2.1. Activity of Soluble Hoveyda–Grubbs Second-Generation Catalyst at Different Conditions
The reactivity of the
Hoveyda–Grubbs second-generation catalyst in the ring-closing metathesis reaction of
1 (
Scheme 1) was evaluated in the presence of different solvents at 25 °C. A full metathesis conversion was achieved in most of the solvents tested. Faster conversion was obtained in tetrahydrofuran (THF) (30 min) (maybe because THF could stabilize the active propagating Ru carbine [
21]) and CH
2Cl
2 (1 h), whereas in others such as hexane, heptane, oluene, DMSO, acetonitrile or ethanol was necessary for incubation at 2 h (data not shown). In a mixture of 20% (
v/
v) acetonitrile in water, the catalyst was not reactive, and directly not soluble in pure water. A second parameter to consider is the effect of the presence of additives. Indeed, it has been demonstrated that the addition of several molecules during the metathesis reaction (such as acetic acid, benzoquinones, or thioureas) [
22,
23,
24] positively affect the final catalytic properties of the organometallic compound. Different additives could result in completely different active species from the same Grubbs catalyst and trigger different catalytic pathways to produce structurally unique products.
In this way, here, the optimization of the metathesis using the homogeneous catalyst in the presence of additives of different nature in THF was evaluated (
Table 1).
The addition of Lewis acids (BF
3-Et
2O, TMSOTf) to the reaction decreased the metathesis conversion catalyzed by soluble Grubbs catalyst between 25% and 30% (entries 2–3,
Table 1). The presence of amines resulted in worse results, showing 40% activity adding pyridine (entry 5,
Table 1). In this case, this could be because of the high vulnerability of the metallacyclobutane intermediate formed in the presence of ethylene, which is rapidly decomposed in the presence of a base, even at room temperature [
25].
However, the use of different ionic liquids as additives slightly improved the activity of the catalyst.
The presence of different polyol polymers did not alter significantly the activity of the catalysts, being positive for their use in the creation of heterogeneous catalysts. Surprisingly, the most improvement of the catalytic performance of the organometallic compound was achieved using phenylboronic acid as additives (entry 14,
Table 1).
Finally, aminoacids and peptides were tested as additives with different results. Aspartic acid dimethyl diester caused a decrease in the activity of the catalysts of 70% in the metathesis reaction, because of the presence of the free amine. However, a cysteine-containing N-acetylated short hydrophobic tripeptide (NHAc–Cys–Phe–Phe–CONH2) slightly improved the rates in the metathesis reaction. Using another more hydrophilic cysteine-containing peptide sequence, the effect was slightly negative.
Recently, the modification of ruthenium catalyst by the introduction of amino acids as ligands produced good results in activity and selectivity in ring-closing metathesis [
26].
Thus, the procedure of activation by the corresponding silver salt of the dipeptide N-terminal protected with tert-butyloxycarbonyl protecting group (BOC), BOC–Ala–Gly–COOH was used. The silver salt was synthesized as previously described [
26] and was incubated as an additive in a solution of Hoveyda–Grubbs catalyst in THF. In this case, unfortunately the new adduct exhibited 75% less activity of the non-modified one (entry 17,
Table 1).
Therefore, with the addition of several additives such as ionic liquids, some polymers or peptides showed positive effects and clearly the catalytic reactivity of the Grubbs catalyst was strongly decreased in the presence of amine-containing molecules and by specific ligand modification with Boc-peptide.
2.2. Preparation of Heterogeneous Catalyst and Application in RCM
After the screening of the RCM conditions of the homogeneous Hoveyda–Grubbs second-generation catalyst, different strategies to generate a heterogeneous catalyst have been used.
Firstly, the Hoveyda–Grubbs second-generation catalyst was immobilized on different functionalized macroporous epoxy acrylic Sepabead resins (
Table 2). Four different supports were used; commercially available ones activated with epoxide, amino or butyl groups and one tailor-made functionalized with phenylboronic acid by modification of epoxy Sepabeads with 4-aminophenyl boronic acid as previously described [
27].
The Ru catalyst offered (1 mg) was rapidly immobilized on the amino-Sepabeads, whereas longer times of incubation were necessary in the other three cases with a final immobilization yield of 50%–60% (
Table 2). The resins free of organometallic compound were tested and no metathesis conversion was observed in any case. However, Ru catalyst was inactive and no metathesis conversion was observed. This is related to the previous effect of amines in the formation of adduct. In this way, the organometallic could react with the amino group by covalent interaction despite being completely inactive [
28]. The same result was found after its adsorption on butyl-Sepabeads, however the immobilization on epoxy-Sepabeads conversed an active heterogeneous catalyst, producing
2 in 53% conversion after 24 h. Also, Ru catalyst immobilized on functionalized phenylboronic acid support catalyzed the RCM with 50% conversion after 96 h. One recycle of the boronic acid-organometallic catalyst was performed, maintaining all the initial activity. Definitely, the reaction of the Ru catalyst immobilized on supports material was lower compared to the homogeneous version, but similar to recent published results [
29].
In this way, a first Ru catalyst–enzyme conjugate was prepared by incubating Hoveyda–Grubbs second-generation catalyst with lipase of C. antarctica (CAL-B) immobilized on butyl-Sepabeads. However, the heterogeneous organometallic biocatalyst was inactive. These results could demonstrate that the Ru catalyst was directly fully absorbed on the butyl-groups on the support and no effect of the immobilized enzyme was detected.
The differences in functional groups on the resin showed that possible adducts between Ru catalyst with epoxides [
30] were more effective than with butyl groups (higher hydrophobic degree).
Considering the results obtained in the immobilization, the preparation of heterogeneous catalyst without the presence of a support material was developed. First, cross-linked enzyme aggregates (CLEA’s) of enzymes were prepared. A novel enzyme–Ru catalyst conjugate was prepared by preparation of the enzyme CLEA, adding Hoveyda–Grubbs second-generation catalyst in the solution. The method consists of the precipitation of the enzyme and a crosslinking process afterwards. Several well-described strategies determine the use of dimethoxyethane (DME) as an excellent solvent as a precipitating agent and glutaraldehyde as a crosslinking agent [
31].
The stability of the Grubbs catalysts in the presence of glutaraldehyde and DME was evaluated in the metathesis reaction of 1 and the activity was not so much affected.
A variety of CLEAs of different proteins were prepared using different protocols, changing the solvent as a precipitant, an amount of protein, an amount of water or co-solvent, or the presence of different additives (data not shown). In all cases, CLEAs with and without the Ru catalyst were prepared. However, only in two cases was metathesis conversion observed. In both cases, the lipase of
C. antarctica B was used as protein (
Table 3). In these conditions, when 1 mg/mL of enzyme solution was used, the formed enzyme–Ru catalyst, CLEA, was active on the metathesis reaction, producing a 14% conversion of
1 after 1 h reaction. Using 5 mg/mL of enzyme, the resulted CLEA barely catalyzed the reaction, with 2% of product in 1 h in THF. The use of CAL-B–CLEA prepared without an organometallic molecule did not catalyze the reaction. Also, no hydrolyzed product was found using this conjugate. However, the low conversion obtained for these CLEAs is caused by the small amount of Ru catalyst present on the solid; less than 10% of the Ru catalyst offered was retained in the heterogeneous catalyst. This could be due to the fast precipitation of the protein which is not enough time for the enzyme–Ru complex interaction.
Therefore, this methodology generated a heterogeneous catalyst but with lower activity compared with the homogeneous one.
A second approach used was the entrapment system between the Hoveyda–Grubbs second-generation catalyst and the lipase by the sol–gel preparation. This strategy refers to a process where the enzyme is mixed with sol–gel solution, followed by a gelation process under the influence of pH and the aging process [
32]. In the sol–gel method, an inert gel network with pore sizes is built by chemical condensation around each enzyme macromolecule, an important difference with the CLEAs strategy where an aggregated structure by covalent chemical modification is formed. Sol–gel organic/inorganic matrices were prepared using tetra-ethyl orthosilicate (TEOS) as a precursor and formic acid to catalyze the hydrolysis and condensation of silicon molecules and ethanol as a solvent. The sol–gel CAL-B without Ru catalyst were previously obtained after 24 h whereas the corresponding sol–gel Ru without protein was finally obtained after 72 h incubation. This latter was repeated, including additives such as polymers or phenylboronic acid but no gelation was observed after 72 h. Then, the conjugation of both (CAL-B and Ru with phenylboronic acid) in a sol–gel was prepared to obtain a solid gel after 24 h.
The activity of the Hoveyda–Grubbs second-generation catalyst was almost conserved in the presence of TEOS, formic acid or ethanol (data not shown).
Thus, the different sol–gel preparations were tested as a catalyst in the metathesis reaction of
1 in THF (
Table 4).
The sol–gel system of lipase or the Ru catalyst alone did not catalyze the reaction (
Table 4), whereas the conjugate system Hoveyda–Grubbs second-generation catalyst-CAL-B exhibited 60% metathesis conversion in 2 h; a slight improvement compared with the results using the soluble catalyst. The catalyst was used two times more, producing 36% of
2 after 2 h in the third use (maintaining 60% activity).
This result could demonstrate that the Ru complex is protected and stabilized by the insertion into the interior space of protein by supramolecular (non-covalent) interactions. CAL-B shows a particularly well defined oxianion cavity, with a hydrophilic area surrounding the active site (Ser105) which generates hydrogen bond interactions and a channel with high hydrophobic degree [
33] where some residues such as I285, L144 and L140 are critical, for example in the location of the hydrophobic part of molecules. Therefore, considering the structure of the Grubbs catalyst, a probable hypothetic location of the catalyst could be into the channel as described in
Figure 1. Combination of hydrophobic interactions of aromatic and isopropyl groups and hydrogen bond interactions of the amino groups could stabilize the conjugation with the protein.
In order to know the structure of the new heterogeneous Ru catalyst–lipase conjugate, Scan Electronic microscopy (SEM) was performed. In
Figure 2, we can observe that a mesoporous structure with homogeneous microspheres containing lipase and Ru catalyst were formed.
One of the most interesting applications of combining enzyme and organometallic complex is performing the reaction in aqueous media.
In this way, the CAL-B–Hoveyda–Grubbs sol–gel was applied in the RCM in 2-(
N-morpholino)ethanesulfonic acid (MES) buffer at pH 6 containing 200 mM NaCl pH 6 (
Table 5). At these conditions, 50% conversion of
2 in 2 h was achieved whereas 30% was achieved using 1 mg of soluble catalyst and no conversion was observed with CAL-B sol–gel without Ru catalysts with the conjugate enzyme–Hoveyda–Grubbs in solution. After 24 h, almost full conversion was obtained (
Table 5).
Thus, a new set of different heterogeneous catalysts were prepared using two different strategies. The first strategy consisted of the immobilization of CAL-B by covalent attachment on an epoxy activated Sepharose (a hydrophilic matrix) and then incubated in a solution containing Hoveyda–Grubbs second-generation catalyst (Epoxy-CAL-B–Hoveyda–Grubbs). The use of this matrix and this immobilization method was selected by considering the previous negative results with butyl-Sepabeads-lipase, avoiding the block of aromatic groups in the protein and the possible adsorption of the Grubbs into the matrix.
In the second strategy, the conjugate CAL-B and Hoveyda–Grubbs second-generation catalyst was prepared by incubation of both mixed, and then this solution was immobilized on epoxy-Sepharose (CAL-B–Hoveyda–Grubbs-Epoxy). Both hybrid catalysts were tested in the ring-closing metathesis of
1 in aqueous solution at 25 °C (
Table 6).
In both cases, no adsorption of Ru catalyst was detected on the Epoxy-Sepharose support. Full conversion in the metathesis of 1 in MES buffer pH 6 containing 200 mM NaCl was observed after 2 h for both immobilized lipase–Ru catalysts and the soluble enzyme–Hoveyda–Grubbs conjugate.
In comparison of both strategy preparations, the second one was the best because similar results were achieved containing half the amount of Grubbs catalyst.
This new organometallic biocatalyst was washed with similar buffer and reused in the reaction, obtaining 80% conversion of 2 at 2 h. This process was repeated and the conversion was 60% after 2 h incubation.