Photoactivated in Vitro Anticancer Activity of Rhenium(I) Tricarbonyl Complexes Bearing Water-Soluble Phosphines

Article

Cite This: Inorg. Chem. XXXX, XXX, XXX −XXX

pubs.acs.org/IC

Photoactivated in Vitro Anticancer Activity of Rhenium(I) Tricarbonyl

Complexes Bearing Water-Soluble Phosphines

Sierra C. Marker, Samantha N. MacMillan, Warren R. Zipfel, Zhi Li, Peter C. Ford,^§

†

‡

§

,†

and Justin J. Wilson*

†

Department of Chemistry and Chemical Biology, Cornell University, Ithaca, New York 14853, United States

Department of Biomedical Engineering, Cornell University, Ithaca, New York 14853, United States

‡

§

Department of Chemistry and Biochemistry, University of California, Santa Barbara, Santa Barbara, California 93106-9510, United

States

*

S Supporting Information

ABSTRACT: Fifteen water-soluble rhenium compounds of

+

the general formula [Re(CO) (NN)(PR )] , where NN is a

3

diimine ligand and PR is 1,3,5-triaza-7-phosphaadamantane

3

(

1

PTA), tris(hydroxymethyl)phosphine (THP), or 1,4-diacetyl-

,3,7-triaza-5-phosphabicylco[3.3.1]nonane (DAPTA), were

synthesized and characterized by multinuclear NMR spectros-

copy, IR spectroscopy, and X-ray crystallography. The

complexes bearing the THP and DAPTA ligands exhibit

triplet-based luminescence in air-equilibrated aqueous sol-

utions with quantum yields ranging from 3.4 to 11.5%.

Furthermore, the THP and DAPTA complexes undergo

photosubstitution of a CO ligand upon irradiation with 365

1

nm light with quantum yields ranging from 1.1 to 5.5% and sensitize the formation of O with quantum yields as high as 70%. In

2

contrast, all of the complexes bearing the PTA ligand are nonemissive and do not undergo photosubstitution upon irradiation

with 365 nm light. These compounds were evaluated as photoactivated anticancer agents in human cervical (HeLa), ovarian

(

A2780), and cisplatin-resistant ovarian (A2780CP70) cancer cell lines. All of the complexes bearing THP and DAPTA exhibited

a cytotoxic response upon irradiation with minimal toxicity in the absence of light. Notably, the complex with DAPTA and 1,10-

phenanthroline gave rise to an IC value of 6 μM in HeLa cells upon irradiation, rendering it the most phototoxic compound in

5

0

this library. The nature of the photoinduced cytotoxicity of this compound was explored in further detail. These data indicate that

the phototoxic response may result from the release of both CO and the rhenium-containing photoproduct, as well as the

1

production of O .

2

INTRODUCTION

abilities to photosensitize the formation of singlet oxygen

■

1

1,9−19

(

O₂).

In hypoxic tumors, the eﬃcacy of these PDT

Chemotherapy is a critical component for the treatment of

nearly all types of cancer. Conventional chemotherapeutic

agents, however, are nondiscriminating cytotoxic compounds

that induce cell death in both cancerous and healthy tissue, a

feature that is manifested in the form of undesirable patient side

eﬀects. One approach to increase selectivity and minimize toxic

side eﬀects is to employ drugs that are activated by light. This

agents is substantially diminished due to the lower concen-

20−23

trations of oxygen present.

An additional limitation of

these porphyrin analogues is their poor aqueous solubility,

which requires clever formulation strategies for in vivo

2

,24,25

administration.

PACT is currently the subject of

5

,26

signiﬁcant preclinical investigation.

PACT agents generally

5

1

−4

operate under an oxygen-independent mechanism, as

irradiation with light triggers release of a cytotoxic chemical

compound.

concept, known as photodynamic therapy (PDT)

or

5

photoactivated chemotherapy (PACT), allows for further

spatiotemporal control of the cytotoxic eﬀects to maximize

cancer cell death, while minimizing healthy cell damage. An

ideal candidate for a PDT or PACT agent should be nontoxic

in the absence of light and therefore only give rise to toxic

27

5

Toward the development of improved PDT and PACT

agents, many research eﬀorts have explored the potential of

1

transition-metal complexes. In addition to functioning as O

2

6

photosensitizers for PDT applications, many inorganic

complexes also undergo photoinduced ligand substitution

eﬀects in regions of light exposure. The concept of PDT is

currently employed in the clinic in the form of the FDA-

approved porphyrin derivatives Photofrin, Foscan, and

7

,8

Verteporﬁn. A potential drawback of the currently used

Received: October 26, 2017

PDT agents is that their mechanism of action stems from their

©

XXXX American Chemical Society

A

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

+

reactions, which make them useful as PACT agents. Following

the concept of PACT, these photosubstitution reactions can be

applied to generate cytotoxic complexes from benign precursors

in an oxygen-independent manner, rendering them suitable for

Scheme 1. Photoreaction of [Re(CO) (NN)(PR )]

3 3

Releasing 1 equiv of CO and a Labile Rhenium(I) Complex,

Which May Both Give Rise to Cytotoxicity in Cancer Cells

a

2

8

29−51

use in hypoxic conditions. Ruthenium PDT

and

2

2,35,40,42,50,52−55

PACT

agents have received the majority of

focus, because they possess ideal photophysical properties for

this application. The diversity of photophysical processes

accessible for other metal ion complexes have prompted

researchers to expand their search for new PDT or PACT

agents.

Complexes of rhenium, for example, possess rich spectro-

scopic and photophysical properties that can be leveraged for

use in PDT or PACT. This concept has been demonstrated

a

S is a solvent molecule.

56

for several rhenium(I) tricarbonyl complexes, which exhibit

triaza-5-phosphabicylco[3.3.1]nonane (DAPTA) because

57−61

potent phototoxic eﬀects in vitro.

All of the potential

these ligands eﬀectively confer water solubility to their

85−87

rhenium-based phototoxic agents to date, however, induce cell

death via production of singlet oxygen and therefore suﬀer from

some of the same limitations as the conventional porphyrin-

based PDT agents. The development of rhenium anticancer

PACT agents that act via alternative mechanisms of action

remains an interesting unmet research objective that could

provide access to new therapeutic modalities. In this study, we

describe our eﬀorts toward this goal, in which we have

developed a series of rhenium(I) tricarbonyl complexes bearing

phosphine ligands that exhibit potent UVA (365 nm) light-

activated toxicity against several types of cancer cell lines.

Although the biological penetration depth of 365 nm light is

too shallow to be useful for in vivo applications, this study

provides a proof of concept for the future development of

rhenium-based PDT and PACT agents.

coordination complexes.

To further screen for cytotoxic

eﬀects of the rhenium photoproducts, we explored ﬁve diﬀerent

diimine ligands, 1,10-phenanthroline (1), 2,9-dimethyl-1,10-

phenanthroline (2), 2,2′-bipyridine (3), 4,4′-dimethyl-2,2′-

dipyridyl (4), and 4,4′-dimethoxy-2,2′-bipyridine (5). We

hypothesized that this diverse set of compounds would possess

a range of photophysical and biological properties, some of

which would be valuable for use in PACT.

Synthesis and Characterization. From the 3 phosphines

and 5 diimine ligands, a total of 15 rhenium tricarbonyl

complexes were synthesized. The synthetic approach followed a

88,89

previously reported procedure (Scheme 2).

In the ﬁrst step,

Scheme 2. Synthetic Scheme for the Preparation of the

+

[

Re(CO) (NN)(PR )] Complexes

3 3 .

RESULTS AND DISCUSSION

Rationale and Approach. Diimine rhenium(I) tricarbonyl

complexes are usually inert to photosubstitution reactions and

exhibit small luminescence quantum yields. In contrast, when

a phosphine is introduced to the inner coordination sphere, the

■

6

2

+

resulting complexes [Re(CO) (NN)(PR )] , where NN =

3

diimine ligand and PR = phosphine ligand, are brightly

3

6

3−66

62,67

luminescent

and photolabile.

Upon irradiation with

UVA light, the CO trans to the phosphine dissociates, forming

+

the dicarbonyl complex [Re(CO) (NN)(PR )(S)] , where S is

a coordinating solvent molecule.

2

3

6

2,68

Complexes such as these

have been investigated as photo-CO-releasing molecules

6

8−74

(photoCORMS),

tools to understand the signiﬁcance of

CO as a biological signaling molecule. We envisioned an

alternative role of these complexes as photoactivated anticancer

agents. The rhenium(I) dicarbonyl photoproduct with a labile

coordination site could interact with biological targets, such as

DNA, in a manner similar to that reported for rhenium(I)

Re(CO) Cl and a diimine ligand were heated to reﬂux in

5

toluene to aﬀord the yellow complexes fac-[Re(CO) (NN)Cl].

3

These complexes were treated with AgOTf to remove the axial

chloride as insoluble AgCl and then allowed to react with the

68

desired phosphine in reﬂuxing tetrahydrofuran (THF) to

aﬀord the compounds PTA-1−PTA-5, THP-1−THP-5, and

DAPTA-1−DAPTA-5 (Scheme 2). The PTA and DAPTA

complexes were isolated via precipitation as the triﬂate salts. In

contrast, all of the THP complexes, except for THP-5, required

additional puriﬁcation by preparative high-performance liquid

chromatography (HPLC). Because the HPLC mobile phase

employed 0.1% triﬂuoroacetic acid, the THP compounds were

isolated as triﬂuoroacetate salts, with the exception of THP-5,

which was isolated as the triﬂate salt upon recrystallization. All

of the complexes exhibited good water solubility, enabling the

preparation of aqueous solutions at concentrations exceeding 1

75−77

tricarbonyl complexes.

In this context, our laboratory has

recently demonstrated that rhenium(I) tricarbonyl complexes

bearing labile aqua ligands exhibit potent in vitro anticancer

activity via a mechanism of action that is distinct from that of

7

8

cisplatin. Thus, we anticipated that the structurally similar

rhenium(I) dicarbonyl phosphine photoproducts would also be

active against cancer cells. Furthermore, the released CO

molecule could elicit an additional cytotoxic eﬀect (Scheme

7

9−84

1).

Such “dual action” chemotherapeutic agents may be

particularly eﬀective against drug-resistant forms of cancer.

In designing our compound library, we employed the

phosphines 1,3,5-triaza-7-phosphaadamantane (PTA), tris-

mM. In contrast, the fac-[Re(CO) (NN)Cl] starting materials

3

7

8

(

hydroxymethyl)phosphine (THP), and 1,4-diacetyl-1,3,7-

are very poorly soluble in water. Among the 15 complexes

B

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

prepared in this investigation, PTA-1 and THP-3 have

complexes. As expected, the PTA and DAPTA complexes

give rise to a signal at −79 ppm, which corresponds to the

outer-sphere triﬂate ion, and the THP complexes display a

signal at −76 ppm, arising from the TFA counterion, except for

THP-5, which exhibits a signal corresponding to the triﬂate

counterion at −79 ppm.

68,83

previously been reported and developed as photoCORMs.

1

The complexes were characterized by H (Figures S1 −S15 in

3

1

the Supporting Information), P{ H} (Figures S18 −S32 in the

1

9

Supporting Information), and F (Figures S35 −S49 in the

Supporting Information) NMR spectroscopy, IR spectroscopy

(

Figures S50 −S64 in the Supporting Information), and

X-ray Crystallography. Single crystals of all 15 complexes

were obtained, and the crystal structures (Figure 1 and Figures

S99 −S102 in the Supporting Information), with the exception

electrospray ionization mass spectrometry (ESI-MS, Figures

S84 −S98 in the Supporting Information). The purity of the

complexes was veriﬁed with elemental analysis and HPLC

68

of the previously reported THP-3, were determined by X-ray

crystallography. Crystal structures for PTA-1, THP-1, and

DAPTA-1 are shown in Figure 1, and selected interatomic

distances and angles for all 14 structures are collected in Table

S2 in the Supporting Information. We note that the crystal

structure of PTA-1 was recently reported, albeit for a diﬀerent

(

Figures S67 −S81 in the Supporting Information). The IR

spectra of the complexes exhibited three distinct CO stretches,

as expected for fac-[Re(CO) (NN)X] (X = H O, Cl, Br)

3

2

7

8,90−93

complexes of C symmetry.

The energies of these

s

stretching frequencies were invariant with respect to the

coordinating diimine ligands. In contrast, a greater dependence

on the nature of the axial phosphine ligand was observed. For

example, PTA-1, THP-1, and DAPTA-1 exhibited pronounced

CO stretches corresponding to the A″ mode at 2030, 2040, and

83

polymorph of this compound; a comparison of this previously

reported structure and ours reveals no signiﬁcant diﬀerences

83

with respect to the geometry of the rhenium complex. All

6

complexes exhibit the expected octahedral geometry for a d

−

1

2

038 cm , respectively, and two less intense stretches

rhenium(I) center. All DAPTA complexes, except for DAPTA-

2, crystallized as the anti isomer, which was the major species

exhibited by NMR spectroscopy. No evidence for rotational

disorder of the DAPTA acetyl groups was observed in the

crystal structures, verifying our conformer assignments. The

Re−P bond lengths were relatively invariant among complexes

bearing diﬀerent diimine ligands. A more pronounced depend-

ence was observed as the phosphine ligand was altered. For

example, the Re−P distances in PTA-1, THP-1, and DAPTA-1

are 2.4343(6), 2.4602(7), and 2.4449(7) Å, respectively. These

corresponding to the two A′ modes at 1940 and 1920, 1930

−1

and 1910, and 1948 and 1914 cm , respectively (Figures S50 −

S64 ). These values are consistent with those for other fac-

+

68,93

[

Re(CO) (NN)(PR )] complexes.

The presence of the

3

axial phosphine shifts the CO stretching frequencies to higher

energies in comparison to the precursor complexes fac-

[

2

Re(CO) (NN)Cl], which exhibit stretching frequencies near

020 and 1880 cm . The CO stretching energy of the

3

−1

88

phosphine complexes is similar to those of cationic complexes

+

such as fac-[Re(CO) (NN)(OH )] (ν_C_O2030, 1930, and

values are in the range expected for [Re(CO) (NN)(PR )]

complexes.

3

2

3

−1

78

+

68,83,98

1

900 cm ) and fac-[Re(CO) (NN)(pyridine)] (ν 2036

Among the 15 complexes, the THP

3

CO

−

1

94

and 1931 cm ). The higher energy CO stretching

frequencies of the phosphine complexes may therefore be a

result of both the π acidity of the trans phosphine ligand and

the overall cationic charge of the complex, features which

decrease π back-bonding to CO π* orbitals.

complexes exhibit a slightly longer average Re−P bond distance

of 2.4574 ± 0.013 Å (the error is the standard deviation over

the ﬁve THP compounds) in comparison to the PTA and

DAPTA complexes, for which the average Re−P distances are

2

.4340 ± 0.0045 and 2.4414 ± 0.0077 Å, respectively. The

The ³¹P{ H} NMR spectra of the PTA and THP complexes

show a single resonance. This resonance is shifted downﬁeld

from that of the free ligand, signifying coordination to the

1

longer Re−P distances found in the THP complexes are most

likely a consequence of the larger cone angle of this phosphine

9

9,100

in comparison to PTA and DAPTA.

The Re−CO

8

5,86,95,96

rhenium center.

For the DAPTA complexes, two

distances trans to the phosphine vary in a manner that is

dependent on the nature of the phosphine as well. In general,

this distance is longest for complexes of PTA and DAPTA and

shortest for those of THP. For example, the Re−CO distances

in PTA-1, THP-1, and DAPTA-1 are 1.962(3), 1.953(3), and

31

signals in the P NMR spectra were observed. We attribute the

two signals to conformational isomers of the DAPTA ligand. As

shown in Chart 1, the relative orientation of the two acetyl

6

8,83

Chart 1. Syn and Anti Conformational Isomers of DAPTA

1.963(3) Å.

The longer trans Re−CO distances in the

DAPTA and PTA complexes may reﬂect a greater π acidity of

this ligand relative to THP. For comparison, the axial Re−CO

+

distances of fac-[Re(CO) (NN)X] complexes, where X is a

3

solvent molecule, range from 1.882 to 1.917 Å and are

therefore approximately 0.036−0.089 Å shorter than those of

78,101,102

the phosphine derivatives.

Photophysical Properties. UV−vis and emission spectra

of all 15 complexes were measured in air-equilibrated pH 7.4

phosphate-buﬀered saline. These data are collected in Table 1,

and the UV −vis spectra of these complexes are shown in

Figures S103 −S107 in the Supporting Information.

Representative spectra of the phen complexes (PTA, THP,

DAPTA)-1 and the bpy complexes (PTA, THP, DAPTA)-3

are given in Figure 2. The overall band shapes and structures of

the absorbance proﬁles appear to be dependent predominantly

on the nature of the diimine ligand rather than the phosphine.

For example, the spectra of PTA-1, THP-1, and DAPTA-1 are

nearly superimposable, as are those of PTA-3, THP-3, and

groups gives rise to anti and syn isomers, which will result in

distinct NMR signals if their interconversion is slow. The

observations of these two isomers in other metal complexes of

97

1

DAPTA was previously reported. The H NMR spectra of the

DAPTA complexes similarly display resonances of two species

that can be attributed to the two conformers. For all of the

DAPTA complexes, the anti isomer was the major species in

1

solution. This assignment is based on the H NMR resonances

of the methyl groups of the DAPTA ligand, which are

1

9

inequivalent only for the anti isomer. Finally, the F NMR

spectra conﬁrm the nature of the counterions of these

C

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 1. X-ray crystal structures of PTA-1, THP-1, and DAPTA-1. Ellipsoids are drawn at the 50% probability level. H atoms bound to C atoms

and counterions are omitted for clarity.

a

Table 1. Photophysical Properties

Φ

_l_u_m, %

λ_m_a_x, nm (ε, M⁻cm⁻¹

1

)

(λ_m_a_x, nm)

Φ_r_x_n, %

0.06

1.4 ± 0.4

τ_a_i_r(μs) τ_N₂(μs)

b

c

compound

PTA-1

THP-1

226 (35900 ± 2600), 225 (22300 ± 600), 274 (26400 ± 1900), 368 (3200 ± 300)

226 (36800 ± 1300), 275 (26800 ± 900), 322 (6400 ± 200), 366 (3600 ± 100)

DAPTA-1 225 (41600 ± 7300), 275 (27200 ± 4900), 323 (5700 ± 900), 367 (3200 ± 400)

228 (40500 ± 1100), 253 (26100 ± 600), 283 (26200 ± 600), 372 (2500 ± 300)

228 (34600 ± 1200), 286 (23400 ± 800), 308 (11900 ± 400), 372 (2100 ± 70)

DAPTA-2 227 (36800 ± 6800), 285 (23800 ± 700), 309 (11500 ± 300), 373 (2100 ± 50)

<0.1 (n.d.)

5.1 ± 1.1 (528)

10.7 ± 0.6 (516) 1.1 ± 0.1

n.d.

1.5

1.9

n.d.

1.0

2.0

n.d.

0.4

n.d.

4.8

5.6

n.d

3.7

7.8

n.d.

0.5

PTA-2

THP-2

<0.1 (n.d.)

0.015

4.5 ± 0.4 (518)

7.2 ± 0.2 (507)

<0.1 (538)

2.0 ± 0.4

1.5 ± 0.1

0.0044

PTA-3

THP-3

248 (24400 ± 800), 307 (11700 ± 500), 319 (12900 ± 500), 349 (3300 ± 300)

222 (18900 ± 200), 249 (20100 ± 1900), 308 (10700 ± 400), 318 (12500 ± 400), 343

6.1 ± 1.7 (536)

1.6 ± 0.1

(3600 ± 700)

DAPTA-3 246 (22600 ± 500), 308 (10700 ± 200), 319 (13000 ± 300), 345 (3400 ± 100)

9.1 ± 3.0 (528)

<0.1 (530)

2.3 ± 0.5

0.021

0.6

n.d.

0.4

0.6

n.d.

0.3

0.4

0.8

n.d.

0.6

1.0

n.d.

0.4

0.6

PTA-4

THP-4

252 (29200 ± 3600), 303 (13600 ± 1800), 315 (14700 ± 2100), 345 (4300 ± 500)

252 (24000 ± 3000), 304 (11800 ± 1500), 315 (13400 ± 1700), 338 (4600 ± 500)

6.5 ± 2.0 (528)

11.5 ± 3.9 (518) 5.5 ± 0.3

<0.1 (520)

3.8 ± 0.9

DAPTA-4 250 (8200 ± 4300), 305 (12800 ± 2000), 315 (14900 ± 2300), 339 (4000 ± 400)

PTA-5

THP-5

227 (36100 ± 2600), 249 (32900 ± 2300), 304 (9800 ± 700), 338 (4400 ± 300)

223 (33900 ± 1000), 251 (30500 ± 900), 303 (8800 ± 300), 337 (4500 ± 80)

0.018

3.4 ± 1.1 (537)

7.1 ± 3.0 (527)

1.2 ± 0.2

1.9 ± 0.1

DAPTA-5 224 (41400 ± 2000), 251 (34100 ± 1900), 303 (9600 ± 500), 332 (4700 ± 300)

a

b

n.d. = not determined due to weak or undetectable luminescence intensity. Luminescence lifetime measured in air-equilibrated pH 7.4 PBS.

c

Luminescence lifetime measured in nitrogen-saturated pH 7.4 PBS.

DAPTA-3. The high-energy (<325 nm) features in these

spectra most likely correspond to intraligand π−π* transitions

and are therefore expected to be only dependent on the nature

of the diimine ligand. The low-energy features correspond to

the metal-to-ligand charge-transfer (MLCT) transi-

with quantum yields ranging from 7.1 to 11.5%. Notably,

complexes of the PTA ligand were eﬀectively nonemissive

under these conditions. The energy of the emission, which

3

109,110

arises from the MLCT state,

varies with both the nature

of the diimine ligand and the phosphine. For example, DAPTA

complexes give rise to higher emission energies in comparison

to the analogous THP complexes (Table 1). This result

indicates that the phosphine ligand plays a larger role in the

6

3,93,103−108

tion,

a state that dominates the photochemical

properties of Re(CO) complexes. Similarly, the energy of the

3

MLCT transition is almost independent of the nature of the

phosphine. This result suggests that the phosphine ligands only

weakly perturb the energies of metal-based donor orbitals, so

that variations in the MLCT bands can be attributed to

diﬀerences in the ligand π* orbitals. For the phen complexes (1

series), the MLCT band is centered at 367 nm; for the bpy

complexes (3 series) this band occurs at 345 nm, consistent

with the higher energy π* orbitals of this less conjugated ligand.

The luminescence properties of the complexes were also

evaluated (Figures S108 −S112 in the Supporting Information).

Representative emission spectra for complexes (PTA, THP,

DAPTA)-1 are shown in Figure 3. All complexes bearing the

THP and DAPTA ligands were emissive in air-equilibrated pH

3

1

MLCT in comparison to the MLCT state that was probed by

the absorption spectra, as described above. The excited state

lifetimes were also measured in both air-equilibrated (τ ) and

air

deoxygenated PBS (τ ) (Table 1 and Figures S113 −S117 in

N2

the Supporting Information), and these range from 0.2 to 2.0

μs for the air-equilibrated PBS solutions, consistent with triplet-

based MLCT excited state lifetimes measured for similar

6

2,93

rhenium complexes.

Notably, the lifetimes of these

complexes are extended under a nitrogen atmosphere,

suggesting that O is an eﬀective quenching agent.

2

Photolysis of the axial CO ligand was investigated by

monitoring the UV−vis spectral proﬁles of the compounds in

pH 7.4 PBS as they were irradiated with 365 nm light (Figures

S118 −S132 in the Supporting Information). Representative

spectral changes of PTA-1, THP-1, and DAPTA-1 upon

irradiation are shown in Figure 4. Whereas all complexes of

7.4 phosphate-buﬀered saline (PBS). The quantum yields for

emission under these conditions ranged from 3.4 to 6.1% for

the THP complexes, depending on the nature of the diimine

ligand. The DAPTA complexes were generally more emissive,

D

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

proceeded cleanly, as evidenced by the presence of well-deﬁned

isosbestic points. Furthermore, gas chromatographic analysis of

three representative compounds showed that 1 equiv of CO is

released per molecule (Table S3 in the Supporting

Information).

Photochemical quantum yields for these compounds were

measured via ferrioxalate actinometry and are given in Table 1.

The nonluminescent PTA complexes were generally not

photoreactive at this excitation wavelength; upper estimates

for their photoreaction quantum yields are less than 0.1%. In

contrast, the DAPTA and THP complexes exhibited quantum

yields ranging from 1.1 to 5.5%. These values are substantially

lower than those previously measured for related rhenium

6

2,68

phosphine complexes in degassed organic solvents;

the use

of aerated aqueous solutions in this case may lead to the lower

quantum yields measured for the compounds described here.

Because PDT agents act by generating singlet oxygen, we

1

investigated the O -sensitization quantum yields (Φ ) for the

2

Δ

1

5 rhenium complexes in air-equilibrated pH 7.4 PBS solutions

Table 2). Upon irradiation with 365 nm light, O was

1

(

2

quantiﬁed via the N,N-4-dimethyl-nitrosoaniline/histidine assay

5

9,111

following previously reported procedures.

yields were determined in reference to those of phenalenone

= 0.98). All PTA compounds exhibited no signiﬁcant

The quantum

(

Φ

Δ

1

production of O upon irradiation. This result suggests that

2

the excited state quenching of these complexes occurs in an O₂-

independent manner. The THP and DAPTA complexes

1

produce O with low to high quantum yields; DAPTA-2

2

exhibits the largest quantum yield of 70%, whereas both THP-4

1

and DAPTA-4 do not give rise to detectable quantities of O

2

upon irradiation.

Figure 2. UV−vis spectra in pH 7.4 PBS solutions for complexes

Reinvestigation of the Photophysical Properties of

THP-3. The photophysical properties of the compound THP-3

were described in a previous publication from one of our

(

PTA, THP, DAPTA)-1 (top) and (PTA, THP, DAPTA)-3

bottom).

68

laboratories. In this earlier study, a luminescence quantum

yield of 18% and a photochemical quantum yield of 21% were

reported. Additionally, the dicarbonyl photoproduct of THP-3

was found to exhibit photoluminescence with a maximum

emission wavelength of 585 nm. Upon reanalysis of THP-3 in

this investigation, we have found substantially lower lumines-

cence and photochemical reaction quantum yields of 6.1% and

1

.6%, respectively. Furthermore, the dicarbonyl photoproduct,

obtained by exhaustive photolysis of THP-3, was found to be

nonemissive. These contradictory results prompted further

investigations of this compound in our laboratories both at

Cornell and at UC Santa Barbara.

As originally reported, THP-3 was isolated as the triﬂate

68

salt, rather than the TFA salt. The triﬂate salt of THP-3 was

independently prepared as previously described and charac-

terized by elemental analysis and X-ray crystallography, which

gave unit cell parameters that matched those in the previous

report. The photophysical properties were investigated using

conditions identical with those employed in the original study

with equipment and facilities at UC Santa Barbara. The

photoluminescence quantum yield was determined using

Figure 3. Emission spectra in pH 7.4 PBS (excitation wavelength 350

nm, concentration 30 μM) for complexes PTA-1 (solid blue line),

THP-1 (long-dash red line), and DAPTA-1 (short-dash green line).

PTA were not photoreactive at this excitation wavelength, the

THP and DAPTA complexes exhibited a marked change in

their UV−vis spectra. We note that, in contrast to our results,

PTA-1 was previously reported to undergo CO loss upon

rhodamine B (Φ = 65%) as a reference standard. Quantum

lum

yields in pH 7.4 PBS in deaerated and air-equilibrated solutions

were determined to be 9.7% and 8.0%, respectively (Figure

S133 in the Supporting Information). The latter value is in

reasonable agreement with that measured at Cornell for the

TFA analogue of THP-3 under similar conditions. Further-

more, after exhaustive photolysis, no photoluminescence was

detected, demonstrating that the triﬂate counterion also does

83

excitation at 365 nm. We only observed photoreactivity of the

PTA complexes when they were irradiated with higher energy

UVB light (280−300 nm). Upon irradiation of the THP and

DAPTA complexes, the bands near 340−360 nm decayed with

the concomitant formation of a new broad feature at lower

energies. In all cases, the photoconversion of these complexes

E

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 4. UV−vis spectra of the photoreactions of PTA-1 (110 μM, left), THP-1 (115 μM, middle), and DAPTA-1 (105 μM, right) in pH 7.4 PBS

over a 4 h irradiation period with 365 nm light.

Table 2. Singlet Oxygen Quantum Yields

that a simple calculation error may account for the error in the

original measurement. This also may be the case for the

luminescence quantum yield, although another source of such

an error would be a partially oxidized rhodamine B standard.

The presence of an impurity in the original samples of THP-3

compound

Φ , %

Δ

PTA-1

<1

THP-1

48.5 ± 0.5

54.0 ± 1

<1

DAPTA-1

PTA-2

68

may explain the claimed photoproduct luminescence. A likely

+

candidate is fac-[Re(CO) (bpy)(OH )] , as this complex is

3

2

THP-2

45.5 ± 0.5

70.5 ± 2.5

<1

+

formed by aquation of the fac-[Re(CO) (bpy)(THF)] , the

3

DAPTA-2

PTA-3

starting material for THP-3. We have independently prepared

and characterized fac-[Re(CO) (bpy)(OH )](OTf) and shown

3

2

THP-3

8.7 ± 0.1

9.1 ± 0.3

<1

it to have an absorption spectrum close to that of THP-3

Figure S135 in the Supporting Information) and to exhibit

DAPTA-3

PTA-4

(

weak luminescence in the region of ∼580 nm (Figure S136 in

the Supporting Information). These results underline the

importance of conﬁrming the purity of all compounds prior to

photophysical measurements due to the sensitivity of these

measurements to small quantities of photoactive impurities.

Density Functional Theory. DFT calculations were

employed to gain a deeper understanding of the photophysical

properties of these compounds. The phen compounds, PTA-1,

THP-1, and DAPTA-1, were explored as a representative

subset of all of the complexes. Geometries were optimized at

THP-4

<1

DAPTA-4

PTA-5

<1

THP-5

15.0 ± 0

12.5 ± 0.5

DAPTA-5

not alter this aspect of THP-3. Finally, the photochemical

reaction quantum yield of this compound was reanalyzed using

a FieldMaxII-TO power meter from Coherent with a PM10 V1

thermopile photodetector (calibrated) to measure the light

intensities. These studies aﬀord a quantum yield of 2.4%

Figure S134 in the Supporting Information), in reasonable

agreement with the value of 1.6% obtained at Cornell. These

photoreaction quantum yields are approximately a factor of 10

112

the BP86 level of theory in the gas phase, and vertical singlet

excited state energies were calculated with time-dependent

DFT (TD-DFT) employing a solvent continuum dielectric

(

113

model for water and the hybrid functional PBE0. A map of

the lowest energy singlet excited states of these complexes is

68

smaller than the value previously reported, and we speculate

Figure 5. State energy diagram of PTA-1 (left), THP-1 (middle), and DAPTA-1 (right) and the corresponding electron density diﬀerence maps for

relevant excited states. Negative (white) and positive (red) electron density illustrates the transitions.

F

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 6. Single-dose (200 μM) cell viability of HeLa cells for all complexes. The compounds were tested in triplicate under both light and dark

conditions. For light-irradiated samples, an irradiation time of 30 min with 365 nm light at a photon ﬂux of (2.38 ± 0.31) × 10⁻¹⁰einstein/s was

employed.

given in Figure 5. For THP-1 and DAPTA-1, the lowest energy

singlet excited states are MLCT in character, as can be

discerned from the diﬀerence density maps that show a shift of

electron density from the metal to the phenanthroline ligand in

the excited state. In contrast, the four lowest energy singlet

excited states for PTA-1 cannot be reasonably described as

MLCT states; the diﬀerence density maps of these excited

states represent N 2p → L π* transitions or ligand-to-ligand

charge-transfers where the N 2p electrons are initially localized

on the PTA phosphine. The ﬁfth excited state of PTA-1 is the

deﬁnitive MLCT, as indicated by the diﬀerence density map.

Because the MLCT is generally the state which mediates the

luminescence and facilitates internal conversion to the

photoreactive ligand ﬁeld excited state, the presence of four

lower energy N 2p → L π* states may enhance nonradiative

and nonreactive decay pathways to the ground state. One can

alternatively think of this as a photoinduced electron transfer

treated with the 15 compounds at a single dose of 200 μM. The

cells were incubated for 4 h with the rhenium compounds and

then irradiated with 365 nm light for 30 min. Another batch of

cells, used as the nonirradiated control, was subjected to the

same conditions in the absence of light. The viability of the cells

in the presence of the 15 diﬀerent compounds under the

conditions of light irradiation and dark are shown in Figure 6

and Table S10 in the Supporting Information. All compounds

at this concentration level, with the exception of THP-2, THP-

5, and DAPTA-5, exhibited little or no toxicity in the dark

(within experimental uncertainty). Upon irradiation, all of the

THP and DAPTA complexes, with the exception of THP-3,

induced signiﬁcant cell death, indicating that these compounds

are potential PDT or PACT agents. Consistent with their lack

of photoreactivity, none of the PTA complexes exhibited

signiﬁcant cytotoxicity upon exposure to light.

On the basis of this initial screening, the most potent

compounds, THP-1, THP-2, THP-5, and all DAPTA

complexes, were further evaluated in full dose−response studies

to determine 50% growth inhibitory concentration (IC )

(PeT) quenching mechanism, which is commonly used to

design ﬂuorescent sensors via modulation of nitrogen lone pair

1

14

orbital energies. The eﬃcient quenching of the MLCT due

to the nitrogen lone pairs of the PTA complexes would,

therefore, explain the lack of observed luminescence for this

class of compounds. Furthermore, the lack of photosubstitution

5

0

values in HeLa cells. These values are collected in Table 3

(Figures S140 −S148 in the Supporting Information). Con-

sistent with the single-dose studies, the compounds were all

nontoxic in the absence of light. Under conditions of light

irradiation at 365 nm, the compounds exhibited enhanced

cytotoxicity. Notably, THP-2 and DAPTA-1 possess IC₅₀

values that are less than 10 μM. The ratio of IC₅₀values in

the dark over those determined upon irradiation give

phototoxicity indices (PI) that are >34 for DAPTA-1 (Figure

7) and >21 for THP-2. The exact phototoxicity indices could

not be determined because the compounds do not induce cell

death in the dark at the highest concentration (200 μM)

screened.

DAPTA-1 and THP-1 were further evaluated in wild-type

(A2780) and cisplatin-resistant ovarian (A2780CP70) cancer

cell lines (Table 4 and Figures S149 −S152 in the Supporting

Information). THP-1 exhibited enhanced cytotoxic eﬀects in

the presence of light in both A2780 and A2780CP70 cell lines.

The IC₅₀values of the irradiated THP-1, however, are

1

is also a consequence of the eﬃcient quenching of the MLCT

state, which must occur more rapidly than intersystem crossing

62

to states that would enable these photoreactivities. We note

that the DAPTA ligand likewise possesses an N 2p lone pair

that may also be available for PeT quenching. We hypothesized

that the presence of the electron-withdrawing acetyl groups

increase the redox potential of the nitrogen lone pair, such that

PeT quenching is less thermodynamically viable. This

hypothesis was further supported by examining the DFT

energies of the nitrogen lone pair orbitals of PTA-1 and

DAPTA-1. The energy of this orbital in PTA-1 is higher than

that of DAPTA-1 by approximately 0.3 eV. As such, the

nitrogen lone pair in DAPTA-1 is less susceptible to oxidation,

rendering PeT quenching less thermodynamically favored.

Photoinduced Anticancer Activity. To rapidly screen for

photoactivated in vitro anticancer activity, HeLa cells were

G

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Table 3. IC₅₀values in HeLa (Cervical Cancer) Cells in the

Absence and Presence of 365 nm Light

cytotoxicity by interacting covalently with biomolecules, in a

similar manner to that for the corresponding tricarbonyl aqua

78

species that we have studied. To explore this possibility, we

synthesized two dicarbonyl analogues, one of DAPTA-1 and

the other of PTA-3, for further evaluation of their anticancer

activity. The dicarbonyl analogue of DAPTA-1 was pursued

because DAPTA-1 is the most potent compound in our library.

In contrast, the dicarbonyl analogue of PTA-3 was investigated

because this compound exhibits no phototoxicity; this

compound would enable us to determine if the dicarbonyl

species could elicit a cytotoxic response in the absence of CO

IC₅₀value (μM) HeLa

a

compound

dark

>200

light

PI

THP-1

26.4 ± 9.2

5.9 ± 1.4

9.6 ± 4.2

19.2 ± 2.9

14.9 ± 3.2

60.3 ± 18.2

68.0 ± 4.3

24.3 ± 9.1

>7.6

>33.9

>20.8

>10.4

>13.4

>3.3

DAPTA-1

THP-2

>200

DAPTA-2

DAPTA-3

DAPTA-4

THP-5

>200

72.6 ± 23.2

>200

1.1

release. The target compounds fac-[Re(phen)(CO) (DAPTA)-

2

DAPTA-5

>8.2

Cl] (DAPTA-1A) and fac-[Re(bpy)(CO) (PTA)Cl] (PTA-

2

a

3

A) were synthesized in two steps from DAPTA-1 or PTA-3

The phototoxic index (PI) is the dark IC₅₀value divided by the light

IC₅₀value. For light-irradiated samples, an irradiation time 1 h with

65 nm light at a photon ﬂux (2.38 ± 0.31) × 10⁻¹⁰einstein/s was

115−118

(Scheme 3). Following previously reported procedures,

3

employed.

Scheme 3. Synthetic Scheme for the Photoproducts DAPTA-

A and PTA-3A

1

DAPTA-1 or PTA-3 was treated with trimethylamine N-oxide

TMAO) and NEt Cl and heated to reﬂux in a mixture of

(

4

CH Cl and CH OH. This reaction induces the liberation of

2

3

the axial CO ligand as CO , providing the open coordination

2

site that is then occupied by the chloride ions in solution to

yield DAPTA-1A or PTA-3A as red solids. The compounds

DAPTA-1A (Figures S16, S33, S65, and S82 in the Supporting

Information) and PTA-3A (Figures S17, S34, S66, and S83 in

the Supporting Information) were characterized by conven-

tional spectroscopic methods. The chlorido complexes were

suﬃciently soluble for biological studies after stirring them in

water for several hours; their dissolution is hypothesized to

occur with concomitant aquation of the chlorido ligand.

Photolyzed solutions of DAPTA-1 give rise to a peak with

the same retention time on HPLC as DAPTA-1A (Figure S153

in the Supporting Information). Furthermore, solutions of

photolyzed DAPTA-1 exhibit the same UV−vis (Figure S154

Figure 7. Dose−response curves for DAPTA-1 in HeLa cells in the

presence (red dashed line) and absence (black solid line) of 365 nm

light irradiation. For light-irradiated samples, an irradiation time of 1 h

with 365 nm light at a photon ﬂux of (2.38 ± 0.31) × 10 einstein/s

was employed.

−10

approximately 6-fold higher in the cisplatin-resistant cell line,

indicating that this photoactivated product is susceptible to

platinum resistance mechanisms. In contrast, DAPTA-1 gave

rise to light-induced IC values of 2.1 and 3.2 μM in the A2780

1

in the Supporting Information) and H NMR spectra (Figure

5

0

and A2780CP70 cell lines, indicating that this compound can

circumvent cisplatin resistance.

Identifying the Cytotoxic Species. Upon photolysis of

S155 in the Supporting Information) as DAPTA-1A,

conﬁrming this compound to be the photoproduct.

From the photolysis of DAPTA-1 in D O (Figure S155 in

2

this class of compounds in water, both CO and the rhenium

the Supporting Information), single crystals of an aqua

+

dicarbonyl complexes fac-[Re(NN)(PR )(CO) (OH )] are

analogue of DAPTA-1A, fac-[Re(phen)(DAPTA)-

3

2

+

released, and singlet oxygen is produced. We hypothesized that

the resulting dicarbonyl rhenium photoproduct could induce

(CO) (OD )] (DAPTA-1A-aqua), precipitated. Crystals of

2

PTA-3A were obtained by vapor diﬀusion of diethyl ether into

Table 4. IC₅₀Values in A2780 (Ovarian) and A2780CP70 (Ovarian Cisplatin-Resistant) Cancer Cells in the Absence and

Presence of 365 nm Light

IC₅₀value (μM)

A2780

light

A2780CP70

light

a

compound

dark

>200

PI

dark

>200

PI

THP-1

4.6 ± 1.4

2.2 ± 1.1

43.5

90.9

29.9 ± 7.7

3.2 ± 0.7

6.7

DAPTA-1

cisplatin

>200

62.5

0.18 ± 0.07

5.14 ± 1.1

a

Phototoxic index (PI) is the dark IC value divided by the light IC₅₀value. For light-irradiated samples, an irradiation time 1 h with 365 nm light at

50

−10

a photon ﬂux (2.38 ± 0.31) × 10 einstein/s was employed.

H

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

Figure 8. X-ray crystal structures of DAPTA-1A-aqua (left) and PTA-3A (right). Bond lengths and angles are reported in Table S2 in the

Supporting Information. Ellipsoids are drawn at the 50% probability level. H atoms bound to C atoms and counterions are omitted for clarity.

N,N-dimethylformamide (DMF). The structures of these

complexes were determined by single-crystal X-ray diﬀraction

(

0

Figure 8). The equatorial Re−CO distances are approximately

.06 Å shorter than those observed in DAPTA-1 and 0.03 Å

shorter than those in PTA-3. The decrease in distance

presumably arises from the loss of a π-accepting CO ligand.

PTA-3A also exhibits a notably smaller Cl−Re−P angle of

168.77(3)°, signiﬁcantly deviating from linearity. The O−Re−P

distance in DAPTA-1A-aqua is also considerably bent at an

angle of 171.70(9)°. This axial ligand bend represents an

interesting deviation from the corresponding angles found in

DAPTA-1 and PTA-1 and other complexes of the type fac-

[

1

Re(CO) (NN)X], with C−Re−X angles between 174 and

3

1

02,119

78°.

A similar rhenium phosphine compound, fac-

[

Re(CO) (NN)(PR )Cl] (PR = P(OEt) ), was reported to

Figure 9. Dose−response curve for DAPTA-1A (black solid line) in

HeLa cells in the absence of light and DAPTA-1 (red dashed line)

under 365 nm light irradiation.

2

3

117

exhibit an Cl−Re−P angle of 173°.

With puriﬁed DAPTA-1A and PTA-3A in hand, the

anticancer activity of these photoproducts was analyzed.

PTA-3A (Figure S156 in the Supporting Information) was

nontoxic in HeLa cells up to 200 μM in the dark. DAPTA-1A

that the phototoxicity of the other compounds studied here

could arise in part from the dicarbonyl product.

The limited toxicity of these dicarbonyl complexes could

possibly be attributed to their diﬀerent cellular uptakes and

intracellular distributions in comparison to the tricarbonyl

phosphine precursors. Upon photolysis in aqueous solution,

dicarbonyl aqua complexes, such as DAPTA-1A-aqua (Figure

(

Figure 9 and Figure S157 in the Supporting Information)

exhibited a mild cytotoxic eﬀect at concentrations >50 μM.

Additionally, solutions of DAPTA-1 after its exhaustive

photolysis gave rise to a cytotoxic response similar to that of

DAPTA-1A (Figure S158 in the Supporting Information),

supporting the toxic eﬀects of this dicarbonyl complex. The

cytotoxicity of PTA-3A and DAPTA-1A were further

investigated under conditions of UVA light irradiation to

probe for additional phototoxicity of these complexes. Neither

complex, however, induced additional cytotoxic eﬀects in the

presence of light. Taken together, these results indicate that the

phototoxicity of DAPTA-1 could arise only in part from the

dicarbonyl photoproduct and that additional cytotoxicity is

elicited through a separate mechanism. The lack of cytotoxicity

of PTA-3A is surprising because related rhenium carbonyl

complexes with a labile coordination site bind covalently to

8

), are formed. On coordination of water to metal centers, the

pK decreases substantially. Depending on their pK values, the

a

photogenerated dicarbonyl complexes could exist either as the

charge-neutral hydroxides or as the cationic aqua complexes at

physiological pH. On the basis of charge, such species would be

expected to have diﬀerent cellular uptake and localization. To

investigate the relative abundances of these species, the pK_a

values of aquated solutions of DAPTA-1A, PTA-3A, and the

photoproduct of THP-1 were determined. These solutions

were titrated with base as the pH and UV−vis spectra were

monitored. Upon basiﬁcation, a shift in the absorbance spectra

from approximately 370 to 430 nm was observed. This shift

arises from the deprotonation of the aqua complex. On the

7

5−77

78,120−122

biomolecules

and induce cell death.

Although we

only investigated DAPTA-1A and PTA-3A, we hypothesize

basis of these data, the pK values for the dicarbonyl aqua

a

I

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

photoproducts of DAPTA-1, THP-1 , and PTA-3 are 9.51,

THP and DAPTA complexes exhibited relatively eﬃcient

quantum yields for luminescence, photosubstitution of CO, and

the production of singlet oxygen. The lack of photoreactivity of

the PTA complexes was attributed to quenching from the

nitrogen lone pair orbitals, via a photoinduced electron transfer

mechanism. The photoreactivities of the THP and DAPTA

complexes were leveraged to generate PDT or PACT agents

that simultaneously released CO and a labile rhenium complex

upon exposure to UV light. The complex DAPTA-1 exhibited

the most potent cytotoxic eﬀects upon light exposure in HeLa

cells and in wild-type and cisplatin-resistant ovarian cancer cells.

The dicarbonyl photoproduct, DAPTA-1A, was only moder-

1

0.07, and 8.75, respectively (Figures S159 −S164 in the

Supporting Information). These values are similar to those

measured for related monoaqua transition-metal com-

plexes,

1

23,124

and they indicate that at physiological pH the

compounds exist as the cationic aqua complexes.

The low cytotoxicity of the rhenium photoproducts,

DAPTA-1A and PTA-3A, prompted us to consider the

1

potential role of O and CO in mediating cell death. Nearly

2

all of the rhenium complexes in this study sensitize the

1

formation of highly toxic O (Table 2). No clear direct

2

correlation between Φ and the phototoxicity exist, however,

Δ

1

indicating that O sensitization is not the only mechanism of

ately cytotoxic, suggesting that an additional mechanism of cell

2

1

activity of these complexes. To further test this hypothesis, the

death may be operable, which could be attributed to O and

2

phototoxicity of DAPTA-1 was investigated under hypoxia,

the released CO. Because no direct correlations between

phototoxicity and both O₂sensitization and CO photo-

1

conditions in which O sensitizers are less eﬀective. HeLa cells

2

were treated with DAPTA-1 with a procedure similar to that

described above, except the cells were subjected to hypoxic

conditions (5% CO , 95% N ) before and after exposure to 365

nm light for 1 h. Under hypoxic conditions, DAPTA-1 induced

moderate phototoxicity characterized by an IC₅₀value of 20 ±

substitution reaction quantum yields could be discerned, we

believe that more complicated and subtle factors are in

operation for this class of compounds. Regardless, these

compounds represent a new type of combined PDT and

PACT agent that operates under a multiaction mechanism.

They can be easily synthesized, are water-soluble, and exhibit

high phototoxic indices. Although this class of compounds is

activated only by UVA light, they represent an important proof

of principle for the development of new phototoxic rhenium

compounds. Ongoing eﬀorts are being pursued to shift the

excitation wavelength of these complexes to the red where in

vivo tissue penetration would be deeper.

2

1

.5 μM (Figure S165 in the Supporting Information). This

value is somewhat larger than that measured under normoxic

conditions (Table 3), suggesting that O sensitization does

1

2

play a role in this compound’s mechanism of action. The lack of

a complete suppression of phototoxicity under these con-

ditions, however, does verify that other factors, namely the

dicarbonyl photoproduct and CO, may also contribute to the

observed activity.

A third possible contributor to the complexes’ phototoxicity

is CO. Understanding the biological eﬀects of CO is an active

area of research. CO at low concentrations can have therapeutic

properties, as it elicits anti-inﬂammatory eﬀects and vaso-

EXPERIMENTAL SECTION

■

Methods and Materials. Rhenium carbonyl was purchased from

Pressure Chemicals (Pittsburgh, PA, USA). Iodobenzene dichloride

was synthesized as previously reported. The diimine ligands 2,2′-

145

1

25−136

dilation.

The application of CO as a cytotoxic agent for

bipyridine (bpy), 4,4′-dimethyl-2,2′-dipyridyl (dmbpy), 4,4′-dime-

thoxy-2,2′-bipyridine (dmobpy), 1,10-phenanthroline (phen), and 2,9-

dimethyl-1,10-phenanthroline (dmphen) were purchased from Sigma-

Aldrich (St. Louis, MO, USA), Alfa Aesar (Heysham, England), or

Beantown Chemical (Hudson, NH, USA) and were used as received.

the destruction of malignant cells has also been ex-

7

2,79−81,84,137−141

plored.

CO most likely kills cells by binding

142

with high aﬃnity to cytochrome c oxidase, thereby inhibiting

the mitochondrial respiration pathway. Inhibition of the

mitochondrial respiration pathway in this manner with CO

reduces cancer cell proliferation by metabolically exhausting

[

Re(CO) (phen)Cl], [Re(CO) (dmphen)Cl], [Re(CO) (bpy)Cl],

3

Re(CO) (dmbpy)Cl], and [Re(CO) (dmobpy)Cl] were synthesized

3

8

8,89

1

43

using a previously reported procedure.

The phosphine ligands

cells. If the cytotoxic eﬀects of these complexes are mediated

solely by CO, then the photoactivated anticancer activity could

potentially correlate to the quantum yield for CO release.

Across the 15 compounds tested, however, we do not observe

this correlation, suggesting that additional factors are in

operation. One possible explanation is that the quantum yields

measured in buﬀer are not representative of those in the

complex biological media, where binding to proteins such as

albumin may alter the values in an unexpected manner. For

example, the photophysical properties of the organic dye Rose

Bengal and TPPS (meso-tetra(4-sulfanatophenyl)porphyrin) are

dramatically altered in the presence of albumin and the cell

1

,3,5-triaza-7-phosphaadamantane (PTA) and tris(hydroxymethyl)-

phosphine (THP) were purchased from Alfa Aesar (Heysham,

England) and Acros (Geel, Belgium), respectively, and were both

used as received. The compound 1,4-diacetyl-1,3,7-triaza-5-

phosphabicylco[3.3.1]nonane (DAPTA) was synthesized from PTA

using a previously reported procedure. All solvents were ACS grade

or higher. All reactions were carried out under ambient atmospheric

conditions without any eﬀort to exclude water or oxygen.

8

6

Physical Measurements. NMR samples were prepared as

solutions using MeOD-d , DMSO-d , DMF-d , or D O as the solvent.

4

6

7

2

NMR spectra were acquired on a Varian Inova 400 MHz spectrometer.

1

H NMR chemical shifts were referenced to residual solvent peaks

versus tetramethylsilane (TMS) at 0 ppm or an external standard, 1,4-

dioxane, at 3.75 ppm in D O. F and P NMR spectra were

referenced using an external standard of KPF in D O ( F, δ −72 ppm

vs CFCl at 0 ppm; P, δ −145 ppm vs H PO at 0 ppm). Samples for

IR spectroscopy were prepared as KBr pellets and were analyzed on a

Nicolet Avatar 370 DTGS (ThermoFisher Scientiﬁc, Waltham, MA,

USA). The HPLC system (LC-20AP) used for puriﬁcation of the Re-

THP complexes (excluding THP-5) consisted of a SPD-20AV UV−vis

detector monitored at 270 and 220 nm (Shimadzu, Japan) and a Epic

Polar preparative column (120 Å, 10 μm, 25 cm × 20 mm; ES

Industries, West Berlin, NJ, USA) at a ﬂow rate of 14 mL/min using a

binary mobile phase containing 0.1% triﬂuoroacetic acid (TFA) and

1

44

membrane environment.

Alternatively, diﬀerential cell

19

31

2

uptake and intracellular localization of these compounds due

to their diﬀerent lipophilicities may play a large role in

mediating the cytotoxic eﬀects of the photoreleased CO and

would explain the clear lack of a correlation between CO

photorelease quantum yield and cytotoxicity.

19

6

2

31

3

4

CONCLUSIONS

■

The structural and photophysical properties of 15 rhenium

tricarbonyl complexes bearing one of 3 water-soluble

phosphines and 5 diﬀerent diimine ligands were investigated.

H O. The method for puriﬁcation consisted of 0−5 min at 10%

2

J

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

7

1

MeOH, followed by a linear gradient of 100% MeOH for over 20 min,

and then 10% MeOH for 1 min. Analytical chromatography was

carried out on a LC-20AT pump with a SPD-20AV UV−vis detector

monitored at 270 and 220 nm (Shimadzu, Japan) using an Ultra

Aqueous C18 column (100 Å, 5 μm, 250 mm × 4.6 mm; Restek,

Bellefonte, PA, USA) at a ﬂow rate of 1 mL/min with a mobile phase

Schlenk cuvette (internal volume of 23.6 mL) was loaded with 6 mL

of the stock solution and a magnetic stir bar and sealed. The UV−vis

spectrum was measured using a Shimadzu UV2401 (PC) spectropho-

tometer. The sample was then exposed to 365 nm light using an Oriel

200−500 W/Hg arc lamp ﬁltered by a 365 nm mercury line

interference ﬁlter. The power of the incident light (50 mW) was

measured using a Coherent FieldMaxII-TO Power Meter. After 1 h,

the UV−vis spectrum was measured again. This was repeated until

there were no further signiﬁcant absorbance changes at the λmax (∼400

nm) of the product, indicating that the photoreaction was complete

(∼3 h).

containing 0.1% triﬂuoroacetic acid (TFA) in H O or MeOH. The

2

method consisted of 5 min at 10% MeOH, followed by a linear

gradient to 100% MeOH over 20 min. High-resolution mass spectra

(

HRMS) were recorded on an Exactive Orbitrap mass spectrometer in

positive ESI mode (ThermoFisher Scientiﬁc, Waltham, MA, USA)

with samples injected as acetonitrile/water solutions with 1% formic

acid. Elemental analyses (C, H, N) were performed by Atlantic

Microlab Inc. (Norcross, GA, USA). UV−visible spectra were

recorded on a Cary 8454 UV−vis instrument (Agilent Technologies,

Santa Clara, CA) or a Beckman Coulter DU800 UV−vis instrument

using 1 cm quartz cuvettes. Lifetime measurements were collected as

described below. Phototoxicity experiments were conducted using a

UVA handlamp elevated 8 cm from the cells to give a photon ﬂux of

A gastight syringe was used to remove a 100 μL sample from the

headspace of the cell through a special screw top ﬁtted with a three-

layer laminated silicone GC septum. This sample was analyzed by gas

chromatography with thermal conductivity detection (GC-TCD)

using an Agilent 6890N gas chromatograph ﬁtted with a Carboxen

1010 PLOT fused silica capillary column (L × I.D. = 30 m × 0.53 mm,

average thickness 30 μm). The GC inlet temperature was 230 °C. The

oven was held at 35 °C for 14 min, then ramped at 20 °C/min to 245

°C, and ﬁnally held at 245 °C for 10 min. The retention time for

carbon monoxide (CO) is about 11.2 min, and that for carbon dioxide

−10

(

2.38 ± 0.31) × 10 einstein/s at 365 nm. Analytical photochemical

measurements were performed using a Newport mercury/xenon arc

lamp. The light output was modulated using a combination of a

Newport heat-absorbing glass ﬁlter (50.8 × 50.8 mm) with infrared

cutoﬀ (Schott KG5 ﬁlter glass), a Newport mercury line bandpass ﬁlter

(CO ) is 21 min. Dioxygen and dinitrogen from ambient air have

2

retention times of about 8.2 and 8.4 min, respectively. The

ChemStation autointegration function was used to calculate the peak

area for CO.

(

25.4 mm, center wavelength 365.0 ± 2 nm), and a Newport visible

absorbing ﬁlter (50.8 × 50.8 mm, center wavelength 340 nm) made of

dark optical glass to isolate monochromatic 365 nm light.

Luminescence quantum yield measurements were carried out on a

Beckman Coulter DU800 UV−vis instrument and Varian Eclipse

Fluorometer.

X-ray Crystallography. Single crystals were grown using vapor

diﬀusion of diethyl ether into solutions of methanol or acetonitrile. For

the compound DAPTA-1A, the crystals were isolated from a solution

The amount of CO in the 100 μL sample was calculated using a

calibration curve made by injecting known amounts of CO from a

Schlenk ﬂask. The moles of CO in the headspace of the Schlenk

cuvette (SC moles of CO) was calculated using the equation

SC moles of CO

moles of CO (syringe) × (SC headspace + injection volume)

=

injection volume

of D O, and those for PTA-3A were collected from vapor diﬀusion of

2

(1)

diethyl ether into DMF. Low-temperature (223 K) X-ray diﬀraction

data for PTA-1, THP-1, DAPTA-1, PTA-2, THP-2, DAPTA-2, PTA-

The volume of the headspace for the Schlenk cuvette (SC headspace)

is calculated by subtracting the solution volume from the total volume

of the Schlenk cuvette.

3, DAPTA-3, PTA-4, THP-4, DAPTA-4, PTA-5, THP-5, DAPTA-5,

DAPTA-1A, and PTA-3A were collected on a Bruker X8 Kappa

diﬀractometer coupled to an ApexII CCD detector with graphite-

monochromated Mo Kα radiation (λ = 0.71073 Å). The structures

For CO, the partition coeﬃcient between the gas phase and water is

∼50 at ambient temperature (41 versus 0.84 mmol/(L atm)); thus,

essentially all of the CO will be in the gas phase. The stoichiometric

ratio (moles CO released/moles of photoreacted complex) was

calculated by dividing the SC moles of CO by the moles of complex

photolyzed. The average and sample standard deviation of three

injections are reported. The results are shown in Table S3 in the

Supporting Information.

146

were solved through intrinsic phasing using SHELXT and reﬁned

2

147

against F on all data by full-matrix least squares with SHELXL

148

following established reﬁnement strategies. All non-hydrogen atoms

were reﬁned anisotropically. Hydrogen atoms bound to carbon were

included in the models at geometrically calculated positions and

reﬁned using a riding model. Hydrogen atoms bound to oxygen were

located in the diﬀerence Fourier synthesis and subsequently reﬁned

semifreely with the help of distance restraints. The isotropic

displacement parameters of all hydrogen atoms were ﬁxed to 1.2

times the U value of the atoms they are linked to (1.5 times for methyl

groups). Details of the data quality and a summary of the residual

values of all the reﬁnements are given in Tables S1 and S2 in the

Supporting Information.

Emission Quantum Yield. The luminescence quantum yields

were measured relative to the standard quinine sulfate (Φ = 0.52, 0.05

M H SO ), which was cross-referenced in our laboratory to harmaline

2

4

1

50

(

Φ = 0.32, 0.005 M H SO ). An excitation wavelength of 350 nm

was used for the samples and standards. The compounds were

measured as solutions in pH 7.4 PBS with the absorbance maintained

2

4

1

50

below 0.1 to prevent inner ﬁlter eﬀects.

At least ﬁve diﬀerent

concentrations of the samples and standards were measured by UV−

vis and ﬂuorescence spectroscopy, and the absorbance at 350 nm was

plotted versus the integrated emission intensity. The slopes of the

resulting lines were used in the equation

Photochemistry. Photochemical reactions were monitored by

UV−vis spectroscopy. Stock solutions of the rhenium compounds at

−2 mM concentrations in PBS (pH 7.4) were diluted in PBS (pH

.4) to a ﬁnal volume of 3 mL at a concentration of approximately 40−

50 μM. The absorbance of these samples was between 0.1 and 0.5 at

65 nm. The solutions were stirred in a 1 cm quartz cuvette and

1

7

1

3

2

S

η

sample sample

2

^Φsample = Φ_r_e_f

irradiated with 365 nm light using the UVA handlamp described

above. At diﬀerent time points, the UV−vis spectra were acquired. The

photoreaction was deemed complete when no further changes in the

UV−vis spectra were observed. For DAPTA-1, the photoreaction was

monitored by HPLC (Figure S153 in the Supporting Information),

S η

ref ref

(2)

where Φ is the quantum yield of the reference, quinine sulfate, S is

ref

the slope of either the sample or the reference, and η is the refractive

index.

1

UV−vis (Figure S154 in the Supporting Information) and H NMR

Photochemical Reaction Quantum Yield. The photochemical

spectroscopy (Figure S155 in the Supporting Information).

reaction quantum yields were determined via potassium ferrioxalate

1

51

CO Release Measurements. These measurements were carried

actinometry.

All solutions were prepared with minimal light

149

out as previously reported. Millimolar stock solutions of DAPTA-1,

DAPTA-4, and THP-3 were prepared by dissolving a known amount

of the respective complex in 10 mL of pH 7.4 PBS. A custom designed

exposure and were stored in the dark. To determine the photon ﬂux

of our lamp, a 1.5 mL volume of potassium ferrioxalate (6.1 mM, 0.05

M H SO ) was irradiated with monochromatic 365 nm light for 30 s

2

4

K

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

with continuous stirring in a 1 cm quartz cuvette. After irradiation, a

0

1

.1 mL aliquot of the ferrioxalate solution was added to 2.9 mL of

,10-phenanthroline (5.7 mM, 0.5 mM H SO ) containing sodium

2

4

acetate buﬀer (1.8 M, pH 5). This procedure was repeated for an

identical potassium ferrioxalate solution kept in the dark to correct for

any thermal reactions. The resulting solution was shaken for 30 s and

then incubated for 1 min. The absorbance was then recorded for both

the irradiated solution and the dark solution. The equation used to

determine photon ﬂux is

ΔAVV N

1

2 A

^q⁼1

000(ΦεV tl)

3

(3)

where N is Avogadro’s number, q is photon ﬂux (photons/s), ΔA is

A

the change in absorbance at 510 nm for the irradiated solution minus

the dark solution, V (mL) is the total volume after dilution, V (mL)

of the compound at the irradiation wavelength. The standard

1

2

is the volume irradiated, V (mL) is the volume of irradiated sample

phenalenone was cross-referenced to the complex fac-[Re-

3

+

diluted into 1,10-phenanthroline, Φ is the quantum yield at 365 nm, ε

(CO) (NNN)] , where NNN = N,N-bis(quinolin-2-ylmethyl)-

3

−1

59

is the extinction coeﬃcient (M cm ), l is path length (cm), and t is

time (s). For potassium ferrioxalate at 365 nm, Φ = 1.25 and ε =

amino]-5-valeric acid (Re-COOH), reported by Leonidova et al.

They reported a singlet oxygen quantum yield of 20% in water. Using

our experimental setup, we measured a value of 33−37%.

−1

−1 151

1

1000 M cm .

Photon ﬂux and quantum yields were cross-

referenced to the actinometer, Reinecke’s salt (Φ = 0.38 at 365 nm),

DFT Studies. All computational calculations were performed using

151

153

to verify the accuracy of our measurements.

version 3.0.3 of the ORCA computational package. The resolution

The quantum yields of the rhenium compounds were determined

by ﬁrst measuring the photon ﬂux of the light source with potassium

ferrioxalate, as described above. The rhenium compound was

irradiated in pH 7.4 PBS with the monochromatic 365 nm light.

The concentrations of these samples were such that the absorbance

was greater than 2 at 365 nm, ensuring that all photons were absorbed

of identity (RI) approximation was used throughout for computational

eﬃciency. Geometry optimizations were performed in the gas phase

1

12

using the functional BP86, the basis set def2-TZVP and its auxiliary

1

54

def2-TZV/J,

and the zeroth-order regular approximation for

1

55−157

relativistic eﬀects (ZORA).

Optimized geometries were

subjected to TD-DFT calculations to determine the nature of the

−A

113

by the sample. For some samples, a correction factor (1 − 10 ) was

applied to the photon ﬂux if the solution was below an absorbance of

relevant excited states. For these calculations, the hybrid functional

1

58

PBE0 was employed,

using the same basis sets and ZORA as

2

, where A is the absorbance of the sample at the irradiation

described above. For the TD-DFT calculations, the SMD solvation

1

59

wavelength, 365 nm. This correction factor accounts for the fact that in

more optically dilute solutions not all photons may be absorbed. After

a given time period of irradiation, the sample was diluted in 2.5 mL of

PBS and the UV−vis absorbance was measured. This diluted sample

was subjected to continued irradiation for 0.5−2 h until the

photoreaction was complete, enabling us to determine the extinction

coeﬃcient of the photoproduct. To determine the quantum yield of

the rhenium compounds, eq 3 was applied to determine Φ, using the

photon ﬂux determined from the potassium ferrioxalate actinometry.

Absorbance change was monitored at 425−450 nm, depending on the

sample.

Lifetime Measurements. Laser excitation for the phosphor-

escence lifetime measurements was provided by pulsing the 405 nm

laser line from a four-line iChrome MLE laser (Toptica Photonics AG,

Munich, Germany). The diode laser in the iChrome was triggered by a

DG535 Digital Delay/Pulse Generator (Stanford Research, Sunnyvale,

CA, USA) at 100 kHz and delivered 100 ns fwhm 405 nm excitation

pulses. The 405 nm pulses were ﬁber-delivered to a sample-ﬁlled

cuvette, and phosphorescence was collected at 90° through a second

ﬁber for delivery to a Bialkali photomultiplier tube (HC125,

Hamamatsu, Bridgewater, NJ, USA) through a 470 nm long-pass

ﬁlter (HQ470lp, Chroma Technology, Bellows Falls, VT). The time-

resolved photon counts were collected in 40 ns time bins using a

SR430 Multichannel scaler (Stanford Research, Sunnyvale, CA). Data

were transferred to a PC via the SR430 GPIB bus and ﬁt to the

standard exponential decay model using MagicPlot Pro software.

Measurements were collected in PBS solutions at 1000 μM. For

deoxygenated measurements, nitrogen gas was bubbled into the PBS

solutions for 15 min, and then the lifetime was determined.

model for water was employed. Twenty singlet excited states were

calculated for PTA-1, THP-1 , and DAPTA-1. All optimized geometry

coordinates are given in Tables S4 −S6 , the lowest energy singlet

excited states are summarized in Tables S7 −S9 , and the frontier

Kohn−Sham orbitals are located in Figures S137 −S139 in the

Supporting Information.

Cell Culture and Cytotoxicity. The HeLa (cervical cancer) cell

line was obtained from American Type Culture Collection (ATCC)

and cultured using Dulbecco’s Modiﬁed Eagle’s Medium (DMEM)

supplemented with 10% fetal bovine serum (FBS). A2780 (ovarian

cancer) and A2780CP70 (cisplatin-resistant ovarian cancer) cell lines

were provided by the Cell Culture Facility of Fox Chase Cancer

1

60

Center

(Philadelphia, PA, USA). These cells were cultured as

monolayers with Roswell Park Memorial Institute (RPMI)-1640

culture media supplemented with 10% FBS. All cell lines were grown

in a humidiﬁed incubator at 37 °C with an atmosphere of 5% CO₂.

Cells were passed at 80−90% conﬂuence using trypsin/EDTA. Cells

were tested monthly for mycoplasma contamination with the

PlasmoTest mycoplasma detection kit from InvivoGen.

The procedure for determining phototoxicity was modiﬁed from the

OECD Guidelines, except MTT was used instead of Neutral Red to

1

61

determine cell viability. All compounds, except for DAPTA-1A and

PTA-3A, were dissolved in PBS at pH 7.4 to prepare a 1−2 mM stock

solution. DAPTA-1A and PTA-3A were solubilized in water after

exhaustive stirring. For single-dose studies, only HeLa cells were used.

For both the single-dose and full-dose responses all cells were grown

to 80−90% conﬂuence, detached with trypsin/EDTA, seeded in 96-

well plates at 2000 cells/well for HeLa cells and 4000 cells/well for

A2780 and A2780CP70 cells in 100 μL of growth media, and

incubated for 24 h. The medium was removed and replaced with fresh

medium (200 μL) containing a single concentration of PTA, THP, or

DAPTA-1−5. The cells were incubated for 4 h, and then the cells were

irradiated with 365 nm light for 30 min (one plate under the same

dosing conditions was kept outside the incubator in the dark as a

control). After exposure to light, the plates were incubated for an

additional 44 h, the medium was removed from the wells, and 3-(4,5-

dimethylthiazol-2-yl)-2,5-tetrazolium bromide (MTT) in DMEM or

RPMI (200 μL, 1 mg/mL) was added. The additional 44 h incubation

Singlet Oxygen Measurements. Indirect singlet oxygen

measurements were performed using the N,N-dimethyl-4-nitrosoani-

line/histidine assay, where histidine is oxidized by singlet oxygen and

the oxidized histidine reacts with N,N-dimethyl-4-nitrosoaniline. All

measurements were performed in PBS solutions containing N,N-

dimethyl-4-nitrosoaniline (25 μM) and histidine (10 mM). The

reaction mixture was irradiated with monochromatic 365 nm light in 1

cm glass cuvettes at 2 min increments, and the bleaching of N,N-

dimethyl-4-nitrosoaniline was monitored at 440 nm. The absorbance

1

11

5

9

L

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

was performed to ensure that the cells were in the logarithmic growth

phase and that the cells had adequate time to regrow after exposure to

the complexes. After 4 h, the MTT/DMEM or RPMI solution was

removed, and the formazan crystals were dissolved in 200 μL of an 8/1

mixture of DMSO and pH 10 glycine buﬀer. The absorbance at 570

nm in each well was measured using a BioTek Synergy HT plate

reader. Cell viability was determined by normalizing the absorbance of

the treated wells to untreated wells. The percent viability data shown

are an average of three independent experiments with six replicates per

concentration. The same procedure was followed as above for the full

dose response, but a 1 h irradiation period was employed instead. Cell

viability was again determined by normalizing the absorbance of the

treated wells to untreated wells. The concentrations of the compounds

versus percent viability were plotted to produce the dose−response

curves, which were analyzed using a logistic sigmoid function ﬁt with

MagicPlot Pro software. The reported IC₅₀values are the average of

three independent experiments with six replicates per concentration

level. For hypoxic conditions, the cells were dosed with the rhenium

complex and then incubated in a hypoxia chamber (Billips-

Rothenburg, Inc., Del Mar, CA, USA), which was purged with an

atmosphere of 95% N and 5% CO for 4 h. The chamber was

dissolved in a minimal amount of methanol (5−10 mL) and ﬁltered

through Celite. The addition of diethyl ether (20−30 mL) precipitated

the desired product as a yellow solid, which was isolated by ﬁltration

1

and dried in vacuo. Yield: 0.298 g (78%). H NMR (400 MHz,

MeOD-d ): δ 9.53 (d, 2H, J = 5.2 Hz), 9.00 (d, 2H, J = 8.3 Hz), 8.34

4

(s, 2H), 8.15 (dd, 2H, J = 5.2, 8.3 Hz), 4.38 (d, 3H, J = 13.1 Hz), 4.26

1

9

(d, 3H, J = 13.1 Hz) 3.64 (s, 6H). F NMR (376 MHz, MeOD-d ,

4

3

1

external standard KPF ): δ −79.56. P{ H} NMR (162 MHz, MeOD-

6

−1

d , external standard KPF ): δ −78.24. IR (KBr, cm ): 3460 m, 2910

4

6

w, 2360 m, 2330 m, 2030 s, 1940 s, 1920 s, 1430 m, 1265 m, 1244 m,

1153 m, 1029 m, 970 w, 950 w, 850 w, 726 w, 637 m, 440 w. ESI-MS

(positive ion mode, CH CN/H O 70/30 and 1% formic acid): m/z

3

2

+

608.08745 ([M] , calcd 608.08558). Anal. Calcd for PTA-1·CH OH

(C H F N O PReS): C, 35.03; H, 3.07; N, 8.88. Found: C, 34.85; H,

3

23

24

3

5

7

3.02; N, 8.81.

Synthesis of [Re(CO) (phen)(THP)]TFA (THP-1). A mixture of

3

[Re(CO) (phen)Cl] (0.100 g, 0.206 mmol) and AgOTf (0.053 g,

3

0.206 mmol) in THF (24 mL) was heated to reﬂux at 75 °C for 3 h in

the dark. The white solid AgCl was removed via vacuum ﬁltration. To

the remaining yellow ﬁltrate was added THP (0.038 g, 0.309 mmol),

and the solution was heated to reﬂux for an additional 15 h. The

reaction mixture was cooled to room temperature, and then THF was

removed by rotary evaporation. The remaining yellow solid was

dissolved in a minimum volume of methanol (5 mL) and ﬁltered

through Celite. The crude yellow oil (0.167 g) was puriﬁed using

preparatory RP-HPLC using a CH OH/H O gradient (0−5 min, 10%

2

removed from the incubator, purged with the hypoxic gas mixture

again, and then irradiated with 365 nm light for 1 h. After this

irradiation period, the cells were further incubated under hypoxia at 37

°

C for 16 h. The plates were then removed from the hypoxia chamber

and allowed to recover under normoxic (5% CO , 21% O ) conditions

2

3

2

at 37 °C for an additional 24 h prior to analysis of cell viability with

MTT, as described above. As an additional control experiment to test

the eﬀect of the UVA light on the cells, cells in the absence of rhenium

complexes were subjected to the same irradiation conditions. The cell

viability of the irradiated cells, relative to control cells that were not

exposed to light, was always >85%, indicating that the direct light

exposure alone has only a minimal eﬀect on the cells.

CH OH; 5−30 min, 10−100% CH OH). Puriﬁed yield: 0.100 g

3

1

(69%). H NMR (400 MHz, MeOD-d ): δ 9.56 (d, 2H, J = 5.2 Hz),

4

8.90 (d, 2H, J = 8.3 Hz), 8.25 (s, 2H), 8.07 (dd, 2H, J = 5.2, 8.3 Hz),

3.75 (t, 2H), 3.75 (s, 6H). ¹⁹F NMR (376 MHz, MeOD-d , external

4

31

1

standard KPF ): δ −76.37. P{ H} NMR (162 MHz, MeOD-d ,

external standard KPF ): δ 2.43. IR (KBr, cm ) 3420 m, 3210 m,

6

4

−1

6

3070 m, 2830 w, 2780 w, 2650 w, 2040 s, 1930 s, 1910 w, 1670s, 1520

w, 1430 s, 1200 s, 1130 s, 1050 s, 888 m, 856 s, 800 s, 723 s, 635 w,

618 w, 538 m, 500 m, 483 w, 436 w. ESI-MS (positive ion mode,

pK Measurements. Stock solutions of DAPTA-1, THP-1, and

a

PTA-3A were made in water at approximately 2 mM and then diluted

to 84−89 μM in water. DAPTA-1 and THP-1 were irradiated with

+

CH CN/H O 70/30 and 1% formic acid): m/z 575.03961 ([M] ,

3

2

3

65 nm light until the photoreaction was complete as evidenced by

calcd 575.03763). Anal. Calcd for THP-1·H O (C H F N O PRe):

2 20 19 3 2 9

UV−vis spectroscopy. PTA-3A was dissolved directly for analysis. The

solutions containing the rhenium complexes were acidiﬁed to an

approximate pH of 2.0 with perchloric acid. These solutions were

titrated with aqueous KOH, and the pH and UV−vis absorbance

spectra were measured at each stage of the titration. The titration

apparatus consisted of a 100 mL water-jacketed glass vessel at 25 °C.

Nitrogen gas was passed through the solution containing 30% KOH

C, 34.05; H, 2.71; N, 3.97. Found: C, 34.02; H, 2.61; N, 4.03.

Synthesis of [Re(CO) (phen)(DAPTA)]OTf (DAPTA-1). A

3

mixture of [Re(CO) (phen)Cl] (0.200 g, 0.42 mmol) and AgOTf

3

(0.106 g, 0.42 mmol) in THF (47 mL) was heated to reﬂux at 75 °C

for 3 h in the dark. The white solid AgCl was removed via vacuum

ﬁltration. To the remaining yellow ﬁltrate was added DAPTA (0.113 g,

0.46 mmol), and the solution was heated to reﬂux for an additional 15

h. The reaction mixture was cooled to room temperature, and THF

was removed by rotary evaporation. The remaining yellow solid was

dissolved in a minimum amount of methanol (5−10 mL) and ﬁltered

through Celite. The addition of diethyl ether (20−30 mL) aﬀorded the

desired product as a yellow solid, which was isolated by vacuum

during the experiment to exclude any CO . The electrode was

2

calibrated using standard buﬀer solutions prior to the experiment.

Carbonate-free solutions of KOH were prepared in freshly boiled

Milli-Q water and standardized against potassium hydrogen

162

phthalate.

The spectral changes resulting from the diﬀerent pH

1

values were used to determine the pK values using the HypSpec

ﬁltration and dried in vacuo. Yield: 0.255 g (73%). H NMR (400

a

162,163

software package.

Species distribution diagrams were plotted

MHz, MeOD-d ): DAPTA complexes show the syn and anti isomers

4

164

using HySS. The titration curves were ﬁt from a wavelength range

of 360 −405 nm. All absorbance spectra and titration curves are in

Figures S159 −S164 in the Supporting Information.

in a 1/5 ratio; however in the aromatic region the isomers are

indistinguishable, δ 9.60 (t, 2H, anti + syn), 9.02 (d, 2H, anti + syn),

8.35 (s, 2H, anti + syn), 8.18 (m, 2H, anti + syn), 5.56 (d, anti), 5.04

(d, syn), 4.75 (d, anti) 4.48 (d, syn), 4.41 (d, anti), 4.12 (m, anti), 3.93

(d, anti), 3.78 (d, anti), 3.64 (s, syn), 3.41 (d, syn), 3.34 (d, anti), 3.04

Synthesis of Re(CO) Cl. A mixture of rhenium carbonyl

5

Re (CO) (1.90 g, 2.91 mmol) and CH Cl (20 mL) was stirred

2

10

2

1

9

until Re (CO) was completely dissolved. Then iodobenzene

(d, syn), 2.95 (d, anti) 1.94 (s, syn), 1.91 (s, anti), 1.80 (s, anti). F

2

10

3

1

dichloride (0.711 g, 2.59 mmol) was added as a solid to the solution,

and the reaction mixture was stirred for 4 h at room temperature. The

desired product, a white solid, was ﬁltered and washed with hexanes

(376 MHz, MeOD-d , external standard KPF ): δ −79.54. P{ H}

4

6

(162 MHz, MeOD-d , external standard KPF ): δ −44.80 ((syn),

4

6

−1

−48.75 (anti). IR (KBr, cm ): 3452 m, 2038 s, 1948 s, 1914 s, 1643 s,

1427 m, 1330 m, 1278 m, 1258 m, 1225 w, 1148 w, 1029 w, 987 w,

888 w, 849 w, 802 w, 720 w, 638 s, 540 m, 498 w, 425 w. ESI-MS

(positive ion mode, CH CN/H O 70/30 and 1% formic acid): m/z

−

1

(

2

20−30 mL). Yield: 1.268 g (68%). IR (KBr, cm ): 3435 w, 2462 w,

157 m, 2035 s, 1958 s, 1002 w, 942 w, 912 w, 590 s, and 555 s.

Synthesis of [Re(CO) (phen)(PTA)]OTf (PTA-1). A mixture of

3

2

+

[

Re(CO) (phen)Cl] (0.224 g, 0.485 mmol) and AgOTf (0.125 g,

680.10846 ([M] , calcd 680.10671). Anal. Calcd for DAPTA-1

3

0

.485 mmol) in THF (50 mL) was heated to reﬂux at 75 °C for 3 h in

(C H F N O PReS): C, 36.23; H, 2.92; N, 8.45. Found: C, 36.35; H,

2.79; N, 8.66.

23

24

3

5

8

the dark. The white solid AgCl was removed via vacuum ﬁltration. To

the remaining yellow ﬁltrate was added PTA (0.0915 g, 0.582 mmol),

and the resulting solution was heated to reﬂux for an additional 15 h.

The reaction mixture was cooled to room temperature, and the THF

was removed using rotary evaporation. The remaining yellow solid was

Synthesis of [Re(CO) (dmphen)(PTA)]OTf (PTA-2). A mixture

3

of [Re(CO) (dmphen)Cl] (0.243 g, 0.473 mmol) and AgOTf (0.121

3

g, 0.473 mmol) in THF (54 mL) was heated to reﬂux at 75 °C for 3 h

in the dark. The white solid AgCl was removed via vacuum ﬁltration.

M

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

To the remaining yellow ﬁltrate was added PTA (0.111 g, 0.71 mmol),

and the solution was heated to reﬂux for an additional 15 h. The

reaction mixture was cooled to room temperature, and then the THF

was removed using rotary evaporation. The remaining yellow solid was

dissolved in a minimum amount of methanol (5−10 mL) and ﬁltered

through Celite. The addition of diethyl ether (20−30 mL) aﬀorded the

desired product as a yellow solid, which was isolated by vacuum

the solution was heated to reﬂux for an additional 15 h. The reaction

mixture was cooled to room temperature, and then the THF was

removed using rotary evaporation. The remaining yellow solid was

dissolved in a minimum amount of methanol (5−10 mL) and ﬁltered

through Celite. The addition of diethyl ether (20−30 mL) aﬀorded the

desired product as a yellow solid, which was isolated by vacuum

1

ﬁltration and dried in vacuo. Yield: 0.523 g (59%). H NMR (400

1

ﬁltration and dried in vacuo. Yield: 0.290 g (78%). H NMR (400

MHz, MeOD-d

4

): δ 9.13 (d, 2H, J = 5.6 Hz), 8.74 (d, 2H, J = 8.3 Hz),

MHz, MeOD-d ): δ 8.80 (d, 2H, J = 8.4 Hz), 8.23 (s, 2H), 8.13 (d,

8.40 (t, 2H, J = 8.3 Hz), 7.81 (t, 2H, J = 5.6 Hz), 4.47 (d, 3H, J = 13.0

4

19

2

3

H, J = 8.4 Hz), 4.33 (d, 3H, J = 13.2 Hz), 4.17 (d, 3H, J = 13.2 Hz),

Hz), 4.38 (d, 3H, J = 13.0 Hz), 3.80 (s, 6H). F NMR (376 MHz,

.42 (s, 6H), 3.34 (s, 6H). ¹⁹F NMR (376 MHz, MeOD-d , external

31

1

MeOD-d

): δ −79.52. P{ H} NMR (162 MHz, MeOD-d , external

4

31

1

−1

standard KPF ): δ −79.56. P{ H} NMR (162 MHz, MeOD-d ,

standard KPF ): δ −77.64. IR (KBr, cm ): 3440 m, 2040 s, 1950 s,

1910 s, 1610 m, 1470 m, 1440 m, 1280 s, 1260 s, 1140 m, 1100 m,

1030 s, 1010 m, 969 m, 946 m, 903 w, 801 m, 770 s, 740 m, 631 s, 614

6

4

−1

external standard KPF ): δ −79.09. IR (KBr, cm ): 3448 m, 3052 w,

6

2

913 w, 2033 s, 1958 s, 1934 s, 1445 m, 1384 w, 1266 s, 1222 w, 1150

m, 1031 m, 1013 w, 972 m, 873 m, 802 w, 743 w, 638 s, 582 w, 509 w,

40 w. ESI-MS (positive ion mode, CH CN/H O 70/30 and 1%

m, 587 m, 573 m, 512 m, 454 m. ESI-MS (positive ion mode,

+

CH CN/H O 70/30 and 1% formic acid): m/z 584.08746 ([M] ,

4

3

2

3

2

+

calcd 584.08558). Anal. Calcd for PTA-3 (C H F N O PReS): C,

formic acid): m/z 636.11922 ([M] , calcd 636.11743). Anal. Calcd for

PTA-2 (C H F N O PReS): C, 36.73; H, 3.08; N, 8.92. Found: C,

20 20

3

5

6

3

2.79; H, 2.75; N, 9.56. Found: C, 32.51; H, 2.76; N, 9.38.

Synthesis of [Re(CO) (bpy)(THP)]TFA (THP-3). A mixture of

24

3

5

6

3

6.66; H, 3.09; N, 8.98.

3

[

Re(CO) (bpy)Cl] (0.100 g, 0.217 mmol) and AgOTf (0.056 g, 0.217

Synthesis of [Re(CO) (dmphen)(THP)]TFA (THP-2). A mixture

3

mmol) in THF (22 mL) was heated to reﬂux at 75 °C for 3 h in the

dark. The white solid, AgCl, was removed via vacuum ﬁltration. To the

remaining yellow ﬁltrate was added THP (0.040 g, 0.326 mmol), and

the solution was heated to reﬂux for an additional 15 h. The reaction

mixture was cooled to room temperature, and then THF was removed

by rotary evaporation. The remaining yellow oil was then dissolved in

a minimal amount of methanol (5 mL) and ﬁltered through Celite.

The crude yellow oil (0.170 g) was puriﬁed using preparatory RP-

of [Re(CO) (dmphen)Cl] (0.209 g, 0.407 mmol) and AgOTf (0.104

3

g, 0.407 mmol) in THF (46 mL) was heated to reﬂux at 75 °C for 3 h

in the dark. The white solid AgCl was removed via vacuum ﬁltration.

To the remaining yellow ﬁltrate was added THP (0.076 g, 0.612

mmol), and the solution was heated to reﬂux for an additional 15 h.

The reaction mixture was cooled to room temperature, and then THF

was removed by rotary evaporation. The crude yellow oil was puriﬁed

using preparatory RP-HPLC using a CH OH/H O gradient (0−5

3

2

HPLC using a CH OH/H O gradient (0−5 min, 10% CH OH; 5−30

min, 10% CH OH; 5−50 min, 10−100% CH OH). Puriﬁed yield:

3

2

3

1

min, 10−100% CH OH). Puriﬁed yield: 0.070 g (45%). H NMR

400 MHz, MeOD-d ): δ 9.17 (d, 2H, J = 5.8 Hz), 8.64 (d, 2H, J = 8.1

Hz), 8.31 (t, 2H, J = 8.1 Hz), 7.72 (t, 2H, J = 6.4 Hz), 3.90 (s, 6H). F

0

.158 g (53%). H NMR (400 MHz, MeOD-d ): δ 8.70 (d, 2H, J = 8.4

3

4

(

Hz), 8.13 (s, 2H), 8.03 (d, 2H, J = 8.4 Hz), 3.44 (s, 6H), 3.37 (s, 6H).

4

1

9

1

3

9

1

F NMR (376 MHz, MeOD-d , external standard KPF ): δ −76.35.

4

6

1

NMR (376 MHz, MeOD-d

4

, external standard KPF

6

): δ −79.34.

P{ H} NMR (162 MHz, MeOD-d , external standard KPF ): δ 6.53.

4

6

31

1

P{ H} NMR (162 MHz, MeOD-d , external standard KPF ): δ 2.50.

−1

4

6

IR (KBr, cm ): 3036 m, 2837 w, 2449 w, 2044 s, 1945 s, 1909 s, 1674

s, 1594 w, 1506 w, 1442 m, 1370 w, 1202 m, 1140 m, 1043 m, 892 w,

−

1

IR (KBr, cm ): 3433 s, 2874 m, 2038 s, 1930 s, 1915 s, 1681 s, 1633

s, 1473 m, 1450 m, 1181 s, 1030 m, 841 w, 769 m, 731 w, 534 w, 497

w, 417 m. ESI-MS (positive ion mode, CH CN/H O 70/30 and 1%

8

59 m, 837 w, 800 w, 781 w, 722 w, 652 w, 619 w, 552 w, 513 w, 499

3

2

w. ESI-MS (positive ion mode, CH CN/H O 70/30 and 1% formic

3

2

+

formic acid): m/z 551.03927 ([M] , calcd 551.03763). Anal. Calcd for

THP-3·H O (C H F N O PRe): C, 31.72; H, 2.81; N, 4.11. Found:

acid): m/z 603.07053 ([M] , calcd 603.06947). Anal. Calcd for THP-

2

18 19

3

2

9

2

·H O (C H F N O PRe): C, 36.02; H, 3.16; N, 3.82. Found: C,

2 22 23 3 2 9

C, 31.68; H, 2.61; N, 4.09.

3

6.03; H, 2.90; N, 3.77.

Synthesis of [Re(CO) (bpy)(DAPTA)]OTf (DAPTA-3). A mixture

3

Synthesis of [Re(CO) (dmphen)(DAPTA)]OTf (DAPTA-2). A

3

of [Re(CO) (bpy)Cl] (0.236 g, 0.51 mmol) and AgOTf (0.131 g, 0.51

3

mixture of [Re(CO) (dmphen)Cl] (0.232 g, 0.451 mmol) and AgOTf

3

mmol) in THF (55 mL) was heated to reﬂux at 75 °C for 3 h in the

dark. The white solid precipitate, AgCl, was removed via vacuum

ﬁltration. To the remaining yellow ﬁltrate was added DAPTA (0.150 g,

(

0.116 g, 0.51 mmol) in THF (54 mL) was heated to reﬂux at 75 °C

for 3 h in the dark. The white solid AgCl was removed via vacuum

ﬁltration. To the remaining yellow ﬁltrate was added DAPTA (0.144 g,

0

.61 mmol), and the solution was heated to reﬂux for an additional 24

0

.59 mmol), and the solution was heated to reﬂux for an additional 15

h. The reaction mixture was cooled to room temperature, and then the

THF was removed using rotary evaporation. The remaining yellow

solid was dissolved in a minimum amount of methanol (5−10 mL)

and ﬁltered through Celite. The addition of diethyl ether (20−30 mL)

aﬀorded the desired product as a yellow solid, which was isolated by

h. The reaction mixture was cooled to room temperature, and then the

yellow precipitate was isolated by ﬁltration and washed with a small

amount of methanol (5 mL) and diethyl ether (20 mL). Yield: 0.319 g

1

(

83%). H NMR (400 MHz, MeOD-d ): DAPTA complexes show the

4

syn and anti isomers in a 1/5 ratio; however in the aromatic region the

isomers are indistinguishable, δ 8.85 (d, 2H, anti + syn), 8.26 (d, 2H,

anti + syn), 8.16 (dd, 2H, anti + syn), 5.51 (d, anti), 4.98 (d, syn), 4.44

1

vacuum ﬁltration and dried in vacuo. Yield: 0.204 g (48%). H NMR

(

400 MHz, MeOD-d ): DAPTA complexes show the syn and anti

4

isomers in a 1/5 ratio; however in the aromatic region the isomers are

indistinguishable, δ 9.21 (t, 2H, anti + syn), 8.77 (d, 2H, anti + syn),

(

d, anti), 3.93 (d, anti), 3.55 (m, anti), 3.37 (d, 3H, anti + syn), 2.98

m, syn), 2.84 (d, anti), 1.90 (s, syn), 1.84 (s, anti), 1.46 (s, 5H, anti).

8

.43 (t, 2H, anti + syn), 7.84 (t, 2H, anti + syn), 5.65 (d, anti), 5.12 (d,

1

3

9

F NMR (376 MHz, MeOD-d , external standard KPF ): δ −79.57.

4

6

syn), 5.01 (d, anti), 4.92 (m, syn), 4.53 (d, anti), 4.22 (m, anti), 4.03

1

P{ H} NMR (162 MHz, MeOD-d , external standard KPF ): δ

4

6

(d, anti), 3.91 (d, anti), 3.80 (s, syn), 3.58 (d, anti), 3.14 (d, anti), 2.01

−1

19

−

41.15 (syn), −45.58 (anti). IR (KBr, cm ): 3440 m, 3070 w, 2040 s,

(s, syn), 1.98 (s, anti), 1.90 (s, anti). F NMR (376 MHz, MeOD-d ,

4

31

1

5

3

970 m, 1930 s, 1640 s, 1510 w, 1420 m, 1380 w, 1340 m, 1270 s,

150 m, 1030 s, 989 m, 890 m, 867 m, 799 m, 705 w, 638 s, 551 m,

13 m, 469 w, 417 w. ESI-MS (positive ion mode, CH CN/H O 70/

external standard KPF ): δ −79.56. P{ H} NMR (162 MHz, MeOD-

6

d , external standard KPF ): δ −44.42 (syn), −48.33 (anti). IR (KBr,

4

6

−1

3

2

cm ): 3457 m, 2965 w, 2880 w, 2037 s, 1958 s, 1918 s, 1646 s, 1473

m, 1419 m, 1331 m, 1265 s, 1155 m, 1024 s, 987 m, 892 m, 776 m,

632 s, 515 m, 422 m. ESI-MS (positive ion mode, CH CN/H O 70/

+

0 and 1% formic acid): m/z 708.13962 ([M] , calcd 708.13801).

Anal. Calcd for DAPTA-2 (C H F N O PReS): C, 37.85; H, 3.29;

27

28

3

5

8

3

2

+

N, 8.17. Found: C, 38.02; H, 3.24; N, 8.09.

30 and 1% formic acid): m/z 656.10825([M] , calcd 656.10671). Anal.

Calcd for DAPTA-3·0.5(CH CH ) O (C H F N O PReS): C,

Synthesis of [Re(CO) (bpy)(PTA)]OTf (PTA-3). A mixture of

3

2

25 29

3

5

8.5

[

Re(CO) (bpy)Cl] (0.564 g, 1.22 mmol) and AgOTf (0.314 g, 1.22

35.67; H, 3.47; N, 8.32. Found: C, 35.17; H, 3.49; N, 8.38.

3

mmol) in THF (123 mL) was heated to reﬂux at 75 °C for 3 h in the

dark. The white solid, AgCl, was removed via vacuum ﬁltration. To the

remaining yellow ﬁltrate was added PTA (0.192 g, 1.22 mmol), and

Synthesis of [Re(CO) (dmbpy)(PTA)]OTf (PTA-4). A mixture of

3

[Re(CO) (dmbpy)Cl] (0.200 g, 0.43 mmol) and AgOTf (0.111 g,

3

0.43 mmol) in THF (47 mL) was heated to reﬂux at 75 °C for 3 h in

N

DOI: 10.1021/acs.inorgchem.7b02747

Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry

Article

the dark. The white solid precipitate, AgCl, was removed via vacuum

Synthesis of [Re(CO) (dmobpy)(PTA)]OTf (PTA-5). A mixture

3

ﬁltration. To the remaining yellow ﬁltrate was added PTA (0.102 g,

of [Re(CO) (dmobpy)Cl] (0.200 g, 0.38 mmol) and AgOTf (0.099 g,

3

0

.65 mmol), and the solution was heated to reﬂux for an additional 15

0.38 mmol) in THF (47 mL) was heated to reﬂux at 75 °C for 3 h in

the dark. The white solid precipitate, AgCl, was removed via vacuum

ﬁltration. To the remaining yellow ﬁltrate was added PTA (0.079 g,

0.50 mmol), and the solution was heated to reﬂux for an additional 15