrophages by 74%. Interestingly, the value found

with any other ROH (transesterification) or RNH₂(transami- for 3 is almost exactly one-half that found for CORM-2, in

doylation) terminal group present at the surface of the cells, agreement with the latter forming 2 equivalents of 3 upon dis-

thus anchoring the Ru complexes as adducts of the formula solution in DMSO.

[

(cell-Y-CO)Ru(CO) L ] and [(cell-Y)Ru(CO) (H O) ] (Y = O, NH)

Taken together, the data presented in this work contribute

to the understanding of the chemistry and mode of action of

2

3

2

3

2

at a later stage following CO loss.

In this way, these CORMs can accumulate eﬃciently at the diﬀerent Ru(CO) Cl L complexes regarding stability, CO and

3

2

surface of cells and actually move from cell to cell by means of CO₂releasing profile and interaction with proteins. A set of

reversible exchanges with carrier molecules in the intercellular diﬀerent ligands with diﬀerent chemical properties has been

medium. This anchoring of the CORM to the cell surface chosen for this systematic characterization. Reasonably stable

enables decomposition or internalization of the Ru scaﬀold to and soluble complexes were obtained. Low cytotoxicity and

take place at or inside the cell. In both cases CO will be readily considerable anti-inflammatory eﬀects are found across all the

available to the cell and its internal targets, e.g. mitochondria. complexes, showing that Ru(CO)₃based CORMs may have a

In the case of internalization, the amount of CO delivered to future role in CO therapy.

the cell can be much higher than that usually achieved by dis-

solved free CO, which rapidly diﬀuses across membranes. This

mechanism is compatible with the observations made with the

CO sensitive fluorescent probe COP-1, showing that CORM-3

Experimental section

General procedures

accumulates at the membrane and CO slowly builds up in the

2

9

cytoplasm.

All work involving animals performed in the Lisbon labora-

tories of Alfama Ltd was done according to the guidelines of

When these CORMs are administered in vivo, a Ru(CO)

2

·

albumin complex is likely responsible for the distribution and the Portuguese animal protection law and the derived guide-

2

6

delivery of CO to other cells in the organism. CORM-3 has lines on the ethical use of animals.

produced in vivo curative results in diseases aﬀecting quite

diﬀerent organs, such as joints, heart, arteries and others.

Gas chromatography with TCD detection

This lack of selectivity is compatible with the Ru(CO) ·albumin The CO release assays were performed in a 7.0 mL Roth®

2

based distribution mechanism. The CO bio-distribution study sample vial equipped with a magnetic stirrer inside and

in Fig. 7 reveals that all the organs studied have increased capped with a PTFE rubber or silicone septa and an aluminum

quantities of CO relative to the controls, with a higher inci- cap. PTFE rubber septa were acquired from Sigma Aldrich®

dence in liver and kidney. However, one cannot really talk and a silicone septa from Roth®. The assays were performed

about a specific accumulation in a given organ and CO from in PBS7.4 or H O, without light, at room temperature and

2

CORM-3 is smeared all over the organism. Importantly, this normal atmospheric air. For compounds 12, 13 and 14 stock

bio-distribution profile is diﬀerent from that obtained by solutions were prepared in acetone, and then aliquots were

Vreman upon stimulation of endogenous heme oxygenase added to PBS7.4, obtaining final solutions of 2 mL with con-

5

4

(

HO). This signals the therapeutic advantage of CORMs over centrations of 25 mM for each compound. These were then

HO since CORMs can direct CO to the organ of interest if prop- placed in the Roth® vials as described above. 250 μL samples

Dalton Trans.

This journal is © The Royal Society of Chemistry 2014

View Article Online

Dalton Transactions

Paper

of the headspace gas were taken with a gas tight Hamilton® to space group P4

3

2

1

2 (see Table 2 for data collection

syringe and were injected into a Thermofinnigan Trace GC statistics).

equipped with a CTR1 column from Alltech™ and a Thermal

Structure determination was accomplished by molecular

Conductivity Detector. The column was in an oven at 36 °C replacement using as a search model the structure with PDB

and the GC was operated at a constant pressure mode code 193L, and several cycles of restrained refinement in

6

5

66

(

3

111 kPa) with He as a carrier gas and a reference gas with a Refmac5 followed by manual model building in Coot

0 mL min⁻flow rate. The detector was set at constant temp- enabled us to produce the final models with a good geometry;

erature (150 °C) and the filament temperature at 250 °C. The PDB_REDO was used for the validation of the final models

injections were made through a packed column injector (PKD) (see Table 2 for refinement statistics).

1

6

7

set at 47 °C and 111 kPa. CO was quantified using a calibration

The coordinates and structure factor have been deposited

curve recorded prior to the reaction course. This was done by in PDB with accession numbers 4UWN, 4UWU and 4UWV (6b,

injecting 250 μL increments of CO up to a final total amount 7 and 8, respectively).

of 2 mL of pure CO gas (carbon monoxide 4.7, purity

DFT calculations

≥99.997%; from Linde) into the system and taking samples

that were injected in the GC.

All calculations were performed using the Gaussian 09 soft-

ware package, and the PBE0 functional, without symmetry

constraints. That functional uses a hybrid generalized gradient

6

8

Gas chromatography with RCP detection

Stock solutions of compounds 12, 13 and 14 were prepared in approximation (GGA), including a 25% mixture of Hartree–

6

9

70

acetone. 1 mL of the solution was added to 1 mL of PBS7.4 Fock exchange with DFT exchange–correlation, given by the

7

1,72

obtaining final solutions of 2 mL with concentrations of Perdew, Burke and Ernzerhof functional (PBE).

The opti-

2

5 μM for each compound. Each sample was prepared in tripli- mized geometries were obtained with the Stuttgart Eﬀective

7

3–75

cate. The final solutions were prepared and immediately Core Potentials and the associated basis set

closed in a 7.0 mL Roth® sample vial equipped with a mag- with an f-polarization function for Ru and a standard 6-31G

netic stirrer inside and capped with a silicone/PTFE septa and (d,p)

augmented

7

6

7

7–80

for the remaining elements. Solvent eﬀects (water)

an aluminum cap. Silicone/PTFE and an aluminum cap were were considered in optimizations of all species involved in the

™

acquired from VWR . The vials were also immediately covered mechanistic studies, using the Polarizable Continuum Model

8

1–84

with aluminum foil to prevent contact with light and the solu- (PCM) initially devised by Tomasi and coworkers

with

tion was stirred at room temperature. Samples (100–200 µL) radii and non-electrostatic terms of the SMD solvation model,

8

5

were taken from the headspace over time (5, 15, 30, 60, 120, developed by Truhlar et al. Transition state optimizations

1

80 and 300 min), with a Gas tight Hamilton® syringe, and were performed with the Synchronous Transit-Guided Quasi-

8

6,87

diluted in another 7.0 mL Roth® vial, capped with a silicone/ Newton Method (STQN) developed by Schlegel et al.,

fol-

PTFE septa and an aluminum cap, and analyzed quantitatively lowing extensive searches of the Potential Energy Surface. Fre-

on a Peak Performer 1 RCP gas chromatograph (GC), which quency calculations were performed to confirm the nature of

allows the CO in the gas to be quantified to concentrations as the stationary points, yielding one imaginary frequency for the

low as 1 to 2 ppb. The reducing compound photometer (RCP) transition states and none for the minima. Each transition

bed and the column were set at constant temperatures, 265 state was further confirmed by following its vibrational mode

and 105 °C, respectively.

downhill on both sides and obtaining the minima presented

The amount of CO was calculated using a calibration curve on the energy profile.

previously obtained in vials of the same volume (7.0 mL),

Cytotoxicity evaluation

using a Linde minican® of CO (30 ppm CO rest in synth. air),

ref. 14960013.

The toxicity of the compounds was evaluated in a murine

macrophage cell line RAW264.7 (ECACC91062702) using the

MTT assay. Briefly, RAW264.7 cells were seeded on 96-well

X-ray crystallography

Hen Egg White Lysozyme (HEWL) has been used as a model plates in DMEM medium (GIBCO, Invitrogen) supplemented

protein for understanding the interaction between CORMs and with 10% FBS (Fetal Bovine Serum; GIBCO, Invitrogen). Cells

proteins.

2

were incubated for 24 hours at 37 °C, 5% CO under a humidi-

Crystals of HEWL have been prepared using NaCl (2–10% fied atmosphere before the addition of the compounds to be

m/v) in acetate buﬀer (pH 4.5) as the precipitating agent. The tested. Compounds were solubilized in water or in 10%

3

crystals appear after 1–2 days with 0.2 × 0.2 × 0.2 mm dimen- DMSO–H

2

O (3, 9, and 11) and added to the macrophage cul-

sions. Soaking procedures have been performed as previously tures at a final concentration of 10, 50 or 100 µM. In the

2

6

described, dissolving the complexes of interest (6b, 7 and 8) control wells, cells were treated with the solvent used to dis-

in a solution containing 12% m/v of NaCl in acetate buﬀer solve the compounds. The final DMSO concentration in cul-

(

pH 4.5) up to a final concentration of 50 mM. After 24 h of tures was never above 0.1%. Cells were incubated for 24 hours

soaking, the crystals were harvested and flash frozen using under the same conditions as those described above. The

0% glycerol as a cryoprotectant. Complete data sets have been culture medium was replaced by a 1 mg mL⁻of MTT solution

1

3

collected using synchrotron radiation and the crystals belong prepared in DMEM-FBS supplemented medium and the cul-

Dalton Trans.

View Article Online

Paper

Dalton Transactions

2

tures were incubated for 1 hour at 37 °C, 5% CO . The super- In vivo bio-distribution of CO

natants were discarded and the formazan crystals produced

were dissolved in DMSO. The plates were incubated with

gentle shaking for 10 minutes and the absorbance of the

medium was read at 550 nm. The absorbance obtained in the

control wells was considered as 100% survival.

CD-1 female mice from Charles-River (6–8 weeks) were treated

with 50 mg kg of CORM-3 (i.p.) in saline. After the respective

times (10, 30 or 60 min) animals were sacrificed and perfused

with 15 mL of PBS. Blood and livers were collected and diluted

−1

50 times or 5 times, respectively, in ice-cold Milli-Q water (liver

tissue was homogenized using a Tissue Tearor™). For CO

quantification, the protocol described by Vreman et al. was fol-

lowed. Briefly, CO was liberated as a gas in a closed vial by

adding 25 µL of water and 5 µL of sulfosalicylic acid (SSA, 30%

LPS stimulation of RAW264.7 cells and nitrite quantification

5

4

Murine macrophages RAW264.7 were seeded into the wells of

a 24-well plate in DMEM medium supplemented with 10%

[wt/vol]) to 30 µL of diluted samples. The vials were incubated

FBS. Cells were incubated for 24 hours at 37 °C, 5% CO under

2

on ice for at least 10 min before being analyzed. The gas in the

headspace of the vials was analyzed quantitatively with a gas

chromatograph (GC) equipped with a reducing-compound

photometry detector (RCP detector) (Peak Laboratories, Moun-

tain View, CA), which allows to quantify CO in the gas at con-

centrations as low as 1–2 parts per billion (ppb). The amount

of CO was calculated using a calibration curve prepared from

CO standards. Control animals were injected with a vehicle

and handled in the same way.

a humidified atmosphere. Compounds were added to the cul-

tures at a final concentration of 10, 50 and 100 µM. At the

same time lipopolysaccharide (LPS; L2880, Sigma) was added

to the cultures at a concentration of 1 µg mL . After 24 hours

the NO production was determined by quantifying the nitrite

in culture supernatants using the Griess reagent (Sigma). In a

−

1

9

6-well plate, 100 µL of the Griess reagent was mixed with the

same volume of the culture supernatant and allowed to react

for 10 minutes. The absorbance was read at 550 nm. A sodium

nitrite standard reference curve was prepared for each assay

for the accurate quantification of nitrite levels in experimental

samples.

Chemical synthesis

All preparations were performed under an atmosphere of nitro-

gen using standard Schlenk techniques. Solvents were dried by

standard procedures, distilled under nitrogen and kept over

Hemolysis index determination

4

Å molecular sieves, except for DMSO that was used as

Red blood cells (RBC) obtained upon centrifugation of sheep

whole blood (in Alsever’s solution; Innovative Research cat. no.

IR1-020N) were used to evaluate the potential of the Ru-com- 13

received (p.a. from Panreac). Microanalyses for CHN were per-

formed at the ITQB, Oeiras, Portugal (by C. Almeida). H and

1

C NMR spectra were recorded on a Bruker Avance III

00 MHz. Chemical shifts are quoted in parts per million. FT-

infrared spectra (KBr pellets) were recorded in a Unicam

Mattson 7000 FTIR spectrophotometer. ATR-FTIR (Attenuated

Total Reflection Fourier Transform Infrared) spectra in solu-

tion (MeOH and PBS) were recorded in a Bruker IFS66/S

plexes to induce RBC hemolysis. A 2% RBC suspension in PBS

4

(100 μL) was distributed in the wells of a 96-well plate. The

complexes were evaluated at concentrations between 0.0078

and 1 mg mL⁻. A 2 mg mL solution of the complexes in

PBS (water solvent complexes) or in 10% DMSO + PBS (non-

water solvent complexes) was prepared followed by 1 : 2 serial

dilutions in the same solvent. These solutions were added

1

−1

spectrometer at room temperature. Ru(DMSO)

4

Cl

2

and

[Ru(CO) Cl were purchased from Strem Chemicals. L-Meth-

3

2 2

]

(100 μL) to the RBC suspension. A 2% RBC solution in water

ionine sulfoxide and 2-hydroxy-4-(methylthio)butyric acid

calcium salt were purchased from Fluka. Pyridine-based

ligands were purchased from Asis Chem. Inc. 1,3,5-Triaza-7-

phosphaadamantane (PTA) and 4-(2-aminoethyl)morpholine

was used as a positive control (RBC lysis). As the negative

control, the 2% RBC suspension was also incubated with the

solvents of each compound. The mixture of complex-RBC sus-

pension was then incubated for 1 h at 37 °C.

The plate was then centrifuged and the absorbance of the

supernatant was measured at 550 nm on a microplate reader

3

were purchased from Sigma Aldrich. The complex Ru(CO) -

35

(

2

DMSO)Cl (3) and the ligand 3,7-diacetyl-1,3,7-triaza-5-phos-

33

phabicyclo[3.3.1]nonane (DAPTA) were prepared using litera-

ture methods. Methyl 2-isocyano-2-methylbutanoate, ethyl

(Bio-Rad). The hemolytic index (HI) was determined using the

formula:

1

2

-isocyanocyclopropanecarboxylate and methyl 2-isocyano-

-methylpropanoate were purchased from GalChimia, Spain.

OD ðcomplex sampleÞ ꢁ OD ðcomplex referenceÞ

OD ðpositive controlÞ ꢁ OD ðnegative controlÞ

ꢂ 100

HIð%Þ ¼

cis,trans,cis- and cis,cis,cis-Ru(CO)

CO) Cl (0.398 g, 0.778 mmol) was dissolved in DMSO

3 mL). The reaction mixture turned orange and was stirred for

2 2 2

(DMSO) Cl (4a, 4b). [Ru-

(

3

2 2

]

where OD (complex reference) = OD of the corresponding 30 min at room temperature. The crude product was extracted

complex solution (endogenous abs.). OD (positive control) = from the DMSO solution with diethyl ether until the extracts

OD of the solution obtained by lysis of the RBC (1% RBC). OD were colourless. The ether filtrates were collected, concen-

(negative control) = OD of the 1% RBC suspension in the com- trated and placed at 4 °C overnight aﬀording an oﬀ-white solid

plexes’ solvent after centrifugation.

(0.269 g; 90%). The spectroscopic data confirm the presence of

both known isomers in the solid and is given here for conven-

A hemolytic index above 10% indicates hemolysis.

Dalton Trans.

View Article Online

Dalton Transactions

Paper

ience. Selected IR (KBr, cm⁻¹) 2077 (s, CuO); 2020 (s, CuO); 2.64; N, 2.73; S, 7.57; Found: C, 23.17; H, 2.79; N, 3.11; S, 7.21.

3

5

1

−1

1

132 (m, Ru–S); 1029 (m, Ru–S); 924 (m, Ru–O). H NMR Selected IR (KBr, cm ): 2133 (s, CuO); 2056 (s, CuO); 1651(s,

1

(CDCl

3

, 400 MHz): δ (ppm) 3.44 (s, 2H, cis, trans, cis-isomer), CvO). H NMR (D

2

O, 400 MHz,): δ (ppm) 4.06 (m, 1H), 3.09

3

.43 (s, 1H), 3.17 (s, 1H), 2.88 (s, 1H), 2.85 (s, 1H, cis, cis, cis- (m, 2H), 2.76 (s, 3H), 2.39 (m, 2H).

1

3

isomer). C NMR (CDCl , 100 MHz): δ (ppm) 185.7 (CO), 45.1

Ru(CO) Cl (3-NC H (CH ) SO Na)

(7). [Ru(CO) Cl ]

3

2

5

4

2 2

3

2 2

(

cis, trans, cis-isomer), 186.1 (CO), 47.5, 42.2, 39.3, 38.9 (cis, cis, (0.634 g, 1.238 mmol) and 3-NC

5

H

4

(CH

2

)

2

SO

3

Na (0.518 g,

cis-isomer).

2.476 mmol) were suspended in acetone (100 mL) and left to

cis,cis,cis-, cis,cis,trans-, and cis,mer-Ru(CO)(DMSO) Cl₂react for 24 h, at room temperature. The solution was filtered,

3

(

1

5). To a stirred orange solution of [Ru(CO)

.035 mmol) in DMSO (15 mL) was added toluene (15 mL). an excess of diethyl ether and a pale yellow solid was precipi-

The reaction mixture was refluxed for 2 h and turned light tated. The solid was washed with diethyl ether (3 × 15 mL),

yellow. The cooled reaction mixture was filtered at room temp- and dried in a vacuum (0.73 g; 63%). Anal. Calcd for RuC10

erature, giving a turbid suspension, which was then filtered. NO SNaCl : C, 25.82; H, 1.73; N, 3.01; S, 6.89; Found: C, 25.40;

To the filtrate, diethyl ether was added and the solution was H, 1.98; N, 2.77; S, 7.12. Anal. Calc. for RuC H NO -

3 2 2

Cl ] (0.530 g, concentrated, transferred via a cannula to a Schlenk tube with

8

H -

6

2

1

0

8

6

placed at −30 °C overnight. An oily yellow residue and a white SNaCl

2 2

.0.15H O: C, 25.67; H, 1.79, N, 2.99; S, 6.85. Selected IR

crystalline product were precipitated. From the cold filtrate, (KBr, cm⁻): 2137 (s, CuO); 2053 (s, CuO). H NMR (D

1

O,

2

another fraction of white crystalline product with yellow con- 400 MHz,): δ (ppm) 8.75 (s, 1H), 8.67 (d, 1H), 8.58 (d, 1H), 8.03

tamination was obtained. This was extracted with ether, (t, 1H), 3.34 (s, 4H).

leaving a yellow powder behind and a white powder started to

3 2 5 4 2 2 3

Ru(CO) Cl (4-NC H (CH ) SO Na) (8). A solution of [Ru-

precipitate from the ethereal solution. After concentration and (CO) Cl ] (0.392 g, 0.765 mmol) in MeOH (20 mL) was added

3

2 2

5 4 2 2 3

cooling, the precipitate was filtered, washed with hexane and to a solution of the pyridine 4-NC H (CH ) SO Na ligand

dried in vacuo aﬀording a white powder (Fraction 1). The (0.320 g, 1.530 mmol) in MeOH (50 mL). The colourless reac-

yellow crude product was crystallised from CH Cl –hexane at tion mixture was stirred for 24 h at room temperature. The

2

−

30 °C aﬀording yellow crystals. Yield: 31%. Anal. Calcd for solution was filtered, concentrated and after the addition of

RuC Cl : C, 19.36; H, 4.18; S, 22.15; Found: (Fraction diethyl ether a white solid was precipitated. The solid was

): C, 19.40; H, 3.84; S, 22.40; Fraction 2: C, 19.83; H, 3.81; S, washed with diethyl ether (2 × 10 mL) and dried in a vacuum

7

H

18

O

4

S

3

2

1

2

−

1

2.32. Selected IR (KBr, cm ): Fraction 1: 2001 (vs, CuO); (0.380 g, 53%). Anal. Calcd for RuC10

H

8

NO

, 400 MHz): Both H, 1.73; N, 3.01; S, 6.89; Found: C, 26.10; H, 2.20; N, 3.21; S,

fractions contained both isomers but in diﬀerent ratios: δ 6.96. Anal. Calc. for RuC H NO SNaCl ·0.1H O: C, 25.72; H,

6 2

SNaCl : C, 25.82;

1

Fraction 2: 2003 (vs, CuO). H NMR (CDCl

3

1

0

8

6

2

−1

(

ppm) 2.75 (s, 1H), 2.78 (s, 1H), 3.17 (s, 1H), 3.45 (s, 2H), 3.52 1.77; N, 3.00; S, 6.87. Selected IR (KBr, cm ): 2137 (s, CuO);

s, 1H) assigned to the cis, cis, cis-isomer; 2.85 (s, 2H), 3.25 (s, 2053 (s, CuO). H NMR (D O, 400 MHz): δ (ppm) 8.69 (d, 2H),

2

1

2

H), 3.43 (s, 2H) assigned to the cis, cis, trans-isomer; 3.41 (s, 8.02 (d, 2H), 3.40 (s, 4H).

H), 3.43 (s, 2H), 3.51 (s, 2H) assigned to the cis, mer-isomer; Ru(CO) Cl (PTA) (9). A solution of [Ru(CO)

, 100 MHz): δ (ppm) 38.5, 39.0, 43.5, 45.3, 0.601 mmol) in MeOH (10 mL) was added to a solution of

6.8, 50.2 assigned to the cis, cis, cis-isomer; 39.1, 43.7, 47.6 1,3,5-triaza-7-phosphaadamantane (0.189 g, 1.203 mmol) in

2

3

2

3 2 2

Cl ] (0.308 g,

1

3

C NMR (CDCl

3

4

assigned to the cis, cis, trans-isomer; 42.9, 46.4, 47.4 assigned MeOH (10 mL), and a white precipitate was readily formed.

to the cis, mer-isomer. While fraction 1 has CO resonances at δ The reaction mixture was stirred for 15 min at room tempera-

=

192.3 and 191.0 ppm, no CO resonances were observed for ture. The precipitate was isolated, washed with MeOH (5 mL)

fraction 2.

and dried in a vacuum (0.280 g, 56%). Anal. Calcd for

L-Ru(CO)

3

Cl

2

(H

3

CSO(CH

2

)

2

CH(NH

2

)CO

2

H)

(6a). [Ru-

N

3

O

3

PCl

2

RuC

9

H

12: C, 26.16; H, 2.93; N, 10.17; Found: C, 25.90;

−

1

(

CO) Cl ] (0.918 g, 1.793 mmol) and L-methionine sulfoxide H, 3.30; N, 10.18. Selected IR (KBr, cm ): 2134 (w, CuO);

3

2 2

3

1

6

0.592 g, 3.586 mmol) were suspended in acetone (100 mL), 2060 (s, CuO); 1994 (s, CuO). P NMR (d -DMSO, 162 MHz):

and left to react for 24 h, at room temperature. The pale yellow δ (ppm) −28.84.

solution was filtered, concentrated, transferred via a cannula Ru(CO) Cl (DAPTA) (10). A solution of [Ru(CO) Cl ]

2 2

3

2

3

to a Schlenk tube with an excess of diethyl ether and an oﬀ- (0.304 g, 0.594 mmol) in MeOH (10 mL) was added to a solution

white solid was precipitated. The solid was washed with of DAPTA (0.272 g, 1.188 mmol) in MeOH (10 mL). The reaction

diethyl ether (2 × 20 mL), and dried in a vacuum (1.25 g; 83%). mixture was stirred for 2 h at room temperature. The solvent

Anal. Calcd for RuC

8 6 2

H11NO SCl : C, 23.07; H, 2.64; N, 2.73; S, was removed and the white resulting solid was washed with

7

(

.57; Found: C, 22.73; H, 2.99; N, 2.95; S, 7.48. Selected IR dichloromethane (2 × 10 mL) and dried in a vacuum (0.432 g,

KBr, cm⁻): 2131 (s, CuO); 2055 (s, CuO); 1651(s, CvO). H 75%). Anal. Calcd for C H Cl N O PRu: C, 29.70; H, 3.32; N,

1

12 16 2 3 5

−1

NMR (MeOD, 400 MHz,): δ (ppm) 4.1–3.8 (m, 1H), 3.1–2.9 (m, 8.66; Found: C, 29.34; H, 3.65; N, 8.47. Selected IR (KBr, cm ):

2

H), 2.70 (s, 3H), 2.5–2.1 (m, 2H).

2135 (w, CuO); 2067 (s, CuO); 2001 (s, CuO); 1635 (s, CvO).

1

D,L-Ru(CO) Cl (H CSO(CH ) CH(NH )CO H) (6b). The com-

H NMR (D O, 400 MHz): δ (ppm) 5.7–3.8 (m, 8H), 3.28 (s, 2H),

3

2

3

2 2

2

31

pound is prepared exactly as 6a above. However, precipitation 2.12 (s, 6H). P NMR (CD

had to be induced at −30 °C and the yield was considerably Ru(CO) Cl (H CS(CH

lower (36%). Anal. Calcd for RuC H NO SCl : C, 23.07; H, thesis: [H CS(CH ) CH(OH)COO] Ca (880 mg, 2.600 mmol)

3

OD, 162 MHz): δ (ppm) 5.87.

3

2

3

2

)

2

CH(OH)CO H) (11). Ligand syn-

2

8

11

6

2

3

2 2

2

Dalton Trans.

View Article Online

Paper

Dalton Transactions

was dissolved in water (30 mL) and the solution was stirred for δ (ppm) 3.88 (s, 3H), 2.12 (m, 2H, 3H), 1.79 (s, 3H), 1.11 (t,

1

3

a couple of minutes, after which time the residue was comple- 3H). C NMR (CDCl , 100 MHz): δ (ppm) 183 (CuO), 181

3

tely dissolved. Sulfuric acid (1 M) was slowly added (2.6 mL, (CvO), 168 (CuN), 69 (CCH

3

C

2

H

5

), 54 (OCH

3

), 33 (CCH

(14). [Ru(CO) Cl ]

2 2

2 3

CH ),

1

equiv.) and the clear solution became turbid after some 25 (CCH

3

), 9 (CH

2 3

CH ).

minutes. The solution was stirred for 90 min after which time

Ru(CO) Cl (CN(c-CCH CH )CO Et

3

2

3

4

the reaction was stopped. A white precipitate (CaSO ) was fil- (0.250 g; 0.49 mmol) was suspended in chloroform (20 mL)

tered oﬀ and washed with a small amount of MeOH. The solu- and treated with ethyl 1-isocyanocyclopropanecarboxylate

tion was dried aﬀording a pale yellow oily residue. The residue (0.208 g; 1.49 mmol). The yellow pale solution was stirred for

was extracted with MeOH, leaving a small amount of a white 5 hours, at room temperature. The solution was concentrated

powder behind (CaSO

4

). The pale yellow filtrate was taken to and diethyl ether (30 mL) was added, and placed at −30 °C

dryness aﬀording a yellow oil that was kept at −30 °C. Yield: overnight. The white solid was isolated and washed with

1

8

2

(

5%. H NMR (400 MHz, CD

3 5 2

OD, RT): δ 4.26 (m, 1H), 2.62 (t, diethyl ether (0.37 g; 96%). Anal. Calcd for RuC10H11NO Cl :

13

H), 2.09 (s, 3H), 2.03 (m, 1H), 1.88 (m, 1H). C NMR C, 30.39; H, 2.30; N, 3.54; Found: C, 30.30; H, 2.21; N, 3.58.

100 MHz, CD OD, RT): δ 177.6 (CO), 70.0, 34.8, 30.6, 15.1. IR Selected IR (KBr, cm⁻): 2249 m (CuN), 2140 vs, 2093 s, 2070

1

3

KBr, cm⁻): 1733 (s, CvO); 1437 (m); 1276 (w); 1224 (m); 1173 vs (CuO), 1738 m (CvO). H NMR (CDCl

1

, 400 MHz,):

3

1

3

m); 1096 (s); 970 (w); 799 (w); 752 (w); 656 (w); 647 (w). δ (ppm) 4.32 (q, 2H), 1.88 (s, 4H), 1.36 (t, 3H). C NMR

Complex 11 was obtained following an identical procedure to (CDCl , 100 MHz): δ (ppm) 183 (CuO), 181 (CvO), 166

3

2 3 2 2 2 2

that described for 6a and 6b; precipitation of the complex was (CuN), 64 (CH CH ), 63 (C(CH CH )), 21 (C(CH CH )), 14

1

induced at −90 °C. Yield: 83%. H NMR (400 MHz, MeOD, (CH

RT): δ 2.65 (m, 1H), 2.09 (s, 3H), 1.55–1.63 (m, 4H). IR (KBr,

2 3

CH ).

−

1

cm ): 2141 (s, CuO); 2077 (s, CuO); 2063 (s, CuO); 1793 (s,

CvO). Anal. Calcd for RuC SCl : C, 23.66; H, 2.48; S,

.89; Found: C, 23.58; H, 2.40; S, 7.93.

Ru(CO) Cl (CNCMe CO Me) (12). [Ru(CO)

.78 mmol) was suspended and stirred in 20 mL of CHCl

8

H

10

O

6

2

Conﬂict of interest

7

0

3

2

3 2 2

Cl ] (0.400 g;

All the authors designated as aﬃliated with Alfama Ltd were

paid by that company to perform the work described herein

and have equity interests (except for G.J.L.B.).

3

.

Methyl 2-isocyano-2-methylpropanoate (3 eq.; 0.298 g) was dis-

solved in 10 mL of CHCl and added at room temperature.

3

After 30 min the solution was completely clear and with a

yellow pale color. The reaction was continued for 5 hours. The

solution was concentrated and on addition of diethyl ether a Acknowledgements

white precipitate was formed. The solid was filtered and dried

We thank Fundação para a Ciência e Tecnologia (FCT), Portu-

in a vacuum (0.67 g, 96%). Anal. Calcd for RuC H Cl NO : %C,

9

2

5

gal (PTDC/QUI-BIQ/117799/2010, PEst-C/EQB/LA0006/2013

and PEst-OE/QUI/UI0100/2013) for funding. J.D.S. and M.F.A.S.

thank FCT for the award of doctoral studentships. A.M., A.C.C.

and P.M.R. thank FCT for the award of postdoctoral fellow-

ships. The National NMR Network (REDE/1517/RMN/2005)

was supported by POCI 2010 (FEDER) and FCT. We also thank

the ESRF for provision of synchrotron radiation facilities

2

8.21; H, 2.37; N, 3.66. Found: C, 27.90; H, 2.53; N, 3.60.

−

1

Selected IR (KBr, cm ): 2247 (s, CuN); 2145 (s, CuO), 2094

1

(s, CuO), 2058 (s, CuO); 1750 (CvO). H NMR (CDCl ,

3

1

3

4

00 MHz): 3.89 (s, 3H, O–CH

3

); 1.85 (s, 6H). C NMR (CDCl

3

):

δ (ppm) 183 (CuO); 182 (CvO); 169 (CuN); 65 (C-(CH

3 2

)

); 54

(OCH ); 27 (C-(CH ) ).

3

3 2

Note: Attempts to recrystallize the compound in a mixture

(

beamlines ID14-1 and ID14-2).

of chloroform and diethyl ether, at −30 °C, gave rise to a

yellow solid. The FTIR is in agreement with this compound

being a dicarbonyl, but its full characterization was not

pursued. Selected IR (KBr, cm ): 2230 m (CuN), 2093 s, 2037

s (CuO), 1758 m (CvO).

−

1

Notes and references

Ru(CO)

.78 mmol) was suspended in chloroform (25 mL) and treated

with methyl 2-isocyano-2-methylbutanoate (0.334 g;

.36 mmol). The yellow pale solution was stirred for 5 hours,

3

Cl

2

(CNCMeEtCO

2

Me) (13). [Ru(CO)

3

Cl

2

]

2

(0.40 g;

1 A. Verma, D. J. Hirsch, C. E. Glatt, G. V. Ronnett and

S. H. Snyder, Science, 1993, 259, 381–384.

2 L. Otterbein, F. Bach, J. Alam, M. Soares, H. Lu, M. Wysk,

R. Davis, R. Flavell and A. Choi, Nat. Med., 2000, 6,

422–428.

3 R. Buelow and J. Woo, WO 2002/078684, 20020401, 2002.

4 R. Motterlini and B. E. Mann, WO 2002/092075, 20020515,

2002.

5 W. Haas, C. C. Romão, B. Royo, A. Fernandes and

I. Gonçalves, WO 2003/066067, 20030203, 2003.

6 R. Motterlini, J. Clark, R. Foresti, P. Sarathchandra,

B. Mann and C. Green, Circ. Res., 2002, 90, E17–E24.

0

2

at room temperature. The solution was reduced in volume and

diethyl ether (30 mL) was added and placed at −30 °C over-

night. A pale yellow solid was isolated (0.11 g; 18%). Recrystal-

lization of the solid in a mixture of chloroform and diethyl

ether, at −30 °C, gave rise to crystals. Anal. Calcd for RuC H -

1

0

11

5 2

NO Cl : C, 30.24; H, 2.79; N, 3.53; Found: C, 30.50; H, 2.98; N,

−

1

3

2

.72. Selected IR (KBr, cm ): 2252 m (CuN), 2149 vs, 2090 s,

1

061 vs (CuO), 1753 m (CvO). H NMR (CDCl , 400 MHz,):

3

Dalton Trans.

View Article Online

Dalton Transactions

Paper

7

J. Clark, P. Naughton, S. Shurey, C. Green, T. Johnson, 27 T. Santos-Silva, A. Mukhopadhyay, J. D. Seixas,

B. Mann, R. Foresti and R. Motterlini, Circ. Res., 2003, 93,

E2–E8.

G. J. L. Bernardes, C. C. Romão and M. J. Romão, Curr.

Med. Chem., 2011, 18, 3361–3366.

8

9

B. E. Mann, Organometallics, 2012, 31, 5728–5735.

I. C. Winburn, K. Gunatunga, R. D. McKernan, R. J. Walker,

28 P. Wang, H. Liu, Q. Zhao, Y. Chen, B. Liu, B. Zhang and

Q. Zheng, Eur. J. Med. Chem., 2014, 74, 199–215.

I. A. Sammut and J. C. Harrison, Basic Clin. Pharmacol. 29 B. W. Michel, A. R. Lippert and C. J. Chang, J. Am. Chem.

Toxicol., 2012, 111, 31–41. Soc., 2012, 134, 15668–15671.

0 Y. Guo, A. B. Stein, W.-J. Wu, W. Tan, X. Zhu, Q.-H. Li, 30 M. Moreno, M. Haukka and M. Kallinen, Appl. Organomet.

1

B. Dawn, R. Motterlini and R. Bolli, Am. J. Physiol. Heart C,

004, 286, H1649–H1653.

1 G. Wang, T. Hamid, R. J. Keith, G. Zhou, C. R. Partridge,

Chem., 2006, 20, 51–69.

31 R. K. Hom and J. A. Katzenellenbogen, Nucl. Med. Biol.,

1997, 24, 485–498.

2

X. Xiang, J. R. Kingery, R. K. Lewis, Q. Li, D. G. Rokosh, 32 A. R. Marques, L. Kromer, D. J. Gallo, N. Penacho,

R. Ford, F. G. Spinale, D. W. Riggs, S. Srivastava,

A. Bhatnagar, R. Bolli and S. D. Prabhu, Circulation, 2010,

S. S. Rodrigues, J. D. Seixas, G. J. L. Bernardes, P. M. Reis,

S. L. Otterbein, R. A. Ruggieri, A. S. G. Gonçalves,

A. M. L. Gonçalves, M. N. D. Matos, I. Bento,

L. E. Otterbein, W. A. Blättler and C. C. Romão, Organo-

metallics, 2012, 31, 5810–5822.

1

21, 1912–1925.

1

2 M. Musameh, C. Green, B. Mann, B. Fuller and

R. Motterlini, J. Heart Lung Transplant., 2007, 26, 1192–1198.

3 B. Chen, L. Guo, C. Fan, S. Bolisetty, R. Joseph, 33 D. Darensbourg, C. Ortiz and J. Kamplain, Organometallics,

M. M. Wright, A. Agarwal and J. F. George, Am. J. Pathol.,

009, 175, 422–429.

2004, 23, 1747–1754.

34 J.-R. Li and X.-H. Bu, Eur. J. Inorg. Chem., 2008, 27–40.

2

1

4 R. Motterlini and L. E. Otterbein, Nat. Rev. Drug Discovery, 35 E. Alessio, B. Milani, M. Bolle, G. Mestroni, P. Faleschini,

2

010, 9, 728–743.

F. Todone, S. Geremia and M. Calligaris, Inorg. Chem.,

1995, 34, 4722–4734.

1

5 B. E. Mann, Top. Organomet. Chem., 2010, 32, 247–285.

6 E. A. Sher, M. Shaklai and N. Shaklai, Adv. Biol. Chem., 36 D. Darensbourg, F. Joo, M. Kannisto, A. Katho,

2

012, 2, 191–197.

J. Reibenspies and D. Daigle, Inorg. Chem., 1994, 33,

200–208.

1

7 A. Marazioti, M. Bucci, C. Coletta, V. Vellecco, P. Baskaran,

C. Szabó, G. Cirino, A. R. Marques, B. Guerreiro, 37 D. N. Akbayeva, L. Gonsalvi, W. Oberhauser, M. Peruzzini,

A. M. L. Gonçalves, J. D. Seixas, A. Beuve, C. C. Romão and

A. Papapetropoulos, Arterioscler., Thromb., Vasc. Biol., 2011,

F. Vizza, P. Brüggeller, A. Romerosa, G. Sava and

A. Bergamo, Chem. Commun., 2003, 264–265.

3

1, 2570–2576.

38 M. Calligaris, Coord. Chem. Rev., 2004, 248, 351–375.

1

8 A. C. Pena, N. Penacho, L. Mancio-Silva, R. Neres, 39 I. Evans, A. Spencer and G. Wilkinson, J. Chem. Soc., Dalton

J. D. Seixas, A. C. Fernandes, C. C. Romão, M. M. Mota, Trans., 1973, 204–209.

G. J. L. Bernardes and A. Pamplona, Antimicrob. Agents 40 E. Alessio, G. Mestroni, G. Nardin, W. Attia,

Chemother., 2012, 56, 1281–1290.

9 R. Cini, S. Defazio, G. Tamasi, M. Casolaro, L. Messori,

M. Calligaris, G. Sava and S. Zorzet, Inorg. Chem., 1988,

27, 4099–4106.

1

2

A. Casini, M. Morpurgo and M. Hursthouse, Inorg. Chem., 41 C. Allardyce, P. Dyson, D. Ellis and S. Heath, Chem.

007, 46, 79–92. Commun., 2001, 1396–1397.

0 D. Valensin, P. Anzini, E. Gaggelli, N. Gaggelli, G. Tamasi, 42 T. Funaioli, C. Cavazza, F. Marchetti and G. Fachinetti,

2

R. Cini, C. Gabbiani, E. Michelucci, L. Messori, H. Kozlowski

and G. Valensin, Inorg. Chem., 2010, 49, 4720–4722.

Inorg. Chem., 1999, 38, 3361–3368.

43 T. R. Johnson, B. E. Mann, I. P. Teasdale, H. Adams,

R. Foresti, C. J. Green and R. Motterlini, Dalton Trans.,

2007, 1500–1508.

44 M. C. Vaney, S. Maignan, M. Ries-Kautt and A. Ducruix,

Acta Crystallogr., Sect. D: Biol. Crystallogr., 1996, 52, 505–

517.

1 M. F. A. Santos, J. D. Seixas, A. C. Coelho,

A. Mukhopadhyay, P. M. Reis, M. J. Romão, C. C. Romão

and T. Santos-Silva, J. Inorg. Biochem., 2012, 117, 285–291.

2 C. C. Romão, W. A. Blättler, J. D. Seixas and

G. J. L. Bernardes, Chem. Soc. Rev., 2012, 41, 3571–3583.

2

3 S. García-Gallego and G. J. L. Bernardes, Angew. Chem., Int. 45 U. Meier, R. Scopelliti, E. Solari and A. Merbach, Inorg.

Ed., 2014, 53, 9712–9721. Chem., 2000, 39, 3816–3822.

4 M. Desmard, R. Foresti, D. Morin, M. Dagouassat, 46 M. Torrent, M. Sola and G. Frenking, Organometallics,

A. Berdeaux, E. Denamur, S. H. Crook, B. E. Mann, 1999, 18, 2801–2812.

D. Scapens, P. Montravers, J. Boczkowski and R. Motterlini, 47 S. Barrows, Inorg. Chem., 2004, 43, 8236–8238.

Antioxid. Redox Signaling, 2012, 16, 153–163.

48 X. Rozanska and R. Vuilleumier, Inorg. Chem., 2008, 47,

2

5 S. McLean, B. E. Mann and R. K. Poole, Anal. Biochem.,

8635–8640.

2

012, 427, 36–40.

49 F. Zhang, L. Zhao, C. Xu and Y. Chen, Inorg. Chem., 2010,

49, 3278–3281.

2

6 T. Santos-Silva, A. Mukhopadhyay, J. D. Seixas,

G. J. L. Bernardes, C. C. Romão and M. J. Romão, J. Am. 50 N. Kuriakose, S. Kadam and K. Vanka, Inorg. Chem., 2012,

Chem. Soc., 2011, 133, 1192–1195. 51, 377–385.

Dalton Trans.

View Article Online

Paper

Dalton Transactions

5

1 Y. Chen, F. Zhang, C. Xu, J. Gao, D. Zhai and Z. Zhao,

J. Phys. Chem. A, 2012, 116, 2529–2535.

2 H. Schulz, A. Goerling and W. Hieringer, Inorg. Chem.,

J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,

R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,

R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,

K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,

J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas,

J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,

GAUSSIAN 09 (Revision A.01), Gaussian, Inc., Wallingford

CT, 2009.

2

013, 52, 4786–4794.

3 G. Laurenczy, L. Helm, A. Ludi and A. Merbach, Helv.

Chim. Acta, 1991, 74, 1236–1238.

4 H. Vreman, R. Wong, T. Kadotani and D. Stevenson, Anal.

Biochem., 2005, 341, 280–289.

5 A. B. Stein, R. Bolli, B. Dawn, S. K. Sanganalmath, Y. Zhu, 69 W. J. Hehre, L. Radom, P. von Ragué Schleyer and

O. L. Wang, Y. Guo, R. Motterlini and Y. T. Xuan, J. Mol.

Cell. Cardiol., 2012, 52, 228–236.

J. A. Pople, Ab Initio Molecular Orbital Theory, John Wiley &

Sons, New York, 1986.

5

6

6 W. Fehlhammer, S. Schrolkamp, M. Hoyer, H. Hartl and 70 R. G. Parr and W. Yang, Density Functional Theory of Atoms

W. Beck, Z. Anorg. Allg. Chem., 2005, 631, 3025–3029. and Molecules, Oxford University Press, New York, 1989.

7 D. Achatz, M. Lang, A. Volkl, W. Fehlhammer and W. Beck, 71 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,

Z. Anorg. Allg. Chem., 2005, 631, 2339–2346. 1997, 78, 1396–1396.

8 S.-I. Murahashi, T. Nakae, H. Terai and N. Komiya, J. Am. 72 J. P. Perdew, Phys. Rev. B: Condens. Matter, 1986, 33, 8822–

Chem. Soc., 2008, 130, 11005–11012. 8824.

9 D. Chatterjee, A. Mitra and R. van Eldik, Dalton Trans., 73 U. Haussermann, M. Dolg, H. Stoll, H. Preuss,

2

007, 943–948.

P. Schwerdtfeger and R. M. Pitzer, Mol. Phys., 1993, 78,

1211–1224.

0 A. F. N. Tavares, M. Teixeira, C. C. Romão, J. D. Seixas,

L. S. Nobre and L. M. Saraiva, J. Biol. Chem., 2011, 286, 74 W. Kuchle, M. Dolg, H. Stoll and H. Preuss, J. Chem. Phys.,

6708–26717. 1994, 100, 7535–7542.

1 S. Werner, N. Szesni, A. Bittermann, M. J. Schneider, P. Hã 75 T. Leininger, A. Nicklass, H. Stoll, M. Dolg and

2

6

Rter, M. Haumann and P. Wasserscheid, Appl. Catal., A,

010, 377, 70–75.

2 H. Ishida, K. Tanaka, M. Morimoto and T. Tanaka, Organo-

metallics, 1986, 5, 724–730.

3 H. Tanaka, H. Nagao, S. Peng and K. Tanaka, Organometal-

lics, 1992, 11, 1450–1451.

4 H. Tanaka, B. Tzeng, H. Nagao, S. Peng and K. Tanaka,

Organometallics, 1992, 11, 3171–3172.

5 G. N. Murshudov, A. A. Vagin and E. J. Dodson, Acta Crys-

tallogr., Sect. D: Biol. Crystallogr., 1997, 53, 240–255.

6 P. Emsley and K. Cowtan, Acta Crystallogr., Sect. D: Biol.

Crystallogr., 2004, 60, 2126–2132.

P. Schwerdtfeger, J. Chem. Phys., 1996, 105, 1052–1059.

76 A. W. Ehlers, M. Bohme, S. Dapprich, A. Gobbi,

A. Hollwarth, V. Jonas, K. F. Kohler, R. Stegmann,

A. Veldkamp and G. Frenking, Chem. Phys. Lett., 1993, 208,

111–114.

77 R. Ditchfie, W. J. Hehre and J. A. Pople, J. Chem. Phys.,

1971, 54, 724–728.

78 W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys.,

1972, 56, 2257–2261.

79 P. C. Harihara and J. A. Pople, Mol. Phys., 1974, 27, 209–

214.

80 P. C. Harihara and J. A. Pople, Theor. Chim. Acta, 1973, 28,

213–222.

2

6

7 R. P. Joosten, K. Joosten, G. N. Murshudov and A. Perrakis,

Acta Crystallogr., Sect. D: Biol. Crystallogr., 2012, 68, 484– 81 E. Cances, B. Mennucci and J. Tomasi, J. Chem. Phys., 1997,

96. 107, 3032–3041.

8 M. J. Frisch, G. W. Trucks, H. B. Schelegel, G. E. Scuseria, 82 M. Cossi, V. Barone, B. Mennucci and J. Tomasi, Chem.

M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, Phys. Lett., 1998, 286, 253–260.

B. Mennucci, G. A. Petersson, H. Nakutsuji, M. Caricato, 83 B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 106,

X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, 5151–5158.

J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, 84 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005,

J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, 105, 2999–3093.

H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, 85 A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys.

F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, Chem. B, 2009, 113, 6378–6396.

K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, 86 C. Y. Peng, P. Y. Ayala, H. B. Schlegel and M. J. Frisch,

K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Comput. Chem., 1996, 17, 49–56.

J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, 87 C. Peng and H. B. Schlegel, Isr. J. Chem., 1993, 33, 449–454.

4

6

Dalton Trans.

Article Doi

A contribution to the rational design of Ru(CO)3Cl2L complexes for in vivo delivery of CO

DOI: 10.1039/c4dt02966f

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c4dt02966f

Full text of DOI:10.1039/c4dt02966f

Products guided by the article

R&D Labs maybe for 583-91-5

Relevant to this article

Hot Product