A New Series of Complexes Possessing Rare “Tertiary” Sulfonamide Nitrogen-to-Metal Bonds of Normal Length: fac-[Re(CO)₃(N(SO₂R)dien)]PF₆ Complexes with Hydrophilic Sulfonamide Ligands

Abstract

Tertiary sulfonamide nitrogen-to-metal bonds of normal length are very rare. We recently discovered such a bond in one class of fac-[Re(CO)₃(N(SO₂R)(CH₂Z)₂)]ⁿ complexes (Z = 2-pyridyl) with N(SO₂R)dpa ligands derived from di-(2-picolyl)amine (N(H)dpa). fac-[M(CO)₃(N(SO₂R)(CH₂Z)₂)]ⁿ agents (M = ^186/188Re, ^99mTc) could find use as radiopharmaceutical bioconjugates when R is a targeting moiety. However, the planar, electron-withdrawing 2-pyridyl groups of N(SO₂R)dpa destabilize the ligand to base and create relatively rigid chelate rings, raising the possibility that the rare M– N(sulfonamide) bond is an artifact of a restricted geometry. Also, the hydrophobic 2-pyridyl groups could cause undesirable accumulation in the liver, limiting future use in radiopharmaceuticals. Our goal is to identify a robust, hydrophilic, and flexible N(CH₂Z)₂ chelate framework. New C₂-symmetric ligands, N(SO₂R)(CH₂Z)₂ with (Z = CH₂NH₂; R = Me, dmb, or tol), were prepared by treating N(H)dien(Boc)₂, a protected diethylenetriamine (N(H)dien) derivative, with methanesulfonyl chloride (MeSO₂Cl), 3,5-dimethylbenzenesulfonyl chloride (dmbSO₂Cl), and 4-methylbenzenesulfonyl chloride (tolSO₂Cl). Treatment of fac-[Re(CO)₃(H₂O)₃]⁺ with these ligands, designated as N(SO₂R)dien, afforded new fac-[Re(CO)₃(N(SO₂R)dien)]PF₆ complexes. Comparing the fac-[Re(CO)₃(N(SO₂Me)dien)]PF₆ and fac-[Re(CO)₃(N(SO₂Me)dpa)]PF₆ complexes, we find that the Re^I–N(sulfonamide) bonds are normal in length and statistically identical and that the methyl ¹³C NMR signal has an unusually upfield shift compared to that in the free ligand. We attribute this unusual upfield shift to the fact that the sulfonamide N undergoes an sp²-to-sp³ rehybridization upon coordination to Re^I in both complexes. Thus, the sulfonamide N of N(SO₂R)dien ligands is a good donor, even though the chelate rings are conformationally flexible. Addition of the strongly basic and potentially monodentate ligand, 4-dimethylaminopyridine, did not affect the fac-[Re(CO)₃(N(SO₂tol)dien)]PF₆ complex, even after several weeks. This complex is also stable to heat in aqueous solution. These results indicate that N(SO₂R)dien ligands form fac-[Re(CO)₃(N(SO₂R)dien)]PF₆ complexes sufficiently robust to be utilized for radiopharmaceutical development.

Introduction

Many fac-[^99mTc(CO)₃L]ⁿ imaging agents with facially coordinated tridentate ligands (L) have been studied^1-8 because of the convenient generation of the fac-[^99mTc(CO)₃(H₂O)₃]⁺ precursor.^9,10 Some of these imaging agents have exhibited satisfactory results in human volunteers and in early patient studies.^4,7,8 Such fac-[^99mTc(CO)₃L]ⁿ agents are more robust and have better pharmacokinetic properties than agents with bidentate ligands.¹¹ The γ-emitting ^99mTc radionuclide has ideal nuclear properties^12,13 for diagnostic applications in nuclear medicine.^1,14-17

The development of ^99mTc radiopharmaceutical agents benefits from an understanding of the chemistry of their Re analogues. The discovery of a straightforward preparation of the fac-[Re(CO)₃(H₂O)₃]⁺ precursor¹⁸ has led to significantly improved aqueous synthetic methods for fac-[M(CO)₃L]ⁿ agents (M = various isotopes of Tc and Re).^{3,16,17,19-21} fac-[Re(CO)₃L]ⁿ complexes serve as excellent structural models for fac-[^99mTc(CO)₃L]ⁿ imaging agents.^4,22-27 Moreover, fac-[^186/188Re(CO)₃L]ⁿ agents themselves are emerging as promising radiopharmaceuticals, owing to their potential usefulness in radiotherapy.^1,12,20,28

New types of tridentate ligands and ligand conjugation methods will expand the likelihood of developing useful new agents with the fac-[M^I(CO)₃]⁺ core (M = ^99mTc, ^186/188Re).^10,29,30 Meeting such goals requires the identification of suitable linker systems with high stability, small size, and core ligands having a tridentate donor framework that does not increase the number of isomers.^31-34 Symmetrical linear tridentate ligands with linkage at the center donor are thus suitable candidates because the generation of racemic or diastereoisomeric mixtures of radiopharmaceuticals can be avoided.³⁵ Recent studies have reported promising biomedical properties for fac-[Re(CO)₃L]ⁿ complexes bearing a tridentate L with three N donors having a substituent replacing the proton at the central sp³ N.^16,29,36,37 However, the conjugation of biologically important groups in these complexes was limited to groups attached via an N–C bond to the central N of ligands with the N(CH₂Z)₂ tridentate ligand framework. Z is commonly an N donor (e.g., 2-pyridyl)^16,36 or a carboxyl group.³ The use of a biologically compatible linking group that did not create an N–C bond would greatly increase the chances of discovering useful agents.

Molecules containing a sulfonamide represent a very important class of biologically active molecules with a wide variety of applications.^39-45 Therefore, we previously set out to explore conjugation that utilizes an N–S bond with the central N being the sulfonamide N.³⁵ The reaction of various sulfonyl chlorides (RSO₂Cl) with di-(2-picolyl)amine (N(H)dpa) afforded N(SO₂R)dpa ligands.³⁵ [Note that we use the normal convention: an N (italic N) designates a substituent location on nitrogen in the name of a compound. In this article, N designates a substituent located on the central or anchoring nitrogen atom of a tridentate ligand.] These N(SO₂R)dpa ligands readily added to fac-[Re(CO)₃(H₂O)₃]⁺ to form fac-[Re(CO)₃(N(SO₂R)dpa)]PF₆ (or BF₄) complexes. We learned that these were the only examples of structurally characterized bearing an N-bound, open-chain tertiary sulfonamide linkage with a normal M–N bond length. These fac-[Re(CO)₃(N(SO₂R)dpa)]PF₆ complexes were the first examples of such structurally characterized complexes with a neutral tertiary sulfonamide donor bound not only to the fac-[Re^I(CO)₃]⁺ core but to any metal center.³⁵ Although the results in that report appeared to serve as proof of principle, the tridentate framework, N(CH₂Z)₂ with Z = 2-pyridyl, is relatively rigid and the resulting geometric constraints could possibly account for the observation of the M–N(sulfonamide) bond of normal length, as found in cases when complicated ligand ring structures fix the bond lengths.³⁵ Thus, the study of fac-[Re(CO)₃(N(SO₂R)(CH₂Z)₂)]ⁿ complexes with more flexible chelate rings is of fundamental importance to coordination chemistry.

The new fac-[Re(CO)₃(N(SO₂R)dpa)]X complexes revealed the feasibility of having a tridentate ligand anchored by a central tertiary sulfonamide N. However, the two 2-pyridyl rings in a potential fac-[M(CO)₃(N(SO₂R)dpa)]ⁿ imaging agent are hydrophobic, an undesirable property in an imaging agent because it is expected to promote liver uptake.¹² Also, the 2-pyridyl groups are electron withdrawing, facilitating decomposition of coordinated N(SO₂R)dpa ligands by strong base.³⁵

With the goals of exploring fundamental coordination chemistry of sulfonamides and of identifying ligands for use in imaging agents, we have now explored a more hydrophilic and more flexible ligand system that is suitable for our new conjugation method. Here we employ a prototypical triamine ligand framework based on diethylenetriamine (N(H)dien).⁴⁶ In the new N(SO₂R)dien ligands, the aromatic 2-pyridyl groups of the dpa moiety are replaced with hydrophilic –CH₂NH₂ groups. The new ligands are stable to base and have the advantage of being small in size, a feature considered to be desirable for bioconjugates.^32,34 All of the new complexes discussed below have the facial geometry, and thus from this point onward we omit the fac- designation when discussing specific compounds.

Experimental Section

Starting Materials

Methanesulfonyl chloride (MeSO₂Cl), 3,5-dimethylbenzenesulfonyl chloride (dmbSO₂Cl), 4-methylbenzenesulfonyl chloride (tolSO₂Cl), N,N,N-triethylamine, trifluoroacetic acid (TFA), N(H)dien, 2-(tert-butoxycarbonyloxyimino)-2-(phenylacetonitrile), 4-dimethylaminopyridine, and Re₂(CO)₁₀ were used as received from Aldrich. Aqueous [Re(CO)₃(H₂O)₃]OTf (OTf = trifluoromethanesulfonate) was prepared by a known method.¹⁸

NMR Measurements

¹H NMR and ¹³C NMR spectra were recorded on a 400 MHz Bruker spectrometer. Peak positions are relative to TMS or to solvent residual peak, with TMS as reference. All NMR data were processed with TopSpin and MestReNova software. NMR data not presented in the Experimental Section can be found in the Results Section or in Supporting Information.

Mass Spectrometric Measurements

High resolution mass spectra were recorded on a Bruker Ultraflex MALDI TOF TOF mass spectrometer and an Agilent 6210 ESI TOF LCMS mass spectrometer.

X-ray Data Collection and Structure Determination

Intensity data were collected at low temperature on a Bruker Kappa Apex-II DUO CCD diffractometer fitted with an Oxford Cryostream cooler with graphite-monochromated Mo Kα (λ = 0.71073 Å) radiation. Data reduction included absorption corrections by the multiscan method, with SADABS.⁴⁷ The structures were determined by direct methods and difference Fourier techniques and refined by full-matrix least squares using SHELXL-97.⁴⁸ All non-hydrogen atoms were refined anisotropically. All H atoms were visible in difference maps, but were placed in idealized positions, except for N–H hydrogen atoms, for which coordinates were refined. A torsional parameter was refined for each methyl group. Compound 6 has two independent formula units in the asymmetric unit.

Synthesis of N,N″-Bis(tert-butoxycarbonyl)diethylenetriamine (N(H)dien(Boc)₂)

The N(H)dien(Boc)₂ ligand was prepared in 96% yield by a known method.^{49 1}H NMR signals (ppm) in CDCl₃: 4.89 (br s, 1H, NH), 3.20 (m, 4H, 2CH₂), 2.72 (t, 4H, 2CH₂), 1.44 (s, 18H, 6CH₃). These ¹H NMR chemical shifts matched the previously reported values.⁴⁹

General Synthesis of N(SO₂R)dien

The following general procedure was employed to obtain the [N(SO₂Me)dienH₂](CF₃CO₂)₂ (1) and N(SO₂R)dien (R = dmb (2), R = tol (3)) ligands: a solution of the sulfonyl chloride (2 mmol) in 30 mL of dioxane was added dropwise over 2 h to a solution of N(H)dien(Boc)₂ (0.61 g, 2 mmol) and triethylamine (0.28 mL, 2 mmol) in dioxane (100 mL) at room temperature. The reaction mixture was stirred at room temperature for 20 h and filtered to remove any precipitate. The solvent was completely removed by rotary evaporation, water (50 mL) was added to the resulting oil, and the product was extracted into CH₂Cl₂ (2 × 25 mL). The CH₂Cl₂ extracts were combined and washed again with water (2 × 25 mL) at pH ∼6. The organic layer was dried with anhydrous Na₂SO₄, and the CH₂Cl₂ was removed by rotary evaporation to yield an off-white solid. The Boc-protected diensulfonamide, N(SO₂R)dien(Boc)₂, was purified if necessary by column chromatography (silica gel column and a mixture of ethyl acetate/hexane). After characterization by ¹H NMR spectroscopy, this N(SO₂R)dien(Boc)₂ product was then deprotected by dropwise addition of trifluoroacetic acid (0.16–0.24 mL, 2–3 mmol) to a CH₂Cl₂ solution (at ∼0 °C) of the compound (5 mL, 1 mmol). The reaction mixture was allowed to warm to room temperature, stirred for 16 h at room temperature, and then filtered; the filtrate was taken to dryness by rotary evaporation. Water at pH ∼8–9 (50 mL) was added to the residue, and the product was extracted into CHCl₃ (2 × 25 mL). The combined CHCl₃ extracts were dried over anhydrous Na₂SO₄, and the solvent was removed by rotary evaporation, leaving a white/off-white powder or, when R = Me, a yellow oil.

General Synthesis of [Re(CO)₃(N(SO₂R)dien)]PF₆

A solution of the N(SO₂R)dien ligand (0.1 mmol) in methanol (2 mL) was added to an aqueous solution of [Re(CO)₃(H₂O)₃]OTf (5 mL, 0.1 mmol). Methanol (1–2 mL) was added to dissolve any precipitate that formed. The pH of the reaction mixture was adjusted to ∼6–7 with 0.5 M NaOH if necessary, and the clear reaction mixture was heated at reflux for 24 h. An excess (0.16 g, 1 mmol) of NaPF₆ was added to the clear solution, and the precipitate that formed within ∼30 min was collected on a filter, washed with water, and air dried.

Synthesis of [N(SO₂Me)dienH₂](CF₃CO₂)₂ ([1H₂](CF₃CO₂)₂)

The use of MeSO₂Cl (0.16 mL, 2 mmol) in the general method described above yielded crude N(SO₂Me)dien(Boc)₂ as a brown oil (0.65 g, 85% yield). ¹H NMR signals (ppm) in DMSO-d₆: 6.92 (br s, 2H, 2NH), 3.14 (m, 4H, 2CH₂), 3.06 (m, 4H, 2CH₂), 2.89 (s, 3H, CH₃), 1.37 (s, 18H, 6CH₃).

The deprotection process described in the general method above afforded N(SO₂Me)dien (1) as a pale-yellow, oily substance (yield: 0.004 g, 18%). ¹H NMR signals (ppm) in CDCl₃: 3.25 (t, J = 6.3 Hz, 4H, 2C(5/6)H₂), 2.93 (s, 3H, CH₃), 2.90 (t, J = 6.4 Hz, 4H, 2C(4/7)H₂); in acetonitrile-d₃: 3.14 (t, J = 6.5 Hz, 4H, 2C(5/6)H₂), 2.86 (s, 3H, CH₃), 2.76 (t, J = 6.5 Hz, 4H, 2C(4/7)H₂); in DMSO-d₆: 3.08 (broad t, 4H, 2C(5/6)H₂), 2.91 (s, 3H, CH₃), 2.67 (broad t, 4H, 2C(4/7)H₂). ¹³C NMR signals (ppm) in CDCl₃: 51.59 (C5/6), 40.70 (C4/7), 37.60 (CH₃).

Compared to ligands with other R groups, isolation for R = Me produced a low yield, owing to the more hydrophilic nature of N(SO₂Me)dien. To increase the amount of material available for synthesis of the complex, the general deprotection process was modified by adding a slight excess of trifluoroacetic acid to obtain the protonated ligand [N(SO₂Me)dienH₂](CF₃CO₂)₂ ([1H₂](CF₃CO₂)₂). A mixture of N(SO₂Me)dien(Boc)₂ (0.38 g, 1 mmol) and trifluoroacetic acid (0.24 mL, 3 mmol) was stirred at room temperature for 16 h, and the resulting mixture was filtered; the white precipitate that was collected on a filter paper was washed several times with CH₂Cl₂ and air dried. The precipitate was then dissolved in methanol (3 mL), and the solution was filtered. CH₂Cl₂ (-10 mL) was added to the filtrate until cloudiness was observed. After 24 h the undisturbed solution yielded thin, colorless, needle-like crystals of the trifluoroacetate salt of the protonated ligand [N(SO₂Me)dienH2](CF₃CO₂)₂ ([1H2](CF₃CO₂)2) (0.29 g, 72% yield). ¹H NMR signals (ppm) in DMSO-d₆: 7.81 (br s, 6H, 2⁺NH₃), 3.37 (t, J = 6.2 Hz, 4H, 2CH₂), 3.04 (s, 3H, CH₃), 3.00 (t, J = 6.3 Hz, 4H, 2CH₂). ESI-MS: m/z calcd for C₅H₁₅O₂N₃S [M + H]⁺: 182.0958; found 182.0957.

Synthesis of N(SO₂dmb)dien (2)

The use of dmbSO₂Cl (0.40 g, 2 mmol) according to the general synthetic procedure above yielded a pale yellow oil, which was purified by dissolving the oil in a minimum (∼1 mL) of ethyl acetate and loading the solution onto a silica gel column. A 1:5 mixture of ethyl acetate:hexane was used to elute the remaining dmbSO₂Cl starting material (UV-vis). The product (UV-vis) was then eluted with a 2:3 (v:v) mixture of ethyl acetate/hexane. Thin-layer chromatography was used to determine the progress of separation. Removal of solvent by rotary evaporation yielded N(SO₂dmb)dien(Boc)₂ as a pale yellow powder (0.87 g, 92% yield). ¹H NMR signals (ppm) in CDCl₃: 7.65 (s, 2H, H2/6), 7.21 (s, H, H4), 5.19 (br s, 2H, NH), 3.32 (m, 4H, 2CH₂), 3.18 (m, 4H, 2CH₂), 2.38 (s, 6H, 2CH₃), 1.45 (s, 18H, 6CH₃).

Deprotection of N(SO₂dmb)dien(Boc)₂ (0.47 g, 1 mmol) with trifluoroacetic acid (0.16 mL, 2 mmol), by the procedure described above, afforded N(SO₂dmb)dien (2) as a white powder (0.26 g, 98% yield). 1H NMR signals (ppm) in CDCl₃: 7.42 (s, 2H, H2/6), 7.20 (s, H, H4), 3.15 (t, J = 6.4 Hz 4H, 2C(5/6)H₂), 2.91 (t, J = 6.4 Hz, 4H, 2C(4/7)H₂), 2.38 (s, 6H, 2CH₃). ESI-MS: m/z calcd for C₁₂H₂₁O₂N₃S [M + H]⁺: 272.1427; found 272.1431.

Synthesis of N(SO₂tol)dien (3)

The general procedure using tolSO₂Cl (0.44 g, 2 mmol) yielded N(SO₂tol)dien(Boc)₂ as a pale yellow oil (0.78 g, 86% yield). ¹H NMR signals (ppm) in CDCl₃: 7.67 (d, J = 8.3 Hz, 2H, H2/6), 7.31 (d, J = 8.1 Hz, 2H, H3/5), 5.18 (br s, 2H, NH), 3.32 (m, 4H, 2CH₂), 3.17 (m, 4H, 2CH₂), 2.43 (s, 3H, CH₃), 1.45 (s, 18H, 6CH₃).

Deprotection of N(SO₂tol)dien(Boc)₂ (0.45 g, 1 mmol) with trifluoroacetic acid (0.16 mL, 2 mmol) afforded N(SO₂tol)dien (3) as a white powder (0.23 g, 92% yield). ¹H NMR signals (ppm) in CDCl₃: 7.70 (d, J = 8.2 Hz, 2H, H2/6), 7.31 (d, J = 8.0 Hz, 2H, H3/5), 3.15 (t, J = 6.4 Hz, 4H, 2C(5/6)H₂), 2.91 (t, J = 6.4 Hz, 4H, 2C(4/7)H₂), 2.43 (s, 3H, CH₃); in acetonitrile-d₃: 7.69 (d, J = 8.4 Hz, 2H, H2/6), 7.39 (d, J = 7.9 Hz, 2H, H3/5), 3.07 (t, J = 6.4 Hz, 4H, 2C(5/6)H₂), 2.79 (t, J = 6.5 Hz, 4H, 2C(4/7)H₂), 2.41 (s, 3H, CH₃). ESI-MS: m/z calcd for C₁₁H₁₉O₂N₃S [M + H]⁺: 258.1271; found 258.1277.

Synthesis of [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4)

The general method above, with the [1H₂](CF₃CO₂)₂ ligand (0.04 g, 0.1 mmol) and [Re(CO)₃(H₂O)₃]OTf (5 mL, 0.1 mmol), afforded [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) as a white crystalline precipitate (0.036 g, 60% yield) after the addition of NaPF6 (∼16 mg). (The pH of the acidic reaction mixture was adjusted to ∼7 with 0.5 M NaOH before the solution was heated at reflux.) ¹H NMR signals (ppm) in DMSO-d₆: 5.65 (m, 2H, NH), 4.27 (m, 2H, NH), 3.57 (s, 3H, CH₃), 3.45 (m, 2H, CH₂), 3.22 (m, 2H, CH₂), 3.10 (m, 2H, CH₂), 2.97 (m, 4H, 2CH₂). 1H NMR signals (ppm) of the R group in acetone-d₆: 3.62 (s, 3H, CH₃); in acetonitrile-d₃: 3.33 (s, 3H, CH₃). Chelate ring ¹H NMR signals for 4-6 in these solvents are reported in the Results Section. ¹³C NMR signals (ppm) in DMSO-d₆: 55.91 (C5/6), 44.86 (C4/7), 32.50 (CH₃). ¹³C NMR signals (ppm) in acetone-d₆: 55.11 (C5/6), 44.46 (C4/7), 30.97(CH₃). MALDI-TOF: m/z calcd for C₈H₁₅O₅N₃SRe [M⁺]: 452.029; found: 452.158.

Synthesis of [Re(CO)₃(N(SO₂dmb)dien)]PF₆ (5)

The use of the general procedure with N(SO₂dmb)dien (2) (0.03 g, 0.1 mmol) afforded [Re(CO)₃(N(SO₂dmb)dien)]PF₆ (5) as a white precipitate (0.047 g, 69% yield) after the addition of NaPF₆ (∼16 mg). ¹H NMR signals (ppm) in DMSO-d₆: 7.72 (s, 2H, H2/6), 7.57 (s, H, H4), 5.72 (m, 2H, NH), 4.28 (m, 2H, NH), 3.41 (m, 2H, CH₂), 3.14 (m, 2H, CH₂), 3.00 (m, 2H, CH₂), 2.65 (m, 2H, CH₂), 2.42 (s, 6H, 2CH₃). ¹H NMR signals (ppm) of the R group in acetone-d₆: 7.77 (s, 2H, H2/6), 7.60 (s, H, H4), 2.46 (s, 6H, 2CH₃); in acetonitrile-d₃: 7.65 (s, 2H, H2/6), 7.55 (s, H, H4), 2.44 (s, 6H, 2CH₃). ¹³C NMR signals (ppm) in acetone-d6: 142.30 (dmb ring C1), 139.54 (dmb ring C4), 130.93 (dmb ring C2/6), 126.71(dmb ring C3/5), 57.83 (C5/6), 46.93(C4/7), 22.09 (2CH3). MALDI-TOF: m/z calcd for C₁₅H₂₁O₅N₃SRe [M⁺]: 542.076; found:542.155.

Synthesis of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6)

The use of N(SO₂tol)dien (3) (0.02 g, 0.1 mmol) as described in the general synthesis afforded [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) as a white precipitate (0.056 g, 84% yield) after the addition of NaPF₆ (∼16 mg). Slow evaporation of a solution of the compound in acetone produced colorless, X-ray quality, needle-shaped crystals. ¹H NMR signals (ppm) in DMSO-d₆: 7.99 (d, J = 8.5 Hz, 2H, H2/6), 7.61 (d, J = 8.1 Hz, 2H, H3/5), 5.70 (m, 2H, NH), 4.28 (m, 2H, NH), 3.40 (m, 2H, CH₂), 3.14 (m, 2H, CH₂), 3.01 (m, 2H, CH₂), 2.66 (m, 2H, CH₂), ∼2.5, overlapped (s, 3H, CH₃). ¹H NMR signals (ppm) of the R group in acetone-d₆: 8.05 (d, 2H, H2/6), 7.65 (d, 2H, H3/5), 2.53 (s, 3H, CH₃); in acetonitrile-d₃: 7.91(d, 2H, H2/6), 7.58 (d, 2H, H3/5), 2.51 (s, 3H, CH3). ¹³C NMR signals (ppm) in DMSO-d₆: 147.80 (tol ring C4), 132.07 (tol ring C2/6), 131.00 (tol ring C3/5), 125.98 (tol ring C1), 56.20 (C5/6), 44.80 (C4/7), 21.69 (CH₃). ¹³C NMR signals (ppm) in acetone-d₆: 148.88 (tol ring C4), 132.70 (tol ring C2/6), 131.58 (tol ring C3/5), 127.05 (tol ring C1), 56.94(C5/6), 46.00 (C4/7), 21.74 (CH₃). MALDI-TOF: m/z calcd for C₁₄H₁₉O₅N₃SRe [M ⁺]: 528.060;found: 528.001.

Challenge Reactions of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6)

Two 5 mM solutions of 6 in DMSO-d₆ (500 μL) were treated separately with 10 molar equiv (0.6 μL, 50 mM) or 2 molar equiv (0.12 μL, 10 mM) of diethylenetriamine and monitored over time by ¹H NMR spectroscopy. Another challenge experiment was carried out with a 5 mM solution of 6 in DMSO-d₆ by first adding 1 molar equiv of 4-dimethylaminopyridine (4-Me₂Npy) and then increasing the 4-Me₂Npy to 10 molar equiv.

Results and Discussion

Synthesis of N(SO₂R)dien and [Re(CO)₃(N(SO₂R)dien)]PF₆

N(SO₂Me)dien (1), N(SO₂dmb)dien (2), and N(SO₂tol)dien (3) were synthesized by coupling N(H)dien with the respective sulfonyl chloride (Figure 1). In this case, unlike in the synthesis of N(SO₂R)dpa ligands,³⁵ additional steps are required to synthesize the N(SO₂R)dien ligands (1–3), owing to the possibility of attack by the terminal N(H)dien amino groups on the sulfur atom of the sulfonyl chloride. Thus, the N(H)dien terminal amine groups were protected with Boc groups. The products were obtained in good yield and purity, as indicated by ¹H NMR spectral data. Reaction of the deprotected ligands with an aqueous solution of [Re(CO)₃(H₂O)₃]OTf afforded [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4), [Re(CO)₃(N(SO₂dmb)dien)]PF₆ (5), and [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) in 60–84% yields. All three complexes were characterized by ¹H and ¹³C NMR spectroscopy and by mass spectrometry (Experimental Section).

Figure 1.

General reaction scheme for the synthesis of N(SO₂R)dien ligands (top) and [Re(CO)₃(N(SO₂R)dien)]⁺ complexes (bottom).

X-ray quality crystals of 4 and 6 were grown by slowly cooling a 5 mM solution of the compound in warm water. Crystals of 5 (grown by slow evaporation of a 5 mM acetone solution) suffered from twinning, space-group ambiguity, and disorder problems in the molecule, which prevented refinement within acceptable values. However, the structural characterization of 5 (using the same procedures as for 4 and 6) confirmed that all three N atoms of N(SO₂dmb)dien are coordinated to Re.

Structural Results

Crystal data and structural refinement details for [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) and [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) are summarized in Table 1, and the ORTEP plots appear in Figure 2 (see Figure 1 for the numbering scheme used to describe the solid-state data). Both 4 and 6 exhibit a pseudo octahedral structure, with the three carbonyl ligands occupying one face; the remaining three coordination sites are occupied by the three nitrogen atoms of the tridentate ligand.

Table 1. Crystal Data and Structure Refinement for [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) and [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6).

	4	6
empirical formula	C8H15N3O5ReS·PF6	C14H19N3O5ReS PF6
fw	596.46	672.55
crystal system	monoclinic	monoclinic
space group	P21/c	P21/c
unit cell dimensions
a (Å)	12.822(3)	26.0233(14)
b (Å)	8.8089(19)	13.0607(7)
c (Å)	14.026(3)	12.2147(7)
β(deg)	94.373(10)	95.831(3)
V (Å3)	1579.6(6)	4130.1(4)
T (K)	100	100
Z	4	8
ρcalc (Mg/m3)	2.508	2.163
abs coeff (mm-1)	8.02	6.15
2θmax (°)	70	71.2
R indicesa	0.018	0.029
wR2 = [I> 2σ(I)]b	0.044	0.055
data collected, Rint	28379, 0.034	51479, 0.033
data/param	6953/239	18802/585

^aR = (Σ║F_o│ - │F_c║)/Σ │F_o│;

^bwR2= [Σ[w(F_o² - F_c²)²]/Σ [w(F_o²)²]]^1/2 in which and w = 1/[σ²(F_o²) + (dP)² + (eP)] and P = (F_o² + 2F_c²)/3.

Figure 2.

ORTEP plots of the cations in [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) (left) and Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) (right). Thermal ellipsoids are drawn with 50% probability.

Selected bond distances and angles of complexes 4 and 6 are presented in Table 2. The Re–C bond distance of the CO group trans to the sulfonamide group is not significantly different from those of the two cis Re–CO bonds, indicating the absence of any trans influence. All Re–N bond distances in 6 are generally similar to the Re–N(sp³) distances observed in [Re(CO)₃L]⁺ complexes with prototypical NNN donor ligands, which range from ∼2.21 to 2.29 Å as the bulk of substituents on the N atoms increases.^46,50 However, the distances from Re to the central N2 (2.2763(14) Å in 4 and 2.2686(19) Å in 6) are significantly longer than the Re-N1 distances (2.2377(14) Å for 4 and 2.221(2) Å for 6) and Re–N3 distances (2.2072(15) Å for 4 and 2.216(2) Å for 6) (Table 2). These Re–N2 distances in 4 and 6 are significantly longer than the Re–N2 distance in [Re(CO)₃(N(H)dien)]PF₆⁴⁶ (2.201(3) Å), but they are comparable to the Re–N2 distance of [Re(CO)₃(N(Me)dien)]PF₆.⁴⁶ The Re–N2 distance in the latter complex (2.250(4) Å) is not statistically different from the Re–N2 distance in [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6, Table 2), suggesting that the bond lengthening in 6 is caused chiefly by steric effects, not electronic effects. Furthermore, the good overlap of the three donor N atoms observed when the three carbonyl carbon (C1, C2 and C3) and Re atoms in the structure of 6 are superimposed with the corresponding atoms in [Re(CO)₃(N(Me)dien)]PF₆⁴⁶ (Supporting Information, Figure S1) demonstrates the lack of any large effect of having a tertiary sulfonamide nitrogen instead of a typical sp³ nitrogen serve as the central anchor of the tridentate ligand.

Table 2. Selected Bond Distances (Å) and Angles (deg) for [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) and [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6).

	4	6		4	6
bond distance			bond angle
Re–N1	2.2377(14)	2.221(2)	N1–Re–N2	77.07(5)	77.54(7)
Re–N2	2.2763(14)	2.2686(19)	N1–Re–N3	86.53(6)	85.83(8)
Re–N3	2.2072(15)	2.216(2)	N2–Re–N3	77.80(5)	77.86(7)
Re–C1	1.9128(16)	1.913(2)	C2–Re–N1	95.72(6)	92.70(9)
Re–C2	1.9109(18)	1.911(2)	C3–Re–N1	92.31(6)	94.99(9)
Re–C3	1.9136(19)	1.917(3)	C2–Re–N3	96.60(7)	98.29(9)
S1–O5	1.4273(13)	1.4280(18)	C1–Re–N3	93.12(6)	91.92(9)
S1–O4	1.4324(13)	1.4339(17)	Re–N2–S1	111.07(7)	111.35(9)
S1–N2	1.7434(13)	1.759(2)	Re–N2–C5	105.93(9)	106.14(13)
			Re–N2–C6	110.01(10)	109.86(13)
non-bonded distance			S1–N2–C5	109.13(10)	108.82(14)
N1–N3	3.046(2)	3.021(3)	S1–N2–C6	108.52(10)	108.18(13)
N1–N2	2.812(2)	2.812(3)	C5–N2–C6	112.19(13)	112.52(18)
N2–N3	2.816(2)	2.818(3)	C1–Re–N2	99.98(6)	100.55(8)
			C3–Re–N2	97.91(6)	96.81(8)
deviation of N2a	0.513(1)	0.514(2)	C1–Re–C2	87.25(7)	89.08(10)
			C3–Re–C2	87.62(7)	87.29(10)
			C1–Re–C3	87.83(7)	87.10(10)

^aDistance to the sulfonamide N from the plane defined by the S and the two C atoms attached to N2 (average of two independent molecules for 6).

The lengthening of the Re–N2 bond in 4 and 6 versus [Re(CO)₃(N(H)dien)]PF₆ is similar to that observed in [Re(CO)₃(N(SO₂R)dpa)]PF₆ complexes (e.g., R = Me and tol) versus the parent [Re(CO)₃(N(H)dpa)]Br complex.^30,35 The similarity of the central Re–N bond distances of the analogous [Re(CO)₃(N(SO₂R)dpa)]PF₆ complexes (average = ∼2.277 Å) and in 4 and 6 (average = ∼2.272 Å) indicates that the donating ability of the sulfonamide N is similar for both ligands.³⁵ Furthermore, all three donor N atoms show good overlap when the C1, C2, C3, and Re atoms in [Re(CO)₃(N(SO₂tol)dien)]PF₆ and [Re(CO)₃(N(SO₂tol)dpa)]PF₆ are superimposed (Figure 3). The central Re–N bond distance of the [Re(CO)₃(N(SO₂Me)dpa)]PF₆ complex (2.2826(16) Å) is not significantly different from this distance in 4. Thus, the normal length of the Re–N(sulfonamide) bonds in [Re(CO)₃(N(SO₂R)dpa)]PF₆ complexes is not a consequence of the rigidity of the 5-membered dpa chelate rings.

Figure 3.

Overlay of Re, C1, C2, and C3 atoms of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) (blue) and [Re(CO)₃(N(SO₂tol)dpa)]PF₆ ³⁵ (magenta) complexes. (The SO₂R groups have been omitted for clarity).

The sulfonamide nitrogen of the N(SO₂tol)dpa ligand in a Pd^II complex is not bound to Pd^II and all bond angles around the sulfonamide nitrogen of the N(SO₂tol)dpa ligand in the bidentate binding mode are close to 120°, indicating sp² hybridization for this N atom.⁵² In contrast, in 4 and 6, all bond angles around the sulfonamide N (N2) are ∼109° (Table 2). This observation, which is consistent with data for [Re(CO)₃(N(SO₂R)dpa)]PF₆ complexes,³⁵ indicates that the sulfonamide nitrogen changes from sp² to sp³ hybridization upon tridentate binding to Re^I.35 This conclusion is confirmed by the fact that the C5–N2 (average = 1.511 Å) and C6-N2 (average = 1.512 Å) bond distances are longer than an average C–N(sp²) distance (∼1.28 Å).⁵³ Data for all of the bond angles and distances involving the sulfonamide N indicate that it is sp³ hybridized.

In [Re(CO)₃(N(SO₂R)dpa)]PF₆ and other complexes that have metal-bound tertiary sulfonamide groups, the distance to the sulfonamide N from the plane defined by the S and the two C atoms attached to N typically ranges from 0.47–0.52 Å.³⁵ Such values are larger than those for the corresponding distance in complexes with an unbound sulfonamide N atom (0.06–0.26 Å).³⁵ In the new complexes, [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) and [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6), this distance (Table 2) is in the range characteristic of a M–N bond.

Most of the known complexes with tertiary sulfonamide ligands exhibit relatively long M–N(sulfonamide) bond distances and short S–N distances, features attributed to the resonance contribution of the lone pair on the sulfonamide N and to the poor electron donation of the sulfonamide N to the metal.^54,55 Thus, both the comparatively short Re–N(sulfonamide) bond distance [2.2763(14) for 4 and 2.2686(19) Å for 6] and the long S–N bond distance in these complexes [1.7434(13) for 4 and 1.759(2) Å) for 6] indicate that the sulfonamide N is a relatively strong donor.

Only one other structurally characterized complex in which a diethylenetriamine derivative is bound to a metal (La^III) via a sulfonamide bond through the N has been reported;⁵⁶ the sulfonamide group in that complex, however, is part of a cyclic heptadentate ligand. Thus, the sulfonamide complexes reported here are the first structurally characterized metal complexes in which the metal is coordinated by the central tertiary sulfonamide N in a linear, highly flexible tridentate ligand.

We have previously discussed at length chelate ring pucker λ or δ chirality and the consequences on structure and NMR spectra of [Re(CO)₃(polyamine)]X complexes.^38,46,50 The cation of [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) and one of the cations of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) (the one shown in Figure 2) have five-membered chelate rings of different chirality (λ or δ). However, the other independent cation in the asymmetric unit of 6 (not shown) has chelate rings of the same chirality. The relative frequency of observing the same or different chiralities in the two chelate rings with a central tertiary sulfonamide N donor (4 and 6) suggests that structures in which the chirality of the rings differs may be slightly more stable than structures having rings of the same chirality. Thus, our finding agrees with cases in which the central N is a classical sp³ N donor, namely that examples in which the chirality of the rings in a given structure differs (λδ or δλ) occur more frequently than those in which the chirality is the same (λλ or δδ).⁴⁶

NMR Spectroscopy

The N(SO₂R)dien ligands and the [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes reported here were characterized by ¹H NMR (1–6) and ¹³C NMR (1, 3–6) spectroscopy, usually in one or more of several solvents (CDCl₃, acetone-d₆, acetonitrile-d₃ and DMSO-d₆). NMR signals were assigned by analyzing the splitting pattern, integration, and data from 2D NMR experiments.

The two methylene ¹H NMR signals in N(SO₂R)dien ligands (1–3) are triplets integrating to four protons in CDCl₃ (Experimental Section and Figure S2). Selected data were obtained in acetonitrile-d₃ or in DMSO-d₆ (Experimental Section and Figure S3). The C(5/6)H₂ triplet for 2 (R = dmb) and for 3 (R = tol) is slightly upfield from the C(5/6)H₂ triplet for 1 (R = Me) in CDCl3 (Figure S2). Shift values for the C(4/7)H₂ triplet are similar for all three ligands. These observations are attributed to the anisotropic sulfonamide aromatic groups, which are close to the C(5/6)H₂ groups but far from the C(4/7)H₂ groups. Similar results were obtained for 1 and 3 in acetonitrile-d₃. The two protons in each methylene group are not sterically similar because the sulfonamide N lies out of the plane defined by the S and the two C atoms attached to N; the apparent magnetic equivalence in the NMR spectra indicates that inversion at the sulfonamide nitrogen leads to time averaging, as discussed previously for the N(SO₂R)dpa ligands.³⁵

The spectra of the respective complexes, on the other hand, are consistent with each of the four protons in the chelate ring giving rise to a multiplet. For example, the methylene group ¹H NMR signals observed in acetonitrile-d₃ changed from two triplets (at 3.07 and 2.79 ppm) for the N(SO₂tol)dien ligand (3) to four multiplets [at 3.51, 3.16 (two overlapped), and 2.59 ppm; see Experimental Section and Supporting Information, Figure S4] for [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6). The least amount of overlap of the multiplets for complexes 4–6 was found in acetone-d₆ (Figure 4). Thus, as expected, the apparent magnetic equivalence of methylene protons in the free ligands is lost in 4–6. Similar changes in signals for methylene protons upon coordination of ligands have been reported for [Re(CO)₃(N(SO₂R)dpa)]PF₆ complexes³⁵ and for [Re(CO)₃(N(H)dien)]PF₆.⁴⁶

Figure 4.

¹H NMR spectra of [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes (4–6) in acetone-d₆ at 25 °C.

We designate the magnetically distinct protons in −CH₂– or −NH₂ groups as endo-H or exo-H protons, on the basis of the orientation of the proton either toward (endo) or away (exo) from the carbonyl ligands (Chart 1). Note that the chelate rings in these complexes are puckered, and the corresponding protons in the two chelate rings are not equivalent in the solid state. However, the chelate rings are fluxional, and time averaging leads to only one signal for each type of CH or NH proton, or six 1H NMR signals for the N(C(5/6)H₂–C(4/7)H₂–NH₂)₂ tridentate framework moiety.

Chart 1.

Front (left) and side (right) views of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6), showing the designation of endo- and exo-CH and endo- and exo-NH protons.

¹H NMR Assignments

¹H NMR signals of the protons in the terminal NH₂ groups of previously reported [Re(CO)₃(L)]ⁿ compounds (L = simple tridentate ligands similar to diethylenetriamine with two terminal −NH₂ groups)^46,50 are well resolved. The endo-NH signal has a more downfield chemical shift than the exo-NH signal; solvent has access to the exposed endo-NH protons, but solvent access to the exo-NH protons is impeded by the chelate rings.^46,50 The signal of the exposed endo-NH protons is shifted downfield by NH-solvent H-bonding. Both the shifts and the characteristic shift separation between these signals (Δ = ∼1.5 ppm in DMSO-d₆) are useful for assigning such NH NMR signals.^46,50

In this work, two broad ¹H NMR signals of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) (5 mM) in DMSO-d₆ at 4.28 and 5.70 ppm decreased in size when D₂O was added, indicating that they are NH signals. The signals are connected by a strong COSY cross-peak, as expected for an NH₂ group (Figure S5). The downfield signal at 5.70 ppm and the relatively upfield signal at 4.28 ppm, each integrating to two protons with Δ = ∼1.5 ppm, can be assigned as endo-NH and exo-NH signals, respectively.^46,50

Similar Δ values of∼1.5 ppm (DMSO-d₆), ∼1.1 ppm (acetone-d₆), and ∼1.0 ppm (acetonitrile-d₃) were observed between the downfield and upfield NH signals for all three complexes studied here (Table 3). This consistency aided in assigning the endo-NH and exo-NH signals in complexes 4–6 (Table 3, Figure 4). Although X-ray quality crystals for compound 5 have not been isolated, the ¹H NMR patterns and shifts of its chelate ring signals in all solvents used match the corresponding ¹H NMR data for the crystallographically characterized compound 6 (Table 3 and Figure 4). These ¹H NMR chemical shifts provide strong evidence that 5 has a structure very similar to that of 6. The NH signals of 6 in different solvents (Figure 5) are shifted downfield as the H-bonding ability of the solvent increases. Thus, significant downfield chemical shifts are observed in acetone-d₆ and DMSO-d₆, in contrast to the relatively upfield chemical shifts observed in acetonitrile-d₃ (Table 3 and Figure 5), owing to the weak interactions of that solvent with the NH groups. The highest Δ value (∼1.5 ppm) occurs in DMSO-d₆ because it has the greatest H-bonding ability among the three solvents.

Table 3. ¹H NMR Shifts (ppm) of exo- and endo-NH and exo- and endo-CH Signals for [Re(CO)₃(N(SO₂R)dien)]PF₆ Complexes in Various Solvents at 25 °C.

Signal	R = Me (4)	R = dmb (5)	R = tol (6)
DMSO-d6
exo-NH	4.27	4.28	4.28
endo-NH	5.65	5.72	5.70
endo-C(5/6)H	3.45	3.41	3.42
exo-C(4/7)H	3.23	3.14	3.14
endo-C(4/7)H	3.10	3.00	3.01
exo-C(5/6)H	2.97	2.65	2.66
acetone-d6
exo-NH	4.45	4.40	4.40
endo-NH	5.54	5.52	5.51
endo-C(5/6)H	3.85	3.81	3.81
exo-C(4/7)H	3.69	3.58	3.55
endo-C(4/7)H	∼3.50	3.50	3.52
exo-C(5/6)H	∼3.50	3.02	3.02
acetonitrile-d3
exo-NH	3.44	3.41	3.37
endo-NH	4.46	4.52	4.51
endo-C(5/6)H	3.50	3.52	3.51
exo-C(4/7)H	∼3.18	∼3.16	∼3.16
endo-C(4/7)H	∼3.18	∼3.16	∼3.16
exo-C(5/6)H	3.00	2.60	2.59

Figure 5.

¹H NMR spectra of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6), illustrating the relative position of NH signals observed at 25 °C in a) acetonitrile-d₃, b) acetone-d₆, and c) DMSO-d₆.

The endo-NH and exo-NH ¹H NMR signals assignments provide a starting point that allows us to exploit our structural data to achieve complete and unambiguous assignment of the ethylene ¹H NMR signals, which are rarely well resolved for coordinated ligands having such chelate rings.⁵⁰ To demonstrate our assignment strategy of assigning the multiplets to a specific endo-CH or exo-CH proton (Chart 1), we use [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) in acetone-d 6 because of the low degree of overlap of signals for the ethylene group (Figure 4). The atom-numbering system used in this discussion appears in Figure 1. Cross-peaks in COSY spectra (Supporting Information, Figure S5) and ROESY spectra (Figure 6 and Supporting Information, Figure S6) identify a multiplet as a C(4/7)H or C(5/6)H signal. In the solid state, the endo-NH protons have a shorter average distance to the endo-C(4/7)H protons (2.20 Å) than to the exo-C(4/7)H protons (2.71 Å), and the exo-NH protons have a shorter average distance to the exo-C(4/7)H protons (2.22 Å) than to the endo-C(4/7)H protons (2.43 Å). Thus, from NH-CH NOE cross-peak intensities, the CH multiplets at 3.52 ppm and 3.55 ppm are assigned to the endo- and exo-C(4/7)H protons, respectively. The average distance from the exo-C(4/7)H protons is shorter to the exo-C(5/6)H protons (2.30 Å) than to the endo-C(5/6)H protons (2.85 Å). In the ROESY spectrum, a strong NOE cross-peak from the exo-C(4/7)H multiplet to the most upfield multiplet at 3.02 ppm thus assigns the multiplet to the exo-C(5/6)H protons. Similarly, an NOE cross-peak from the endo-C(4/7)H multiplet to the most downfield multiplet at 3.81 ppm assigns this multiplet to the endo-C(5/6)H protons. The tosyl methyl singlet (2.53 ppm) of 6 has an NOE cross-peak to the tosyl doublet at 7.65 ppm, unambiguously assigning it to the tosyl H3/5 protons and the other tosyl doublet to H2/6. The latter is more downfield (8.05 ppm), consistent with the proximity of the H2/6 protons to the sulfonamide group, shown in previous work on the [Re(CO)₃(N(SO₂tol)dpa)]PF₆ analogue to have an electron-withdrawing inductive effect.³⁵ A strong H2/6-to-endo-C(5/6)H NOE cross-peak and a very weak H2/6-to-exo-C(5/6)H cross-peak is consistent with the shorter H2/6-to-endo-C(5/6)H average distance (2.60 Å) compared to the H2/6-to-exo-C(5/6)H average distance (2.81 Å), thus confirming the assignment of the endo-C(5/6)H and exo-C(5/6)H signals (Table 3).

Figure 6.

¹H-¹H ROESY spectrum of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) in acetone-d₆ at 25 °C. An expanded version of this figure is in Supporting Information (Figure S6).

¹H NMR Shift Interpretation

The exo-C(5/6)H signal is more upfield for 5 and 6 than for 4 in acetone-d₆ (Figure 4). In 6 (Figure 2 and Chart 1), the relative average distance from the centroid of the tosyl ring to the exo-C(5/6)H protons (which point toward the center of the ring) is shorter (3.45 Å) than the distance (4.09 Å) to the endo-C(5/6)H protons (Chart 1). Thus, the anisotropic effect of the aromatic ring of the R group explains the unusually upfield shift of the exo-C(5/6)H multiplet at 3.02 ppm in acetone-d₆ for 5 and 6. This explanation is supported by the fact that the exo-C(5/6)H signal is not so upfield for 4, a complex in which R lacks an aromatic ring. This same trend, with the methylene endo-CH signal downfield from the exo-CH signal and the latter signal appearing more upfield when R has an aromatic ring (tol) than when R = Me, was reported for the [Re(CO)₃(N(SO₂R)dpa)]X complexes.³⁵

Coordination of a ligand normally leads to downfield ¹H NMR shift changes attributable to the metal inductive effect. Upon coordination to Re^I of the free N(SO₂R)dien ligands 1 and 3 in acetonitrile-d₃, the downfield shift for signals of the C4/7 protons is ∼ 0.4 ppm. For example, the C(4/7)H₂ triplet for 1 shifts from 2.76 ppm downfield to ∼3.18 ppm for the overlapping C(4/7)H signals of [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4). A corollary of the relationships of the distance of the exo-C(5/6)H protons to the aromatic groups mentioned above for 6 is that the endo-C(5/6)H protons are farther from R, and the shift of the endo-C(5/6)H signal for 4 and 6 is not influenced by the R group. Hence, compared to the C(5/6)H₂ triplets of 1 and 3, the endo-C(5/6)H signals of 4 and 6 are shifted downfield (∼0.4 ppm to ∼3.5 ppm in acetonitrile-d₃). Remarkably, upon coordination of N(SO₂Me)dien to form 4, the exo-C(5/6)H signal (expected to shift downfield by ∼0.4 ppm owing to inductive effects) is shifted slightly upfield (0.14 ppm) compared to the free ligand triplet, even though R = Me is not anisotropic. Presently there are not enough other complexes with assigned signals for us to interpret the factors influencing shift, but the upfield shift may be caused by the SO₂ group anisotropy. Also, although the reaction of free amines with acetone prevents comparison using acetone-d₆, in DMSO-d₆ (the other solvent rather different from acetonitrile-d₃ that we used for the complexes) the same pattern of shift changes was observed upon coordination of 1 to form 4, including the slight upfield shift of the exo-C(5/6)H signal (Figure S3). Thus, solvent effects are probably not causing the unexpected upfield shift.

The methyl ¹H NMR signal of N(SO₂Me)dien (1) shifted downfield upon coordination of 1 to form [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) in both acetonitrile-d₃ (0.47 ppm) and DMSO-d₆ (0.66 ppm) (Experimental Section and Supporting Information, Figure S3). These downfield shifts can be attributed to the Re^I inductive effect. We previously attributed a similar downfield shift (∼0.76 ppm) observed for [Re(CO)₃(N(SO₂Me)dpa)]PF6 to the Re^I inductive effect.³⁵ A slightly greater downfield shift (∼1.03 ppm) relative to the free ligand reported for the [Re(CO)₃(N(Me)dien)]PF₆ complex in DMSO-d₆⁴⁶ can be explained by the transmission of the inductive effect through just two bonds (N–C and C–H) versus three bonds in 4 (N–S, S–C, and C–H).

¹³C NMR Shift Interpretation

Unlike ¹H NMR signals, the shifts of ¹³C NMR signals are not very dependent on solvent effects (e.g., as discussed for the NH signals) or on aromatic ring anisotropic effects (e.g., as discussed for exo-C(5/6)H signals). Instead, ¹³C NMR signals are influenced more significantly by other factors, with signals of aliphatic carbons most frequently being shifted downfield by metal inductive effects. Analysis of ¹³C NMR shifts can often provide insight into the effect of a diamagnetic metal center on the ligand. Therefore, we undertook the unambiguous assignment of the ¹³C NMR signals of selected ligands and complexes as described in Supporting Information. The ¹³C NMR signals of the chelate ring carbons shift downfield, as expected, by 4 to 5 ppm upon coordination of the.tridentate ligands 1 and 3 to form complexes 4 and 6, and the shifts for the C4/7 and C5/6 signals of 4 are very similar to the shifts for these signals of 6 (data for acetonitrile-d₃ are reported in Table S1).

The methyl group ¹³C NMR signal for the free N(SO₂Me)dien ligand (1) appears at 37.87 ppm, but it is shifted considerably upfield for [Re(CO)₃(N(SO₂Me)dien)]PF₆ (4) (33.02 ppm, Table S1 in Supporting Information). The almost 5 ppm upfield change in shift when 1 forms 4 is unusual in that ¹³C NMR shifts are influenced mainly by through-bond effects, such as the metal inductive effect expected to produce a downfield shift. For example, a significant ∼14 ppm downfield ¹³C NMR shift was observed for the methyl group of [Re(CO)₃(N(Me)dien)]PF₆ (56.48 ppm) compared to the methyl group in the free N(Me)dien ligand (42.72 ppm) in acetonitrile-d₃.⁵¹ Because of the unexpected upfield direction of the shift change of the methyl ¹³C NMR signal of 4, we measured the ¹³C NMR shift of the methyl signal of [Re(CO)₃(N(SO₂Me)dpa)]PF₆ in DMSO-d₆ (32.88 ppm), a value ∼6.5 ppm upfield compared to the methyl signal in the free N(SO₂Me)dpa ligand (39.24 ppm).⁵¹ We attribute the observed unusual upfield shift for 4 to the fact that the sulfonamide N undergoes an sp²-to-sp³ rehybridization upon coordination to Re^I in both 4 and [Re(CO)₃(N(SO₂Me)dpa)]PF₆.

Challenge Reactions of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) in DMSO- d₆

No change was observed in the ¹H NMR signals of 6 (5 mM), even after six months, indicating that 6 is robust. However, addition of diethylenetriamine to 5 mM solutions of 6 to create solutions that were 50 mM or 10 mM in diethylenetriamine led to the elimination of all peaks for 6 by the next day and after about 6 days (t_1/2 ∼28 h), respectively. In both cases, the final ¹H NMR spectrum exhibits peaks for the N(SO₂tol)dien ligand (3) and for the known [Re(CO)₃(N(H)dien)]PF₆ complex.⁴⁶ No signals indicating any intermediates were observed. The complete displacement of the coordinated N(SO₂tol)dien ligand of 6 by 10 mM diethylenetriamine establishes that although the Re–N bonds involving the central donor of N(SO₂R)dien ligands have normal lengths, the donor ability of the central sulfonamide nitrogen atom is lower than that of the central nitrogen atom of diethylenetriamine.

A related study³⁸ was reported for the neutral Re(CO)₃(tmbSO₂-dien) complex. The coordinated unsymmetrical monoanionic NNN donor ligand, tmbSO₂-dien^- (Chart 2), employed in that study has a 2,4,6-trimethylbenzenesulfonyl (tmbSO₂) group linked to one of the terminal N atoms of diethylenetriamine. The terminal donor of the free ligand prior to coordination is a secondary sulfonamide of the type (RSO₂)R'NH. When 10 molar equiv of diethylenetriamine was added to a solution of Re(CO)₃(tmbSO₂-dien) (3 mM, DMSO-d₆), no coordinated tmbSO₂-dien- ligand was displaced, even after the solution was heated at ∼60 °C.³⁸ Thus, the deprotonated monoanionic tmbSO₂-dien^- ligand is clearly a better ligand than neutral N(SO₂tol)dien.

Chart 2.

The tmbSO₂-dien^- monoanionic ligand.³⁸

If the central tertiary sulfonamide donor in N(SO₂tol)dien were coordinated weakly enough to the Re in [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes, a good monodentate ligand might be able to substitute for the coordinated sulfonamide group. Therefore, we conducted a challenge experiment in which the strongly basic, potentially monodentate 4-Me₂Npy ligand was added to [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6). No ¹H NMR spectral changes were observed over many days, even with a tenfold excess of 4-Me₂Npy. Such evidence indicates that the central sulfonamide bond is strong enough to stay coordinated to Re, even in the presence of an excess of a strong monodentate ligand. Also, the results show that the [Re(CO)₃(N(SO₂tol)dien)]PF₆ complex is resistant to strong base, an advantage over [Re(CO)₃(N(SO₂Me)dpa)]PF₆, which decomposes readily in the presence of base.³⁵

Solubility of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) in Water

A mixture of 5 mg of [Re(CO)₃(N(SO₂tol)dien)]PF₆ (6) and 450 μL of D₂O gave a clear solution when heated in a boiling water bath for 30 min. The ¹H NMR spectrum of this solution, which compares favorably with spectra of 6 in other solvents (Figure 4), has four methylene multiplets, whereas the N(SO₂tol)dien ligand has two methylene triplets in D₂O. These results established that 6 dissolved unchanged to an extent much greater than is required for radiopharmaceuticals and that 6 is robust even in a hot aqueous solution. Furthermore, another experiment conducted using 5 mg of 6 also showed complete dissolution (observed visibly and by ¹H NMR spectroscopy) of 6 in a 450 μL:20 μL of D₂O:DMSO mixture (95.7%:4.3%). In contrast, similar NMR experiments carried out using the previously reported [Re(CO)₃(N(SO₂Me)dpa)]PF₆ complex³⁵ show that even the [Re(CO)₃(N(SO₂R)dpa)]PF₆ complex having the smallest R group is insoluble in D₂O or in D₂O:DMSO (95.7%:4.3%).

Summary and Conclusions

From the present results with [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes, in which the chelate rings are less rigid, we conclude that the M–N bond of normal length observed was not a result of the rigid tridentate framework in the [Re(CO)₃(N(SO₂R)dpa)]X³⁵ complexes. In both series of complexes, the methyl ¹³C NMR signal of the R = Me member of the series exhibited a very unusual upfield shift for an aliphatic carbon signal upon coordination of the ligand. This result, attributed to the similar sp²-to-sp³ rehybridization of the sulfonamide N upon coordination to Re^I in both series of complexes, further establishes that coordination of the sulfonamide N is not influenced by the rigidity of the N(SO₂R)dpa chelate rings. The decomposition of the [Re(CO)₃(N(SO₂R)dpa)]X complexes by base³⁵ led us to hypothesize that base was attacking the coordinated N(SO₂R)dpa ligand, most likely by deprotonating the CH2 group, and that the low electrophilicity of the Z = CH₂NH₂ group of the N(CH₂CH₂NH₂)₂ framework would confer stability toward base. The stability of the new [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes toward base supports these hypotheses and allowed us to conduct a challenge reaction study with the basic N(H)dien ligand. N(H)dien replaced the coordinated N(SO₂tol)dien ligand in [Re(CO)₃(N(SO₂tol)dien)]PF₆, indicating that the neutral sulfonamide N central donor in N(SO₂tol)dien is a somewhat weaker donor than the central traditional sp³ N donor in N(H)dien. Nevertheless, the new [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes are long lived, even in the presence of base, and are relatively robust to heat treatment. As expected, the moderately hydrophilic character of the Z = CH₂NH₂ group of the N(CH₂CH₂NH₂)₂ framework also confers water solubility on the [Re(CO)₃(N(SO₂R)dien)]PF₆ complexes. The aqueous solubility of the new ligands and complexes is much higher than necessary for radiopharmaceutical kit formulation. Also, there are perceived advantages in using small chelate ligands when constructing bioconjugates, and the N(CH₂CH₂NH₂)₂ framework is relatively small. Thus, the results obtained here suggest that N(SO₂R)dien ligands should be explored in the development of radiopharmaceuticals, including bioconjugates.