Formation of a Metal-to-Nitrogen Bond of Normal Length by a Neutral Sufonamide Group within a Tridentate Ligand. A New Approach to Radiopharmaceutical Bioconjugation

Abstract

We demonstrate that a tertiary sulfonamide group, N(SO₂R)R′₂, can re-hybridize to form a M–N bond of normal length even when the group is in a linear tridentate ligand, such as in the new tridentate N(SO₂R)dpa ligands derived from di-(2-picolyl)amine (N(H)dpa). N(SO₂R)dpa ligands were used to prepare fac-[Re(CO)₃(N(SO₂R)dpa)](PF₆ or BF₄) complexes. Structural characterization of the new complexes established that the tertiary sulfonamide nitrogen atom binds to Re with concomitant sp²-to-sp³ re-hybridization, facilitating facial coordination. The new fac-[Re(CO)₃(N(SO₂R)dpa)]X structures provide the only examples for any metal with the sulfonamide as part of a noncyclic linear tridentate ligand and with a normal metal-to-nitrogen(tertiary sulfonamide) bond length. Rare previous examples of such normal M–N bonds have been found only in more constrained situations, such as with tripodal tetradentate ligands. Our long-term objectives for the new tridentate N(SO₂R)dpa ligands are to develop the fundamental chemistry relevant to the eventual use of the fac-[M^I(CO)₃]⁺ core (M = ^99mTc, ^186/188Re) in imaging and therapy. The sulfonamide group uniquely contributes to two of our goals: expanding ways to conjugate the fac-[M^I(CO)₃]⁺ core to biological molecules and also developing new symmetrical tridentate ligands that can coordinate facially to this core. Tests of our conjugation method, conducted by linking the fac-[Re^I(CO)₃]⁺ core to a new tetraarylporphyrin (T(N(SO₂C₆H₄)dpa)P) as well as to a dansyl (5-(dimethylamino)naphthalene-1-sulfonyl) group, demonstrate that large molecular fragments can be tethered via a coordinated tertiary sulfonamide linkage to this core.

Introduction

Improved methods of ligand design are needed for the synthesis of radiopharmaceuticals containing the fac-[^99mTc^I(CO)₃]⁺ core; such agents have received widespread attention in the past two decades.^1-6 Many factors have contributed to this interest, including the convenient generation of the fac-[^99mTc(CO)₃(H₂O)₃]⁺ precursor,^7,8 the stability and kinetic inertness of fac-[^99mTc(CO)₃L]ⁿ imaging agents when L is a facially coordinated tridentate ligand,^1,4,9 and the potential that novel ligands bound to this core will lead to agents with new or superior clinical utility.¹⁰ Such fac-[^99mTc(CO)₃L]ⁿ agents have been shown to possess desirable biological characteristics in human volunteers,^4,11 with fac-[^99mTc(CO)₃(NTA)]^2- (NTA = nitrilotriacetate)¹¹ showing promise as a renal imaging agent in early patient studies.¹² Historically, the study of nonradioactive Re analogues has played an important role in guiding ligand design and in providing a basis for understanding the chemistry of ^99mTc diagnostic agents and ^186/188Re therapeutic agents.^1,13,14 Likewise, many fac-[Re(CO)₃L]ⁿ analogues have more recently facilitated the fac-[^99mTc(CO)₃L]ⁿ biomedical studies by serving as models that allow insight into the chemistry and biomedical characteristics of these short-lived radioactive ^99mTc imaging agents.^4,15-20

Symmetrical linear tridentate ligands that do not induce asymmetry at the metal center are desirable for avoiding racemic or diastereoisomeric mixtures of radiopharmaceuticals. For example, di-(2-picolyl)amine (N(H)dpa) forms a symmetrical complex, fac-[Re(CO)₃(N(H)dpa)]X.²¹ Promising biomedical properties have been reported for fac-[Re(CO)₃(N(R)dpa)]ⁿ complexes bearing a substituent linked by a C–N bond at the central sp³ N, replacing the amine proton of N(H)dpa.^10,22,23 In referring to such ligands in this article, we use N(substituent)dpa, with the italic N designating the central nitrogen. Ligand design requires that the central donor of the tridentate ligand, such as the central sp³ N in N(H)dpa or N(R)dpa, accommodates facial coordination. A thioether sulfur atom achieves this geometry,^16,24,25 but substituents imparting desirable biodistribution or targeting characteristics cannot be attached at a central sulfur donor atom.

Full exploitation of the fac-[M^I(CO)₃]⁺ (M = ^99mTc, Re) core relies on new types of tridentate ligands and ligand conjugation methods.^8,21,23 In this study, we explore conjugation utilizing an N–S bond by preparing new N(SO₂R)dpa ligands (R = Me, tmb, dmb, and Me₂Nnap; see Figure 1 for ligand sketches and abbreviations). We show that treatment of fac-[Re(CO)₃(H₂O)₃]⁺ with such open-chain tridentate N(SO₂R)dpa ligands affords fac-[Re(CO)₃(N(SO₂R)dpa)]X complexes (Scheme 1), although the central N-donor atom is part of a tertiary sulfonamide group (N(SO₂R)R′₂). X-ray structural characterization shows that this nitrogen atom is a strong donor to Re. Our study of fac-[Re(CO)₃(N(SO₂R)dpa)]X complexes provides the only examples of structurally characterized complexes with a tertiary neutral sulfonamide donor bound to the fac-[Re^I(CO)₃]⁺ core. Furthermore, these compounds are the only examples of metal complexes in which any such open-chain tertiary sulfonamide linear tridentate ligand possesses a M–N(sulfonamide) bond of normal length, regardless of the metal center. Indeed, these are among the very few crystallographically characterized complexes containing tertiary sulfonamide N donors with any metal.^26-33 In previously reported sulfonamide complexes, the bound sulfonamide is either constrained within a ring system or has elongated bonds.^26-33

Figure 1.

Ligands mentioned or used in this study: di(2-picolyl)amine (N(H)dpa); N,N-di(2-picolyl)methanesulfonamide (N(SO₂Me)dpa); N,N-di(2-picolyl)-2,4,6-trimethylbenzenesulfonamide(N(SO₂tmb)dpa); N,N-di(2-picolyl)-3,5-dimethylbenzenesulfonamide (N(SO₂dmb)dpa); N,N-di(2-picolyl)-4-methylbenzenesulfonamide (N(SO₂tol)dpa); and N,N-di(2-picolyl)-5-(dimethylamino)naphthalene-1-sulfonamide (N(Me₂Nnap)dpa).

Scheme 1.

Synthesis of [Re(CO)₃(NSO₂R)dpa)]⁺ Complexes.

In the design of new ligands, the nature of the geometric constraints induced by the central donor atoms is an important factor. The ligand should be able to coordinate facially. As an example, the propensity of thioethers to enforce a facial coordination mode^34,35 was utilized in designing renal agents, including agents bearing lanthionine dipeptide ligands.^4,16,17,36 These NSN-donor dipeptides with amino acids linked via a thioether group were used to study small fac-[^99mTc(CO)₃L]ⁿ agents in humans because the more traditional peptides coordinate with a deprotonated nitrogen, which enforces the meridional coordination mode.⁴ The sulfonamide group explored here has an additional advantage over the thioether or peptide N-donor groups in being able to serve as a juncture point for conjugation. We validated the versatility of the sulfonamide method by conjugating the dansyl group or a tetraarylporphyrin biological targeting group to the fac-[Re^I(CO)₃]⁺ core. (From this point on, we omit the fac- designation when discussing specific compounds because all compounds have this geometry.)

Experimental Section

Starting Materials

Methanesulfonyl chloride (MeSO₂Cl), 2,4,6-trimethylbenzenesulfonyl chloride (tmbSO₂Cl), 3,5-dimethylbenzenesulfonyl chloride (dmbSO₂Cl), 4-methylbenzenesulfonyl chloride (tolSO₂Cl), 5-(dimethylamino)naphthalene-1-sulfonyl chloride (Me₂NnapSO₂Cl, dansyl chloride), di-(2-picolyl)amine (N(H)dpa), and Re₂(CO)₁₀ were used as received from Aldrich. [Re(CO)₃(H₂O)₃]OTf (OTf = trifluoromethanesulfonate) was prepared by a known method.³⁶ [Re(CO)₃(CH₃CN)₃]BF₄ and meso-tetrakis(4-chlorosulfonylphenyl)porphyrin (TClSO₂PP) were synthesized as described elsewhere,^37,38 and their ¹H NMR spectra were in agreement with those reported.

NMR Measurements

¹H NMR spectra were recorded on a Bruker 400 MHz spectrometer. Peak positions are relative to tetramethylsilane (TMS) or solvent residual peak with TMS as reference. All NMR data were processed with TopSpin and Mestre-C software.

X-ray Data Collection and Structure Determination

Intensity data were collected at low temperature on a Nonius Kappa CCD diffractometer fitted with an Oxford Cryostream cooler and graphite-monochromated Mo Kα (λ = 0.71073 Å) radiation. Data were collected at 90 K for 1 and 2, but for 4 (T = 200 K) and 5 (T = 220 K), higher temperatures were chosen to avoid phase changes that occurred with twinning. Data reduction included absorption corrections by the multi-scan method, using HKL SCALEPACK.³⁹ All X-ray structures were determined by direct methods and difference Fourier techniques and refined by full-matrix least squares by using SHELXL97.⁴⁰ All non-hydrogen atoms were refined anisotropically. All H atoms except those on water molecules in 2 and 5 were visible in difference maps, but were placed in idealized positions. Those on water were not located and are likely disordered. A torsional parameter was refined for each methyl group. For 1, the PF₆^- anion was disordered into two positions treated with equal populations. For 5, the BF₄^- anion site exhibited both substitutional and orientational disorder. The site was occupied by bromide with 6.8(3)% population, and the BF₄^- had two orientations with relative population 59(3):34(3)%. Structure 2 contains two independent formula units in the asymmetric unit, only one of which is shown in Figure 2. Crystal data and refinement details are included in Table 1.

Figure 2.

ORTEP plots of the cations in [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1), [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2), [Re(CO)₃(N(SO₂tol)dpa)]PF₆ (4), and [Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ (5). Thermal ellipsoids are drawn with 50% probability.

Table 1.

Crystal Data and Structure Refinement for [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1), [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2), [Re(CO)₃(N(SO₂tol)dpa)]PF₆ (4), and [Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ (5)

	1a	2	4	5
empirical formula	C16H15N3O5ReS·PF6	C24H23N3O5ReS·PF6·H2O	C22H19N3O5ReS ·PF 6	C27H24N4O5ReS· 0.93(BF4)·0.07(Br)·4H2O
fw	692.54	814.70	768.63	861.58
crystal system	monoclinic	triclinic	monoclinic	monoclinic
space group	C2/c	P1̄	C2/c	C2/c
unit cell dimensions
a (Å)	20.549(2)	14.353(2)	15.102(2)	33.684(4)
b (Å)	20.297(2)	15.010(2)	10.4922(15)	8.1471(10)
c (Å)	12.5771(10)	16.028(3)	34.186(4)	23.635(3)
α (deg)	90	110.298(8)	90	90
β (deg)	123.039(4)	115.806(7)	100.598(4)	97.278(6)
γ (deg)	90	96.429(9)	90	90
V (Å3)	4397.5(7)	2770.6(7)	5324.5(12)	6433.8(14)
T (K)	90	90	200	220
Z	8	4	8	8
ρcalc (g/cm3)	2.092	1.953	1.918	1.779
abs coeff (mm-1)	5.78	4.61	4.78	4.01
2θmax (°)	71.2	55.8	55.0	52.2
R b [I > 2σ(I)]	0.026	0.048	0.029	0.037
wR2 c	0.055	0.105	0.087	0.088
Res. dens (eÅ-3)	-1.36, 1.58	-1.29, 1.41	-1.10, 1.20	-0.96, 1.09
data/param	10038/332	13140/764	5756/353	6317/456

^aThe asymmetric unit of 1 contains one [Re(CO)₃(N(SO₂Me)dpa)]⁺ cation and half of two crystallographically independent PF₆^– anions; one anion lies on an inversion center, and the other anion lies on a twofold axis and is disordered.

^bR = (Σ‖ F_o| - | F_c‖)/Σ|F_o|;

^cwR2 = [Σ[w(F_o² - F_c²)²]/Σ[w(F_o²)²]]^1/2, in which w = 1/[σ² (F_o)² + (dP)² + (eP)] and P = (F_o² + 2F_c²)/3; d = 0.021, 0.0463, 0.041, and 0.0399, and e = 0.6452, 0, 0, and 2.129 for 1, 2, 4, and 5, respectively.

Synthesis of [Re(CO)₃L]PF₆ and [Re(CO)₃L]BF₄ Complexes

The following general procedure was employed to obtain the N(SO₂R)dpa ligands: a solution of the sulfonyl chloride (0.9–1.5 g, 5 mmol) in 25 mL of dioxane was added dropwise over a period of 2 h to a solution of N(H)dpa (1.80 mL, 10 mmol) in 100 mL of dioxane at 20 °C. The reaction mixture was stirred at room temperature for 24 h and then filtered to remove any precipitate before the dioxane was completely removed by rotary evaporation. Water (30 mL, pH ∼5) was added to the resulting oil, and the product was extracted into CH₂Cl₂ (2 × 25 mL). The CH₂Cl₂ extracts were combined, washed with water (2 × 25 mL), and taken to dryness to yield an oil, which was used to synthesize [Re(CO)₃L]PF₆ and [Re(CO)₃L]BF₄ in the general procedure outlined here. The preparation utilized 0.1 mmol of the reactants. An aqueous solution of the ligand (0.1 mmol in 2 mL) was treated with aqueous [Re(CO)₃(H₂O)₃]⁺ (0.1 mmol in 3 mL). Methanol (2–3 mL) was added to dissolve any precipitate that formed, and the clear reaction mixture was heated at reflux for 12 h. A slight excess of NaPF₆ or NaBF₄ was added to the clear solution, and the resulting precipitate was collected on a filter, washed with water, and air dried. (If the pH of the final reaction mixture was below 6, it was adjusted to 7 before the addition of NaPF₆ or NaBF₄. For N(SO₂tmb)dpa and N(SO₂Me₂Nnap)dpa, methanol (∼1 mL) was used initially to dissolve the ligand).

[Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1)

The general ligand preparative method described above, with MeSO₂Cl (0.39 mL) and N(H)dpa (1.80 mL), yielded the crude N(SO₂Me)dpa ligand as a deep red oil (1.40 g, 76% yield). ¹H NMR signals (ppm) in DMSO-d₆: 8.50 (d, J = 6.5 Hz, 2H, H6/6′), 7.75 (t, J = 7.6 Hz, 2H, H4/4′), 7.34 (d, J = 7.8 Hz, 2H, H3/3′), 7.27 (t, J = 6.5 Hz, 2H, H5/5′), 4.48 (s, 4H, CH₂), 3.11 (s, 3H, CH₃). In the 0.1 mmol preparative method described above, N(SO₂Me)dpa (28 mg) was used. [Re(CO)₃(N(SO₂Me)dpa)]PF₆ formed as a white precipitate (38 mg, 54% yield) after the addition of NaPF₆ (∼ 15 mg). Slow evaporation of a solution of this compound in acetone produced colorless, needle-like crystals that were characterized crystallographically. ¹H NMR signals (ppm) in DMSO-d₆: 8.89 (d, J = 5.4 Hz, 2H, H6/6′), 8.09 (t, J = 6.7 Hz, 2H, H4/4′), 7.57 (d, 2H, H3/3′), 7.53 (t, J = 6.5 Hz, 2H, H5/5′), 5.44 (d, J = 15.8 Hz, 2H, CH₂), 5.13 (d, J = 15.8 Hz, 2H, CH₂), 3.87 (s, 3H, CH₃).

[Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2)

The general method described above, with tmbSO₂Cl (1.10 g) and N(H)dpa (1.80 mL), yielded N(SO₂tmb)dpa as a pale orange oil (1.75 g, 92% yield). ¹H NMR signals (ppm) in DMSO-d₆: 8.43 (d, J = 6.4 Hz, 2H, H6/6′), 7.65 (t, J = 7.9 Hz, 2H, H4/4′), 7.22 (t, J = 6.1 Hz, 2H, H5/5′), 7.09 (d, J = 7.8 Hz, 2H, H3/3′), 6.97 (s, 2H, H3/5), 4.53 (s, 4H, CH₂), 2.53 (s, 6H, CH₃), 2.23 (s, 3H, CH₃). In the 0.1 mmol preparative method, N(SO₂tmb)dpa (38 mg) afforded [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ as a white precipitate (51 mg, 64% yield) after the addition of NaPF₆ (∼15 mg). Slow evaporation of a solution of the compound in CHCl₃ produced colorless, block-shaped crystals that were characterized crystallographically. ¹H NMR signals (ppm) in DMSO-d₆: 8.90 (d, J = 5.4 Hz, 2H, H6/6′), 8.05 (t, J = 7.8 Hz, 2H, H4/4′), 7.64 (d, J = 7.8 Hz, 2H, H3/3′), 7.49 (t, J = 6.6 Hz, 2H, H5/5′), 7.46 (s, 2H, H3/5), 5.22 (d, J = 15.9 Hz, 2H, CH₂), 4.53 (d, J = 15.9 Hz, 2H, CH₂), 2.78 (s, 6H, CH₃), 2.43 (s, 3H, CH₃).

[Re(CO)₃(N(SO₂dmb)dpa)]BF₄ (3)

The general method described above, with dmbSO₂Cl (1.02 g) and N(H)dpa (1.80 mL), yielded N(SO₂dmb)dpa as a pale orange oil (1.45 g, 86% yield). ¹H NMR signals (ppm) in DMSO-d₆: 8.37 (d, J = 5.6 Hz, 2H, H6/6′), 7.67 (t, J = 6.8 Hz, 2H, H4/4′), 7.35 (s, 2H, H2/6), 7.28 (d, J = 7.8 Hz, 2H, H3/3′), 7.21 (s, 1H, H4), 7.19 (t, J = 5.9 Hz, 2H, H5/5′), 4.54 (s, 4H, CH₂), 2.29 (s, 6H, CH₃). In the 0.1 mmol preparative method, N(SO₂dmb)dpa (37 mg) afforded [Re(CO)₃(N(SO₂dmb)dpa)]BF₄ as a white precipitate (55 mg, 67% yield) after the addition of NaBF₄ (∼15 mg). ¹H NMR signals (ppm) in DMSO-d₆: 8.88 (d, J = 5.4 Hz, 2H, H6/6′), 8.02 (t, J = 7.7 Hz, 2H, H4/4′), 7.96 (s, 2H, H2/6), 7.72 (s, 1H, H4), 7.51 (d, J = 8.4 Hz, 2H, H3/3′), 7.46 (t, J = 6.4 Hz, 2H, H5/5′), 5.25 (d, J = 16.2 Hz, 2H, CH₂), 4.55 (d, J = 16.2 Hz, 2H, CH₂), 2.51 (s, 6H, CH₃, overlap with solvent residual peak). Anal. Calcd for C₂₃H₂₁BF₄N₃O₅ReS: C, 38.10; H, 2.92; N, 5.80. Found: C, 37.87; H, 2.83; N, 5.88.

[Re(CO)₃(N(SO₂tol)dpa)]PF₆ (4)

The general method described above, with tolSO₂Cl (0.95 g) and N(H)dpa (1.80 mL), yielded N(SO₂tol)dpa as a deep brown oil (1.52 g, 86% yield). ¹H NMR signals (ppm) in DMSO-d₆: 8.35 (d, J = 4.8 Hz, 2H, H6/6′), 7.70 (d, J = 8.2 Hz, 2H, H2/6), 7.65 (t, J = 7.7 Hz, 2H, H4/4′), 7.36 (d, J = 8.0 Hz, 2H, H3/5), 7.25 (d, J = 7.8 Hz, 2H, H3/3′), 7.19 (t, J = 4.9 Hz, 2H, H5/5′), 4.49 (s, 4H, CH₂), 2.39 (s, 3H, CH₃). In the 0.1 mmol preparative method, N(SO₂tol)dpa (35 mg) afforded [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ as a white precipitate (50 mg, 65% yield) after the addition of NaPF₆ (∼15 mg). Slow evaporation of a solution of the compound in CHCl₃ produced colorless, block-shaped crystals that were characterized crystallographically. ¹H NMR signals (ppm) in DMSO-d₆: 8.87 (d, J = 5.4 Hz, 2H, H6/6′), 8.21 (d, J = 8.4 Hz, 2H, H2/6), 8.02 (t, J = 7.8 Hz, 2H, H4/4′), 7.76 (d, J = 8.4 Hz, 2H, H3/5), 7.48 (d, J = 7.9 Hz, 2H, H3/3′), 7.47 (t, 2H, H5/5′), 5.58 (d, J = 16.0 Hz, 2H), 4.53 (d, J = 16.0 Hz, 2H), 2.57 (s, 3H, CH₃).

[Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ (5)

The general method described above, with dansyl chloride (1.42 g) and N(H)dpa (1.8 mL), yielded N(SO₂Me₂Nnap)dpa as a pale orange oil (1.96 g, 91% yield). ¹H NMR signals (ppm) in DMSO-d₆: 8.44 (d, J = 8.5 Hz, 1H), 8.37 (d, J = 4.5 Hz, 2H, H6/6′), 8.20 (d, J = 8.6 Hz, 1H), 8.16 (d, J = 7.3 Hz, 1H), 7.58 (t, 2H, H4/4′), 7.54 (m, 2H), 7.23 (d, J = 7.6 Hz, 1H), 7.18 (t, 2H, H5/5′), 7.12 (d, J = 7.8 Hz, 2H, H3/3′), 4.72 (s, 4H, CH₂), 2.82 (s, 6H, CH₃). In the 0.1 mmol preparative method, N(SO₂Me₂Nnap)dpa (43 mg) afforded [Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ as yellow block-like crystals (38 mg, 44% yield) after the addition of NaBF₄ (∼15 mg). The product was characterized crystallographically. ¹H NMR signals (ppm) in DMSO-d₆: 8.90 (m, 3H), 8.71 (d, J = 7.7 Hz, 1H), 8.58 (d, J = 8.7 Hz, 1H), 7.98 (t, J = 8.1 Hz, 2H, H4/4′), 7.92 (t, J = 7.9 Hz, 1H), 7.78 (t, J = 8.2 Hz, 1H), 7.47 (t, J = 6.6 Hz, 2H, H5/5′), 7.42 (d, J = 7.6 Hz, 2H, H3/3′), 7.39 (d, J = 7.9 Hz, 1H), 5.64 (d, J = 15.7 Hz, 2H), 4.53 (d, J = 15.8 Hz, 2H, CH₂), 2.91 (s, 6H, CH₃).

T(N(SO₂C₆H₄)dpa)P

A solution of TClSO₂PP (0.4 g, 0.39 mmol) in CH₂Cl₂ (50 mL) was treated with N(H)dpa (3.7 mL, 1.98 mmol), and the reaction mixture was stirred at room temperature for 24 h. Impurities in this mixture were then extracted with water (3 × 25 mL); the organic layer was dried over anhydrous Na₂SO₄ and the solvent removed under vacuum. The purple residue was crystallized from CH₂Cl₂/hexane, and the purple crystals were washed with hexane (0.32 g, 49% yield). ¹H NMR signals (ppm) in DMSO-d₆: 8.87 (s, 8H, βH), 8.55 (d, J = 4.1 Hz, 8H, H6/6′), 8.37 (d, J = 8.3 Hz, 8H, oH), 8.24 (d, J = 8.4 Hz, 8H, mH), 7.83 (t, 8H, H4/4′), 7.49 (d, J = 7.9 Hz, 8H, H3/3′), 7.34 (t, J = 6.2 Hz, 8H, H5/5′), 4.90 (s, 16H, CH₂), -2.96 (s, 2H, NH). ESI-MS m/z): [M + H]⁺ = 1660.4881, [M + 2H] ²⁺ = 830.7465. Calcd for [M + H]⁺ = 1660.4886, [M + 2H] ²⁺ = 830.7443.

[[Re(CO)₃]₄T(N(SO₂C₆H₄(dpa)P](BF₄)₄ (6)

[Re(CO)₃(CH₃CN)3]BF₄ (28 mg, 0.058 mmol) was added to a solution of T(N(SO₂C₆H₄)dpa)P (20 mg, 0.012 mmol) in a mixture of chloroform/acetone (25 mL/5 mL). The reaction mixture was heated at reflux for 16 h and then reduced to dryness by rotary evaporation. The maroon residue obtained was quickly washed with CH₂Cl₂, dissolved in acetone, and the resultant solution layered with hexane to give a maroon precipitate (15 mg, 40% yield). ¹H NMR signals (ppm) in DMSO-d₆: 9.19 (s, 8H, βH), 9.01 (d, J = 5.4 Hz, 8H, H6/6′), 8.84 (s, 16H, o and mH), 8.14 (t, J = 7.9 Hz, 8H, H4/4′), 7.71 (d, J = 8.0 Hz, 8H, H3/3′), 7.58 (t, J = 6.8 Hz, 8H, H5/5′), 5.98 (d, J = 16.4 Hz, 8H, CH2), 4.97 (d, J = 16.4 Hz, 8H, CH2), -2.81 (s, 2H, NH).

Results and Discussion

Synthesis of N(SO₂R)dpa and [Re(CO)₃(N(SO₂R)dpa)]X (X = PF₆ or BF₄)

Potentially tridentate ligands with R = Me, tmb, dmb, tol, and Me₂Nnap groups at the central N (Figure 1) were synthesized in good yield by coupling N(H)dpa with the relevant sulfonyl chloride. NMR spectral data indicated that the products were uncontaminated with by-products and could be used without further purification (cf. Experimental Section and Figure S1, Supporting Information). Aqueous [Re(CO)₃(H₂O)₃]OTf³⁶ was used to prepare [Re(CO)₃(N(SO₂R)dpa)]X complexes 1-5 (Scheme 1). NMR spectra of the new complexes are illustrated below or in the Supporting Information (Figure S2). No change was observed in the ¹H NMR signals of 1 and 4 (5 mM, DMSO-d₆), even after six months.

Synthesis of the Porphyrin, T(N(SO₂C₆H₄)dpa)P, and the 4:1 Re Adduct, [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄

We have utilized our synthetic approach reported for T(R¹R²NSO₂C₆H₄)P [R¹ = N-py-n-CH₂ (n = 2 or 4) and R² = alkyl)]⁴¹ in order to prepare T(N(SO₂C₆H₄)dpa)P, a porphyrin bearing four tertiary N(SO₂)dpa moieties (Scheme 2). Because the porphyrin is not water soluble, we employed the [Re(CO)₃(CH₃CN)₃]BF₄ precursor³⁷ to form [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄. In this 4:1 adduct, each fac-[Re^I(CO)₃]⁺ moiety is coordinated to three N-donor atoms (of a peripheral tertiary dpa moiety) linked to the porphyrin aryl group. In previous work, Alessio and co-workers utilized the [Re(CO)₃(DMSO)₃]⁺ precursor to synthesize [Re(CO)₃(bipyridine)(DMSO)]⁺,⁴² which was then used to prepare 1:1 and 4:1 adducts between Re^I(CO)₃(bipyridine) units and meso-tetrakis(4-pyridyl)porphyrin (TpyP(4)).⁴³ In these Re-porphyrin adducts, only one Re–N bond, involving a peripheral meso-pyridyl group, links the Re moiety to the porphyrin. The [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄ adduct prepared in this study is expected to maintain its integrity better because each Re moiety is linked by three donor atoms.

Scheme 2.

Synthesis of T(NSO₂C₆H₄)dpa)P.

Structural Results

Crystal data and details of the structural refinement for complexes 1, 2, 4, and 5 are summarized in Table 1. The Re complexes reported here exhibit a pseudo octahedral structure (Figure 2), with the three carbonyl ligands occupying one face. The remaining three coordination sites are occupied by three nitrogen atoms of the tridentate ligands. The atom numbering systems in Figure 2 are used to describe the solid-state data.

The unusual and most interesting structural features of these complexes involve the sulfonamide containing the central nitrogen (N2), which is bound to Re. First, we present information on other structural aspects, beginning with a consideration of bond lengths. In [Re(CO)₃L]⁺ complexes with prototypical NNN donor ligands, Re–N(sp³) bond distances (generally ∼2.23–2.29 Å)^44,45 are longer than Re–N(sp²) bond distances (generally 2.14–2.18 Å).^37,46,47 The Re–N(sp²) bond distances involving the pyridyl groups (N1 and N3, Figure 2) of [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1), [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2), [Re(CO)₃(N(SO₂tol)dpa)]PF₆ (4), and [Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ (5) are similar and fall within a normal range (Table 2). For the most part, the Re–N(pyridyl) bond distances are not significantly different from the values reported for [Re(CO)₃(N(H)dpa)]Br (2.177(5) and 2.183(5) Å)²¹ and for [Re(CO)₃(N(CH₂CO₂Et)dpa)]Br (2.161(6) and 2.174(5) Å).²¹ In addition, superimposing the donor N and the Re atoms in [Re(CO)₃(N(CH₂CO₂Et)dpa)]Br with the corresponding atoms in 1 reveals very good overlap of the dpa moieties of both compounds (Figure S3, Supporting Information). Thus, any effect of having a tertiary sulfonamide nitrogen, rather than a traditional sp³ nitrogen, anchoring the two chelate rings is minimal.

Table 2.

Selected Bond Distances (Å) and Angles (deg) for [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1) and [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2), [Re(CO)₃(N(SO₂tol)dpa)]PF₆ (4), and [Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ (5)

	1	2	4	5
bond distances
Re–N1	2.1736(17)	2.175(5)	2.179(3)	2.170(4)
Re–N2	2.2826(16)	2.279(5)	2.275(3)	2.272(4)
Re–N3	2.1948(18)	2.175(5)	2.173(3)	2.181(4)
Re–C1	1.928(2)	1.909(7)	1.937(4)	1.925(6)
Re–C2	1.905(2)	1.893(7)	1.907(5)	1.900(6)
Re–C3	1.936(2)	1.936(8)	1.938(5)	1.945(6)
S1–O5	1.4288(17)	1.433(5)	1.416(3)	1.425(4)
S1–O4	1.4299(15)	1.428(4)	1.437(3)	1.420(3)
S1–N2	1.7552(17)	1.755(5)	1.752(3)	1.776(4)
bond angles
N1–Re–N2	77.16(6)	76.98(18)	75.86(12)	78.08(14)
N1–Re–N3	80.12(7)	86.49(19)	78.85(12)	78.42(16)
N2–Re–N3	76.25(6)	73.96(18)	77.58(11)	76.37(15)
C2–Re–N1	94.33(8)	100.3(3)	96.83(16)	95.53(19)
C3–Re–N1	96.45(8)	90.4(2)	95.52(15)	96.96(19)
C2–Re–N3	96.55(8)	94.6(3)	97.25(17)	96.5(2)
C1–Re–N3	96.18(8)	96.1(2)	96.56(14)	97.37(19)
Re–N2–S1	111.00(8)	109.8(2)	111.68(15)	111.91(19)
Re–N2–C9	107.70(11)	109.4(3)	104.1(2)	108.3(3)
Re–N2–C10	108.47(12)	104.1(3)	110.1(2)	106.5(3)
S1–N2–C9	109.69(13)	110.9(4)	109.5(2)	109.1(3)
S1–N2–C10	108.22(12)	113.5(4)	109.8(3)	109.4(3)
C9–N2–C10	111.8(2)	108.9(5)	111.5(3)	111.5(4)

Consistent with a re-hybridization of the sulfonamide N atom from sp² to sp³, the Re–N2(sp³) bond distance involving the central sulfonamide group is significantly longer than the Re–N1(sp²) and Re– N3(sp²) bond distances involving the pyridyl groups for all four compounds (Table 2). The Re–N2 bond distance (2.2826(16) Å) of [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1) (Figure 2), is comparable to that of the unquestionably sp³-hybridized N atom (2.232(4) Å) of the central donor group of [Re(CO)₃(N(CH₂CO₂Et)dpa)]Br.²¹ These bond distances are slightly longer than that (2.187(4) Å) of the parent complex, [Re(CO)₃(N(H)dpa)]Br,²¹ indicating that replacing the NH proton with a larger substituent leads to bond lengthening. This result is attributable mainly to steric rather than electronic effects. Thus, these comparisons clearly indicate that the sulfonamide N is a relatively strong electron donor because the SO₂ group on N2 in 1 is more bulky than the CH₂ group on N2 in [Re(CO)₃(N(CH₂CO₂Et)dpa)]Br.

In [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1) (Figure 2), the angles of the sulfonamide nitrogen are close to 109.5° (Table 2), another result clearly illustrating that the hybridization of the sulfonamide nitrogen has changed from sp² to sp³ upon binding to Re. Similar conclusions apply to the other structurally characterized complexes in Figure 2. However, the bond angles of [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2, Table 2) show on average a greater deviation from values for other complexes, and this deviation appears to be clearly attributable to the stacking of the tmb rings (3.8–3.9 Å contact distances) in the crystal (Figure S4, Supporting Information). This result and features of structures in the literature (see below) suggest that the angles of the sulfonamide group are quite pliable, a feature that can lead to ligands which exhibit coordinative flexibility.²⁷

Fewer than twenty complexes are known in which tertiary sulfonamides have been shown to bind to metals;^26-33 in all of these reported compounds the sulfonamide N is positioned advantageously by ligand rings. The evidence for angles close to 109.5° (Table 2) clearly indicates that binding to Re^I results in sp²-to-sp³ re-hybridization of the tertiary sulfonamide nitrogen. This finding that all angles are close to the tetrahedral value is a common feature in the complexes of many metals in various oxidation states (Ni^II, Co^II, Mn^II, Cu^II, La^III)^26,27,32,33 with a bound sulfonamide. One useful approach to appreciating the sp³ nature of the sulfonamide nitrogen is to determine the distance by which the N lies out of the plane defined by the S atom and the two C atoms attached to N. For complexes 1, 2, 4, and 5, this out-of-the-plane distance ranges from 0.486–0.515 Å. Likewise, for complexes with other metals and differing widely in the ligands and geometries, values of 0.38–0.51 Å are typically found.^26,27,32,33 In contrast, for complexes in which the sulfonamide nitrogen is not bound, the out-of-the-plane distances in reported structures typically range from 0.06–0.125 Å.^26,27

New complexes 1, 2, 4, and 5 are the only examples of complexes with a tridentate ligand having a coordinated tertiary sulfonamide with a typical M–N(sp³) distance. In almost all cases, the metal-bound tertiary sulfonamide nitrogen and the donors of the adjacent chelate rings occupy the face of an octahedron^26,28,33 or have a similar arrangement in complexes with other geometries.^29,30 In other words, these three N-donor atoms and the metal atom typically do not define a common plane. However, because the sulfonamide group is pliable, when the R group attached to sulfur possesses a suitably positioned donor atom, the ligand can coordinate in a tripodal tetradentate fashion.²⁷ In fact, only in such rare cases is the M–N(sulfonamide) bond length as short as a typical M–N(sp³) distance.²⁷ In these rare tripodal tetradentate cases, the sulfonamide N and the two flanking N donor atoms linked to it coordinate meridionally, not facially.²⁷ In the only other case of meridional coordination, the M– N(sulfonamide) bond length is long (2.298 Å).²⁷ A ten-coordinate square antiprismatic La^III complex with one macrocyclic ligand derived from cyclen, having one sulfonamide N donor and three amine N donors, has a long La^III–N(sulfonamide) bond, and the coordination mode is between facial and meridional.³²

For compounds 1, 2, 4, and 5, the relatively short metal-to-nitrogen(sulfonamide) bond (adjusting for the Re^I characteristics) correlates with a relatively long N–S bond. The N–S distances found in this work are 1.755–1.776 Å. In the rare examples in which the M–N(sulfonamide) bond length reflects a strong bond (enforced by the tripodal nature of the ligand), the N–S distance is ∼1.73–1.76 Å.²⁷ For more common cases in which the M–N(sulfonamide) bond length is much longer than the typical value for the M–N distance of that metal in its oxidation state, the N–S distance is shorter, ∼1.7 Å or less.²⁷ For complexes in which the sulfonamide is not coordinated, the N–S distance is about 1.62–1.64 Å.²⁷ Thus, although the bond angles appear to reflect sp²-to-sp³ re-hybridization of the sulfonamide nitrogen, weakly coordinated or non-coordinated sulfonamide groups have N–S bonds possessing considerable double-bond character.²⁷ The N–S bond length in the square antiprismatic La^III complex³² is consistent with this relationship.

NMR Spectroscopy

T(N(SO₂C₆H₄)dpa)P and all complexes and ligands reported here were characterized by NMR spectroscopy in DMSO-d₆ (NMR data and assignments are presented in the Experimental Section). In the free N(SO₂R)dpa (R = Me, tmb, dmb, tol, and Me₂Nnap) ligands, the signal for the methylene groups is a singlet (Figures 3 and S1). The protons in each group are magnetically equivalent because the sulfonamide group can achieve a time-averaged planar geometry with both methylene groups in the pseudo plane.

Figure 3.

Comparison of the ¹H NMR spectra of ligands (N(SO₂Me)dpa (a), N(SO₂Me₂Nnap)dpa (c)) and the corresponding Re complexes ([Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1) (b), [Re(CO)₃(N(SO₂Me₂Nnap)dpa)]BF₄ (5) (d)) in DMSO-d₆ at 25 °C.

The methylene groups of [Re(CO)₃(N(SO₂R)dpa)]X complexes 1–5 are related by a mirror plane. However, unlike the case of the free ligand, the two protons on each methylene group are not magnetically equivalent (cf. Figure 2); they project toward and away from the carbonyl ligands and are designated as endo- and exo-CH protons, respectively (Chart 1). The methylene protons give rise to two doublets (Figures 3 and S2) rather than one singlet, as found for the free ligand (Figures 3 and S1).

Chart 1.

Designation of the methylene group endo-CH and exo-CH protons in complexes 1-6.

Bidentate coordination of the N(SO₂R)dpa ligand via the two pyridyl N atoms also prevents the tertiary sulfonamide group from achieving a time-averaged planar geometry. Methylene doublets for N(SO₂R)dpa ligands in the bidentate binding mode were observed for the pseudo square planar complexes, Pd^II(N(SO₂tol)dpa)Cl₂⁴⁸ and Pt^II(N(SO₂R)dpa)Cl₂.⁴⁹ (These Pt^II complexes contain the N(SO₂R)dpa ligands used in the current study, Figure 1.) The geminal coupling constants also were similar; for example, J = 15.8 Hz for 1 and J = 15.0 Hz for Pt(N(SO₂Me)dpa)Cl₂. Furthermore, no trend was observed in the shifts of the methylene ¹H NMR signals. Thus, the methylene signals of the coordinated N(SO₂Me)dpa ligand are not useful for assessing binding mode (bidentate or tridentate).

The methyl ¹H NMR signal of [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1) is shifted downfield by 0.76 ppm relative to the N(SO₂Me)dpa ligand; however, this signal shifted downfield by only 0.10 ppm for Pt(N(SO₂Me)dpa)Cl₂.⁴⁹ The downfield shift for 1 can be explained best by the inductive effect resulting from a direct N–Re bond (Figure 2) compared to the absence of a N–Pt bond in Pt(N(SO₂Me)dpa)Cl₂. Thus, for the N(SO₂Me)dpa complexes, the methyl group ¹H NMR shift appears to provide a good spectroscopic handle for assigning binding mode.

COSY spectra were useful in assigning the aromatic signals of the free N(SO₂Me)dpa ligand and [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1) in DMSO-d₆ (Figures S5 and S6, Supporting Information). All of the aromatic signals for 1 were shifted more downfield than the corresponding signal of the free ligand. For example, the most downfield doublet (8.50 ppm) of the free N(SO₂Me)dpa ligand is the pyridyl H6/6′ signal (consistent with the close proximity of the H6/6′ protons to a pyridyl nitrogen); the H6/6′ signal has an even more downfield shift (8.89 ppm) in 1, as expected from the Re^I electron-withdrawing inductive effect on the coordinated pyridyl group. The pyridyl H6/6′ signal (9.24 ppm) of Pt(N(SO₂Me)dpa)Cl₂ is even more downfield than that of 1. The same trend in H6/6′ shifts is seen when comparing the spectra of [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2) and Pt(N(SO₂tmb)dpa)Cl₂. We attribute the greater downfield shifts in the Pt complexes to a greater electron-withdrawing inductive effect of Pt^II as compared to Re^I. In DMSO-d₆ for L = 3-(4′-methylpyridin-2′-yl)-5,6-dimethyl-1,2,4-triazine,⁵⁰ the pyridyl H6 signal appeared more downfield for PtLCl₂ (9.35 ppm)⁵⁰ than for [Re(CO)₃L(CH₃CN)]BF₄ (8.97 ppm),⁴⁹ a result indicating that Pt^II does indeed have a generally greater inductive effect than Re^I. This greater effect for Pt^II than for Re^I adds further confidence to our analysis that the small downfield shift for the Me ¹H NMR signal for Pt(N(SO₂Me)dpa)Cl₂ compared to that for 1 indicates that the sulfonamide N binds to Re^I but not to Pt^II; otherwise, the Me ¹H NMR signal for the Pt^II compound would be more downfield than that for the Re^I compound.

For [Re(CO)₃(N(SO₂Me)dpa)]PF₆ (1), we could not use the weak NOE cross-peaks from the methyl signal to both methylene doublets to assign unambiguously the doublets to exo-CH and endo-CH protons (Chart 1). However, the NOE cross-peak from the pyridyl H3/3′ signal (7.57 ppm) was larger to the methylene doublet at 5.13 ppm than to the doublet at 5.44 ppm (not shown). In the molecular structure of 1, nonbonded distances of ∼2.44 and 2.45 Å between the pyridyl H3/3′ and the exo-CH protons are shorter than the nonbonded distances of 3.02 and 3.09 Å between H3/3′ and the endo-CH protons. This information, combined with the size of the NOE cross-peaks, indicates that the more upfield signal (5.13 ppm) is the exo-CH doublet for 1.

Except for the distorted [Re(CO)₃(N(SO₂tmb)dpa)]PF₆ (2) complex, all of the nonbonded distances from the pyridyl H3/3′ protons of structurally characterized complexes are shorter to the exo-CH protons (∼2.5 Å) than to the endo-CH protons (∼3.0 Å). In a ROESY spectrum of [Re(CO)₃(N(SO₂dmb)dpa)]PF₆ (3) in DMSO-d₆ (Figure S7), an NOE cross-peak from the CH doublet at 4.55 ppm to the H3/3′ signal assigns this doublet to the exo-CH protons. An NOE cross-peak between the dmb H2/6 signal (7.96 ppm) and the CH doublet at 5.25 ppm assigns this doublet to the endo-CH signal. Thus, the exo-CH and endo-CH assignments are mutually confirmed because the nonbonded distances from the H2/6 protons in [Re(CO)₃(N(SO₂tol)dpa)]PF₆ (4) to the exo-CH and endo-CH protons are 3.06 and 2.59 Å, respectively.

Changes in the NMR signals in DMSO-d₆ for T(N(SO₂C₆H₄)dpa)P on formation of [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄ (6) (Figure 4) establish beyond question that the formulation of 6 is correct. The methylene group singlet for T(N(SO₂C₆H₄)dpa)P becomes two doublets in 6. In the ROESY spectrum in DMSO-d₆ (not shown), an NOE cross-peak from the doublet at 4.97 ppm to the pyridyl H3/3′ signal assigns that doublet to the exo-CH protons, and the NOE cross-peak from the other methylene doublet (5.97 ppm) to the phenylene signal (8.84 ppm) assigns this doublet to the endo-CH protons (Chart 1).

Figure 4.

Comparison of the ¹H NMR spectra of T(N(SO₂C₆H₄)dpa)P (bottom) and [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄ (6) (top) in DMSO-d₆ at 25 °C.

Coordination of the four Re moieties to T(N(SO₂C₆H₄)dpa)P leads to the apparent singlet signal (8.84 ppm, Figure 4) for the clearly inequivalent phenylene ortho and meta protons of 6 (Scheme 2). We attribute this interesting spectral feature to a similar chemical shift of the signals of the phenylene group, leading to non-first-order spectral features in this aromatic region. Binding of the porphyrin sulfonamide groups to Re leads to a larger downfield shift change (by 0.60 ppm) of the m-H (H2/6) signal (for the phenylene protons closest to Re) versus the smaller downfield shift change (by 0.37 ppm) of the phenylene o-H (H3/5) signal. This result fully supports the proposed structure.

The ¹H NMR spectrum of [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄ (6) in DMSO-d₆ (Figure 4) reveals other effects of Re coordination to the pyridyl nitrogen. The ¹H NMR signal for the pyrrole βH (Scheme 2) is ∼0.3 ppm more downfield than the βH signal of the free T(N(SO₂C₆H₄)dpa)P ligand. The pyridyl H6/6′ signal (J = 5.4 Hz) of 6 appears at 9.01 ppm versus 8.55 ppm for T(N(SO₂C₆H₄)dpa)P. The ¹H NMR shifts of the pyridyl signals of 6 are very similar to those of the crystallographically characterized complex 1, another indication of the presence of a Re–N(pyridyl) bond. Furthermore, the two methylene doublets for 6 have J = 16.4 Hz, a value very similar to that for 1 (J = 15.8 Hz). As is true for the smaller complexes, the two doublets for [[Re(CO)₃]₄T(N(SO₂C₆H₄)dpa)P](BF₄)₄ can be attributed to the fact that the CH₂–N(SO₂R)–CH₂ grouping cannot achieve the time-averaged planarity required for magnetic equivalence.

In general, the relatively downfield and upfield methylene doublets are assigned to the endo-CH and the exo-CH signals, respectively, for complexes 1–6. However, there is some variability in the exact shift values (cf. Experimental Section, Supporting Information). The shift dependence on the nature of R is complicated, but we can explain the shift dependence within the series of R groups that have methyl-substituted benzene rings (tol, dmb, and tmb). The exo-CH signals all have nearly identical shifts (Figure S2) and the endo-CH signals for tol and dmb also have nearly identical shifts. However, the endo-CH signal for the tmb complex is relatively upfield. The two ortho methyl groups of tmb appear to position the ring so that it has an anisotropic shielding of the endo-CH proton. This positioning is found in the molecular structure and may be the reason for the stacking interactions found in the solid (Figures S4 and S8).

Effects of Bases

As mentioned above, these complexes are stable for long periods in DMSO-d₆. However, when basic ligands were added to a 5 mM solution of [Re(CO)₃(N(SO₂Me)dpa)]PF₆ in acetonitrile-d₃ (4-dimethylaminopyridine, 4-methylimidazole, or 3,5-lutidine) or in DMSO-d₆ (4-dimethylaminopyridine), complicated ¹H NMR spectral changes attributable to a decomposition reaction were observed. For 4-dimethylaminopyridine, decomposition was observed in the first spectrum (recorded after 15 min); however, for 3,5-lutidine, decomposition was first noticeable in a spectrum recorded after 24 h. When triethylamine (a strong base but a very poor ligand) was added to a similar 5 mM solution of [Re(CO)₃(N(SO₂Me)dpa)]PF₆, immediate decomposition was observed. All spectra contained a very similar set of peaks attributable to decomposition products. These results are consistent with base-catalyzed decomposition of the coordinated N(SO2Me)dpa ligand. In contrast, adding 1 molar equiv of triethylamine to a 5 mM acetonitrile-d₃ solution of the N(SO2Me)dpa ligand itself produced no changes in the ¹H NMR spectrum, even after two weeks. These observations suggest that the ligand decomposition is promoted by the electron-withdrawing inductive effect of the Re^I(CO)₃ metal center.

The absence of any need to protect the two terminal pyridyl donor nitrogens is an advantage of using the dpa framework for bioconjugation; however, the electron-withdrawing pyridyl group may contribute to the base-catalyzed decompostion. We have thus redesigned the ligand framework for future studies.

Summary and Conclusions

The [Re(CO)₃(N(SO₂R)dpa)]X (X = PF₆ or BF₄) complexes synthesized and characterized as a prelude to radiopharmaceutical studies with novel N(SO₂R)dpa ligands provide structural evidence establishing that the Re–N bond formed between a tertiary sulfonamide N and Re can have a normal length. Because there are distinctive NMR spectral features associated with such binding in the N(SO₂R)dpa ligands, it is clear that the tertiary sulfonamide N binds to Re even in cases such as [Re(CO)₃(N(SO₂dmb)dpa)]PF₆ (3), for which an X-ray structure could not be obtained.

Monodentate tertiary sulfonamides do not bind metals. Thus, as illustrated in the examples here and in previous reports,^28-31 tertiary sulfonamides bind metals via the N atom only when geometrical restraints are enforced by the ligand so as to place the N atom in a favorable position to form a bond. The tertiary sulfonamide nitrogen adopts stereochemical and structural features consistent with sp³ hybridization. As a result, the N atom accommodates a facial tridentate geometry when the sulfonamide anchors two chelate rings. Of note, sulfonamide binding in the [Re^I(CO)₃L]⁺ complexes described in this study is facilitated by sp²-to-sp³ re-hybridization of the sulfonamide nitrogen. In such complexes, the tertiary sulfonamide appears to be a relatively good donor, as judged by the near-normal Re–N bond length and the relatively long N–S bond. In most cases reported in the literature, the M–N bond is long, and the N– S bond is relatively short.^26,27,32

The conjugation approach described here has potentially wide applications. This prospect is supported by our results showing that a sulfonamide can be used to conjugate the fac-[Re^I(CO)₃]⁺ core unit to a dansyl group or a porphyrin. We anticipate no limitation in conjugating other molecules to such a core or in extending this chemistry to ^99mTc analogues. Our results demonstrate that the tertiary sulfonamide linkage may be utilized for tethering biologically important molecules, and we are confident that the chemistry with the fac-[Re^I(CO)₃]⁺ core reported here can be readily extended to the fac-[^99mTc^I(CO)₃]⁺ core.