Luria Bertani (LB) and Luria-Agar (LA) culture media were used, with the latter prepared by incorporating agar (15 g/L) (Conda Laboratories SA, Madrid, Spain) into LB medium. Isopropyl-β-d-thiogalactopyranoside (IPTG; Invitrogen Life Technologies, Carlsbad, CA, USA) was used to induce lipase expression. Affinity columns (HiTrap Chelating HP; GE Healthcare, Uppsala, Sweden) were used to purify LipC12. The Escherichia coli TOP10 strain (Invitrogen Life Technologies) was used for plasmid storage at -80 °C and the BL21(DE3) strain (Novagen, Madison, WI, USA) was used for expression. The immobilization support was Immobead 150P (Sigma-Aldrich, Merck, St. Louis, MO, USA), with a particle size of 0.15–0.5 mm. Triethylamine, n-heptane, n-hexane (99.5 %), t-butanol and toluene were from Vetec (Duque de Caxias, RJ, Brazil), ethanol (99.5 %), glycerin, hydrochloric acid, acetic acid, isopropanol and Tween 80 were from Synth (Diadema, SP, Brazil), tris(hydroxymethyl)aminomethane and imidazole were from Neon Química (Suzano, SP, Brazil). Cetyltrimethylammonium bromide (CTAB), Coomassie R-250, triolein (65 %), oleic acid (90 %) and kanamycin were from Sigma-Aldrich, Merck. Commercial olive oil (Gallo brand) was bought at a local supermarket. All other reagents, such as salts used for solution preparation, surfactants and reaction substrates, were of analytical grade.

LipC12 was produced and purified according to Glogauer et al. (28), with slight modifications. E. coli BL21(DE3), carrying the plasmid pET28a-lipC12, was cultivated at 37 °C in 800 mL of LB medium containing kanamycin (50 μg/mL), in 2-liter Erlenmeyer flasks. When the absorbance at λ=600 nm reached value of 0.600, IPTG was added (to give a concentration of 0.5 mmol/L) for induction and the culture was incubated for another 16 h, at 20 °C. The broth was centrifuged (CR21E; Hitachi, Tokyo, Japan) (4000×g) at 18 °C for 10 min, giving a cell pellet that was then resuspended in 50 mL of lysis buffer (50 mmol/L Tris-HCl, pH=7.5, 500 mmol/L NaCl, 10 mmol/L imidazole) and sonicated in an ice bath using a SONICATOR® XL 2020 (Heat Systems-Ultrasonics Inc., New York, NY, USA; twelve 20-second pulses of 90 W, with 30-second intervals). Cell debris was removed by centrifugation (15 000×g) of the crude extract at 4 °C for 15 min. The supernatant was purified with a HiTrap column, previously loaded with Ni²⁺ and equilibrated with lysis buffer. After loading with the His-tagged protein, the column was eluted stepwise, with increasing concentrations (50 to 500 mmol/L) of imidazole in a buffer containing 50 mmol/L Tris-HCl (pH=7.5) and 500 mmol/L NaCl. Two column volumes of buffer were passed at each imidazole concentration. The fractions that contained proteins were analyzed by SDS-PAGE and then pooled and dialyzed with 50 mmol/L Tris-HCl buffer (pH=7.5) containing 150 mmol/L NaCl and 10 mmol/L CaCl₂.

The purified LipC12 preparation had a protein concentration of 2.3 mg/mL, determined by the bicinchoninic acid (BCA) method (29) using a kit (Pierce Biotechnology, Rockford, IL, USA), and had a specific olive-oil-hydrolyzing activity of (1948±88) U/mg (mean±standard deviation, N=5). The cell-free crude extract of LipC12 had a specific olive-oil-hydrolyzing activity of (2774±139) U/mg (N=5) and a total protein concentration of 14.5 mg/mL. Densitometry analysis of an SDS-PAGE gel showed that LipC12 represented 29.4 % of the total protein (i.e. 4.26 mg/mL). A 10-µL aliquot of 0.01 % (m/V) sodium azide was added to the enzyme solution, which was then stored at 4 °C. The specific activity of this LipC12 solution remained constant during the studies.

LipC12 was immobilized through covalent binding on Immobead 150P. Two different immobilization solutions were used: i) crude cell-free extract that had been centrifuged to eliminate cell debris, and ii) a solution of purified enzyme. The optimized protocol of Madalozzo et al. (30) was used, with minor adaptations.

Dry Immobead 150P beads were used, without pretreatment. A mass of 0.1 g of support was added to 5 mL of LipC12 solution at pH=7.5. This represented a protein loading of 10 mg/g for the purified preparation and a protein loading of 200 mg/g for the crude extract. The suspension was incubated on an orbital shaker (TE-421; TECNAL, Piracicaba, SP, Brazil) at 150 rpm for 6 h at 4 °C. The olive-oil-hydrolyzing activity of the lipase-containing supernatant offered for immobilization was followed during this incubation. After the incubation, the immobilized preparation was washed (3×) with 50 mL of 50 mmol/L Tris-HCl buffer (pH=7.5). It was then recovered by filtration through qualitative filter paper (Whatman n° 15), desiccated for 16 h under partial vacuum at 4 °C and stored at -20 °C.

The immobilization efficiency (IE/%) achieved at the end of the 6-hour incubation was calculated as follows:

where A_i is the olive-oil-hydrolyzing activity (U) of the supernatant before the addition of Immobead 150P and A_f is the olive-oil-hydrolyzing activity (U) that remains in the supernatant after the incubation.

The activity retention (AR/%) was calculated as follows:

In this equation, A_O represents the triolein-hydrolyzing activity of the immobilized preparation (which was measured in n-hexane) and A_T represents the theoretical olive-oil-hydrolyzing activity of the immobilized preparation. This theoretical value was calculated as follows:

where m_ip and m_fp are the masses of protein in the supernatant before and after immobilization, respectively. SA_O is the specific triolein-oil-hydrolyzing activity (U/mg of protein), measured in n-hexane, of the free enzyme. If the lipase is activated upon immobilization, then AR values can be above 100 %.

Control experiments were done with the enzyme solution being incubated under the immobilization conditions for 24 h, but without the supports. The supernatants in these control experiments showed no loss of enzymatic activity.

To confirm that LipC12 was covalently immobilized on Immobead 150P, desorption studies were done. Immobilized LipC12 (20 mg) was placed in Eppendorf tubes that contained 2 mL of a 2 % (m/V) sodium dodecyl sulfate (SDS) solution in distilled water. After a 30-minute incubation in boiling water (31), the immobilized enzyme was removed by filtration, and the protein content in the filtrate was determined using the BCA method (29). No protein was detected in the filtrate of LipC12 immobilized from the purified fraction, whereas LipC12 immobilized from the cell-free crude extract gave a low concentration of protein of 0.06 mg/mL, below the lower level of sensitivity of the method. These results show that LipC12 was effectively attached to the support through covalent bonding.

Two bioimprinting strategies were used with LipC12 immobilized onto Immobead 150P, using either purified LipC12 or the cell-free crude extract containing LipC12: (i) pre-immobilization, with the bioimprinting agents added to the solution of free LipC12 prior to immobilization, and (ii) post-immobilization, with the bioimprinting agents added to immobilized LipC12.

Pre-immobilization bioimprinting was done in sealed 25-mL Erlenmeyer flasks, each containing enzyme solution (either 3 or 4 mL; independently of the volume, the solution contained 1 mg of purified LipC12 or, in the case that the crude extract was used for immobilization, 20 mg of protein) and 1 or 2 mL of the bioimprinting agents (Table 1). The flask was incubated at 4 °C on an orbital shaker (TE-421; TECNAL) at 150 rpm, initially for 15 min, and then for a further 6 h after the addition of 0.1 g of Immobead 150P. The immobilization efficiency (IE, Eq. 1) and retention of activity (AR, Eq. 2) were then determined, based on the olive-oil-hydrolyzing activity, measured in aqueous medium. The immobilized derivative was then washed with 5 mL of 50 mmol/L Tris-HCl buffer, pH=7.5, filtered under vacuum (through Whatman filter paper n° 15), and dried for 24 h at 4 °C in a vacuum desiccator. After drying, it was washed twice more with t-butanol (5 mL each time) and dried once more for 24 h under vacuum in the desiccator. Finally, it was stored in Eppendorf tubes at -20 °C until use.

Post-immobilization bioimprinting was done with 0.1 g of dried immobilized LipC12 preparation (obtained using the standard LipC12 immobilization procedure described above). This preparation was added to a sealed 25-mL Erlenmeyer flask with 4 mL of Tris-HCl buffer (50 mmol/L, pH=7.5) and 1 mL of the bioimprinting solution (Table 1). The flask was incubated on a shaker (150 rpm) at 4 °C for 15 min. The immobilized derivative was then recovered by filtration (Whatman filter paper n° 15) under vacuum. After being washed twice with t-butanol (5 mL each time) to remove the bioimprinting agents, it was dried for 24 h at 4 °C in a vacuum desiccator and then stored in Eppendorf tubes at -20 °C until use.

Two bioimprinted immobilized derivatives, prepared through immobilization of LipC12 from a crude extract, were reused over multiple cycles of triolein hydrolysis in n-hexane: one derivative underwent pre-immobilization bioimprinting with 147 nmol oleic acid (OA5), while the other underwent post-immobilization bioimprinting with 29.4 nmol oleic acid (OA1). After each cycle, the immobilized derivatives were recovered by vacuum filtration (Whatman filter paper n° 15) and washed twice with n-hexane (5 mL each time). They were then dried at 4 °C for 16 h in a vacuum desiccator and added to fresh reaction medium for the next cycle. The activities are reported as percentages of the absolute conversion that was achieved in the first cycle.

These tests were done with only φ(water)=2 % added to the reaction medium, as bioimprinting tends to be ineffective at high water volume fractions, which allow the lipase molecule sufficient flexibility for it to revert to the structure that it had prior to the bioimprinting (22).

Statistical analysis and graphing were done using OriginPro (36). The values presented in the figures and tables of this work correspond to mean values±standard deviation. Mean values were compared using Student’s t-test, with the aid of Microsoft Excel analysis tools; p-values lower than 0.05 were interpreted as indicating a significant difference.

Two pre-immobilization bioimprinting strategies were tested. In one strategy, purified LipC12 preparations were bioimprinted with the bioimprinting agents listed in Table 1 and then immobilized on Immobead 150P; this will be referred to as the “pre-pure” strategy. In the other strategy, crude cell-free extracts containing LipC12 were bioimprinted with the same bioimprinting agents and then immobilized on Immobead 150P; this will be referred to as the “pre-crude” strategy.

With the pre-pure strategy, the most effective bioimprinting agent was OA5 (R_H=828 %) (Fig. 1a). The next best bioimprinting agents were hydrophilic solvents, with R_H values of 785 % for methanol and 767 % for ethanol. Good results were also obtained with n-heptane (R_H=673 %).

With the pre-crude strategy, again, the most effective bioimprinting agent was OA5 (R_H=936 %) (Fig. 1b). The hydrophobic solvents gave the next best results, with R_H values of 837 % for n-heptane and 828 % for toluene. Good results were also obtained with the hydrophilic solvents, with R_H ranging from 720 to 748 %.

Although good results were obtained with OA5, treatment with OA1 (in which the oleic acid amount is a fifth of that in OA5) gave significantly lower R_H values, 235 % for the pre-crude strategy and 530 % for the pre-pure strategy. Comparing these two R_H values obtained with OA1, the lower value obtained for the pre-crude strategy might be due to adsorption of a significant proportion of the limited amount of oleic acid on the non-LipC12 proteins in the crude extract.

For all pre-immobilization bioimprinting treatments, the immobilization efficiencies (IE) were 100 % (Table S1), indicating that the presence of bioimprinting agents did not affect the immobilization of LipC12. All treatments showed higher activity retention values than the control (AR=204 % for the pre-pure strategy and AR=261 % for the pre-crude strategy), suggesting activation of LipC12 through the bioimprinting treatment (Table S2). Note that the AR values are directly proportional to the R_H values, as they are calculated based on the theoretical activity of LipC12 on the support and the measured activity of the immobilized preparation (see Eq. 2).

Two post-immobilization bioimprinting strategies were tested. In one strategy, purified LipC12 was immobilized on Immobead 150P and the immobilized preparation was then bioimprinted with the agents listed in Table 1; this will be referred to as the “post-pure” strategy. In the other strategy, a crude cell-free extract containing LipC12 was immobilized on Immobead 150P and the immobilized preparation was then bioimprinted with the same bioimprinting agents; this will be referred to as the “post-crude” strategy. For both strategies (post-pure and post-crude), the immobilization efficiency (IE) was 100 % (Table S1). The AR values were 260 % for the post-pure strategy and 261 % for the post-crude strategy (Table S2).

With the post-pure strategy, the most effective bioimprinting agent again was OA5 (R_H=1211 %). The next best results were obtained with ethanol (R_H=611 %) and n-heptane (R_H=352 %) (Fig. 2a). With OA1, the R_H value was only 200 %.

With the post-crude strategy, the most effective bioimprinting agent was OA1 (R_H=1356 %) (Fig. 2b). Two hydrophilic solvents gave the next best results, t-butanol (R_H=1216 %) and methanol (R_H=1024 %). With OA5, the R_H value was 923 %. With the solvents methanol, ethanol, t-butanol, n-heptane and toluene, there was a tendency for the R_H values obtained in the post-crude strategy (ranging from 600 to 1200 %) to be higher than the corresponding values obtained in the post-pure strategy (ranging from 100 to 600 %).

In our experiments reported above, both hydrophilic solvents (methanol, with logarithm of partitition coefficient log P=-0.77; ethanol, log P=-0.31; t-butanol, log P=0.57) and hydrophobic solvents (toluene, log P=2.73; n-heptane, log P=4.66) activated immobilized LipC12. However, there is no clear correlation between the degree of activation and the log P values of the solvents. Hydrophobic solvents typically interact with the lid domain of lipases, promoting its opening and enhancing catalysis, similar to the phenomenon of interfacial activation (6). The mechanism of activation of lipases by pretreatment with polar solvents is less clear, but the activation of LipC12 by methanol and ethanol is not surprising, as prior work showed that free LipC12 was significantly activated by preincubation (for 48 h at 4 °C) in 30 % aqueous solutions of methanol and ethanol, with relative activities of 1561 and 588 % respectively (28).

Some insight into the effects of solvents is given by the work of Liu et al. (6), who pretreated immobilized Pseudomonas cepacia lipase (PS) with pure organic solvents of varying log P values, molecular structures, and functional groups. These treatments induced changes in the secondary structure of PS: immobilized PS pretreated with all solvents had decreased contents of α-helices and β-turns and increased contents of β-sheets and random coils, regardless of whether the solvent activated the enzyme or not. However, it should be noted that Liu et al. (6) used pure organic solvents for the pretreatment, whereas we used around 20 % solvent in Tris-HCl buffer.

In the bioimprinting experiments above, the relative activity of immobilized LipC12 (R_H) was evaluated based on the hydrolysis of triolein in n-hexane. We selected the bioimprinting agents that performed best in this hydrolysis study (i.e. OA1 and OA5) and used them to evaluate the effect of bioimprinting on the esterification activity, namely the esterification of oleic acid with ethanol, in n-hexane.

For both pre-immobilization bioimprinting and post-immobilization bioimprinting, the relative esterification activity (R_E) was higher when LipC12 was immobilized from the crude extract than when it was immobilized from a purified preparation (Fig. S3). The best result, R_E=345 %, was obtained with the pre-crude strategy, with OA5 as the bioimprinting agent. The next best result, R_E=308 %, was obtained with the post-crude strategy, with OA1 as the bioimprinting agent. The low R_E values obtained when LipC12 was immobilized from the purified preparation contrast with the high R_H values obtained with these same preparations in the previous experiments. The fact that the bioimprinting was done with one of the products of triolein hydrolysis, namely oleic acid, might have contributed to the high R_H values. In the case of the ester synthesis reaction R_E, it was not possible to investigate bioimprinting with the product (i.e. the ester), as the bioimprinting is done in aqueous medium and the lipase would hydrolyze the ester.

Since the best results for triolein hydrolysis were obtained using LipC12 immobilized from the crude extract, for both pre-immobilization bioimprinting (R_H=936 % for OA5) and post-immobilization bioimprinting (R_H=1356 % for OA1), these bioimprinted immobilized preparations were evaluated for their reusability in successive cycles of triolein hydrolysis in n-hexane. Although there was experimental error of the order of ±10 %, both preparations retained essentially the same activity over seven reaction cycles (Fig. 3). In other words, the increased activity conferred by bioimprinting was maintained throughout the reuse cycles.

Although there is a significant body of work concerning bioimprinting of immobilized lipases, our study is the first that compares pre-immobilization bioimprinting and post-immobilization bioimprinting. Previous works have used either pre-immobilization bioimprinting, or post-immobilization bioimprinting, or “simultaneous immobilization and bioimprinting” (in which the bioimprinting agent is added during the immobilization step itself). Further, our study is the first to compare bioimprinting of a lipase immobilized from a pure preparation with that of a lipase immobilized from a crude extract. Our results from this part of the study, in which we obtained the best relative hydrolytic activity with the post-crude strategy (R_H=1356 %, for OA1), are especially encouraging, since immobilization from the crude extract avoids the costs of purifying the enzyme prior to immobilization.

Table 2 (7, 8, 22, 23, 25-27, 37–41) shows previous studies in which immobilized lipases have been bioimprinted with fatty acids, with the results being evaluated in hydrolysis and esterification reactions. In this table, relative activities are defined in a manner analogous to Eq. 4, with the activity of the bioimprinted preparation being divided by that of the immobilized preparation without bioimprinting and then multiplied by 100 to give a percentage. When the results of relative activities were not provided by the authors, they were calculated from the activity values given in the paper. Most authors have used commercial enzymes. Also, most authors have immobilized the lipases by adsorption (Table 2), using either hydrophobic supports, such as Accurel MP-1000 (25) and NKA resin (38), or hydrophilic supports, like the ion exchange resin D152H (27). Bioimprinting prior to immobilization is the most used strategy. However, some authors have added bioimprinting agents during the immobilization itself, especially when immobilizing by encapsulation or producing crosslinked enzyme aggregates (CLEAs) (7, 8, 13, 37).

Our best relative hydrolytic activity, 1356 %, obtained with post-immobilization bioimprinting, is higher than the relative hydrolytic activities that have been reported previously for pre-immobilization bioimprinting and simultaneous immobilization and bioimprinting (Table 2). Notably, there have been no prior reports on hydrolytic activities obtained with post-immobilization bioimprinting. The most closely related study is that of Sánchez et al. (25). They bioimprinted LipC12 with oleic acid for 60 min prior to immobilization on various supports, including Immobead 150, Accurel MP 1000, polypropylene powder, Nanomer I.44P (a nanoclay containing approx. 40 % dimethyl dialkyl amine by mass), and chitosan. The immobilized preparations were utilized in both the hydrolysis of soybean oil and the esterification of oleic acid with 1-pentanol. The bioimprinting treatment significantly increased the conversions for almost all their immobilized preparations, with the most notable enhancements occurring in those derived from hydrophobic supports. However, Sanchez et al. (25) only reported conversions after 5 h and they did not report the relative initial activities of bioimprinted LipC12 as we did in the current work. Also, the reactions that they used to evaluate the effect of bioimprinting on hydrolysis and esterification are different from those that we used. These differences in the strategy for evaluating the effects of bioimprinting makes it difficult to compare our results directly with theirs. In any case, in their work, bioimprinted and non-bioimprinted LipC12 gave no difference in conversions for the hydrolysis of soybean oil (i.e. the relative conversion at 5 h was 100 %), whereas in our work post-immobilization bioimprinting gave a relative activity of 1356 %. Likewise, in their work, bioimprinted LipC12 gave a relative conversion at 5 h of 128 % for the esterification of oleic acid with 1-pentanol, whereas in our work post-immobilization bioimprinting gave a relative activity of 308 % for the esterification of oleic acid with ethanol. This comparison shows that post-immobilization bioimprinting is a good alternative for improving LipC12 activity.

For relative esterification activities (Table 2), our best value of 345 % for post-immobilization bioimprinting with the higher oleic acid amount (OA5) is among the highest reported in the literature, comparable to those obtained for silica gel entrapment (350 %) (41) and sol-gel matrix entrapment (348 %) (39), both involving simultaneous bioimprinting and immobilization onto CLEAs. The only study reporting results for post-immobilization bioimprinting was done with Candida rugosa lipase (CRL) immobilized on polypropylene powder (22), with the bioimprinting agents consisting of a low concentration of a commercial blend of fatty acids (C₁₄-C₁₆), ethanol, and buffer at pH=7.0, resulting in a relative esterification activity of 168 %.

Although our bioimprinting strategies enhanced LipC12 activity in the current work, we used a specific combination of lipase, substrates, immobilization support, bioimprinting agent, and activity assay. Further research is necessary to confirm whether these strategies are effective for a broad range of lipases.