Short‐chain alcohols inactivate an immobilized industrial lipase through two different mechanisms

Broadly used in biocatalysis as acyl acceptors or (co)‐solvents, short‐chain alcohols often cause irreversible loss of enzyme activity. Understanding the mechanisms of inactivation is a necessary step toward the optimization of biocatalytic reactions and the design of enzyme‐based sustainable processes. The functional and structural responses of an immobilized enzyme, Novozym 435 (N‐435), exposed to methanol, ethanol, and tert‐butanol, are explored in this work. N‐435 consists of Candida antarctica lipase B (CALB) adsorbed on polymethacrylate beads and finds application in a variety of processes involving the presence of short‐chain alcohols. The nature of the N‐435 material required the development of an ad hoc method of structural analysis, based on Fourier transform infrared microspectroscopy, which was complemented by catalytic activity assays and by morphological observation by transmission electron microscopy. The inactivation of N‐435 was found to be highly dependent on alcohol concentration and occurs through two different mechanisms. Short‐chain alcohols induce conformational changes leading to CALB aggregation, which is only partially prevented by immobilization. Moreover, alcohol modifies the texture of the solid support promoting the enzyme release. Overall, knowledge of the molecular mechanisms underlying N‐435 inactivation induced by short‐chain alcohols promises to overcome the limitations that usually occur during industrial processes.

ity of hydrophobic substrates and the chemo-, regio-, and enantioselectivity of enzymes are often increased. [4][5][6][7] It has been reported that in some cases nonconventional media can impair enzyme activity and stability. [8][9][10] Enzyme inactivation can be irreversible, that is, due to protein denaturation and aggregation, or reversible, due to competition between the solvent and substrate molecules to bind the active site. [8,11] Immobilization of the biocatalyst is one of the strategies used to counteract the adverse effects of solvents, and often provides additional advantages, for example it might contribute to improve relevant biochemical features such as enzyme stability, selectivity, and specificity. [12,13] Among immobilized industrial enzymes, Novozym 435 (N-435) is the biocatalyst most widely used in fine chemistry, polymer synthesis and modification, glycerides modification and biodiesel production. [14] In this formulation, the Candida antarctica lipase B (CALB) is immobilized by interfacial activation on a Lewatit VP OC 1600 macroporous support. [15] The exploitation of N-435 was not compromised by the reduction of the catalytic activity triggered by solvent-induced support alteration [14] and/or by hydrophilic molecules such as water and glycerin. [16] These observations point out that the interaction between solvents and N-435 is a key parameter. Nonetheless, only a few seminal works investigated about the impact of solvents on the structure and activity of immobilized CALB. [17][18][19] In this work, we analyze in detail the effects of short-chain alcohols (methanol -MeOH; ethanol -EtOH -and tert-butanol -tBuOH) of particular interest in biocatalysis because they can be used both as acyl acceptors and as a (co)-solvents in synthetic reactions such as the production of biodiesel. [17][18][19][20][21][22][23] Our results indicate that alcohols at high concentration inactivate N-435 through the formation of protein aggregates and by the enzyme release from the solid support.

Effects of short-chain alcohols on N-435
To study the effect of MeOH, EtOH, and tBuOH, 20 mg of N-435 were incubated in milliQ water, in the absence and in the presence of 15%,

Effects of short-chain alcohols on free CALB
Free CALB was recombinantly produced in Escherichia coli cells and purified from the culture supernatant as described in Mangiagalli et al. [24] Purified CALB was lyophilized in a freeze-dryer and stored at 4 • C. Lyophilized CALB was resuspended at the concentration of 1.5 mg ml -1 in sodium phosphate buffer 10 mM, pH 7, in the absence and in the presence of 15%, 45%, and 90% alcohols. Samples were centrifuged for 5 min at 10,000 g; the supernatant was filtered using a 0.22-μm filter (Euroclone, Pero, Italy), transferred to a clean tube and incubated at 25 • C for 24 h. At the end of incubation, several aliquots from each sample were withdrawn and analyzed to assess the effects of alcohols on hydrolytic activity, solubility, and secondary structure of CALB.

Hydrolytic activity assays
The hydrolytic activity of N-435 and free CALB was evaluated on Activity measured after incubation in the absence of alcohol was considered to be 100%.

Transesterification activity assays
The transesterification activity of N-435 was determined in a mixture containing triolein (0.388 mg) and MeOH (42.7 mg) in a molar ratio 1:3. Reactions were incubated at 37 • C and mixed at 500 rpm for 48 h in a thermal shaker (Eppendorf, Hamburg, Germany). The methyl ester produced was quantified using FTIR spectroscopy as described in Natalello et al. [25] Briefly, 5 μl of the organic phase were deposited onto the diamond, single reflection, attenuated total reflection (ATR) device spectra. [25] The transesterification activity of N-435 was normalized using the following equation: Activity measured after incubation in the absence of alcohol was considered to be 100%.

Fourier transform infrared microspectroscopy (micro-FTIR) analyses
To study the secondary structure of immobilized CALB, N-435 beads were deposited on an infrared-transparent BaF 2 window and crushed by applying a second BaF 2 window ( Figure 1). Bead bits were then measured in transmission mode by an infrared microscope Varian 610-IR coupled to the infrared spectrometer Varian 670-IR (Varian, Australia Pty Ltd., Mulgrave VIC, Australia) and equipped with an MCT detector. The variable microscope aperture was adjusted to ≈100 μm × 100 μm and the following conditions were employed: 2 cm -1 spectral resolution, 25 kHz scan speed, 512 scan coadditions, triangular apodization. The absorption spectrum of the immobilized proteins was obtained by subtracting the spectrum of Lewatit VP OC 1600 resin (Lanxess, Koln, Germany), measured under the same experimental conditions, from the spectrum of N-435 beads. In particular, the following resin peaks were employed to optimize subtraction: ≈1730, ≈1603, and ≈1587 cm -1 . When necessary, spectra were corrected for vapor absorption. The second derivative spectra were obtained after the Savitsky-Golay smoothing of the absorption spectra, using Resolutions Pro Software (Varian Australia). For the characterization of free CALB secondary structure, 2-3 μl of protein were deposited on a BaF 2 support and measured after solvent evaporation, following the same experimental conditions described above.

Solubility assay
The solubility of free CALB after exposure to alcohol was assessed by SDS-PAGE. After incubation with alcohol, samples (total protein, 40 μl) were centrifuged at 4 • C for 10 min at 10,000 g. An equal volume (10 μl) of supernatant (soluble proteins) and total protein from each sample was analyzed on 14% SDS-PAGE. The densitometric volume of the protein band was determined with Image Lab software (Bio-Rad, California, USA). At each condition, relative solubility was calculated with reference to total protein.

TEM analysis
N-435 beads were sectioned in 150-nm slices using an ultramicrotome

Assessment of CALB release from the solid support
The release of CALB from the solid support was assessed by Bradford protein assay, hydrolytic activity assay and FTIR analysis of the supernatants collected after N-435 exposure to alcohol (see Section 2.2).
The amount of CALB in the supernatant was determined by the Bradford protein assay (Bio-Rad, California, USA) using serum albumin as the standard. The hydrolytic activity was measured using pNPO as a substrate, as described in Section 2.4. An equal volume (≈90 μl) of supernatant from each sample, including the negative control, was lyophilized, resuspended in 4 μl of water and analyzed by ATR-FTIR spectroscopy. Two microliters of these samples were deposited on the diamond, single reflection, ATR device (Quest Specac, UK) and the spectrum was collected after solvent evaporation, under the conditions described in Section 2.5.

Statistical analysis
All the experiments were repeated six times. Statistical analysis was performed using OriginPro 2020 (OriginLab Corporation, Northampton, USA). p-values were determined by using an unpaired two-tailed t-test. Second derivative spectra of the amide I band region were obtained by averaging at least 8 spectra for each condition.

RESULTS
To study the secondary structure of N-435 and monitor alcohol induced changes, we developed a method based on micro-FTIR, which allows overcoming the interference of the solid support in the spectral regions of interest [26,27] (Figure 1A). N-435 beads are fragmented between two BaF 2 windows, and the fragments are visually inspected by a microscope to select those of thickness suitable for micro-FTIR analyses. Among them, spectra were collected from fragments characterized by high protein absorption. Compared to a previously published procedure, [26] our approach allows direct characterization of immobilized CALB, avoiding the use of synchrotron light and inclusion of beads in paraffin wax. The FTIR spectrum of CALB (in black in Figure 1B) was obtained by subtracting from the N-435 spectrum (in red) that of Lewatit VP OC 1600 resin (in blue). The resulting spectrum displays well resolved amide I (C═O stretching) and amide II (C-N stretching and N-H bending) bands, both due to the vibrations of the peptide bond. [28,29] The second derivative spectrum of CALB in the amide I region shows two main peaks at 1662 cm −1 assigned to α-helices (with possible contributions of random coils) and at 1640 cm −1 , assigned to native β-sheet structures. These spectral components were found in the FTIR spectra of free [24] (black line in Figure 4B and Figure S1) and immobilized CALB [26] and are consistent with the α/β fold of the protein. [30] Although our method does not provide information on the spatial distribution of the enzyme on the solid support, [26] it proved useful for studying the secondary structure of immobilized CALB and was applied to monitor the structural changes induced by short-chain alcohols.

Effects of short-chain alcohols on CALB
The effects of MeOH and EtOH on the structure and function of lipases is of great interest due to their use as acyl acceptors in enzymatic synthesis reactions, [8,[17][18][19] as MeOH is widely used in biodiesel production, although it is a competitive inhibitor of free CALB, [24,[31][32][33][34] and EtOH is of growing relevance in renewable process design. [35] Nonetheless, their effects on N-435 are still unclear and unexplored.
In the absence of alcohol, N-435 shows a specific hydrolytic activity of 1.61 ± 0.12 au mg -1 and a specific transesterification activity of 6.42 ± 0.49% ME /mg, both taken as 100% of relative activity in Figure 2. At 15% MeOH, the relative activity in terms of hydrolysis and transesterification are 74.41 ± 6.93% and 79.92 ± 4.62%, respectively (Figure 2A and B). Upon incubation in 45% and 90% MeOH, the relative activities of N-435 drop to 49.82 ± 6.93% and 51.08 ± 6.41% for the hydrolysis and 46.66 ± 5.67% and 48.09 ± 7.18% for transesterification (Figure 2A and B). Micro-FTIR analysis indicates that incubation in 15% MeOH does not affect the secondary structure of immobilized CALB ( Figure 3A). By contrast, incubation in 45% and 90% MeOH induces protein aggregation, as indicated mainly by the presence of a peak at ≈1625 cm -1 , which is typical of intermolecular β-sheet structures [28,29] (Figure 3A). A similar behavior was also observed for free CALB, although accompanied by a more remarkable loss of structure, hydrolytic activity, and solubility ( Figure 4A, D, and G). Overall, immobilization seems to mitigate the detrimental effects of MeOH, in fact after exposure to 90% MeOH, the relative activity of N-435 is 2-fold higher than that observed for free CALB.
Hydrolytic and transesterification activities are slightly affected by incubation in 15% EtOH (relative activities: 77.41 ± 6.05% and 80.10 ± F I G U R E 2 Effects of short-chain alcohols on the activity of N-435. Relative activity after exposure of N-435 to MeOH (A, B), EtOH (C, D), and tBuOH (E, F). Hydrolytic activity was measured using pNPO as a substrate at the concentration of 1 mM, while transesterification activity was assayed using a mixture containing MeOH and triolein in a 1:3 molar ratio. Mean values of six independent measurements are represented with error bars indicating standard deviations. Statistical analyses were performed using unpaired two-tailed Student's t-test, n.s.: not significant p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.
Tert-butanol is one of the most used (co)-solvents in lipasecatalyzed reactions, since this class of enzymes is not active on tertiary alcohols. [22,35] tBuOH is widely used in biodiesel production due to its ability to dissolve glycerol, which would otherwise accumulate on the surface of N-435 beads, causing inactivation and promoting bead aggregation during the reaction. [16,[21][22][23]36] The effects of tBuOH on the activity and structure of immobilized CALB are more pronounced than those exerted by MeOH and EtOH. Indeed, immobilized CALB undergoes aggregation already at 15% tBuOH ( Figure 3). In 90% tBuOH, the residual hydrolysis and transesterification activities are 34.03 ± 5.41% and 35.03 ± 6.04%, that is, significantly lower than those obtained with MeOH and EtOH at the same concentration. These negative effects were also observed on free CALB (Figure 4), which completely loses its activity and solubility in the presence of 90% tBuOH ( Figure 4C, F, and I). Interestingly, after exposure to 15% tBuOH, free CALB retains its activity, solubility and secondary structure ( Figure 4C, F, and I).

F I G U R E 3
Effects of short-chain alcohols on the secondary structure of N-435. Micro-FTIR spectra were collected after exposure to MeOH (A), EtOH (B), and tBuOH (C). Second derivative spectra of the amide I band region were obtained by averaging at least 8 spectra for each condition. The shadowed area refers to the standard deviation of the data. Peak assignment is reported in Table S1.

Effects of short-chain alcohols on CALB release from the solid support
Our results indicate that in the presence of the tested alcohols, N-435 undergoes inactivation by aggregation, to varying degrees depending on the alcohol's chemical-physical properties and concentration.
At this point, we wondered if the inactivation of N-435 was entirely ascribable to aggregation of the enzyme in situ, or whether the alcohol may also induce release of CALB from its support.
To monitor the release of CALB from the resin, N-435 beads were exposed to alcohol and the supernatants were collected and analyzed.
The amount of released CALB was determined by the Bradford assay ( Figure 5A-C). Under all conditions tested, alcohol causes a progressive release of CALB, with more pronounced effects of tBuOH and EtOH. FTIR absorption spectra of lyophilized supernatants show the presence of two main components, corresponding to the protein signal (peaks of amide I and amide II) and the matrix signal ( Figures 5D-F and S2). We observed that the amide I and II bands are better resolved at high alcohol concentrations, indicating an increase in protein concentration, in agreement with the Bradford assay. Analysis of second derivatives of amide I band obtained upon treatment with 45% and 90% alcohol indicates that the released protein is either denatured or forms aggregates, with tBuOH showing the most remarkable effect ( Figure S3). The specific activity of supernatants collected after alcohol exposure decreases as alcohol concentration increases ( Figure S4), due to the progressive release of partially active CALB from its support.
The effect of 0% and 90% alcohol on the N-435 matrix was analyzed by TEM to assess possible morphological changes. The macroporous texture of the Lewatit VP OC 1600 support exposed to 90% MeOH appears remarkably compact ( Figure 5G and H). In contrast, larger pores are present after incubation in tBuOH and EtOH ( Figure 5I and J). The support alterations observed in this work are similar to those reported in literature, and obtained in other conditions. [18,37,38] In addition, it has been reported that the Lewatit VP OC 1600 matrix can dissolve, leading to the release of acyl donors such as benzoic acid, sorbic acid, and 2-hydroxyethyl sorbate, which are contaminants giving potentially undesired products. [39] Overall, the release of the enzyme from N-435 is suggested to depend on the conformational effects induced by the alcohol on both CALB and its support.

DISCUSSION
In this work, the molecular mechanisms of N-435 inactivation induced by alcohols were studied by combining experimental data obtained by spectroscopic, biochemical, and microscopic techniques. Although this issue is of broad interest, technical difficulties due to the interference of the solid support result in a few works available on this topic. [17][18][19] The method presented here has the advantage of being simple, straightforward, and applicable to other non-conventional media and enzyme preparations as well.
CALB activity in organic media was proved sensitive to water content. [31,[40][41][42][43][44] The proposed inactivation mechanism gives a prominent role to the thickness of the hydration shell whose increase results in limiting the access of the substrate to the active site. [41,43,44] In our system, N-435 samples were freeze-dried after incubation to completely remove traces of water and alcohols. Based on this procedure, it is reasonable to assume that N-435 inactivation is mainly due to structural effects.

F I G U R E 4
Effects of short-chain alcohols on free CALB. Relative activity after exposure of free CALB to MeOH (A), EtOH (B), and tBuOH (C). The hydrolytic activity was measured using 1 mM pNPO as a substrate. Micro-FTIR spectra of CALB were collected after exposure to MeOH (D), EtOH (E), and tBuOH (F). The second derivative spectra of the amide I band region were obtained by averaging at least 8 spectra for each condition. The shadowed area refers to the standard deviation of the data. Peak assignments are reported in Table S1. Solubility of free CALB after incubation in MeOH (G), EtOH (H), and tBuOH (I). Mean values of six independent measurements are represented with error bars indicating standard deviations. Statistical analyses were carried out using unpaired two-tailed Student's t-test, n.s.: not significant p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.
Our results indicate that MeOH and EtOH at the highest concentrations used in this work induce the aggregation of free and immobilized CALB, causing irreversible loss of activity, whereas CALB well tolerates MeOH and EtOH up to 15%. Therefore, data highlight the importance of finding a trade-off between the alcohol concentration required by a given reaction and the enzyme performance and strengthen the stepwise addition of alcohols during the process as a possible solution to counteract the irreversible inactivation of N-435 induced by high alcohol concentrations. [36,[45][46][47] On the other hand, the inactivation of N-435 by tBuOH is more pronounced than those induced by MeOH and EtOH and not consistent with the beneficial effect (i.e., increase of product yields) reported when tBuOH is used as a solvent in biodiesel transesterification reactions. [16,[21][22][23]36] We cannot exclude that the glycerol produced during transesterification stabilizes CALB and prevents its aggregation, as reported for other proteins [48,49] and would hypothesize that the outcomes of tBuOH in transesterification reactions could be due to a combination of negative (i.e., inactivation of the enzyme) and positive (i.e., better dissolution of MeOH and glycerol) effects.
CALB release from the solid support, observed in the presence of tested alcohols, might be due to the weakening of the enzyme-support interactions and/or to changes of the solid support itself. The phe-F I G U R E 5 Effects of short-chain alcohols on the release of CALB from the solid support. Bradford assay (A-C). The mean values of six independent measurements are represented with error bars indicating standard deviations. Statistical analyses were carried out using unpaired two-tailed Student's t-test, *p < 0.05, **p < 0.01, ***p < 0.001. ATR-FTIR analyses (D-F) performed on lyophilized supernatants, resuspended in water, after incubation in MeOH (D), EtOH (E), and tBuOH (F). Am. I and Am. II: amide I and amide II bands; M: peaks typical of the Lewatit VP OC 1600. One of six independent experiments is shown. TEM analysis (G-J) on N-435 incubated for 24 h at 37 • C in the absence and in the presence of 90% alcohols. One of three independent experiments is shown.
nomenon of enzyme leaching is commonly observed when lipases are immobilized by interfacial activation on hydrophobic support as in the case of N-435 and has to be considered with caution when planning the conditions for the use of this enzyme in industrial application. [14,[45][46][47] The overall scenario that can be drawn from the data reported suggests that short-chain alcohols induce inactivation of N-435 through two different mechanisms: i) conformational changes that induce aggregation of CALB on the surface of the solid support; ii) detachment of the enzyme from its support. The higher susceptibility to alcohol inactivation of free and released CALB indirectly suggests that immobilization protects the enzyme from harmful alcohols effects.
On this ground, the design of biocatalytic processes should consider the complexity of the interaction between short-chain alcohols and immobilized CALB. Indeed, in addition to the inactivation mecha-nisms described in this work there is evidence for influence of water activity, [31,[40][41][42][43][44] competitive inhibition [24,31,32,34] and effects on the local conformational dynamics. [24,33,50,51] In conclusion, the knowledge of the multitude of factors and mechanisms underlying enzymatic inactivation by short-chain alcohols promises to contribute to both the optimization of biocatalytic reactions and the development of new sustainable processes.