1. After the trans-azophenol in the illuminated part of the organic phase is converted to the cis-azophenol (eq 1), it reacts with the aqueous NaOH forming a Bu₄N⁺cis-A⁻ ion pair in the organic phase and forcing a Pic⁻ from the organic phase into the illuminated aqueous compartment (eq 2). The Bu₄N⁺cis-A⁻ then reverts to Bu₄N⁺-trans-A⁻ (eq 3) which then abstracts a proton from the H₂O in the dark aqueous compartment, generating an OH⁻ and allowing a Pic⁻ ion to enter the organic phase as Bu₄N⁺Pic⁻ (eq 4). The net result of each cycle is the appearance of one Pic⁻ and the disappearance of one OH⁻ from the illuminated aqueous compartment and the disappearance of one Pic⁻ and the appearance of one OH⁻ in the dark aqueous phase. The liquid membrane consisted of two 100-mL, round-bottomed flasks joined by a bridge for the supernatant toluene solution (see Figure 2). Into the two round-bottomed flasks was poured 50 mL each of an aqueous solution containing 10⁻² M NaOH, 1.18×10⁻³ M Bu₄NPic, and 1.5 M Na₂SO₄. A toluene solution containing 6.17×10⁻⁴ M 2-hydroxy-3,5,6-trichloro-4'-methylazobenzene was then poured carefully on top to cover the two aqueous phases. The whole apparatus was immersed in a 25.0 °C constant temperature bath, the two flasks were subjected to vigorous stirring by means of two magnetic stirrers, and the left side of the liquid membrane was illuminated with a 275-W incandescent GE sunlamp. At the beginning of the experiment the contents of the two aqueous compartments were identical. At intervals aliquots of the two aqueous phases were assayed for picrate content. Figure 3 shows the increase in the picrate concentration of the illuminated aqueous compartment of the liquid membrane and the concomitant decrease in the picrate concentration on the dark side. As can be seen from eq 1-4 the proton flux is exactly equal to changes in the picrate concentrations and is about 2×10⁻⁸ equiv per hour. 2. Dioscorine (4) is the main alkaloid found in the tropical yam Dioscorea hispida Dennstedt. We have previously established that this novel isoquinuclidine alkaloid is derived from nicotinic acid (1) and acetic acid. A biogenetic scheme for dioscorine was considered which involved a condensation between 3,6-dihydronicotinic acid (2) and a branched eight carbon unit (3) derived from four acetate units, one of the terminal carboxyl groups being ultimately lost in the formation of dioscorine. In order to probe this proposed biogenetic scheme, feeding experiments have been carried out with [6-¹⁴C,2-³H]nicotinic acid and [6-¹⁴C,6-³H]nicotinic acid. Both these precursors were incorporated into dioscorine with complete retention of ³H relative to ¹⁴C (Table I). An alternative biosynthetic scheme, illustrated in Scheme I was considered in which nicotinic acid was activated by conversion to its betaine, trigonelline (5). Nucleophilic attack of the acetate-derived fragment at C-6 of trigonelline affords 6. Reduction of the dihydropyridine ring with a shift of a double bond and decarboxylation of the β-keto acid yields 7. Compound 9 arises by decarboxylation of the β-iminium carboxylic acid. The isoquinuclidine ring system is then formed by reaction of the enamine in 9 with the ketone, generating the hydroxy acid 8. Reduction of the iminium ion and lactone formation then affords dioscorine. This new scheme differs from the previous one in that the bond which ultimately becomes 1-6 in dioscorine is formed before the 4-5 bond. This scheme also accommodates the alkaloid dumetorine (10) which has been found in a related species Dioscorea dumetorum. This new hypothesis has now been tested by feeding [methyl-¹⁴C,6-²H,³H]trigonelline to D. hispida plants. Deuterium was introduced at C-6 in the hopes that its incorporation into dioscorine could be monitored by ²H NMR. A good incorporation (Table I) of the radioactive isotopes into dioscorine was obtained with almost complete retention of ³H relative to ¹⁴C. Before examination of the ²H NMR of this labeled dioscorine, it was necessary to unequivocally assign the ¹H NMR spectrum of the alkaloid. This was done by examination of its 2D-HETCOR and 2D-COSY NMR. The HETCOR spectrum revealed that the hydrogen at C-1 (52.23 ppm in the ¹³C NMR) has a chemical shift of 2.36 ppm (a pentet in an expanded spectrum). This 2D spectrum also indicated that the signals for C-6 and C-9 had been previously misassigned, the correct assignments being at 40.77 and 39.34 ppm, respectively. The 2D-COSY spectrum confirmed the relative stereochemistry depicted in structure 4, the rigid nature of the isoquinuclidine ring resulted in different chemical shifts for the geminal hydrogens on positions C-3, C-6, C-7, and C-8. The ²H NMR of the labeled dioscorine (100 mg in 458 mg of unenriched chloroform) exhibited only two significant peaks above the natural abundance level, one at 7.26 ppm (due to natural abundance of deuterium in the chloroform) and the other at 2.36 ppm. By integration of these two peaks it was established that the ²H enrichment at C-1 of dioscorine was 0.134%. Since the deuterium enrichment at C-6 in the administered trigonelline was 92.5%, this represents a specific incorporation of ²H of 0.14%, in excellent agreement with the specific incorporation of ¹⁴C. Trigonelline thus serves as a direct precursor of the isoquinuclidine ring of dioscorine. 3. In contrast to the extensive coordination chemistry of Co(II), monomeric Rh(II) complexes have proven elusive by virtue of their propensity for dimerization (e.g., Rh₂(OAc)₄) and/or disproportionation to Rh(III) and Rh(I). Recently we have been exploring the coordination chemistry of crown thioethers such as 9S₃ (1,4,7-trithiacyclononane), 18S₆, and 24S₆ with a view toward stabilization of low oxidation and spin states, in the expectation that the unusual electronic structures induced by these ligands would confer unusual reactivity as well. Thioether complexes of rhodium attract particular interest owing to their potential parallel to industrially important rhodium phosphine complexes. We report herein our synthetic, physical, and structural investigation of [Rh(9S₃)₂]³⁺, the first reported homoleptic thioether complex of Rh(III), and its reduction to a rare example of a stable monomeric Rh(II) complex. Reaction of rhodium(III) triflate (prepared by reaction of RhCl₃·3H₂O with silver triflate) with 2 equiv of 9S₃ in MeOH gives a pale orange solution that upon concentration and cooling deposits colorless needles (yield: 48%). Anal. Calcd for RhC₁₅H₂₄S₉F₉O₉: C, 19.78; H, 2.66; found (Oxford microanalytical service) C, 19.44; H, 2.58. ¹H NMR (300 MHz, CD₃CN, TMS, δ) 3.52 (m) at room temperature. Recrystallization from MeOH gave crystals suitable for X-ray diffraction measurements. The molecular structure of [Rh(9S₃)₂]³⁺ (Figure 1) shows a rigorously centrosymmetric RhS₆ coordination sphere in which the metal ion nestles between two 9S₃ rings. Two of the unique Rh-S distances are somewhat longer (2.345 (3) and 2.348 (3) Å) than the third (2.331 (2) Å); similar Rh-S bond lengths have been reported for a dimethyl sulfide complex of Rh(III). Interestingly, these distances closely approach those very recently found for [Ru(9S₃)₂]²⁺ (in which Ru-S distances range from 2.331 (1) to 2.344 (1) Å), despite the difference in charge between the two complexes. Intraligand dimensions and torsional angles differ insignificantly from those found either in other complexes of this ligand or, indeed, in the free ligand itself. Electrochemical studies of [Rh(9S₃)₂]³⁺ reveal an extraordinary result. Cyclic voltammetry in MeNO₂ on a glassy carbon electrode (Figure 2) shows two quasi-reversible one-electron processes at -309 and -721 mV versus SCE (ΔEₚ = -71 and -98 mV, respectively; v = 50 mV/s). Controlled potential coulometry establishes that both processes entail transfer of one electron (n = 1.05 and 0.97, respectively). Preparative electrolysis at -500 mV affords a straw-colored solution that exhibits an EPR spectrum (at 298 K) with g = 2.046 without resolved ¹⁰³Rh (I = 1/2, 100%) hyperfine splitting. This spectrum is consistent with a monomeric Rh(II) complex (d⁷, S = 1/2) and confirms the stability of the one-electron-reduced product.