Selective hydrogenation of cinnamaldehyde and phospholipids in aqueous-organic biphasic systems with ruthenium(II) complex catalysts

2.1 Hydrogenation of liposomes

Soybean lecithin (4 mg) was dissolved in 0.4 ml chloroform. The chloroform was evaporated by an argon stream and 6 ml ion-exchanged water or 6 ml phosphate buffer (pH 6.93) was added. The mixture was sonicated for 3×2 min (Branson Sonifier 250, Branson Ultrasonics, Danbury, CT 06810, USA) with cooling between the 2 min intervals. The liposome obtained this way was deoxygenated by several vacuum-argon refill cycles; [RuH(OAc)(mtppms)₃] or [RuH(OOc)(mtppms)₃] was added under inert conditions and the reaction mixture was stirred under hydrogen for the desired time. For reactions under H₂ pressure, homemade heavy-wall glass hydrogenation reactors were used. Analysis of the reaction mixtures was done by gas chromatography (HP 5890 Series II GC, SP2330 30×0.25 mm, FID, carrier gas nitrogen, column temperature 190°C, isotherm, temperature of the injector and detector 250°C) after extraction and transesterification of the lipids to methyl esters (HCl/MeOH).

Unsaturated fatty acids in soybean lecithin are found only in the C:18 series (18:1 oleic acid; 18:2 linoleic acid; 18:3 linolenic acid) therefore the unsaturation of the lipid can be calculated as unsaturation=A_18:1+2×A_18:2+3×A_18:3 (A=peak area). Extent of hydrogenation was characterized by the ratio (percentage) of remaining unsaturation to the unsaturation of the initial sample, x=100×(unsaturation of hydrogenated lipid)×(unsaturation of control)^-1. Alternatively, conversion is defined as conv=100-x.

2.2 Hydrogenation of cinnamaldehyde

A total of 13 mg (0.01 mmol) [RuH(OAc)(mtppms)₃] was placed into a glass pressure reactor sealed by a rubber septum, which was subsequently deoxygenated. Deoxygenated ion-exchanged water (3 ml) or 0.2 M aqueous acetate buffer was added under argon, followed by 100 μl (0.8 mmol) cinnamaldehyde. The reaction mixture was vigorously stirred at a temperature of 80°C. After the desired time, the reactor was cooled, 0.5 ml of the reaction mixture was mixed with 1 ml of water and 1 ml toluene, vortexed, centrifuged and the organic phase was filtered through MgSO₄. The dry organic phase was analyzed by gas chromatography (Agilent 7890A Gas Chromatograph; Chrompack HP-5 30 m×32 mm×0.25 µm; FID; carrier gas nitrogen, temperature program: 130°C/6 min, 60°C/min to 250°C, 250°C for 5 min).

Turnover frequency (TOF) is defined as TOF=(mol converted cinnamaldehyde)×(mol catalyst)^-1×h^-1.

3.1 Synthesis of [RuH(OOc)(mtppms)₃]

This compound was obtained by the method previously employed for the synthesis of RuH(OAc)(mtppms)₃] [14]. ¹H- and ³¹P-NMR data are agreement with an Ru(II)-monohydride with an octahedral structure, where the three phosphorus ligands are meridional and the hydride is in the cis-position to all three. The solid state IR spectrum shows characteristic vibration frequency for a terminal Ru-hydride at 1996 cm^-1and the ν(OCO, asym) and ν(OCO, sym) vibrations are displayed at 1532 cm^-1and 1435 cm^-1, respectively, yielding a Δν=ν(OCO, asym)−ν(OCO, sym)=97 cm^-1. These data are in good agreement with those found earlier for the analogous carboxylato complexes [RuH(OAc)(PPh₃)₃] [15] and [RuH(OAc)(mtppms)₃] [14] with Δν values of 77 cm^-1and 93 cm^-1, respectively. It is generally accepted [15] that Δν~120 cm^-1 refers to bidentate coordination of carboxylato ligands, and, indeed, for [RuH(OAc)(PPh₃)₃] this was confirmed by single crystal X-ray diffraction structure determination [16]. The NMR characteristics are also consistent with a trans-dihydride structure with three meridional phosphines and a monodentate carboxylato ligand, as was suggested in the case of a related formato complex, trans-[Ru(H)₂(OOCH)(mtppms)₃] [17]. Nevertheless, based on the IR properties, we prefer the formulation as a monohydride; further studies are underway to decide the exact structure of the compound.

Whereas [RuH(OAc)(mtppms)₃] dissolves well in water [14], [RuH(OOc)(mtppms)₃] is less soluble; however at concentrations required for catalysis (e.g., 4 mg(ml)^-1) it gives clear solutions.

3.2 Hydrogenation of cinnamaldehyde with [RuH(OAc)(mtppms)₃] catalyst

Hydrogenation of α,β-unsaturated carbonyl compounds is an important reaction in fragrance and flavor industries and cinnamaldehyde (1) is often used for testing catalysts with regard to their selectivity in hydrogenation of C=C or C=O bonds [18, 19]. Products of the hydrogenation (Scheme 1) can be the saturated aldehyde, 3-phenylpropanal (2), the unsaturated alcohol, 3-phenyl-2-propen-1-ol (3) and the saturated alcohol, 3-phenylpropan-1-ol (4) of which 3 is the most desired compound.

Scheme 1

Possible products of cinnamaldehyde hydrogenation.

Since cinnamaldehyde is only slightly soluble in water, reactions using an aqueous solution of the catalyst are, in fact, biphasic. According to the literature, aqueous organic biphasic hydrogenation of cinnamaldehyde with Rh(I)-phosphine catalysts usually affords the C=C hydrogenation product 2 [18] while Ru(II)-phosphine catalysts can be 100% selective to the unsaturated alcohol [20, 21] 3, however, the selectivity strongly depends on the pH of the aqueous phase [21].

We have found that [RuH(OAc)(mtppms)₃] catalyzed the hydrogenation of cinnamaldehyde under mild conditions (80°C, 4–7 bar H₂) both in water and in aqueous acetate buffer, as solvent, at various pH values. The pressure dependence of the composition of the reaction mixture is shown in Figure 1. At 7 bar H₂ after 4 h, the conversion of 1 was 52.6% which corresponds to a TOF of 10.5 h^-1. This TOF is close to the TOF (14 h^-1)which was obtained with {[RuCl₂(mtppms)₂]₂}+3 mtppms at 1 bar H₂ pressure [21]. It is seen from the Figure, that the major product is the most valuable unsaturated alcohol, 3; furthermore, its yield increased upon increasing the hydrogen pressure, and reached 48% under 10 bar H₂. Such a beneficial effect of hydrogen pressure on the selectivity of cinnamaldehyde hydrogenation has already been observed with similar Ru(II)-mtppms catalysts [22]. It is also important, that 2 and 4 are formed in <10% yield, each, even at 10 bar H₂ pressure.

Figure 1

Effect of pressure on the hydrogenation of cinnamaldehyde with [RuH(OAc)(mtppms)₃] catalyst. ◆ cinnamaldehyde, ■ 3-phenyl-2-propanal, ▲ 3-phenyl-2-propan-1-ol, ● 3-phenyl-2-propen-1-ol. Conditions: 0.01 mmol catalyst, 0.8 mmol cinnamaldehyde, 3 ml water, 80°C, 4 h.

Knowing the pH-sensitivity of cinnamaldehyde hydrogenations, we carried out reactions in 0.2 M acetate buffer solutions of pH 3.6, 4.6 and 5.6. Acetate buffer seemed a reasonable choice, since the catalyst contains an acetato ligand; unfortunately, it does not allow the reactions to be studied at higher pH values. The results are shown in Figure 2.

Figure 2

Effect of pH on the hydrogenation of cinnamaldehyde with [RuH(OAc)(mtppms)₃] catalyst. ◆ cinnamaldehyde, ■ 3-phenyl-2-propanal, ▲ 3-phenyl-2-propan-1-ol, ● 3-phenyl-2-propen-1-ol. Conditions: 0.01 mmol catalyst, 0.8 mmol cinnamaldehyde, 3 ml acetate buffer, 80°C, 2 h, P(H₂)=7 bar.

The hydrogenations of cinnamaldehyde in acetate buffer were run only for 2 h, so the conversion at pH 5.5 (23.8 %) is only slightly less than half the conversion in water after 4 h (26.4 %) (see also Figure 3 later). In the studied range, the pH has a small but important effect on the rate (Figure 2) and product distribution: the yield of 3 increased to 14.8% at pH 5.5. The yield of 2 varied from 3.3% (pH 3.6) to 6.2 % (pH 5.5), while that of 4 remained below 2%, so that at pH 5.5 cinnamyl alcohol was dominant (64%) among all hydrogenation products.

Figure 3

Time-course of the hydrogenation of cinnamaldehyde with [RuH(OAc)(mtppms)₃] catalyst. ◆ cinnamaldehyde, ■ 3-phenyl-2-propanal, ▲ 3-phenyl-2-propan-1-ol, ● 3-phenyl-2-propen-1-ol. Conditions: 0.01 mmol catalyst, 0.8 mmol cinnamaldehyde, 3 ml phosphate buffer (pH 7.0), 80°C, P(H₂)=7 bar.

The time course of the hydrogenation was investigated with 7 bar H₂ pressure using a 0.2 M phosphate buffer of pH 7.0 as the solvent (Figure 3). The reaction proceeded with a steady rate and high selectivity. After 5 h at 59.4% cinnamaldehyde conversion, the ratio of 3 to (2+3+4) was 75.4%. After 4 h, the conversion was 33.7% (TOF=6.7 h^-1), so phosphate buffer slightly decreased the rate of the reaction relative to the case of water or acetate buffer as solvents; such an unfavorable effect of phosphate buffer has already been detected [23] with similar water-soluble catalysts.

In summary, it can be concluded that in aqueous biphasic systems under mild conditions, [RuH(OAc)(mtppms)₃] proved to be an active and selective catalyst for the hydrogenation of cinnamaldehyde, a representative α,β-unsaturated carbonyl compound. The major product in all cases was cinnamyl alcohol (3), a valuable chemical for flavor and fragrance industries.

3.3 Hydrogenation of soybean lecithin with [RuH(OAc)(mtppms)₃] and [RuH(OOc)(mtppms)₃] catalysts

Hydrogenation of soybean lecithin liposomes was studied with both catalysts under mild conditions. Depending on their composition, liposomes are stable only around physiological temperatures, therefore we choose 40°C as the maximum reaction temperature. Both ion-exchanged water and phosphate buffer (pH 6.93) were used as solvents. The results are expressed as remaining unsaturation of the lipids as defined in the experimental section and are summarized in Table 1.

With phosphate buffer as the solvent, at 25°C and 1 bar H₂ pressure [RuH(OAc)(mtppms)₃] proved to be a very inefficient catalyst (entry 1); however, at 40°C, approximately 8% reduction of total unsaturation was observed (entry 2). This slow reduction of the C=C double bonds is in accord with the selectivity of these catalysts towards hydrogenation of the C=O function in cinnamaldehyde. Furthermore, the C=C double bonds of lipids are buried inside the double layer of the vesicles and are not freely available to the catalysts [24].Nevertheless, it should be added here that modification of the fluidity of biomembranes does not require removal of a high percentage of unsaturation, since the cells may disrupt easily as a consequence of increased rigidity of their boundary membranes [25]. In the light of this requirement, 5–8% saturation of the constituent lipids is already sufficient to trigger physiological processes, such as, for example, de novo synthesis of unsaturated fatty acids to counteract the undesirable fluidity change [26]. At 1 bar H₂ pressure, but otherwise identical conditions, [RuH(OOc)(mtppms)₃] was even less active, although the hydrogenation proceeded steadily in time (entries 3–5). Increasing the H₂ pressure to 5 bar resulted in a significant increase of conversion up to approximately 10% (entry 6). A similar increase of conversion was observed with [RuH(OAc)(mtppms)₃] at 25°C (entries 1 and 7). Unbuffered, ion-exchanged water was found to be a better solvent for hydrogenation of soybean liposomes. With [RuH(OAc)(mtppms)₃] catalyst at 40°C, we found 82.1% remaining unsaturation in water (entry 9); this compares favorably to the 91.6% observed for hydrogenation in phosphate buffer (entry 2). The rate-decreasing effect of phosphate buffer was also observed in hydrogenations of cinnamaldehyde (see above). Finally, increase of the H₂ pressure increased the conversion in water, too, and the lowest remaining unsaturation, 78.7% was determined at 5 bar and 40°C (entry 10). In our earlier experiments [26] we did not see any harmful effect of increased hydrostatic pressure on live cells or protoplasts at P(H₂)~10 bar.

In summary, the carboxylato complexes [RuH(OAc)(mtppms)₃] and [RuH(OOc)(mtppms)₃] seem to be suitable for hydrogenation of biomembranes under biologically acceptable conditions. Obviously, a great deal of optimization work is still to be done, including the determination of time course of hydrogenations at various H₂ pressures, the effect of other buffers, osmotics and salts present in culture media, survival response of cells to hydrogenation conditions, etc.). Such experiments are in progress in our laboratory and the results will be presented in due time.