今天推送的文章是发表在Science上的 “Directed evolution of enzymatic silicon-carbon bondcleavage in siloxanes”,通讯作者是来自加州理工学院化学与化学工程系的Frances H. Arnold教授和Dow Silicones Corporation（陶氏有机硅有限公司）的Dimitris E. Katsoulis研究员。
Linear and cyclic volatile methylsiloxane (VMS) is an artificial compound with material properties such as high backbone flexibility and low surface tension that can be used in many consumer applications, from detergents and defoamers to lotions, shampoos, and conditioners (Figure 1). Cyclic VMS is also an important raw material for the synthesis of silicone polymers. In order to meet consumer and feedstock demand for siloxanes, VMS is produced at approximately one million tonnes per year. However, the social benefits of VMS must be balanced against its potential environmental pollution, bioaccumulation, and toxicity. Considering the prevalence, practicability, and potential problems of VMS, it is of great significance to degrade VMS by Si-C bond cleavage.
Due to its high thermal stability and lack of functional groups, the degradation of VMS is difficult. The hydrolysis of Si-O bonds results only in the formation of species to form silanols and siloxanediols, while complete degradation requires cleavage of relatively inert Si-C bonds. The chemistry to achieve VMS degradation is limited to a few examples, including TiO2 photocatalysis, pyrolysis, and atmospheric oxidation of hydroxyl radicals. Typically, these Si-C bond cleavage reactions are caused by one or more oxidation of siloxane methyl groups. Studies in higher organisms have found that VMS is metabolized into a range of products, including metabolites indicative of Si-C cleavage, which is thought to occur after a C-H hydroxylation event. Enzymatic oxidation unlocks a Si-C bond cleavage mechanism that adapts to a wide range of environmental and process conditions. However, no enzyme has been found to be able to hydroxylate siloxane C-H bonds or promote Si-C bond cleavage. The authors hypothesize that cytochrome P450 can oxidize both unactivated alkyl C-H bonds and equally strong siloxane C-H bonds. Therefore, this paper explores the potential of these enzymes as a starting point for enzyme-catalyzed Si-C bond cleavage.
Figure 1
1
Discovery and directional evolution of Si–C bond cleavage activity of siloxanes
Cytochrome P450BM3 is a self-contained enzyme formed by the fusion of a heme domain with a NADPH-dependent reductase domain. This soluble bacterial enzyme has been engineered to catalyze a variety of unnatural hydroxylations. To test the hypothesis that enzymatic C-H hydroxylation promotes Si-C bond cleavage, the authors evaluated the ability of a series of cytochrome P450BM3 variants to oxidize siloxane C-H bonds in hexamethyldisiloxane. Among them, a previously unpublished P450BM3 variant evolved to silane and siloxane Si-H hydroxylation, named LSilOx1 (linear siloxane oxidase, generation 1), which was chosen as the starting point for the evolution of Si-C bond cleavage activity (Figure 2A). LSilOx1 has 13 amino acid substitutions relative to wild-type P450BM3, and wild-type P450BM3 does not have siloxane hydroxylation or Si-C bond cleavage activity.
Figure 2
After several rounds of directed evolution, Si–C bond cleavage activity was enhanced, where activity is defined as the ratio of silanol 5 concentration to enzyme concentration (Figure 2A). To expand the range of enzymatic VMS Si–C bond cleavage to more complex substrates, the authors also studied linear siloxane2. Random mutagenesis of LSilOx4 and enzymatic reactions in glass 96-well plates enabled the identification of the improved variant, LSilOx5 (Figure 2B). The variant LSilOx5 also exhibits activity against cyclic siloxane 3 and thus serves as a suitable starting point for directed evolution (Figure 2C). The evolution of siloxanes 2 and 3, which are representative of linear and cyclic siloxanes, has led to different lineages of LSilOx and CSilOx (cyclic siloxane oxidase) variants, demonstrating the ability to enhance the Si–C bond cleavage activity of different scaffolds. The three final variants, LSilOx4, LSilOx7, and CSilOx3, retained more than 50% activity when formulated as lyophilized lysates. Whereas, wild-type enzymes are inactive against siloxanes 1-3. These data suggest that enzymes can cleave Si–C bonds under mild conditions—an activity that is not possible in any previously reported chemical catalyst and has not been reported in enzymes. Although the overall Si–C bond cleavage activity is currently moderate, this suggests that biological activity on this non-native substrate is possible and can be enhanced. Further engineering and research into the prevalence of this activity will result in more powerful siloxane-degrading biocatalysts. The current level of activity also enables the study of the mechanism of Si–C bond cleavage.
2
Nature of enzymatic Si–C bond cleavage
Using siloxane 1 and hydroxyethylene glycol 4 as model systems, a series of experiments were performed to explore the mechanism of Si–C bond cleavage. Using purified LSilOx4, the authors determined that the enzymatic conversion of hydroxyethylene glycol 4 to silanol 5 was NADPH-dependent (Figure 3A) and oxygen-dependent. This suggests that the oxidation of Fe-heme is involved in the cleavage of Si–C bonds. The removal of the FAD domain of LSilOx4 resulted in a 2.6-fold loss of Si–C bond cleavage activity of the variant LSilOx4ΔFAD (Figure 3B). These results are consistent with the involvement of enzymatic oxidation in Si–C bond cleavage. The authors then turned their attention to revealing the fate of the cleaved methyl groups. The authors used ABTS and purpald colorimetric methods to detect methanol and formaldehyde, respectively. The results showed formaldehyde as a by-product of the enzymatic reaction. Figure 3D shows the enzymatic reaction time course with hydroxyethylene glycol 4, and the authors hypothesize that there are traces of the GC-MS peak being formylsiloxane 10 (Figure 4). To investigate this possibility, the authors performed Swern oxidation of hydroxyglycol 4, which resulted in the in-situ synthesis and characterization of formylsiloxane by 1H-29si heteronuclear polybond correlation (HMBC) and proton nuclear magnetic resonance (1H NMR)10. Subsequently, the authors obtained high-resolution mass spectrometry data for formylsiloxane10 produced by an enzymatic reaction with hydroxyethylene glycol 4 after 5 minutes of ethyl acetate extraction. With this evidence, it can be speculated that the initial C-H hydroxylation of siloxane 1 is followed by a second oxidation to formylsiloxane 10, which is converted to silanol 5, ostensibly by [1,2]-Brook rearrangement and hydrolysis (Figure 4C). In summary, tandem oxidation of siloxane methyl groups unlocks the enzymatic hydrolysis mechanism of Si-C bonds in VMS and serves as an entry pathway for future biodegradation efforts.
Figure 3
Figure 4
3
Unlock unknown degraded activities in nature
A large body of directed evolution literature has shown that even trace amounts of activity can be amplified, resulting in powerful biocatalysts for transformations unknown to the biological world, including forging and breaking enzymes that break Si–C bonds. The enzyme-catalyzed cleavage of the Si–C bond of siloxanes reported in this paper is a proof of principle and represents the first step in the biodegradation of siloxanes, which are not currently considered biodegradable. Legislation appears to have restricted the use of VMS, including octamethylcyclotetrasiloxane, which the authors have demonstrated to be active. By engineering an enzyme capable of cleaving Si-C bonds, the authors took a critical step towards the eventual biodegradation of VMS.